sample="quota" bates="505876974" isource="rjr" decade="1980" class="ni" date="19850311" Interoffice Memorandum To: Mr. Jerry Lawson From: D. C. Kay Subject: Effect of Physical and Chemical Properties of Substrates on Substrate Performance (A Non-Experimental Approach) Date: March 11, 1985 I. Introduction An investigation of the influence of various chemical and physical substrate properties on the performance of substrates in a TGA device has been carried out. This investigation involved searching the literature to learn what is known about the properties of glycerol, the interactions of a liquid such as glycerol with a solid surface an d about various substrate materials which we are considering using in a TGA model. This knowledge was then applied to the TGA system to predict what characteristics would be desirable to enable a substrate to meet the requirements of a substrate needed for a TGA device. The requirements of a substrate for a TGA device have been defined. They are: 1. to hold a humectant which forms an aerosol when heated gas is passed over the adsorbed or absorbed humectant. 2. to hold the humectant tightly at conventional storage warehouse temperature so that the humectant does not migrate over time to other parts of the TGA device such as the fuel source and paper wrappings. 3. to be able to release enough humectant at the temperatures in the capsule so that a quantity of aerosol, similar to that of conventional cigarettes , is produced. 4. not to catalyze breakdown of the humectant , to which produce s offtasting aerosol. In recent years, much information has been reported in the literature concerning materials which we are considering as candidates for a TGA substrate, i.e. graphites, activated carbons, and metal oxides (alumina, silica and clays). The impetus for gaining this knowledge is the wide use of these materials as catalysts, catalyst supports and as adsorbents in purification systems. Although these substrate materials are used in the similar applications, there are many differences in the structure of these materials and the interaction which can occur between an adsorbing solid substrate (adsorbent) and the humectant which will interact with the solid surface (adsorbate). For a particular system, one type of adsorbent will react with the humectant, differently than another type of adsorbent. Also, within a particular class of adsorbents, i.e. activated carbons, graphites and metal oxides, wide differences in adsorption and desorption performance can be seen. Currently at RJR, we are using B-3 (glycerol) as a humectant, so we will assume that the humectant is glycerol. In this report, the chemical and physical properties of glycerol will be considered first, so that later in the report, we can relate the properties of glycerol alone to the properties of glycerol adsorbed on a substrate surface. After the properties of glycerol are considered, adsorption phenomenon common to these substrates will be reviewed and then finally, the influence of the chemical and physical properties of these substrates, sectioned into the following classes will be considered. The four classes are: 1. Graphites 2. Activated Carbons 3. Metal Oxides (In particular, alumina) 4. Clays (in particular vermiculite) These classes were selected because of the differences in physical and chemical properties between the classes and because of the possibility of using these type of materials in the final TGA product. II. Glycerol In the simplest form, the structure of glycerol can be written: Glycerol has been selected as a humectant for the TGA system because glycerol is inexpensive and because much is known about glycerol and the formation of aerosols from glycerol. Glycerol forms an aerosol which is almost odorless, and which visually resembles the aerosol released when a conventional cigarette burns. When considering glycerol as a humectant, the structure, physical and chemical properties of the humectant must be considered. These properties are outlined below. A. Structure of Glycerol It is important to consider the detailed chemical structure of glycerol when considering its adsorption or absorption by porous substrates. Structural information is needed to estimate the smallest pores into which glycerol could penetrate. The bond angles and bond distances for glycerol have been determined by Sundaralingam et al. (1965) and are given below. The actual size of glycerol cannot be determined simply by looking at bond angles and bond distances, because a molecule such as glycerol could preferentially be made in many conformations, i.e. various arrangements in three dimensional space arrived at by rotation about single bonds. The energy of the conformations would be influenced by intramolecular and intermolecular hydrogen bonding. A diagram showing some of the potential conformations for 1, 2 propanediol are shown in Figure 1. Van Koningsveld (1968) determined the structure of glycerol by using x-ray diffraction and determined that the preferred conformation (lowest energy) was the structure shown in Figure 2. Van Koningsveld (1968) also determined the crystal structure of glycerol. From the crystal structure, the dimensions of the glycerol molecule can be obtained. The cell dimensions are: A systematic study which was carried out by Kiselev and coworkers (1959, 1961a, b, 1966a, b, 1968) found that the effect of pore narrowing: 1. Had the greatest influence on ontthe adsorption of large molecules such as the longer chain n-alkanes ; the effect becomes apparent for such hydrocarbons as n-, iso, and cyclopentane, n-butane, and benzene when the pore neck diameter decreases to less than 60-40 A. 2. Is much less pronounced with the small molecules of nitrogen, whose adsorption due to dispersion forceas is supplemented by interaction between their pi electrons and the strongly patronized hydrogens of the silanol groups. 3. Is also much less pronounced with methanol whose adsorption is due partly to hydrogen bonding, which is not sensitive to pore narrowing: in fact, the effect with nitrogen and methanol becomes appreciable only when the pore neck diameter is below 30A. 4. Is least with water, which is sorbed mainly through hydrogen bonding. Based on the studies by Kiselev and coworkers, it is unreasonable to believe that the non-specific interactions of the smaller pores have much effect on the offtaste that is detected when smoking a TGA device. Glycerol would interact with the surface of the substrate through specific interactions, i.e. hydrogen bonding. (The sites for hydrogen bonding on the individual classes of substrates are given in the later sections of this report where each class of substrate is discussed individually). The van der Waals force contributions to the adsor ption in the small pores would not be very significant. Also because we do not desorb all of the glycerol from the TGA substrate when the device is used, it is likely that most of the ultramicroscopes are not even being emptied. The effect of pores on the performance of a TGA substrate is probably more related to surface area. For materials with constant pore volume, but different surface areas, the materials with the greater surface area will have smaller pores. Thus, the fact that the material has smaller pores means that more adsorbate can interact with the surface of the material. If indeed the surface of the material is catalyzing the breakdown of glycerol more offtaste products will be detected in the aerosol. Also, glycerol will not migrate as readily from the materials which has the small pores, and thus the larger surface area. As mentioned above in the section describing the influence on surface area on migration, the more that the glycerol interacts with the substrate surface, the less there is the tendency for the glycerol to migrate. The fact that PG-60 a material with little surface area (<1m2/g) gives no offtaste and an activated carbon with a surface area of 1000 m2/g gives a fair amount of offtaste, supports this above proposal. Also the glycerol migrates from the pg-50 but the tendency for the material to migrate from an activated carbon is much less. Thus, to summarize, there are many physical substate properties that contribute to offtaste. The amount of surface area and thus surface active sites which can interact with the sglycerol, seems to be very important. The effect of the potential in the small pores seems to be small, however the fact that when the pores are smaller, more adsorbate can interact with the substrate surface (assuming constant pore volume) seems to be a more reasonable explanation for the effect of pore size. IV. Solid Substrates - General Chemical Properties A. The effect of pH One of the chemical properties which the substrates have in common, is that their surface can either be acidic or basic. The pH of a substrate is the pH of a water extract obtained under prescribed conditions. The effect of the pH of the substrate on the adsorbed glycerol is essential the effect of the extractable hydrogen or hydroxy ions. If these ions are readily extractable, they may catalyze breakdown of the glycerol in the same way the acids and bases were shown to catalyze reactions of glycerol in the first section of the report. pH may be also giving a measure of the surface oxide density. If the surface sites are acidic, a more acid extract would indicate that there are a greater number of oxide surface sites. This would indicate a greater number of sites for the adsorption of glycerol. B. Physical Characteristics Unlike the other substrates which we are considering as substrate candidates, vermiculite does not possess pores to achieve its highly absor ptive properties. Vermiculite has a high ability to adsorb glycerol, because glycerol not only adsorbs on the outer surface of the vermiculite but adsorbs in the interlayers as well. The external surfaces parallel to the silica and alumina layers and the interlayer surfaces of the vermiculite can be thought of as a plane possessing a net negative charge balanced by cations as shown: C. Chemistry 1. General Chemistry A general chemical formula for natural vermiculite is: formulaname The Mg2+ and Ca2+ ions serve to balance the charge difficiency due to lattice substitutions and are largely exchangeable ions. 2. Surface Chemistry Because the sites giving rise to the negative charge are due to substitutions within the lattice and not to broken bonds, the interaction between glycerol and the surface oxygen groups can be due to hydrogen bonding. formulaname Because of the three glycerol hydroxy groups and the possibility to interact strongly with the glycerol surface. It is likely that the glycerol is adsorbed on each face of the interlayers. The charged planar surface sites of vermiculite which account for about 80% of the adsorption properties of vermiculite are not pH dependent. However, pH dependent sites can be found on the edge of the vermiculite crystals. These sites are pH dependent as follows: formulaname D. Interaction with Glycerol van Olphen (1977) reports that many organic compounds with a dipole character (which includes glycerol) are adsorbed on the layer surfaces and probably also on the edge surfaces of a clay in analogy with the behavior of a clay with water. As is the case with the adsorption of water on clays, it is not known to what extent the polar groups of the organic molecule associate with the counterions of the clay and to what extent they are hydrogen-bonded to the oxygen surfaces. According to van Olphen infrared data indicates that the hydrogen-bonding interaction is the significant interaction. Eltantawy (1977) studied the adsorption o f glycerol on tovermiculite. In this study, it was found that the adsorption of glycerol was dependent on the counterion in the interlayer and that only one layer of glycerol adsorbed in the interlayer. No explanation of the types of interactions which take place between glycerol and vermiculite was given. Vermiculite as a TGA Substrate Vermiculite is a suitable candidate for a TGA substrate. The physical and chemical nature of this material are much different that the other materials which we are considering for use as TGA substrates. Vermiculite does not possess pores in the same sense a activated carbon or alumina, but adsorbs adsorbates such as glycerol in the interlayers of the clay structure. Glycerol can probably desorb more readily from the interlayers than glycerol can from pores. There should be a statistical advantage to the desorption of glycerol from vermiculite, because glycerol can desorb for the f our sides of the interlayers whereas for activated cartbon pores, glycerol can only desorb from the pore opening. In the interlayers, there is also little steric hindrance to desorption of the glycerol because other glycerol molecules in the interlayers act as pillars to allow the easy desor ption of glycerol. Although the exact nature of the bonding of the glycerol in the interlayers is not known, the interaction can only occur between oxides and not hydroxyl groups, because there are not hydroxyl groups in the interlayers. VIII. Summary Much has been learned about the reactivity of glycerol, adsorption of glycerol on solid surfaces, the chemistry of the substrate surfaces and the interaction of glycerol with the various substrates. Generally, it has been found that there are two main substrate properties that dictate how the substrate will interact with the glycerol. These two are the ease in glycerol reaching and being released from the active surface sites, which is related to pore size and surface area, and the chemistry of the substrate surface. To obtain a suitable TGA substrate we need to balance the tendency of the substrate to catalyze reactions involving glycerol with the tendency of the substrate to adsorb the glycerol via physical interaction in order to prevent migration of the glycerol in the TGA device.