Document ga2BJBb68KvMO2jE0o9kBvY4J

GRADE Zonol'fe Construction Products Division TO: a. N. Crawford DATE: August 31, 1984 FROM: j. R. Hindman SUBJECT: Commentary on the Chatfield & Lewis paper, "Development & Application of Analytical Techni que for Measurement of Asbestos . Fibers in Vermiculite" cc : W J. McCa.ig, H C.Duecker, INTRODUCTION As you requested, I have studied the data and conclusions presented by Chatfield and Lewis concerning ''Vermiculite Sc rolls". Since their conclusions are based on elemental analysis by energy dispersive x-ray ( EDX ) methods and crystal structural parameters obtained by selected area electron diffraction (SAED ), I should clarify some aspects of the magnesium rich sheet sili cate minerals for you first. SOME BASIC CRYSTAL STRUCTURE Phy1losi1icates are a class of silicate minerals that are characterized by sheets of SiO*--* and A1CU-5 tetrahedra linked together in a two dimensional array. This is usually referred to as the tetrahedral layer and it is normally attached on the apex side to a layer of di- or trivalent cations and hydroxyl ions in what is referred to as the octahedral layer. The cations in octahedral coordination are surrounded by six hydroxyl ions, or a combination of the oxygen atoms (which are the apices of the (Si,Al)CU tetrahedra) and hydroxyls. In Figure 1, the basic tetrahedral sheet (Si^Oio) and octahedral layer, as brucite Mg(0H>2, are shown. The symbols of connected triangles and connected parallelograms are used later for the sake of clarity. Ionic substitution in the tetrahedral sheets is usually limited to A1 for Si. However, in the octahedral layer one may find Mg, Fe, A1 and rarely Li. In Figure 2, the simplified structures of phlogopite mica, magnesium chlorite, magnesium vermiculite, and chrysotile are shown. By using structural symbols one is able to see that the four structures are elegant in their simplicity. In phol- gopite the octahedral layer is sandwiched between two tetrahedral sheets. The strong charge imbalance caused by 25X of the tetra hedral cations being Al*3 is compensated for by potassium ions between the tetrahedral-octahedral-tetrahedral sheets. (If there were no aluminum in the tetrahedral layer, there would be no charge imbalance, no potassium, and the structure would be that of talc ). Although phlogopite does not commonly alter to chlorite, the structure of talc-ch lorite is shown because of its resembl- ance to that of vermic ulite. In vermiculite the layer which conta ined potassium befor e alteration of phlogopite now contains 10002151 GRACE TO: A. N. Crawford DATE: August 31, 1964 Page 2 Zonolite Construction Products Division exchangeable cations surrounded by water molecules in octahedral coordination. In chlorite, non-exchangeable magnesium octahe- drally surrounded by hydroxyls occupy this layer. The two struc tures are similar in size. Chrysolite is the simplist structure with layers of one tetrahedral sheet bound to the octahedral sheet and nothing between the layers. The resulting layer to layer distance is about 7.3 A, compared to 10 A for phlogopite and 15 A for vermicu* 1 ite. SOME BASIS CRYSTAL CHEMISTRY In the table below are listed some chemical data for magne sium phyllosilcates. Chatfield and Lewis do not report analyses for Na, K, Ca or Fe' and so we will not consider these elements. Remember that ,A1 can substitute in both tetrahedral (Si) and octahedral (Mg) positions. *_Mg____ X A1 X Si Phlogopite Chlorite Mg-Vermiculite NH*-Ve.rmicul i te Chrysotile 17.48 26.32 17.74 15.75 30.50 6.47 nil 4.11 4.08 nil 20.19 20.27 20.19 20.01 23.49 In the chemical evaluation of the vermiculite scrolls, Chatfield and Lewis measure peak areas of energy dispersive x-ray spectra. While this type of data can be used for qualitative and semi-quantitative work, there are two drawbacks which should be kept in mind: EDAX won't detect lithium, and sodium data is affected by the use of copper-TEM grids. Both Li and Na can enter into the vermiculite structure. IDENTITY OF VERHICULITE SCROLLS There are two primary criteria for identifying a mineral phase: chemical composition and crystal structure. Remember in Figure 2 one could see that phlogopite, vermiculite and chrys olite share common building blocks - the tetrahedral and octahed ral layers. Thus, it is expected that x-ray or electron diffrac tion photographs should share common features. Of primary import ance is the repeat distances along the "zero layer". Chrysolite must show a reflection at 7.3 A and vermiculite should show a reflection at 14-15 A. The repeat distance in the scrolls 10002152 GRACE TO: A. N. Crawford DATE: August 31* 1984 Page 3 Zonolile Construction Product* Division is 10.2 A which would indicate a mica structure. There is another reflection appearing between the (002) and the (020) chrysotile positions which could be diagnostic if identified. The chemical data obtained by EDAX is not as satisfying as one would like. The Mg/Si and Al/Si peak area ratios obtained from the scrolls are too low for vermiculite and phlogopite, and don't match other magnesium phyllosilicates. If the EDAX data is acceptable and the assumption of a mica structure is correct, then we have a fascinating enigma. The enigma can be explained if the Palabora mill uses froth flotation in its mineral processing. It is well known that ammonium ions can replace cations in the exchangeable layer of vermiculite. The ammonium vermicu lite crystal structure is smaller than that of magnesium or calcium vermiculite and is similar to that of biotite or phlo gopite. The probability of finding NH*-" in a cationic flotation circuit is quite good. It is not uncommon to add soluble silicate compounds to reagent schemes. At Libby fluorosi1icic acid is used; and sodium silicate is commonly added at other locations. When one consid ers the size of a vermiculite scroll one can see how a coating of silicate-organic reagent could add Si in a non-crystalline form and add an EDAX invisible coating of organic molecules. Whether or not this coating would cause a vermiculite platelet to curl is not clear, but this organic layer ''jelly rolled" into the scroll would account for the low electron scattering power mentioned by Chatfield and Lewis. In summary, vermiculite scrolls may be ammonium vermiculite containing residual flotation reagent. EXFOLIATION STUDY On page 331, Chatfield and Lewis referred to a study of fibers/scrolls collected from samples after expansion in a muffle furnace. The results of this study would be interesting since chrysotile decomposes to olivine, Mg2Si0 at about 600 C. One would also expect that a major change in NH-s-vermiculite might take place before reaching that temperature. 10002153 GRADE TO: A. N. Crawford DATE: August 31, 1984 Page 4 Zonolile Construction Product* Division CONCLUSIONS Chemical and electron diffraction data presented by Chatfield and Lewis indicate that vermiculite scrolls are similar to a mica structure and may be NH<* vermiculite. The NH^-vermiculite scrolls could be formed by the action of NH**- ions and soluble silicate reagent components present in the Palabora mill. This analysis, if correct, would have obvious implications for LTEV products and fiber counts. The electron diffraction data presented by Chatfield and Lewis do not indicate the presence of chrysotile in those vermiculite scrolls studied. J . R. H I NDlTAN : ch Attachments: Figures and 2 10002154 FIGURE 1. THE BASIC BUILDEC BLOCKS OF SHEET SILICAIES: THE TETRAHEDRAL LAYER (LEFT) AND THE OCTAHEDRAL LAYER (RIGHT) * O c' 9o oooP 60006060600 o o ,o o .c >: JBH 8/31/84 10002153 FIGURE 2. THE GENERAL STRUCTURES OF TOUR MAGNESIUM PHYLL0SILICA1E MINERALS i c JR1I 8/31 /84