Method development in thermal analysis. Part 1:

Transcription

1 1/2005 Information for users of METTLER TOLEDO thermal analysis systems Dear Customer, The next two editions of UserCom will focus on systematic method development. Measurement curves are usually much easier to interpret when optimum method parameters are used. Factors such as the heating rate, type of crucible, and furnace atmosphere have a decisive influence on the quality of measurement results. We are pleased to announce an important new development for customers interested in the measurement of sorption processes. The new TGA Sorption Analyzer System allows you to perform measurements under closely defined conditions of relative humidity and temperature. 21 TA Tip Method development in thermal analysis. Part 1: Dr. Markus Schubnell Introduction The development and validation of methods is of major importance in today s quality assurance systems. The starting point is usually a trial method that is then optimized and validated in several iterative steps. The final result is a validated method that is used for SOPs (SOP: Standard Operating Procedure). The development and validation of a measurement procedure is time-consuming and costly. This means it is important to start off with a good trial method right from the beginning. The following article attempts to systematize the development of thermoanalytical methods and discusses the most important aspects involved. Figure 1 presents an overview of this process. Contents TA Tip - Method development in thermal analysis. Part 1: 1 New in our sales program - Controlled relative humidity interface for the TGA/SDTA85x e 5 Applications - Model free kinetics 6 - Determination of the adsorption and desorption of moisture in pharmaceutical substances 9 - Investigation of the temperature stability of polymer additives and their decomposition products by TGA-MS and TGA-FTIR 11 - Determination of glass transition temperatures of powder disks by TMA 13 - Phase transitions of lipids and liposomes 16 Dates - Exhibitions 19 - Courses and Seminars 19

2 Physical properties Specific heat capacity Coefficient of expansion Young s modulus Choosing the measuring technique: DSC TMA (dilatation, penetration, DLTMA } tension, bending) TGA DMA (shear, tension, compression, bending) TOA (thermo-optical analysis) Physical transitions Melting and crystallization Vaporization, sublimation, drying Glass transition, softening Polymorphism (solid-solid transitions) Liquid crystals Purity determination Chemical properties Decomposition, pyrolysis, oxidative stability Composition, content (moisture, fillers), ash Kinetics, reaction enthalpy Cross-linking, vulcanization (process parameters) DSC TGA/SDTA TMA/SDTA DMA/SDTA Sample preparation Choosing the crucible Choosing the temperature program Choosing the atmosphere After the measurement Evaluation Validation Validated method Figure 1. Procedure for developing a thermoanalytical method. Table 1. Overview of the application possibilities of different thermoanalytical techniques. means very suitable, means less suitable. Step 1: Choosing the right measurement technique The analytical task Method development begins with precisely defining the information you hope to get from an analysis of the sample. Typical questions could for example be At what temperature does the glass transition occur? Does the sample exhibit polymorphism? How pure is my product? What is the moisture content of my sample? and so on. Depending on the analytical task and the information required, you first have to decide which measurement technique to use. Table 1 presents an overview of the application possibilities of various thermoanalytical measurement techniques. Sensitivity The most important considerations at this point are basic questions that have to do with later validation of the method: Is the sensitivity of the method good enough to obtain the desired information? What possible consequences arise from the sensitivity of the method, e.g. with regard to sample size or heating rate? What accuracy can I expect to achieve? Is the accuracy sufficient for my purposes? Do any interfering effects have to be taken into account? Are the effects more serious with one measurement technique than with another (robustness of method)? To answer these questions, one needs to understand the operating principles of the instruments and to have had practical experience. Particularly important is information on the signal-to-noise ratio, the long-term stability or drift, and measurement reproducibility. Example: What is the smallest mass loss step that can be resolved by TGA? Answer: The decisive point in this case is the signal-to-noise ratio of the measurement signal (balance and surroundings). As a rule of thumb, a measurable mass change should be at least four times 2

3 greater than the background noise signal. Assuming that the noise is 1 µg, the minimum step height is 4 µg. If the sample mass is 10 mg, this means that mass changes of the order of 0.4 per thousand can be measured. However, the accuracy with which such a small mass step can be measured also depends on the width of the step. An example of a very low mass loss step is shown in Figure 2. Comment: Here, under the accuracy of a measurement (related terms also used are trueness or bias), we mean the closeness of agreement between the mean value of a set of results and the accepted true value of the quantity. Accurate results of course require good precision. Precision is the closeness of agreement between independent measurement results on identical samples. It is a measure of the scatter or spread of the measurement results and is usually expressed as the standard deviation (often calculated as the relative standard deviation or the coefficient of variation). In the case of a purity determination, for example, it was known that the true value for the degree of purity of the sample was 98.4%. The result of a purity determination by DSC was 98.6 ± 0.1%. The deviation from the true value (i.e. the accuracy) is therefore 0.2%, and the precision 0.1%. Whether this is acceptable or not depends on your own requirements and is a matter to be considered in the validation. Measurement mode Once you have decided to use a particular measurement technique (here DSC, TGA, TMA or DMA), the next question concerns the measurement mode in which the instrument is to be operated. Table 2 summarizes various instrument-specific measurement modes and their use. Step 2: Sampling and sample preparation The most important points concerning sampling and sample preparation can be summarized as follows: Is the sample representative of the total amount? To obtain reliable results, you may have to measure several samples and compare the results. Sample processing: To obtain optimum thermoanalytical results, samples often have to be mechanically processed (e.g. Figure 2. Calcination of 28 µg calcium carbonate in mg aluminum oxide: The CaCO 3 loses CO 2 up to about 600 C (calcination). From stoichiometric considerations, a step of 12.3 µg is expected from the decomposition reaction of calcium carbonate. Measurement technique DSC Special measurement modes TM DSC (ADSC, IsoStep ) Use Separation of changes in c p from non-reversing events (vaporization, crystallization, chemical reactions) TGA MaxRes For automated optimized temperature resolution of neighboring mass changes TMA/DLTMA DMA Dilatation (low load on sample) Penetration (large load on sample) Tension Bending Tension Compression Shear Bending Mode to measure the coefficient of thermal expansion Particularly suitable for the analysis of thin films (glass transition, melting temperature, film thickness) For fibers and films, shrinkage Glass transition of filled materials and other stiff samples Above all for fibers and thin films Foams, elastomers Elastomers, most thermoplastics, powder, pastes Fiber-reinforced plastics, thermoplastics, thermosets Table 2. Special measurement modes for different TA techniques and their applications. cut, ground, polished, etc.). This affects samples both mechanically and thermally: It may in some cases lead to undesirable changes in a sample (e.g. with polymorphous substances). Thermal pretreatment: Annealing at a suitable temperature eliminates the thermal history of the sample. The information obtained then relates solely to the material under investigation. If the sample is not annealed, information about the conditions under which it was produced, i.e. its thermal history, can also be obtained. Sample geometry and sample size. The sample should be no larger than is nec- 3

4 essary to determine the result with the desired accuracy (i.e. as large as necessary but as small as possible). Insertion of the sample in the crucible (DSC, TGA) or in the instrument (TMA, DMA). See Table 3 for specific details. Typical questions about sample size How much sample do I need to determine a residue of 1% with an accuracy of ±1%? Answer In this case, the accuracy of the thermogravimetric measurement depends mainly on the reproducibility of the blank curve. For example, if the reproducibility is 10 µg in the temperature range considered, then these 10 µg represent 1% of the measured residue. The residue must therefore be at least 1 mg. Since the residue is about 1% of the original sample, a sample mass of at least 100 mg must be used. What is the sample thickness needed to determine a coefficient of thermal expansion (CTE) of 20 ppm/k over a temperature interval of 10 K with an accuracy of ±2%? Answer On heating the sample through 10 K (ΔT = 10 K), the thickness of the sample with a CTE of 20 ppm/k increases by 200 ppm (CTE = ΔL/ΔT 1/L 0 ). Assuming that the change in thickness of the sample in this temperature interval can be determined with an accuracy of about 20 nm, the change in sample thickness over this 10-K interval must be about 1 µm (ΔL = 1 µm). This change corresponds to 200 ppm of the original sample thickness (L 0 ). This must therefore be at least 5 mm. DSC/TGA TMA DMA Optimum contact of the sample with the crucible (thermal conductivity) Sample must not move within the crucible Sample must not react with the crucible Surfaces of the sample should ideally be flat and parallel For dilatometry, use a quartz glass disk between the sample and the probe to distribute the force exerted by the probe uniformly over the sample Table 3: Important aspects of sample preparation for different TA techniques. Step 3: Choosing the crucible (DSC and TGA only) The most important considerations regarding the choice of crucible are the Volume of the crucible (sample mass; with the TGA, also gas exchange). Heat capacity and thermal conductivity of crucible: this influences the resolution (separation of thermal events) and sensitivity of the DSC or SDTA signal). Crucible material: the sample must not react with the crucible material. Further aspects are presented in Figure 3. For a detailed discussion, see UserCom 5. Recommendations: For DSC, we recommend the use of the small 20-µL aluminum crucibles: these crucibles have the lowest heat capacity and give the best sen- Geometry of the sample must be known exactly Sample must be properly mounted in the clamp Possibly adjust force on the sample at the start temperature The thermocouple must not touch the sample or the furnace and should always be placed in the same position sitivity and time resolution. For the TGA, we recommend the 30-µL alumina crucible as standard crucible. If the temperature range of the measurement is below 600 C, and if a reaction with the sample is not expected, the 40-µL aluminum crucible can also be used for TGA. The advantages are the excellent thermal conductivity and a much better SDTA signal due to its low heat capacity. In addition, the crucible can be disposed of after use and does not have to be cleaned. Besides the standard crucibles, we also offer a variety of special crucibles manufactured from different materials (gold, platinum, copper, sapphire and Pyrex glass) for different conditions (normal, medium and high pressure). They are available in a number of different sizes. Figure 3. Factors influencing the choice of crucible. 4

5 New in our sales program Controlled relative humidity interface for the TGA/SDTA85x e A TGA/SDTA85x e module can be expanded with a controlled relative humidity interface and a relative humidity generator (VTI RH-200) to form a system that allows adsorption and desorption processes to be studied from room temperature to 90 C. A heated transfer line ensures that the humidified and temperature controlled air does not cool or condense before it reaches the furnace chamber. A small humidity sensor (Rotronic) is available as an option to accurately determine the relative humidity inside the furnace chamber. The relative humidity generator itself has a dew point sensor for exact humidity control. Figure 1. Cross-section of the controlled relative humidity interface. Left: the end of the heated transfer line. Above this is the furnace outlet through which the humidity sensor projects into the furnace chamber. Right: the crucible is just visible. The relative humidity influences the processibility, storage stability, usability and numerous other properties of materials. Adsorption behavior is therefore studied in very many application fields and industries. Measurements at defined levels of relative humidity are performed on widely different materials such as plastics, pharmaceuticals, chemicals, metals, and foodstuffs. Figure 2. The VTI RH-200 Relative Humidity Generator with the heated transfer line ( 5

6 Applications Model free kinetics Ni Jing Introduction The kinetics of chemical reactions can be easily determined from DSC or TGA measurements. The METTLER TOLEDO STAR e software offers three different software options: the classical n th order kinetics and two so-called model-free methods. The n th order kinetics approach assumes that the activation energy is constant throughout the entire reaction. The reaction rate, dα/dt, is given by the equation where α is the conversion, K 0 is the pre-exponential factor, E a the activation energy, R the gas constant, T the temperature and n the order of the reaction. An n th order reaction is therefore completely described by the parameters n, E a and K 0. This model is, however, at best only suitable for simple reactions. In more complex reactions in which several reaction steps proceed in parallel, or in which the reaction is not completely chemically controlled, n th order kinetics fails. In such cases, model free kinetics is an excellent alternative to describe the reaction kinetics. This study shows how model free kinetics can be used to evaluate DSC and TGA measurements and make predictions about the isothermal behavior of reactions (i.e. determine conversion as a function of time at a certain temperature, or the time needed at a given temperature to reach a certain conversion). Basic principles of model free kinetics Model free kinetics assumes that the activation energy does not remain constant during a reaction but changes. Furthermore, it assumes that the activation energy at a particular conversion is independent of the temperature program (the iso-conversion principle ). In the STAR e software, two different versions of model free kinetics have been implemented as software evaluation options: the standard model free kinetics (MFK, Model Free Kinetics) and the advanced model free kinetics (AMFK, Advanced Model Free Kinetics). The difference between these two approaches is to be found in the numerical algorithms that are used for the evaluation. Furthermore, with standard MFK, only dynamic heating curves can be evaluated. With AMFK, however, curves can be evaluated that have been measured with any temperature program, including isothermal measurements. Experimental details As raw data, model free kinetics requires at least three DSC or TGA curves measured with different temperature programs. First of all, it is always best to make a trial run in order to first get a general impression of the reaction profile, that is, the temperature range, reaction enthalpy and mass loss. Relatively slow heating rates (e.g. 5 K/min) should be used for materials that release large amounts of energy during the reaction (e.g. explosives). For safety reasons it is advisable to use high-pressure crucibles and low sample masses (< 5 mg). With other materials, heating rates of 10 or 20 K/min can be used for the trial run. For MFK, a minimum of three experiments must be performed at different heating rates. Four or five measurements are however recommended to improve the reliability of the calculated predictions. The heating rates used should increase in large steps such as 1, 2, 5, 10, and 20 K/min. The use of smaller steps, for example 5, 6, 7, 8, 9 K/min, reduces the accuracy of the results. At increasing heating rates, chemical reactions take place at higher temperatures. Accordingly, the temperature range must match the heating rate. Particularly when high heating rates are used, for example with curing reactions, (undesired) decomposition reactions frequently occur toward the end of the reaction under investigation. In such cases, lower heating rates should be used. As mentioned above, isothermal temperature programs can be used for AMFK. As a rule, the temperatures used are below the peak temperature. The reason for this is that above the peak temperature the reaction is very rapid, which makes correct evaluation of the curves difficult. A fundamental problem with isothermal DSC experiments is that at the beginning of the experiment the initial deflection and chemical reaction overlap. In this case, the DSC curve of a second run of the same sample (which should then no longer show any reaction) measured under otherwise identical conditions can be used as the baseline. Just as with MFK, at least three measurements must be made with different temperature programs for AMFK. Four measurements are however also recommended for AMFK; dynamic measurements can be combined with isothermal measurements. With AMFK, the rate at which the measurement values are recorded should be set to the maximum value of 10 measurement values per second. As far as sample preparation is concerned, it is important to make sure that all the samples are prepared in exactly the same 6

7 way. In particular, the sample mass should be about the same for all samples. And obviously the crucibles used must not react with the sample. Furthermore, it should be noted that the size of the hole in the crucible lid could influence reactions in which gases are liberated. Evaluations The typical revaluation procedure used for MFK and AMFK is illustrated in the example shown in Figure 1 (for DSC measurements). Step 1 (Fig. 1 above left): Display of DSC measurement curves. As already mentioned, the peak maximum shifts to higher temperatures at increasing heating rates. The reaction enthalpy however is independent of the heating rate and should therefore be constant within about ±10%. At first sight, this does not appear to be the case in the heat flow curves displayed (the area under the reaction peak is clearly larger at high heating rates than at low heating rates). The reason for this is that the reaction enthalpy is evaluated from the curve plotted as a function of time, whereas in Figure 1 the curves are shown as a function of temperature. It should also be noted that the measured reaction enthalpy depends strongly on the type of baseline used. As far as possible the same baseline type should be used for all measurements. If the reaction enthalpy changes in a systematic way, this could also indicate the overlap of different processes. The statements made here about DSC curves, are in the same sense valid for TGA curves; instead of the reaction peaks, mass loss steps occur that should agree within about ±5% with each other. Step 2 (Fig. 1, above right): The conversion curves are calculated from the measured curves using the DSC/Conversion or TG/Conversion routines. Here again, attention must be paid to the baseline, which should not intersect the measurement curve at any point. With TGA measurements, the problem of baselines is not so serious. Often, instead of the TGA curve, its first derivative, the so-called DTG curve, is used. In this case, the conversion curve Figure 1. Procedure for MFK and AMFK analysis: Step 1: Measurement curves at different heating rates (2.5, 5, 10, 20 K/min); the peak area (DSC measurements) or the step height (TGA measurements) should be constant to ±10% and ±5% respectively. Step 2: Calculation of the conversion curves. Step 3: Calculation of the activation energy as a function of the conversion. Step 4: Predictions using MFK or AMFK. The reaction investigated was the curing of an epoxy resin. is calculated in the same way as the DSC curve. With conversion curves, one should make sure that the curves for the different heating rates, in particular for low and high degrees of conversion do not intersect. If this occurs, it can usually be avoided by using other limits or another baseline. If a reaction exhibits several clearly separated reaction steps (e.g. the decomposition of calcium oxalate monohydrate), it is better to analyze each step individually. Such multi-step reactions make themselves apparent because in the conversion curves the conversion stagnates at a certain temperature (the conversion curve runs horizontally). Step 3 (Fig. 1, below right): The conversion-dependent activation energy is now calculated from the conversion curves. Since the numerical procedures are different, evaluations with MFK and AMFK yield different activation energy curves (see Fig. 2). The algorithms used for the calculation of predictions are different for MFK and AMFK. For this reason, activation energies calculated using AMFK cannot be used to calculate predictions using MFK. Although MFK and AMFK yield different activation energies, the predictions made by MFK and AMFK do not significantly differ (see Fig. 2, below right). Step 4 (Fig. 1, below left): The conversiondependent activation energy can now be used to make predictions, for example for the conversion of a reaction at a particular temperature as a function of time. To do this, the details of the desired temperatures and conversions are entered in the Kinetics/Conversion Parameter Settings or Kinetics/Iso-Conversion Parameter Settings menu. Using this information, the software then calculates the conversion curves at different temperatures as a function of time and presents the results graphically and in tabular form. If possible, the predictions obtained should then be checked by performing another experiment under appropriate conditions. An example is shown in Figure 3. Here, temperature programs were used consisting of three segments, namely a dynamic segment with isothermal segments before and after (see lower left diagram). Using AMFK, the conversion-dependent activa- 7

8 tion energy (above right) was calculated from the three measurement curves (above left) and then finally the predictions (below right) for the degree of conversion as a function of time at three different temperatures. A comparison of the predictions at 25 C with a corresponding measured curve shows very good agreement between the experiment and the AMFK predictions. Conclusions Model free kinetics allows even complex chemical reactions to be analyzed using just a few DSC or TGA measurements and without making any assumptions about a particular reaction model. The STAR e software has implemented two equivalent procedures (MFK and AMFK software options). From the mathematical point of view, the two approaches differ in the numerical algorithms used. The main difference for the user is that MFK can only evaluate heating curves, but AMFK can in addition evaluate isothermal or combinations of isothermal and dynamic measurements. Both procedures usually yield somewhat different activation energies. However, with regard to predictions, MFK and AMFK give very similar results. Figure 2. Comparison of MFK and AMFK: Left: Initial data for MFK and AMFK. Although MFK and AMFK clearly yield different activation energy curves (above right), the predictions made with the two procedures do not significantly differ (below right). Figure 3. DSC curves (above left) with temperature programs consisting of three segments (below left). In this case, kinetic analysis can only be performed with AMFK. An isothermal experiment was performed at 25 C (below right, red dashed curve) in order to check the predictions made by AMFK (below right, black curves). The figure shows that prediction and measurement agree very well. 8

9 Determination of the adsorption and desorption of moisture in pharmaceutical substances Dr. Matthias Wagner The behavior of substances with regard to drying, moisture uptake, and moisture content has become a topic of major importance because moisture can often have adverse effects on the properties of materials and products. The new TGA Sorption Analyzer System provides a very convenient method for studying such phenomena. The advantages of the technique are illustrated in the following article using amiloride hydrochloride dihydrate as an example. Introduction Recent surveys among TA users have confirmed that one of the current trends in modern thermal analysis is to control the gas atmosphere surrounding the sample. This can involve the use of reactive gases, the application of vacuum or pressure, or setting different levels of relative humidity (RH). In particular, investigations at defined relative humidity are becoming more and more important. This article describes the application of the new TGA Sorption Analyzer System to study a pharmaceutically active substance. An application study from the aroma/foodstuffs industry was published in UserCom 17 in 2003 [1]. The relative humidity influences the processibility, storage stability and usability of many materials such as pharmaceutical products (active ingredients, and fillers like lactose), plastics (nylon), construction materials (cement), metals (iron/ rust formation), explosives (dynamite) and foodstuffs (potato chips). This makes it necessary to investigate material properties at defined levels of relative humidity or to measure the humidity dependence of the material. A sample exposed to high relative humidity at room temperature tends to take up moisture. Products stored in contact with the open air may take up or lose moisture, depending on the relative humidity. Among other effects, the uptake of moisture can also influence mechanical properties, as anyone who has left potato chips in the open for a few days knows. In this case, moisture acts as a plasticizer and shifts the glass transition of the potato chips to below room temperature; the chips are then soft and no longer crisp [2]. The study of the behavior of materials as a function of relative humidity is particularly important with pharmaceutical preparations. This begins early on in the processing stage. A spray-dried powder can, for example, cause immense problems if it becomes moist and blocks the supply lines and dispensing devices, possibly leading to a shutdown of production. And if the finished medication takes up moisture due to inadequate packaging while in stock in the drug store, the shelf life of the product is obviously reduced. Furthermore, increased moisture content can also lead to major changes in the structural properties of the drug and reduce its bioavailability and therapeutic effect. One possible reason for such a change due to the uptake of moisture is the recrystallization of the active substance. This phenomenon is referred to as pseudopolymorphism, and the term pseudopolymorph refers to the compounds formed, which are known as hydrates or solvates. These are produced when the crystalline form changes as a result of the incorporation of water or solvent molecules into the crystal lattice. Stoichiometric hydrates (e.g. mono-, di-, tri-hydrates, etc.) are often stable compounds in which water is strongly bound as so-called water of crystallization. In contrast, moisture can also be merely adsorbed on the surface, in which case the water is only weakly bound. Hydrates and anhydrates (i.e. the anhydrous form that does not contain any water of crystallization) behave differently and can have different medicinal properties. It is important to identify and characterize pseudopolymorphs because they can be separately patented just like polymorphs [3]. This matter is usually investigated early on in the development phase. Experimental details TGA is a quantitative method and is therefore ideally suited to study the drying or moisture uptake of a substance, or to determine its moisture content. To reliably set up a defined relative humidity in the furnace chamber requires an instrument system consisting of a computer-controlled relative humidity generator, a heated transfer line to maintain the humidified air flow at a defined temperature, an interface on the TGA instrument, and if necessary, an optional humidity sensor inside the furnace chamber. The present study was carried out using a VTI RH-200 relative humidity generator and a Rotronic HygroClip SC04 humidity sensor with the TGA/SDTA851 e (large furnace). The system was used to investigate the influence of relative humidity on pure (> 98 %) amiloride hydrochloride dihydrate, a derivative of the diuretic drug amiloride (Fig.1). H 2 N Cl N NH 2 N Figure 1. Chemical structure of anhydrous amiloride (N-amidino-3,5-diamino-6-chloropyrazinecarboxamide). O N NH 2 NH 2 Diuretics are drugs that help to remove excess water from the body by increasing the amount that is excreted as urine. Diuretic drugs are used in the treatment of a variety of disorders including hypertension (high blood pressure) and conditions in which there is excessive accumulation of fluid in the body or body tissues (ascites, edema). Amiloride hydrochloride dihydrate (approx. 14 mg of fine yellow powder) was weighed into a 150-µL platinum crucible and inserted into the TGA instrument at 25 C. The sample was heated to 125 C at 5 K/min, held at this temperature for 30 minutes and then cooled down to 25 C 9

10 again at 5 K/min. A relative humidity (RH) of 5% was maintained during this phase of the experiment. The sample was then held isothermally at 25 C and exposed to increasing levels of RH from 5 to 95% in steps of 10% with first and final steps of 5%. Afterward, the RH was successively decreased stepwise down to 10% RH. The measurements were performed using a gas flow of 100 ml/min. Results Different standard procedures are available to characterize the loss on drying depending on the method described in the particular pharmacopeia. Figure 2 shows the mass loss curve of the test substance performed according to USP 26 guidelines (United States Pharmacopeia) [4]. In this method, a sample of approximately 10 mg is heated from room temperature at a heating rate of 10 K/min. The mass loss between room temperature and 200 C must not be less than 11.0% and must not be greater than 13%. Responsible for mass loss is the initial evaporation of a very small amount of weakly bound water up to about 80 C followed by the elimination of both molecules of water of crystallization between 80 and 140 C. The dehydration can be clearly seen in the SDTA curve as a two-step endothermic effect. Evaluation of the TGA curve yields a mass loss of 11.8%. From about 280 C onward, the substance melts with simultaneous decomposition. The measured melting point of 292 C (onset) is about 1.5 K less than the value given for the anhydrate in the Merck Index [5]. This is no doubt due to the formation of decomposition products that contaminate the substance and hence lower the observed melting point. The curves in Figure 3 show the initial dehydration at 125 C, the rehydration, and moisture adsorption and desorption at 25 C. The dotted curve depicts the temperature program, the red stepped curve the changes in relative humidity, and the continuous black curve the change in mass of the sample. The sample (initial mass mg) was first heated to 125 C, held at this temperature for 30 minutes and then cooled back down to 25 C. At the end of this period, the recorded mass was constant and the sample has lost approximately 1.7 mg or Figure 2. The TGA and simultaneously recorded SDTA curves of amiloride hydrochloride dihydrate. In the upper diagram, the loss on drying was evaluated according to USP 26 [4]. Figure 3. Uptake and release of moisture measured using a sample of amiloride hydrochloride dihydrate. The dihydrate can only be converted to the anhydrate at a higher temperature in a dry atmosphere. 11.5% mass. The relative humidity in the sample chamber was then increased in a series of steps of 5 and 10% over a period of 800 minutes. Sufficient time was allowed each time for the sample mass to equilibrate and stabilize. The first small step in the mass curve corresponds to uptake of moisture at 5% RH. The mass curve shows the increase in sample mass for each increase in relative humidity. The original mass of the sample is reached at about 50% RH; rehydration is then complete. Each further increase in sample mass now corresponds to the adsorption of moisture on the surface of the sample, that is, to the uptake of weakly bound water. Following this, the RH was then decreased in steps. A corresponding stepwise change (decrease) in mass was again observed. This experiment is recorded in the right part of Figure 3 and on an expanded scale in Figure 4. With adsorbed moisture, the 10

11 bound water from a pharmaceutical active substance such as amiloride hydrochloride dihydrate. The equipment including the heated transfer line and the optional humidity sensor enabled well-defined humidity conditions to be set up in the TGA furnace chamber. This coupled with the high sensitivity and stability of the TGA balance allowed even the smallest losses in mass to be resolved. All in all, the new TGA Sorption Analyzer System is a reliable and sensitive instrument that can be used to accurately determine the adsorption and desorption behavior of many different types of substances. Figure 4. Release of adsorbed (weakly bound) water from amiloride hydrochloride hydrate (expanded section of the curve shown in Fig. 3) mass attained at each value with decreasing RH corresponds to the value obtained at the same RH with increasing RH. The dihydrate however can only be converted to the anhydrate at a higher temperature in a dry atmosphere. The period in which the RH was decreased is displayed on an expanded scale in Figure 4. In particular, one should note the excellent stability of the balance signal, which allows the very small mass loss steps to be accurately measured. Summary The results clearly demonstrate that the TGA Sorption Analyzer System is able to reliably measure the drying process, the adsorption of strongly bound and weakly bound water, and desorption of weakly Literature [1] M. Schudel, J.B. Ubbink, Ch. Quellet, Measurement of dynamic water vapor sorption processes by modified TGA, METTLER TOLEDO TA UserCom 17, 1/2003, pp 7-9 [2] Y.H. Roos, Thermal Analysis, State Transitions and Food Quality, Thermal Analysis Seminar Presentation, Cork (Ireland), 2004 [3] J. Bernstein, Polymorphism in Molecular Crystals, Clarendon Press, Oxford 2002, Chapter 10 [4] United States Pharmacopeia, Vol. 26, 2003, p 114 [5] The Merck Index, 10 th edition, 1983, p 60 (No. 406) Investigation of the temperature stability of polymer additives and their decomposition products by TGA-MS and TGA-FTIR Dr. P. Fux, Ciba Specialty Chemicals Inc., Switzerland; Cyril Darribère, Mettler Toledo GmbH Introduction The processing of plastics and their additives puts high demands on temperature accuracy in the various processing steps as well as on temperature homogeneity within the material being processed. Because of the typically long processing times at high temperatures, there is always the possibility that products begin to decompose during production. In this particular case, an unpleasant smell was noticed during the processing of a plastic. A sample was analyzed by TGA followed by evolved gas analysis in order to measure the temperature range in which decomposition occurred and to identify the volatile compounds produced. The gaseous products were first analyzed using a mass spectrometer (MS) and then later with a Fourier transform infrared spectrometer (FTIR). These two techniques allow volatile decomposition compounds and gaseous elimination products to be characterized and in some cases identified. The following example describes the identification of ammonia as a decomposition product. It demonstrates the power of such combined methods and their importance for product development in an early phase of research. Experimental details TGA-MS measurements were performed using a large-furnace TGA/SDTA851 e connected online to an Inficon Thermostar mass spectrometer. The sample was heated 11

12 from 30 C to 400 C at 10 K/min. A 900-µl alumina crucible with pierced lid was used in order to measure as much sample as possible. The purge gas was argon (90 ml/min). In a preliminary experiment, the ion masses in the range 10 < m/z < 150 were analyzed. Based on the results of this experiment, ions of several different m/z values were selected and their intensities individually monitored in a second experiment (multiple ion detection, MID). The MS ion intensities recorded were m/z 15, 16, 17, 18, which are characteristic for water and ammonia. TGA-FTIR measurements were performed using a large-furnace TGA/SDTA851 e connected online to a Nicolet Nexus FTIR spectrometer. The purge gas was nitrogen (50 ml/min). Once again, the samples were heated in a 900-µL alumina crucible with pierced lid. Infrared (IR) spectra were continuously scanned at a resolution of 4 cm -1 and 16 successive scans repeatedly averaged. Figure 1. TGA mass loss curve and MS ion intensities. The peaks in the MS ion curves indicate the release of water (m/z 17 and 18) and ammonia (m/z 15, 16, 17). Discussion of the TGA-MS results Up until 400 C, three mass loss steps can be identified in the TGA curve (Fig. 1). Between 50 and 150 C, the m/z 17 and 18 ion intensities indicate the presence of traces of water. Evaluation of the TGA curve yields a mass loss of about 0.2%. During the second mass loss step (1.8%), peaks are observed in the m/z 15, 16, 17, 18 ion curves. These peaks indicate the release of water (m/z 17 and 18) as well as ammonia (m/z 15, 16, 17). The overlap of the fragment ions of water and ammonia make the interpretation of the mass spectra rather difficult. Figure 2. TGA, DTG and Gram-Schmidt (GS) curves, and chemigram for ammonia gas. The wavenumber range cm -1 was chosen for the ammonia chemigram. Discussion of the TGA-FTIR results The TGA results of the TGA-FTIR analysis are of course the same as those obtained using TGA-MS. The TGA-FTIR technique generates a large number of FTIR spectra. These are usually processed by calculating a so-called Gram-Schmidt (GS) curve. This is obtained through automatic integration of the infrared absorption of each spectrum over the entire wavenumber range. The Gram-Schmidt curve corresponds to the IR absorption intensity of the gaseous products arriving in the heated cell from the TGA as a function of sample temperature (or time) and is therefore directly related to the concentration of these decomposition products. Figure 2 shows the TGA and DTG (first derivative) curves together with the Gram-Schmidt curve. Two mass loss steps at 250 C and 390 C can be identified in the TGA curve (the mass-loss step at about 60 C due to moisture in the TGA-MS measurement cannot be seen with the ordinate scale used). In principle, the FTIR spectra corresponding to these temperatures could be recalled and interpreted in order to identify the gases that were in fact evolved. Spectral interpretation, however, requires considerable experience and presupposes ideas about the composition of the gas mixture. A simpler approach is to continuously integrate certain wavenumber ranges and display the signal as a function of time or temperature. The resulting curves are known as chemigrams. The wavenumber ranges chosen are those characteristic 12

13 for certain compounds or for particular functional groups in possible compounds. Chemigrams simplify the task of processing the complex IR spectra obtained from gas mixtures. In this example, the wavenumber range cm -1 (characteristic for gaseous ammonia) was chosen. The chemigram for this range is shown in Figure 2 and indicates that ammonia is very likely responsible for the three steps that occur at about 250 C, 320 C and 390 C. Conclusions TGA provides quantitative information on the mass changes of a sample as a function of temperature. The processes that occur can only be completely understood if qualitative information about the evolved decomposition products is available. This information can be obtained by connecting the TGA online to a gas analyzer such as a mass spectrometer (MS) or an FTIR spectrometer (FTIR). In this case, MS and FTIR analysis of the gaseous decomposition products indicated that emission of ammonia can be expected at sample temperatures above about 200 C. In addition, the analysis showed that ammonia is evolved in three steps. The results obtained from the FTIR measurements are consistent with those obtained from the MS analysis. However, based on the MS data alone, it would have been more difficult to identify the threestep emission process of ammonia. Care must however be exercised when using chemigrams because other compounds can have absorption bands in the same wavenumber region. In many cases, therefore, the precise nature of compounds evolved in decomposition processes can only be characterized through the combination of both MS and FTIR data. TGA-FTIR is particularly suitable for detecting small molecules that are difficult to interpret by MS. Note: For detailed information on evolved gas analysis and the TGA-MS and TGA-FTIR techniques, please consult the METTLER TOLEDO Thermal Analysis Collected Applications booklet entitled Evolved Gas Analysis (Order No ). Determination of glass transition temperatures of powder disks by TMA Matthée, K. and Dr. K. Knop, Institut für Pharmazeutische Technologie der Heinrich-Heine-Universität, Düsseldorf, Germany Introduction The glass transition temperature (T g ) of a polymer film can easily be determined by TMA. In contrast, the reliable determination of the T g of a powder sample by TMA is more difficult, especially compared with DSC. Powders and fine shavings can however be measured by first pressing disks of the material in a special die (Fig. 1). The disks can be pressed relatively easily at a defined and reproducible pressure by filling the special die described above with a known mass of powder and then placing it in a manually operated press whose spindle is connected to a torque wrench (Fig. 2). The pressure applied to the powder is varied by changing the torque setting. The influence of the applied pressure (torque) and the mass of the powder on the T g was investigated using TMA. The results were compared with T g measurements of films cast from the same polymer. Pharmacoat 606, a low viscosity hydroxypropyl methyl cellulose (HPMC) from the family of non-ionic water-soluble cellulose ethers, was chosen as a model substance. Such cellulose ethers are used in many different types of pharmaceutical applications, for example as film coatings for drug formulations. Experimental details and results Figure 1. The die used for preparing disks from powders and shavings (dimensions are in mm, barrel and plunger are made of tool quality steel). Figure 2. Manual press and die with torque wrench. Preparation of disks Disks were prepared using sample masses of 5, 10, 15, 20, 25 and 30 mg and torques of 2 and 10 Nm (ten disks each, n = 10). The thickness of each disk was measured using a micrometer and the values plotted as a function of sample and torque. As expected, the thickness increases linearly with increasing sample mass (Fig. 3). 13

14 Thickness in mm Sample mass in mg Figure 3. Thickness of disks as a function of sample mass. In addition, the effect of different values of the torque on the thickness of disks prepared with a sample mass of 30 mg was studied. As expected, the curve asymptotically approaches a limiting value corresponding to complete compaction of the material (Fig. 4). Sample mass HPMC [mg] Torque 2 Nm Torque 10 Nm T g1 T g2 Table 1. Glass transition (T g1 ) and penetration temperatures (T g2 ) as a function of torque (proportional to the applied pressure) and sample mass (mean value ± standard deviation, n = 5). T g1 T g ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 2.4 Thickness in mm Torque in Nm Figure 4. Thickness of disks as a function of torque (applied pressure). TMA measurements The experiments were performed using a METTLER TOLEDO TMA40 measuring cell and the results evaluated with the STAR e software. The glass transition was measured in the penetration mode (flat probe with small surface area, 1 mm 2 ) with a force of 0.2 N in the temperature range 0 to 250 C at a heating rate of 10 K/min. Nitrogen (200 ml/min) was used as purge gas. The disks were stored for at least 24 h over silica gel before measurement to ensure uniform moisture content. The first two series of experiments investigated the relationship between the glass transition temperature and the sample mass at constant torque values of 2 or 10 Nm. The mean value obtained (n = 5) was calculated. The onset temperature of the step was used for the evaluation (Table 1 and Fig. 5). The transitions T g1 (glass transition/softening, [1,2]) and T g2 (melting/decomposition, [1,2]) occur as distinct steps in all the measurements (Fig. 5). With masses Figure 5. Above: Typical TMA curve of a powder disk with the first step due to softening at the glass transition (T g1 ), and the second step due to complete penetration of the probe (T g2 ). Below: Typical TMA curve of a cast film with T g0 as additional softening step. Torque [Nm] after preparation T g1 T g2 after storing for 3 months T g1 T g ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 1.5 Table 2. Glass transition (T g1 ) and penetration temperatures (T g2 ) as a function of torque (proportional to the applied pressure) and storage time (mean value ± standard deviation, n = 5). 14

15 of 15, 20, 25 and 30 mg the transitions occur reproducibly at about 158 and 230 C respectively, whereas with 5 and 10 mg the values are somewhat lower. These two disks were in fact difficult to handle irrespective of the torque used to press them. The disks broke apart easily on handling with tweezers. Disks prepared using a sample mass of 30 mg with both torque values were most easy to handle. This mass was therefore used as standard for the next series of experiments in which different pressures were applied with torque values of 2 to 10 Nm. The mean (n = 5) values were again determined from five measurements (Table 2). The results showed that pressure had no significant influence of the glass transition Film thickness [mm] T g0 temperature except that with a torque 2 Nm the glass transition was slightly lower. In another experiment, the disks were stored over silica gel for three months and then measured again to determine the possible influence of storage time on the glass transition temperature. A comparison of the values obtained from the freshly made disks with those from disks stored for three months showed that there was no significant difference in the measured glass transition temperatures. The values obtained for the pressed powders were compared with the values measured using films of pure HPMC. These films were prepared by dissolving HPMC in water, casting films and then drying them at 60 C in a drying oven. The results are summarized in Table 3 and Figure 5. T g1 T g Mean value 90.1 ± ± ± 3.8 Table 3. Glass transition (T g1 ) and penetration temperatures (T g2 ) of cast films of different thickness. The film thickness varied markedly across any given film. Instead of two clear steps, sometimes three steps occurred that show a rather gradual change in slope. The T g0 (first softening step before the glass transition) was not observed in every measurement. The values obtained lie between 68.6 C and C, which gives a mean value of 90.1 C. The T g1 and T g2 transitions yield a mean value of C and C respectively. The values do not seem to depend on the film thickness. Conclusions In summary, one can say that the determination of the glass transition temperature of HPMC using powder disks is a good alternative to using cast films. Both methods yield similar values as far as the glass transition temperatures (T g1 ) and the melting or decomposition temperatures (T g2 ) are concerned. It is however easier to press disks of constant thickness than it is to cast uniform films. The reproducibility of the values obtained with the HPMC disks is appreciably better. The applied pressure and sample masses used to prepare powder disks can be varied within wide limits, without any significant effect on the observed glass transition temperature values. Literature [1] Okhamafe A. O., York P.; Journal of Pharmaceutical Sciences 77, (1988) [2] Johnson K. et al.; International Journal of Pharmaceutics 73, (1991) 15

16 Phase transitions of lipids and liposomes Dr. Sophia Hatziantoniou, Prof. Dr. Costas Demetzos, University of Athens, Greece; Dr. Matthias Wagner The phase transition temperatures (T m ) and transition enthalpies (ΔH) of bilayer membranes formed from dipalmitoylphosphatidylcholine (DPPC) were measured by differential scanning calorimetry (DSC). These bilayers served as model membranes. The effect of several different pharmaceutically active flavonoids on membrane fluidity was also studied. The DSC results revealed close-lying thermal effects that depended on the structure of the flavonoid. A relationship between the flavonoid interaction with the model membranes and its ability to change the ordered lipid structure of the DPPC was observed. Introduction Phospholipid molecules consist of a polar head linked to two long acyl groups (e.g. DPPC, Fig. 1. 1). When dispersed in water, lipids align themselves with their polar heads toward the water to form micelles (clusters), liposomes (microscopic concentric spheres or vesicles) or other structures. Liposomes are produced when some phospholipids aggregate to form double layers of molecules and then close to form bilayer membranes. Both unilamellar (SUVs, small unilamellar vesicles) and multilamellar (MLVs, multilamellar vesicles) structures can be produced into which drug molecules can be incorporated. This has led to the use of liposomes as drug delivery systems for medical applications (via intravenous injection). Liposomes resemble the membrane of a living cell and are used as models for cell membranes. Flavonoids are polyphenolic compounds with a diphenylpropane structure (C 6 -C 3 -C 6 ) that exhibit different biological effects, e.g. anti-inflammatory, bactericidal, antiviral and fungicidal action, tumor growth reduction, protective effect for the liver, antitumor- and antioxidation activity [1]. In this study, several flavonoids have been incorporated into liposomes. These compounds are able to penetrate cell walls and their ability to alleviate cardiac diseases has been described in the literature [2]. The Figure 1. Chemical structures: 1 DPPC, 2 quercetin, 3 rutin, 4 isoscutellarein, and 5 the glycoside of isoscutellarein. interaction of flavonoids with liposomes has been investigated to gain a better understanding of their interaction with cell membranes and to tailor-make liposomes as controlled-release drug delivery systems. The DSC heating curves of lipid bilayers show two phase transitions: first a small peak, the so-called pretransition, followed by a second (main) transition. DPPC bilayers exist in the gel phase (L β ) at temperatures below 35 C, whereas above 42 C they are present in the liquid crystalline phase (L α ). The pretransition corresponds to a reorganization of individual lipid molecules in the lipid bilayer. Between 35 and 42 C, the phospholipid bilayer is in the P β or so-called rippled phase. Following the pretransition at 35 C, several conformational changes occur in the lipid molecules as well as changes in the geometry of the bilayers [5]. The most notable change is transgauche isomerization, which changes the acyl chain conformation (see Fig. 2). The number of gauche conformational isomers influences the fluidity. This depends not only on the temperature but also on the penetration of active substance molecules 16

17 trasound in an ice bath for two periods of 20 minutes with a pause of 5 minutes. The resulting vesicles were then held at room temperature for 30 minutes. Unbound flavonoids were removed at room temperature by gel filtration chromatography using a saturated Sephadex (G-75) column equilibrated with HPLC grade water (ph 5.6). T < T p gel T p < T < T m rippled phase T > T m liquid crystalline Figure 2. Schematic presentation of the alignment of acyl chains in a saturated phospholipid such as DPPC to form a double layer of molecules: gel in quasi-crystalline state (T<T p ), rippled phase (T p <T<T m ), and in the liquid crystalline state (T>T m ). present, which changes the lipid environment. With a given polar head, the transition temperature depends on the type and length of the acyl chains the longer the chain, the higher the temperature. Pure DPPC bilayers exhibit characteristic transitions with low enthalpy changes and a sharp main transition (T m ), both of which occur at the expected transition temperatures of 35.1 and 41.3 C. DSC is a useful technique to study the thermal effects of additives in bilayer membranes and has already been used in the past to study flavonoid-biomembrane interactions [3]. Materials and methods Liposome preparation and flavonoid incorporation Pure liposomes Liposomes were prepared using the thin film hydration method [4]. DPPC (66.06 mg) was weighed into a round-bottom flask and dissolved in chloroform. The solvent was distilled off at -40 C leaving a thin film on the inner surface of the flask. The lipid film was hydrated with HPLC grade deionized water (3 ml) at 51 C for one hour, which led to the formation of MLVs (multilamellar vesicles). A 300-µl portion of this material was used for the DSC measurements. The remainder of the sample was subjected to two cycles, each lasting 20 min with an interval of 5 min in a probe sonicator in an ice bath. This resulted in the formation of SUVs (small unilamellar vesicles). Some 500 µl of this was lyophilized and used for other DSC measurements. The MLVs and SUVs were prepared according to the method described above and afterward freeze-dried. Hydration of the stored powder regenerated the liposome suspensions. Liposome-flavonoid mixtures For the flavonoid-incorporation study, the lipid film was also prepared by dissolving DPPC (0.090 µmoles) in chloroform. Quercetin (Fig. 1. 2) (0.020 µmoles) was dissolved in ethanol. Rutin (Fig. 1. 3) (0.020 µmoles), isoscutellarein (Fig. 1. 4) (0.020 µmoles) and isoscutellarein glycoside (Fig. 1. 5) (0.020 µmoles) were dissolved in methanol. Each flavonoid solution was then added to a lipid solution and the organic solvent distilled off under reduced pressure. The samples were stored overnight in a desiccator. The lipid film was hydrated through the addition of HPLC grade water (3 ml) and the MLVs produced by vigorously shaking the suspensions in a water bath above T m, the transition temperature from gel to the liquid crystalline phase of the lipid (41 C), and stirring for 75 minutes. The resulting liposome suspensions consisting of MLVs were frozen in a dry ice bath (solid CO 2 /n-butanol) and thawed by heating to 41 C in a water bath. The size of the vesicles was reduced by performing a series of 15 freezing/thawing cycles: The large unilamellar vesicles (LUVs) that resulted were subjected to ul- DSC measurements The MLV and SUV samples were measured using a METTLER TOLEDO DSC822 e equipped with an HSS7 sensor and a Julabo intracooler. Pure MLV and SUV samples were prepared by dispersing the powdered liposomes in the appropriate quantities of bidistilled water. DPPC and flavonoid mixtures with the following dry masses DPPC/2 (5.7 mg), DPPC/4 (5.9 mg), DPPC/3 (5.9 mg) and DPPC/5 (7.3 mg) were weighed into 40-µL aluminum crucibles together with the same mass of deionized water. The crucibles were then hermetically sealed with an aluminum lid and placed in the DSC. An aluminum crucible with lid containing the same amount of water was used as the reference. The sample was held isothermally for 5 minutes at 0 C, heated to 70 C at 5 K/min and then cooled to 0 C at 5 K/min. This was repeated with a further 3 heating and 2 cooling segments with the same parameters. The DSC was calibrated for temperature and enthalpy with 4-nitrotoluene (melting temperature 51.4 C) and indium as reference substances. The enthalpies and characteristic temperatures were evaluated using the METTLER TOLEDO STAR e software. Results and discussion Phase transitions of pure liposomes Figure 3 shows the phase transition measurements of pure MLVs (above) and pure SUVs (below). The concentrations of the sample solutions were mol% for the MLVs and 0.029% for the SUVs. The SUVs can be distinguished from the MLVs by the peak temperature of the pretransition. The figure also shows that the high sensitivity of the DSC s HSS7 sensor allows the very weak pretransition at 35 C to be detected even when the concentration of the sample solution was very low. 17

18 Thermal effects of flavonoids in phospholipid bilayers The DSC technique was used to compare the thermotropic properties of the flavonoids with structures 2, 3, 4, 5 (Fig. 1.) in DPPC bilayer membranes. The DSC curves were evaluated for T onset, T m and ΔH. As can be seen in Table 1, measurements of fully hydrated DPPC bilayers with incorporated flavonoids yielded broad phase transitions with enthalpies of 42.5 J/g (isoscutellarein glycoside) to 35.1 J/g (rutin), and phase transition temperatures for the main transition, T m, between 41.9 C (quercetin) and 40.2 C (isoscutellarein glycoside). Figure 3. Phase transitions of MLV and SUV liposomes; the insert shows the pretransition of SUV on an expanded ordinate scale. Corresponding flavonoid in Figure 1 Sample T onset ( C) T m ( C) ΔH (J/g) pure DPPC DPPC/QUERCETIN DPPC/RUTIN DPPC/ISOSCUTELLAREIN DPPC/ISOSCUTELLAREIN GLYCOSIDE Table 1. Transition (T m ) and onset temperatures (T onset ), transition enthalpies (ΔH) of the measured MLV preparations. Each sample had an original molar DPPC/flavonoid ratio of 9:2. Table 1 summarizes the characteristic values of T onset, T m and ΔH for the flavonoid interaction with DPPC model bilayers. The results indicate that all preparations have about the same phase transition temperatures T m. The ΔH values show that flavonoid 3 has the greatest effect with regard to lowering ΔH. This can be explained by assuming that the lipophilic part of flavonoid 3 penetrates more deeply into the lipid bilayer than for example flavonoid 5, which interacts less with phospholipids. This was also predicted using data obtained from molecular modeling and computer calculations [5]. The observations furthermore agree with incorporation rates measured for both glycosides that showed greater efficiency for flavonoid 3 (71%) than for flavonoid 5 (36%). The observed interference caused by flavonoids 2, 3 and 4 is also an interesting result. The effect of these flavonoids was to lower ΔH, even down to 35.1 J/g for flavonoid 3 (ΔH for pure DPPC bilayers is 45 J/g, Table 1), whereas flavonoid 5 only had a marginal influence. Figure 4 shows several typical measurement curves. Figure 4. Typical DSC curves of DPPC/flavonoid mixtures. Summary DSC studies on the interaction of flavonoids with model DPPC membranes and on the relationship of the flavonoid structure to the lipid environment showed that the efficiency of incorporation depends on the structure of the flavonoid. The results indicate that the presence of a sugar group, 18

19 or even just a number of hydroxyl groups in different positions in the flavonoid structure, plays a role in the incorporation of flavonoids in liposomes and the interaction of flavonoids with DPPC membranes. The liposome formulations with flavonoids were also tested for their activity against three human cancer cell lines. In the case of quercetin, the formulation showed a lower degree of growth inhibition compared with the free quercetin; improved growth inhibition was however observed with the isoscutellarein formulation [5]. Literature [1] Narajana K.R., Reddy M.S., Chaluvadi M.R., Krishna D.R. (2001) Bioflavonoids classification, pharmacological, biochemical effects and therapeutic potential. Indian Journal of Pharmacology 33: [2] Gordon M. H., Roedig-Penman A. (1998) Antioxidant activity of quercetin and myricetin in liposomes CPL 97: [3] Saija A., Bonina F., Trombetta D., Tomaino A., Montenegro L., Smeriglio P., Castelli F. (1995) Flavonoid-biomembrane interactions: a calorimetric study on dipalmitoylphosphatidylcholine vesicles. International Journal of Pharmaceutics 124: 1-8. [4] Mayer L.D., Ti L.C.L., Bally M.B., Mitilenes G.N., Ginsberg R.S., Cullis P.R. (1990) Characterization of liposomal systems containing doxorubicin entrapped in response to ph gradients. Biochimica et Biophysica Acta 1025: [5] Goniotaki M., Hatziantoniou S., Dimas K., Wagner M., Demetzos C. (2004) Encapsulation of naturally occuring flavonoids into liposomes. Studies on their physicochemical properties andon their biological activity against human cancer cell lines. Journal of Pharmacy and Pharmacology, (accepted). Dates Exhibitions, Conferences and Seminars Veranstaltungen, Konferenzen und Seminare IRC 2005 June 7-9, 2005 MECC Maastricht (Netherlands) VI Congreso Nacional de Materiales Compuestos June 27-29, 2005 Valencia (Spain) MEDICTA 2005 July 2-6, 2005 Thessaloniki (Greece) Khimia 2005 September 5-9, 2005 Moscow (Russia) STK September 8-9, 2005 Basel (Switzerland) XVII Convegno Nazionale AIM September 11-15, 2005 Naples (Italy) NATAS 2005 September 17-21, 2005 Universal City (California, USA) Expoquimia 2005 November 14-18, 2005 Barcelona (Spain) TA Customer Courses and Seminars in Switzerland Information and Course Registration: TA-Kundenkurse und Seminare in der Schweiz Auskunft und Anmeldung bei: Frau Esther Andreato, Mettler-Toledo, Analytical, Schwerzenbach, Tel: , Fax: , esther.andreato@mt.com Courses / Kurse SW Basic/TMA/DMA Basic (Deutsch) 19. September, 2005 SW Basic/TMA/DMA Basic (English) September 26, 2005 DMA Advanced/TGA (Deutsch) 20. September, 2005 DMA Advanced/TGA (English) September 27, 2005 TGA-MS/DSC Basic (Deutsch) 21. September, 2005 TGA-MS/DSC Basic (English) September 28, 2005 DSC Advanced/TGA-FTIR (Deutsch) 22. September, 2005 DSC Advanced/TGA-FTIR (English) September 29, 2005 SW Advanced (Deutsch) 23. September, 2005 SW Advanced (English) September 30, 2005 TA-Kundenkurse und Seminare in Deutschland Für nähere Informationen wenden Sie sich bitte an: Frau Petra Fehl, Mettler-Toledo GmbH, Giessen, Tel: , petra.fehl@mt.com Anwenderworkshop DSC 11./ Giessen Fachseminar «Thermoanalytische und rheologische Messmethoden für die Materialcharakterisierung in Qualitätssicherung, Produktentwicklung sowie F&E» September 2005 Weitere Informationen zu diesen Veranstaltungen finden Sie unter: Corsi e Seminari di Analisi Termica in Italia Per ulteriori informazioni Vi preghiamo di contattare: Simona Ferrari, Mettler-Toledo S.p.A., Novate Milanese, Tel: , Fax: , simona.ferrari@mt.com DSC base 7 Giugno Settembre 2005 Novate Milanese DSC avanzato 8 Giugno Settembre 2005 Novate Milanese TGA 9 Giugno Settembre 2005 Novate Milanese TMA 10 Giugno Settembre 2005 Novate Milanese 19

20 Cours et séminaires d Analyse Thermique en France Renseignements et inscriptions par Christine Fauvarque, Mettler-Toledo S.A., Av. de la pépinière, Viroflay Cedex, Tél: , Fax: DMA/TMA et le logiciel STAR e 3 octobre 2005 Viroflay (France) TGA et logiciel STAR e 6 octobre 2005 Viroflay (France) DSC les bases et logiciel STAR e 4 octobre 2005 Viroflay (France) Logiciel STAR e 7 octobre 2005 Viroflay (France) DSC avancé et logiciel STAR e 5 octobre 2005 Viroflay (France) La cinétique appliquée à l analyse thermique Prof. Nicolas SBIRRAZZUOLI 22 septembre 2005 Viroflay (France) Durabilité des matériaux polymères: applications de la DSC Prof. Mohamed BABA 13 octobre 2005 Viroflay (France) L analyse thermique dans le domaine de la sécurité des procédés Prof. Francis STOESSEL 20 et 21 octobre 2005 Viroflay (France) DSC à température modulée : théorie, applications, dangers Prof. Jean-Pierre GROLIER 27 octobre 2005 Viroflay (France) L analyse thermique dans les domaines pharmaceutique, cosmétique et biologique Prof. Jean-Marie LETOFFE 17 novembre 2005 Viroflay (France) DSC, thermo-microscopie, cristallisation en solution. Application à l industrie pétrolière Prof. Jean-Marie LETOFFE 18 novembre 2005 Viroflay (France) Méthodes thermiques d analyse (DSC, TGA et TMA) appliquée à la caractérisation des polymères Prof. Jean-Marie LETOFFE 23 et 24 novembre 2005 Viroflay (France) TA Customer Courses Benelux For more details of training courses and seminars please contact: Netherlands: Hay Berden, Tel: or hay.berden@mt.com Belgium: Annick Van Hemelrijck, Tél: or annick.vanhemelrijck@mt.com DMA User Forum Benelux June 10, 2005 Lovaine La Neuve (B) Introductory Seminar New Products September 28, 2005 Tiel (NL) Cursos y Seminarios de TA en España Para detalles acerca de los cursos y seminarios, por favor, contacte con: Francesc Catala, Mettler-Toledo S.A.E., Tel: , francesc.catala@mt.com Seminario de aplicationes TA: Octubre 18, 2005 Barcelona Seminario de aplicaciones TA: Octubre 25, 2005 Madrid Seminario para usuarios STAR e : Octubre 19, 2005 Barcelona Seminario para usuarios STAR e : Octubre 26, 2005 Madrid TA Customer Courses and Seminars in the UK For details of training courses and seminars, please contact: Rod Bottom, Mettler-Toledo Ltd, Leicester, Tel: , Fax: , rod.bottom@mt.com DSC BASIC 25 October, 2005 Leicester DSC BASIC 26 October, 2005 Leicester TA Customer Courses and Seminars in the USA and Canada For information, please contact: Tom Basalik at , Fax: or by to: tom.basalik@mt.com TA User Training October 11-12, 2005 Columbus (Ohio) TA Customer Courses and Seminars in Japan For details of training courses and seminars please contact: Yoshiaki Suzuki at Mettler-Toledo Japan, Tel: , Fax: , or by to: yoshiaki.suzuki@mt.com STAR e Infoday Seminar June15, 2005 Tokyo Technical Center STAR e Infoday Seminar June17, 2005 Osaka Branch Editorial team Dr. J. Schawe, Dr. R. Riesen, J. Widmann, Dr. M. Schubnell, Dr. M. Wagner, Dr. D. P. May, Ni Jing, Marco Zappa, Urs Jörimann, Physicist Chem. Engineer Chem. Engineer Physicist Chemist Chemist Chemist Material Scientist Electr. Engineer METTLER-TOLEDO GmbH, Analytical, Sonnenbergstrasse 74, CH-8603 Schwerzenbach, Switzerland Kontakt: urs.joerimann@mt.com, Tel: , Fax: Internet: 05/2005 Mettler-Toledo GmbH, ME , Printed in Switzerland 20