UBC Undergraduate Research

Variations in wetting angles for carbonate-silicate melts Huang, Katherine Mar 31, 2014

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
52966-Huang_Katherine_EOSC_449_2015.pdf [ 1.34MB ]
Metadata
JSON: 52966-1.0053618.json
JSON-LD: 52966-1.0053618-ld.json
RDF/XML (Pretty): 52966-1.0053618-rdf.xml
RDF/JSON: 52966-1.0053618-rdf.json
Turtle: 52966-1.0053618-turtle.txt
N-Triples: 52966-1.0053618-rdf-ntriples.txt
Original Record: 52966-1.0053618-source.json
Full Text
52966-1.0053618-fulltext.txt
Citation
52966-1.0053618.ris

Full Text

     VARIATIONS IN WETTING ANGLES FOR CARBONATE-SILICATE MELTS     by     KATHERINE HUANG     A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  BACHELOR OF GEOLOGICAL SCIENCES (HONOURS)   in   THE FACULTY OF SCIENCE     This thesis conforms to the required standard   ………………………… Dr. Kelly Russell  THE UNIVERSITY OF BRITISH COLUMBIA  (Vancouver)  MARCH 2014        ii  ABSTRACT   The sessile drop method was used to measure the wetting angles of carbonate-silica melt droplets as a proxy for surface tension. Melts were synthesized by sintering pure anhydrous Na2CO3 powder with SiO2 powder. The final weight percentages of Na2O, SiO2, and CO2 of each melt were calculated. Samples of approximately 0.03g of each melt composition were melted on a substrate of alumina at 1000°C until the droplets had fully relaxed. The droplets were quenched at room temperature and wetting angle measurements were taken. Droplets of melt compositions of 62 wt% SiO2 and less produced wetting angles of 3 to 9°. Droplets of melt composition of 71 wt% produced drastically higher wetting angles of 23°. These results indicate that SiO2 content does not significantly affect the wetting angle of the Na2SiO3- SiO2 system until SiO2 composes 71 wt% of the of the melt composition. The phase diagram of the SiO2-Na2CO3 system is used to estimate the varying amounts of liquid and solid Na2CO3, solid Na2SiO3, liquid Na2Si2O5, and solid SiO2 in each melt. Due to its molecular structure and high number of non-bridging oxygens/tetrahedral (NBO/T) Na2CO3 likely is the controlling component of the surface tension of the melts below 71 wt% SiO2. Once the melt composition exceeds 62 wt% SiO2,, Na2Si2O5 controls the surface tension of the melt via the introduction of a bridging oxygen that affects the interactions between the liquid molecules and increases the wetting angle.                 iii  TABLE OF CONTENTS  COVER PAGE............................................................................................................................................... i ABSTRACT .................................................................................................................................................. ii TABLE OF CONTENTS ............................................................................................................................. iii ACKNOWLEDGEMENTS .......................................................................................................................... v INTRODUCTION ........................................................................................................................................ 1 PAST LITERATURE ................................................................................................................................... 3 Contact Angle and Surface Tension .......................................................................................................... 3 Methods .................................................................................................................................................... 4 Sessile Drop Method ............................................................................................................................. 4 Improved Sessile Drop Method ............................................................................................................. 4 Contactless Techniques ......................................................................................................................... 5 Effects on Surface Tension and Contact Angle ........................................................................................ 6 Effects of Temperature .......................................................................................................................... 6 Effects of Atmosphere ........................................................................................................................... 7 Effects of Melt Time .............................................................................................................................. 7 Effects of Drop size ............................................................................................................................... 8 Relating Contact Angle to Surface Tension .............................................................................................. 8 ESTABLISHING AN EXPERIMENTAL METHOD FOR BOROSILICATE AND SODA LIME MELTS ....................................................................................................................................................... 10 Methodology ........................................................................................................................................... 10 Melt Preparation ................................................................................................................................. 10 Substrate Preparation ......................................................................................................................... 10 Experimental Procedure ..................................................................................................................... 11 Wetting Angle Measurements .............................................................................................................. 11 Effects of Variables on Wetting Angles ................................................................................................. 13 Effects of Temperature ........................................................................................................................ 13 Effects of Melting Time ....................................................................................................................... 14 Effects of Melt Volume ........................................................................................................................ 15 Effects of Solid Substrate .................................................................................................................... 17 iv  Experimental Parameters for Carbonate-Silica Experiments .................................................................. 19 CARBONATE-SILICA EXPERIMENTS ................................................................................................. 20 Carbonate Melt Structure ................................................................................................................... 20 Carbonate-Silica Melt Preparation .................................................................................................... 22 Carbonate-Silica Mixture Products .................................................................................................... 22 Experimental Procedure ..................................................................................................................... 24 Results and Discussion ........................................................................................................................... 25 Changes in Surface Tension as a Function of Melt Composition ....................................................... 28 Geological Implications .......................................................................................................................... 33 CONCLUSIONS......................................................................................................................................... 37 REFERENCES ........................................................................................................................................... 38 APPENDIX ................................................................................................................................................. 41 Appendix 1. ............................................................................................................................................. 41 Appendix 2 .............................................................................................................................................. 47               v  ACKNOWLEDGEMENTS     I would like to thank my supervisor, Dr. Kelly Russell, for continuous support, guidance, and help. Thank you for taking me as a student and inspiring me to take a special interest in volcanology and experiments to better understand the earth and its magmatic processes.  Thank you to Amy Ryan for teaching me everything I know about the CESL Lab and the ongoing help with my results, research, and writing my thesis.  Thank you as well to Jörn Unger for all your help in the machine shop and to Corky Smith for kindly taking me to the Twin Sisters quarry.               INTRODUCTION  The exact compositions of kimberlite magmas are not certain but are known to be volatile-rich, have low SiO2 contents (<25-35%), contain high amounts of MgO (15-20%), and are characterized by low densities and viscosities (Kono et al., 2014; Russell et al., 2012). Russell et al. (2012) propose that carbonatitic/near carbonatitic melts are the parent melts of kimberlite melts. At high pressures (>2.5GPa), partial melting of carbonated mantle sources often produces carbonatite melts (Wyllie, 1980).  These partial melts near the solidus at the base of or below the cratonic mantle lithosphere are composed of up to 40 wt% CO2, 10 wt% H2O and <10 wt% SiO2 dissolved in the melt (Dasgupta et al., 2013).  As the melts travel through the mantle towards the crust, orthopyroxene from mantle xenoliths is assimilated which increases the silica content of the melt.  The transition from carbonatitic to silicic melt composition is coupled with a decrease in CO2 solubility. Although the solubility of CO2 in silicic melts increases with pressure, CO2 exsolution is driven initially by depressurization and then by the addition of SiO2. Exsolution of CO2 as a volatile phase forces the melt to ascend rapidly due to an increase in buoyancy. During ascent, the magma continues to incorporate orthopyroxene, which ultimately causes the melt to preferentially crystallize olivine (Russell et al., 2012). Because a kimberlite eruption has never been observed and the origin of kimberlite magma is poorly understood, kimberlite parent magma is modeled to be similar to carbonate-rich magmas.  In this study, I experimentally investigate the extent to which transport properties of melts vary as a function of composition. My overarching goal is to investigate how melts change as a function of increasing SiO2. The experiments are on carbonate and siliceous-carbonate melts and are designed to measure wetting angle as a proxy for surface tension.  Contact angle measurements can be used to study the wetting of a system since variations in wetting angle with composition have the potential to provide information on melt structure and polymerization (Amirfazli and Neumann, 2004). Numerous studies have shown an increase in wetting angle (𝜃) with an increase in liquid surface tension (𝛾𝑙𝑣). The contact angle of a liquid is a complex property that is dependent on 3 interfacial energies: 𝛾𝑠𝑣, 𝛾𝑠𝑙, and, 𝛾𝑙𝑣 (solid-vapour, solid-liquid, and liquid-vapour, respectively). Variables 2  such a surface heterogeneity, surface roughness, melt temperature and melt drop size also affect contact angle (Amirfazli and Neumann 2004). As such, these factors must be highly controlled or they will inhibit the ability to determine an accurate relationship between composition, which in turn affects interfacial energies, and contact angle. Therefore, I run several trial experiments with borosilicate and soda lime on a constant substrate of alumina and then dunite to prove that contact angle is dependent on substrates and my methods (time, temperature, and size). The two silicates are used because of their well-known physical and chemical properties and compositions. Once the experimental method is developed and tested on these melts, I then apply the method to carbonate-silica compositions. The study is concluded with a discussion of the implications of these results for geological processes.                 3  PAST LITERATURE   Previous studies are reviewed to outline the empirical relationships defined between external factors and surface tension, and to compare my methods to those of other techniques and for different melts. Later, the results from these studies are compared with mine to observe the differences due to factors that were not considered in my experiments. Results are also compared to prove that the relationships I find are valid despite certain factors not being constrained.  Contact Angle and Surface Tension  The Young equation (1) states that the contact angle of an infinitely large liquid drop in equilibrium on a solid substrate is determined by three interfacial tensions:     𝑐𝑜𝑠𝜃∞ =𝛾𝑠𝑣−𝛾𝑠𝑙𝛾𝑙𝑣        (1)     where   is the contact angle at the intersection of the liquid-solid and liquid-vapor interfaces andsv , sl , and lv  are solid-vapour, solid-liquid, and liquid-vapour interfacial tensions, respectively.  The contact angle of a drop of a specific size is dependent on several other factors and can be expressed as: 𝑐𝑜𝑠𝜃 = 𝑐𝑜𝑠𝜃∞ −𝜎𝛾𝑙𝑣1𝑅       (2)            where σ is the line tension, R is the drop radius, and   is the contact angle at the solid-liquid-vapour intersect. The final drop shape is determined by the liquid surface tension (γ), line tension (σ), and the volume of the drop (Yuan and Lee, 2013; Fatallohi, 2013). Line tension is defined as the tension at the point where solid, liquid, and vapour interfaces meet at the contact angle and determines to tendency of the solid and liquid to adhere (Amirfazli and Neumann, 2004; Fatallohi, 2013). Line tension increases surface area and decreases contact angle, whereas 4  surface tension decreases surface area and increases contact angle. The relationship between these two values will determine the final drop shape and contact angle, as shown in Figure 1.  Figure 1. Three possible shapes of a droplet reflecting the interplay of surface tension (  ) and adhesion coefficient (σ) (Fatallohi, 2013).   Equations (1) and (2) are only valid under ideal conditions and do not consider external effects such as gravity, oxidation, or pressure. Factors that affect surface tension have been extensively studied using several methods  Methods  Sessile Drop Method  The sessile drop method was is a popular method used to measure contact angle and compute surface tension. The general protocol for this method is to prepare high purity samples, prepared as small cylinders, then heat the samples in a sealed furnace with pure Ar gas (Yuan et al., 2002; Lie et al., 2003). Contact angles can be easily measured from the droplet. More recent experiment use photographs taken of the droplet from inside the furnace at different time intervals. Surface tension and contact angle can be calculated from these photographs using the droplet contour and the known density of the material (Yuan et al., 2002).   Improved Sessile Drop Method  Hamuyuni et al. (2012) and Chang and Lin (2011) used an improved sessile drop method to measure surface tension of molten sulfides and glass melts, respectively. Hamuyuni et al. (2012) used 5mm diameter sulfide powders on a substrate of dense alumina. Samples were 5  heated in a furnace with a pure Ar gas atmosphere to inhibit reactions between the liquid and air. Chang and Lin (2011) followed a similar protocol with a small piece of glass of a certain mass. Surface tension measurements for both studies were taken from inside the furnace when the system reached equilibrium at melting temperature. The improvement from previous use of this method involes different way to calculate surface tension. High-resolution image sensors capture to a clear and undistorted drop shape profile and use the Young-Laplace equation (3):  pRR  2111        (3)  where p  is the pressure change across the surface and 1R  and 2R  are curvature radii at the surface of each point (Change and Lin, 2011; Fatallohi and Hajirahimi, 2013). Hundreds to thousands of points are taken from the edge of the drop profile and fitted with the Young-Laplace equation to calculate surface tension. An improved sessile drop method used by Fujii et al. (2006) also employs high-resolution technology for more accurate measurements and surface tension calculation by automatic image processing. However, rather than the traditional method of placing a solid composition on a substrate and heating it, they use an electromagnetic levitator to drop molten silicon and take measurements immediately after the droplet contacts the substrate. This method provides comparable results to the non-contact oscillating drop method also used in their study, as well as the levitating drop method used by Millot et al. (2008).   Contactless Techniques  Because many molten substances, particularly silicon, react with a substrate and cause measurement errors, more recent methods to study surface tension involve contactless techniques (Rhim and Oshaka, 2000). Millot et al. (2008) use a levitating drop method on molten silicon. Pure Ar and N are used to control the atmosphere to O2<3ppm. Surface tension can easily be calculated from this method. Results were similar to those of the sessile drop method. However, specific conditions, such as those employed by Fujii et al. (2006), are required for the sessile drop method to have consistent results.  6  Effects on Surface Tension and Contact Angle  Effects of Temperature Goleus et al. (1996) calculated surface tensions of borosilicate melts of various bulk compositions from 900-1450ºC. Results showed a decrease in surface tension with an increase in temperature for all borosilicate compositions. This means the system has a negative temperature coefficient. Results of the calculations compared to experimental values correlated well, with a standard deviation of +/-11.6 x 10mN/m. Huang et al. (2001) ran experiments to determine the wetting of borosilcate glass on a substrate of alumina silicate and saw a decrease in the value of contact angles from 1300ºC to 1600ºC (Figure 2).           Figure 2. Contact angles of borosilicate on mullite at (from left to right) 1410ºC, 1500ºC, and 1600ºC. Huang et al. (2001).   A negative temperature coefficient was also found for molten silicon by Zhou et al. (2003), Fuji et al. (2005), Kimura et al. (1997), and Shishkin and Basin (2004). Most of these studies found similar results (Figure 3), although variations exist and are later discussed to be due to other factors that were controlled differently.                Figure 3. Surface tension of molten silicon as a function of temperature from Zhou et al. (2003) (left) and Fujii et al. (2005) (right). 7  Effects of Atmosphere  Chang and Lin (2011) recorded surface tension values of 303+/-20mN/m for glass melts in air and 319+/-9mN/m in a controlled Ar atmosphere at 850ºC. These results correlate well with similar experiments (Parikh 1958; Clare et al., 2003). The surface tension of silicon is highly sensitive to O2, which, along with H2O, may react chemically with the glass (Fujii et al. 2005; Chang and Lin 2011). For silicon without a controlled atmosphere, Zhou et al. (2003) observed a 4% decrease in surface tension compared to silicon at oxygen partial pressures of   10-19 MPa. Since Ar is an inert gas, experiments run in an Ar atmosphere with no O2 particles likely provide more accurate values of the true surface tension. However, Parikh (1958) found a negligible change in surface tension for silica under vacuum, dry N2, and H2O saturated air at 1100-1300ºC. He also studied the effects of various atmospheres on soda lime glasses and found that surface tension of soda lime decreased when H2O was added to the atmosphere at 550ºC.  Maintaining a low partial pressure environment requires purified Ar gas and an oxygen partial pressure of around 10-19 MPa (Li et al., 2003). However, for the experiments with borosilicate and soda lime, quantitative values of surface tension are not necessary. Furthermore, the ability to control the atmosphere is not within the scope of this project. As such, the assumption is made that if atmospheric conditions are kept constant, qualitative relationships between temperature, time, drop size, substrate, and composition will be consistent.   Effects of Melt Time   At a constant temperature, melt time did not have a significant effect on contact angle of borosilicate for 3 hours and longer. Li et al. (2006) found that surface tension did not change with time once a melt reached its intended temperature. These results are consistent at several temperatures. For pure silicon glasses, there was negligible weight loss due to evaporation during sessile from experiments by Chang and Lin (2011). The chemical similarities between borosilicate and silicon allow one to safely assume negligible borosilicate evaporation in this experiment, thus my results suggest that contact angle will remain the same indefinitely once the final drop shape is reached.   8  Effects of Drop size Amirfazli et al. (1998) observed a linear relationship between cos  and R1 for liquid tin on silica substrate, where   is the contact angle and R  is the radius of the drop (Figure 4). These results agree with Equation (2) which correlates contact angle with drop radius.     Figure 4. Change in contact angle with increasing drop radius of liquid Sn on silica at 900ºC (left) and interpretation of the results with respect to the Equation (2) (right) (Amirfazli, 1998).   Relating Contact Angle to Surface Tension  The sessile drop and improved sessile drop methods require analysis of the drop shape, including its dimensions and curvature. The technology in this study limits the ability to accurately calculate surface tension. Therefore, a relationship between surface tension and contact angle must be made.  David and Neumann (2014) calculated the liquid surface tensions and contact angles for several liquids (Figure 5). The results suggest a strong direct relationship between the two variables, which agree with experimental results.  9   Figure 5. Relationship between calculated cosine of contact angle and calculated liquid surface tension of non-hydrogen-bonding, organic liquids (compared to hydrogen-bonding liquids) suggests that measuring contact angle can be a valid method to quantify liquid surface tension (David and Neumann, 2014).  Likewise, I assume a direct relationship between an increase in contact angle   and an increase in liquid surface tension (or increasing cos  as an inverse relation to liquid surface tension) when interpreting results.         10  ESTABLISHING AN EXPERIMENTAL METHOD FOR BOROSILICATE AND SODA LIME MELTS   Due to the complexity and number of factors influencing liquid contact angle, the first set of experiments were run with borosilicate and soda lime on alumina to prove that contact angle is dependent on time, temperature, and size, and then on dunite to prove that contact angle is also dependent on substrate.  The constant substrate of ceramic alumina plates is used to remove any inconsistencies in physical properties of the surface. The two silicates melts were chosen because of their well-known physical and chemical properties and compositions.  Methodology  Melt Preparation Borosilicate and soda lime samples from were cored with a 0.5cm or 1cm diameter drillbit then cut to a height of 0.5cm or 1cm with an isomet low-speed saw. Diameter and height were then measured with a caliper and weighed with a Sartorius scale. These measurements were used to calculate volume, surface area, and density of each core. Since I did not have access to a 0.25cm drillbit, smaller samples were prepared using a mass that corresponds to a volume of a 0.25x0.25cm core (approximately 0.03g). All size and mass measurements are in Appendix 1.  Substrate Preparation  The two types of substrate were used for the experiments were alumina disks and dunite. Circular alumina disks are 42mm diameter and 2mm thick.  Dunite samples were taken from the olivine quarry at Twin Sisters. The samples were chosen based on high modal percentage of forsterite, large mineral grains, and minimal to no alteration. The dunite rock was cut into half inch thick slabs of approximately 4”x3” square diameter with slab thickness to 1/2" using a rock saw. The slabs were then polished on one side 11  of the large surface using a coarse-grained and then finer-grained aluminum oxide powder. This is again to reduce error in the measurement of contact angle by removing surface features of inconsistent texture. Several theoretical models and experiments have shown that heterogeneity and roughness of the solid surface can affect the contact angle-drop size relationship (Amirfazli et al. 1998). The dunite slabs were gradually preheated in the furnace from 789 to 1050ºC for 1 hour to avoid fracturing.  Experimental Procedure  Samples were placed in the center of the substrate and loaded into the furnace, which was preheated for 30 minutes to ensure that it was at the intended temperature. The sample was then removed from the oven and quenched at room temperature. Since the equipment limits the ability to measure the contact angle of the melt in the furnace, the experiments and calculations are done with the assumption that contact angle does not change upon quenching.   Wetting Angle Measurements  A Theta optical tensiometer instrument was used to measure the static contact angle between the melt and the substrate using the sessile drop method. The machine measures contact angle and surface tension by fitting the Young-Laplace equation around the droplet (Biolin 2012; Hamuyuni et al. 2012). However, many of the droplets were too big for the machine to measure surface tension, so contact angle was the only measurement that can be made. Measuring the static contact angle was appropriate for these samples because this method requires a droplet standing on a surface without a change in the three-phase boundary (Biolin 2012).  The optical tensiometer records 12 measurements per second for a total of 120 recordings from which mean angle and standard deviation are calculated (Table 1). Each sample was analyzed from 8 different sides and the standard deviation between all sides was calculated. Between these values and the machine standard deviations, the larger value was taken as the final standard deviation. The final angle used was the average from all sides.   12  Table 1. Example of wetting angle ( ) measurements from an optical tensiometer. Rows 1-8 are averages and standard deviations from 120 measurements for each side of the sample. For this sample, the machine error (side 7) is larger than the standard deviation between each side, so +/-1.45 is the final standard deviation used.   Side Mean  S.D. 1 149.078+-0.0415 2 148.227+-.0437 3 132.3+-.12 4 129.2+-1.29 5 146.899+-.041 6 145.987+-.0413 7 134.3+-1.45 8 139.29+-.15 Average  147.54775 S.D 1.373   For samples too large for the optical tensionmeter, contact angles were measured by taking photos from an oblique view at a constant distance away from the sample and analyzed using the angle measurement function on ImageJ (Figure 6). Methods to calculate average angle and standard deviations are the same, except 4 hand measurements from each side were taken rather than 120. All mean and standard deviation values are in Appendix 1.                 Figure 6. Photos taken at an oblique view being analyzed in ImageJ. The value shown corresponds to a single measurement of 145° on one side of a borosilicate droplet Bo11 (top) and a single measurement of 35° on one side of a soda lime droplet SL9 (bottom).     13  Effects of Variables on Wetting Angles  Effects of Temperature  Borosilicate and soda lime cores of size 0.5cm diameter by 0.5cm height were melted at 1000ºC, 1050ºC, and 1100º. Measurements of wetting angles ( ) are in Table 2. The temperature for all experiments did not exceed 1100ºC because it has been suggested that a temperature high enough to cause olivine instability may cause the olivine to dissolve into the melt, thus decreasing the relative silica content of the melt, which would affect later experiments that use dunite as the substrate (Wanamaker and Kohlstedt 1991).   Table 2. Effect of temperature on contact angles for borosilicate and soda lime melts. Sample size was 0.5x0.5cm.    At 1000ºC, the borosilicate did not fully relax, as this temperature is not high enough above the material’s softening point to allow for full structural relaxation. Wetting angle decreased drastically between 1050 ºC and 1100 ºC for both borosilicate and soda lime (Figure 7). A slight increase in contact angle between 1000 ºC and 1050 ºC may be due to measurement or procedural error. Since temperature affects wetting angle at this temperature range, all subsequent borosilicate and soda lime experiments were run at 1050ºC in order to stay constant and to be operating well above the liquidus of both materials. Sample No. Composition Substrate Temp (ºC) Time (hr) θ S.D.Bo15 Borosilicate Alumina 1000 5Bo7 Borosilicate Alumina 1050 5 140.3 0.9 TimeBo19 Borosilicate Alumina 1100 5 101 2 3hrSL2 Soda Lime Alumina 1000 5 51.6 2.6 3hrSL1 Soda Lime Alumina 1050 5 39 2 5hrSL3 Soda Lime Alumina 1100 5 5 0.7 3hrDid not relax14    Figure 7. An increase in temperature causes a decrease in contact angle for soda lime and borosilicate.  Effects of Melting Time   Borosilicate and soda lime rods were melted at 1050ºC at various times. This was to ensure that both materials can structurally relax at the parameters that the experiments are operating at, and to determine if melting time affects contact angle after it has structurally relaxed. Because borosilicate has the highest viscosity of all the materials in the experiments, the time required to relax will theoretically be the longest of all the experiments. Therefore, if the borosilicate fully relaxes, it was assumed that all other compositions used would relax under the same melting conditions.   Contact angles of borosilicate were not affected by melting time in this time range (Figure 8). Contact angles of soda lime were 33.4º+/-0.4 and 38.96º+/-2 for 3 hours and 5 hours, respectively (Table 3). The cylindrical starting material had a contact angle of approximately 90º. Assuming that the contact angle continuously decreases during structural relaxation, the slight 15  change in contact angle with time is likely due to error since the angle uncharacteristically changes from 90º to 33º and then to 39º.  Table 3. Effect of melt time on contact angles for borosilicate and soda lime of size 0.5x0.5cm.      Figure 8. Melt time does not have a significant effect on the contact angle at 1050ºC for this time range.   Effects of Melt Volume  To determine if the size of the samples had an effect on the contact angle, experiments were run where core size was the only manipulated variable. Borosilicate cores of dimensions Sa ple No. Co position Substrate Temp (ºC) Time (hr) θ S.D. Sample NoBo10 Borosilicate Alumina 1050 3 141 2 xBo6 Borosilicate Alumina 1050 4 139.3 1.7 xBo7 Borosilicate Alumina 1050 5 140.3 0.9 xSL9 Soda Lime Alumina 1050 3 33.4 0.4 YesSL1 Soda Lime Alumina 1050 5 39 216  1cm diameter by 1cm height, 0.5cm diameter by 0.5cm height, and 0.25cm diameter by 0.25cm height on alumina substrate were melted. The same setup was used for soda lime cores. My results show that borosilicate produced the same contact angle for all droplet sizes used (Table 4). However, droplet size had a significant effect on contact angle for soda lime between 1x1cm and 0.5x0.5cm (Figure 9). At sizes smaller than 0.5x0.5cm, wetting angle did not change. Therefore, subsequent experiments were performed with samples of a volume corresponding to a 0.25x0.25cm cylinder rod.   Table 4. Wetting angles of soda lime and borosilicate at different sizes.      Figure 9. Wetting angles for borosilicate and soda lime at of sizes 0.25x0.25cm, 0.5x0.5cm, and 1x1cm on alumina substrate at 1050ºC for 5 hours.  Sample No. Composition Substrate Diameter (cm) Temp (ºC) Time (hr) θ S.D.Bo11 Borosilicate Alumina 0.25 1050 5 147 1.37Bo7 Borosilicate Alumina 0.5 1050 5 140.33 0.91Bo4 Borosilicate Alumina 1 1050 5 140 4SL12 Soda Lime Alumina 0.25 1050 5 36.37 2.49SL1 Soda Lime Alumina 0.5 1050 5 38.96 2SL5 Soda Lime Alumina 1 1050 5 60 3.4117  Effects of Solid Substrate  Once I was confident in my results, I ran the experiments that were determined to be ideal (1050ºC, 5hr, 0.25cm diameter) on olivine to evaluate how substrate affects wetting angle.  The change in wetting angle between the two substrates was higher for borosilicate than soda lime (Figure 10). It is evident from large standard deviation values of experiments run on dunite that there was a high degree of asymmetry; moreso for the borosilicate melts than the soda lime (Table 5). Figure 11 shows how borosilicate and soda lime droplets relax on dunite. Note that the optical tensiometer can only take wetting angle measurements for symmetrical drops, so angles from borosilicate drops had to be measured manually using ImageJ. The borosilicate melts appear to be pulled in various directions according to fractures in the dunite and grain boundaries. This is logical because grain boundaries provide an easier passageway for melts to migrate, and melt migration is also controlled by wetting properties and volume of the melt (Schafer and Foley, 2001). Furthermore, a study by Wanamaker and Kohlstedt (1991) found that the wetting angle of silicate melts on single crystals of olivine is dependent on the orientation of the olivine crystal face. The dunite substrate consists of too many factors affecting results which interferes with the primary question of this project and is therefore not used for further experiments with carbonate and silicate.  Table 5: Results of experiments run on alumina disks for borosilicate and soda lime of mass of 0.03g.    Sa ple No. C mposition Substrate Temp (ºC) Time (hr) θ S.D.Bo11 Borosilicate Alumina 1050 5 147 1.37B 13 Borosilicate Dunite 1050 5 84.5 3.23Bo17 Borosilicate Dunite 1050 5 83.83 8.42SL12 Soda Lime Alumina 1050 5 36.37 2.49SL14 Soda Lime Dunite 1050 5 18.57 1.02SL16 Soda Lime Dunite 1050 5 19.02 0.95SL18 Soda Lime Dunite 1050 5 19.16 1.7218   Figure 10. Wetting angles on dunite are plotted against wetting angle on alumina substrate for borosilicate and soda lime. Proximity of the each date set to the trendline of x=y shows how much wetting angle differs between substrates.          Figure 11. Images taken from the sessile drop method feature of the optical tensiometer show highly asymmetrical wetting angles for borosilicate on dunite (left). Soda lime (right) relaxes symmetrically on dunite.      19  Experimental Parameters for Carbonate-Silica Experiments    The results of the borosilicate and soda lime experiments were analyzed to determine the ideal parameters to use for the following carbonate-silica experiments to ensure that melt composition is the only factor that causes a change in wetting angle. A constant temperature of 1000°C, which is above the liquidus of the carbonate-silica melts, were used. Because melting time was determined to not affect wetting angle, droplets were melted until they reach full relaxation, depending on the composition. Samples of approximately 0.03g were melted into sessile droplets. Alumina disks were used as a constant substrate because dunite introduces complications into the system and makes measuring the wetting angle difficult.                  20  CARBONATE-SILICA EXPERIMENTS   The purpose of these experiments is to explore for changes in the surface tension of Na-carbonate melts as they become more siliceous. The expectation is that there may be a measurable shift in surface tension properties where a certain SiO2 content is reached, indicating a profound change in melt structure (i.e. degree of polymerization). The experimental melts were synthesized by melting varying weight percentages of sodium carbonate and silica powders. Changes in surface tension properties of the experimental melts were monitored by measuring the wetting angles of sessile drops.  Wallace and Green (1988) experimentally determined carbonatite melts to be rich in Ca, Mg, Fe, Na, and K with a small amount of dissolved silicate. Their experimental melts quenched to form mineral assemblages of dolomite and Na-Mg carbonate minerals. Although most carbonate melts that have been discovered are Ca-rich, there are types of carbonate melts that are alkali-rich, including those produced by the famous Oldoinyo Lengai volcano (Jones et al. 2013). Alkali-rich carbonate melts are dominated by Na2O, K2O, CaO, and CO2 (Jones et al. 2013). Na2CO3 is easy to work with experimentally because of its low melting temperature and stability at low pressures. For simplicity and ease, I used Na2CO3 in my experiments as a representation of the carbonate component in carbonate-silica melts.   Carbonate Melt Structure  Carbonate melts consists of ionic compounds of a CO32- anion and a metal cation. These ionic melts are considered relatively structureless due to the inability to polymerize. On a molecular level, the CO32- anion lacks an unpaired orbital to allow for covalent bonding and thus can only form weak ionic bonds (Jones et al., 2013). The strength of the electronegativity of the metal cation affects the intramolecular bond strength of the carbonate ion and thus has a major control on the stability of the molecule (Genge et al. 1994). Spectroscopy on quenched carbonate glasses suggest that the ionic bonding of a metal cation (Na+) with CO32- could support a framework structure for carbonate melts, but such a structure would not be rigid enough to form glass due to the weakness of the ionic bonds (Genge et al. 1994). Furthermore, low viscosities 21  and densities and high mobilities of carbonate melts are favourable to spontaneous crystallization which prevents glass formation.   In contrast, silicate melts are covalently bonded polymers characterized by a network structure that strongly resists rearrangement and therefore quench to glass (Jones et al. 2013).  The SiO2-Na2O system is well studied (Jones et al., 2005). Powder mixtures of SiO2 and Na2CO3 undergo the reaction in Equation (4) at 1000ºC.  Na2CO3(s) + SiO2(s)  Na2CO3(l) + CO2(g) + Na2SiO3(s)        (4)  This reaction is initiated by the melting of Na2CO3 (Tm=851ºC). Liquid Na2CO3 wets the SiO2 grains, causing the mixture to vigorously react, release CO2 gas, and create a new hybrid Na2CO3-SiO2 melt (Jones et al., 2013). At high SiO2 contents the resultant mixture will comprise a CO2-depleted Na2O-SiO2-CO2 melt coexisting with Na2SiO3 crystals. If there is enough SiO2 in the system, the CO2-producing reaction involving molten Na2CO3 will go to completion and the final melt will not contain any dissolved CO2. If there is still SiO2 in the melt after all of the Na2CO3 has reacted, then SiO2 will begin to react with the solid Na2SiO3 crystals to form liquid Na2Si2O5, as shown in Equation (5).   SiO2(s) + Na2SiO3(s)  Na2Si2O5(l) + Na2SiO3(s) + SiO2(s)                   (5)  Note that the reactions in Equation (4) and Equation (5) only occur above the Na2Si2O5 melting point and below the Na2SiO3 melting point (Table 6).   Table 6. Melting points of eutectics and compounds in the Na2O-SiO2 system. These temperatures have been proven to be valid for Na2CO3. Adapted from Jones et al. (2005).    x Na2O T  (ºC)  Eutectic/Compound0.255 793   Eutectic between Na2Si2O5 and SiO20.333 874   Na2Si2O50.372 846   Eutectic between Na2SiO3 and  Na2Si2O50.5 1089  Na2SiO322  Carbonate-Silica Melt Preparation  Carbonate-silica mixtures were measured out by weight percentage. Solid slabs of HPFS Fused Silica (SiO2) were crushed to powder using an agate mortar and pestle. The carbonate used was anhydrous sodium carbonate (Na2CO3) powder of 99.999% purity from Sigma Aldrich. Initial mixtures contained less wt% SiO2 than the intended final mixture to compensate for CO2 loss during melting. For the melt of a final composition of 30 wt% SiO2, I initially mixed only 25 wt% SiO2 with 75 wt% Na2CO3. The materials and a crucible were weighed separately and then together to ensure there was no mass loss during the weighing process. The powders were stirred, poured into the alumina crucible, melted at 1000ºC for 3 hours, and then cooled in a desiccator to inhibit moisture absorption. The crucible with the sample was then re-weighed. Assuming that the only mass loss was due to CO2, the composition of the sample was re-calculated. The fused mixture was then removed from the crucible, crushed, mixed, poured into a new alumina crucible, and sintered for a second time for 3 hours at 1000ºC to increase homogenization of the mixture. The mixture was re-weighed and the composition was re-calculated again, assuming the second mass loss was due to CO2. The same process was carried out for melts of varying Na2CO3-SiO2 compositions. Re-calculated compositions of my melts after CO2 loss are listed in Table 7. Equation (6) describes a potential deleterious reaction between Na2CO3 melts and the alumina crucibles and disks (Meher et al., 2010)  Al2O3 + Na2CO3  2NaAlO2 + CO2                            (6)  Pure Na2CO3 powder was only sintered until it was completely melted (12 minutes) to inhibit the reaction from Equation (6).   Carbonate-Silica Mixture Products   There was no significant CO2 loss from the melting of pure Na2CO3. For melts of 11 wt% and 37 wt% SiO2, there were significant CO2 losses from the first and second sintering processes. This is likely because the reaction is Equation (4) will not consume all of the SiO2 if some of the SiO2 is not in physical contact with the Na2CO3 (Jones et al., 2005). 23  Table 7. Melt compositions calculated by measuring the samples before and after fusing and attributing all weight loss to CO2. The final calculated compositions are reported for the products resulting from two sequential fusions.    Sample Label SiO2Na2CO3 (g) Total (g)Carb 5 0.00000 7.79610 7.79610CS4 0.48870 4.71150 5.20020CS2 1.71231 4.04283 5.75514CS3 1.34710 1.90100 3.24810CS6 2.61570 2.70800 5.32370CS7 3.40740 2.38600 5.79340CS8 3.07500 1.34070 4.41570Mechanical Mixtures (wt. %)SiO2Na2O CO2TOTAL0.0000 58.4771 41.5229 100.009.3977 52.9816 37.6207 100.0029.7527 41.0786 29.1687 100.0041.4735 34.2246 24.3019 100.0049.1331 29.7455 21.1214 100.0058.8152 24.0837 17.1011 100.0069.6379 17.7549 12.6072 100.001st Mass Loss(g) SiO2Na2O CO2TOTAL0.056 0.0000 58.9002 41.0998 100.000.7314 10.9358 61.6530 27.4112 100.000.9413 35.5705 49.1111 15.3184 100.000.79 54.8025 45.2239 -0.0264 100.001.1269 62.3261 37.7325 -0.0586 100.000.9946 71.0053 29.0752 -0.0805 100.000.5914 80.4069 20.5005 -0.9074 100.002nd Mass 2nd Mass Loss(g) (g) SiO2Na2O CO2Sample Label 7.7961 0 0.0000 58.9002 41.0998 Carb 54.0451 0.1877 11.4680 64.6530 23.8791 CS40.8301 0.0301 36.9089 50.9589 12.1322 CS22.339 0.0025 54.8611 45.2723 -0.1334 CS33.5209 0.0059 62.4307 37.7959 -0.2265 CS63.8873 0.005 71.0967 29.1127 -0.2094 CS73.6767 0.004 80.4944 20.5229 -1.0173 CS8Recalculated Compositions after CO2 Loss (wt%)Recalculated Compositions after 2nd CO2 Loss (wt%)Starting MixturesFused Samples24   After the melt was crushed and re-mixed, some of the unreacted SiO2 was in contact with the Na2CO3 and thus able to react and form CO2 gas to be exsolved. This suggests that melts of 11 wt% and 37 wt% SiO2 are composed of the reaction products from Equation (4) (liquid Na2CO3 and solid Na2SiO3). There were no significant CO2 losses during the second sintering process for melts of 54 wt%, 62 wt%, 71 wt% and 80 wt% SiO2. This suggests that the initial composition contained enough SiO2 to react with all of the Na2CO3 and that all of the CO2 gas was driven off during the first sintering process. Therefore, melts of 54 wt% SiO2, 62 wt% SiO2, 71 wt% SiO2, and 80 wt% SiO2 are likely composed of the reaction products from Equation (5) (liquid Na2Si2O5, solid Na2SiO3, and solid SiO2). My melt compositions are plotted on the SiO2-Na2CO3 phase diagram in Figure 12.            Figure 12. Phase diagram of the SiO2-Na2CO3 system. The red line indicates the temperature at which my compositions were sintered (1000ºC) and my melt compositions are plotted. Adapted from Hrma (1985).    Experimental Procedure  Small pieces of each mixture of approximately 0.03g were melted on alumina disks to create sessile drops for wetting angle measurements. Because the melts of 0 wt% and 11 wt% SiO2 have high Na2CO3 contents, the droplets react with the alumina substrate (Equation (6)). Therefore, the pure Na2CO3 and the 11 wt% SiO2 droplets were removed from the furnace immediately after the melt had structurally relaxed (~1-2 minutes).  25  This process could not be used for the droplets of 37 wt% SiO2 droplets because of the significantly longer relaxation timescales for the melts. However, insignificant weight loss of the droplets after relaxation indicates that, even if the melt did indeed contain Na2CO3, it did not react with the alumina substrate. The droplets of 37 wt% SiO2, 55 wt% SiO2, 62 wt% SiO2, and 80 wt% SiO2 were heated for 1-2 hours. Droplet weights and melt times are listed in Appendix 2.   Results and Discussion  Pure Na2CO3 droplets produced a wetting angle of θ =3±0.4º. Droplets of 11 wt% SiO2, 37 wt% SiO2, 55 wt% SiO2, and 62 wt% SiO2 produced average wetting angles of 2.5±0.5º, 8.5±2.1º, 8.15±2.5º, and 5.25±2.4º, respectively (Table 8). All measurements are in Appendix 2.  Note that in Table 8 and all subsequent references to melt composition, the values of xSiO2, xNa2O, and xCO2 weight percentages are normalized from the values in Table 7 so that the values of CO2 wt% are not negative. The wetting angles increases substantially to an average of 22.9±3.4º for droplets of 71 wt% SiO2 (Figure 13). Figure 14 shows the profiles of the drops of 62 wt% SiO2 and 71 wt% SiO2 in which the change in wetting angle is evident. Droplets of 80 wt% SiO2 did not relax into a round sessile drop so wetting angle measurements were not taken (Figure 15).         Figure 14. Profiles of droplets of 62 wt% SiO2 (left) and 71 wt% SiO2 (right) forming symmetrical sessile drops on alumina substrate with different wetting angles.  Figure 15. Profile of a droplet of 80% SiO2 forming an asymmetrical sessile drop on alumina substrate. 26        Figure 13. Contact angles of compositions with varying bulk compositions of silica and carbonate.   27   Experiment No.Temp (ºC)TimexSiO2 (wt%)xNa2O (wt%)xCO2 (wt%)θS.D.Average θS.D of θAvgKH_14_2810002min058.941.130.430.4KH_14_3110001min11.4764.6523.882.50.52.50.5KH_14_2710003hr36.9150.9612.138.52.18.52.1KH_14_3310001hr54.845.209.80.98.15KH_14_4710001hr54.845.2010.50.92.479919354KH_14_5010001hr54.845.205.10.6KH_14_5110001hr54.845.207.21.28.22.5KH_14_3610001hr62.337.706.52.45.25KH_14_5210001hr62.337.7040.55.31.81.767766953KH_14_3910001hr7129019.80.5KH_14_4910001hr7129026.50.6KH_14_5410001hr7129022.51.722.93.422.93333333KH_14_4610001hr802003.370954365KH_14_5610002hr80200Did not relaxDid not relaxTable 8.  Wetting angles on alumina substrate for droplets with re-calculated xSiO2-xNa 2O-xCO2 melt compositions. All experiments are run on alumina substrate.  xSiO2, xNa 2O, and xCO2 weight percentages are normalized from the values in Table 7 so the values of CO2 wt% are not negative.  28  Droplets of 11 wt% SiO2 relaxed into flat, homogeneous drops.  A phase separation of components was observed for droplets of 37 wt% SiO2. One of the phases is a melt that relaxes into a very flat droplet on the alumina, and the other phase is a droplet with a higher wetting angle that doesn’t fully relax. This correlates well with my assumption that these drops are composed of liquid Na2CO3 and solid Na2SiO3, which are likely the two components that separated. The phase separation creates an asymmetrical droplet, which is the reason for the higher variance of the wetting angle data. Because the phase separation affects the shape of the droplet, the wetting angle may not be an accurate representation of the surface tension and thus the data is considered invalid. Melts of 55 wt% SiO2, 62 wt% SiO2, and 71 wt% SiO2 created symmetrical droplets. The relatively high variance values of the wetting angles are due to varying wetting angles between droplets of each melt. This is attributed to possible external factors such as variations in the smoothness and heterogeneity of the alumina substrate, atmospheric fluctuations, and/or heterogeneity of the melts.  Changes in Surface Tension as a Function of Melt Composition   Previous studies report values of surface tension for Na2CO3 melts near their  liquidus of ~216mN/m (Hseih and Selman, 2010; Wolff, 1994; Igarashi et al., 1992). If wetting angle is the only physical property being tracked as a proxy for surface tension, then adding SiO2 to Na2CO3 does not greatly affect the surface tension in my experiments until the droplet has a composition of over 62 wt% SiO2.  I propose the explanation that Na2CO3 is the controlling component of the wetting angles of the Na2CO3-SiO2 system up to 62 wt% SiO2. Na+ atoms in Na2SiO3 are not strongly bonded to any particular atom, much like Na+ atoms in Na2CO3 (Ching and Murray, 1983). There are two types of oxygen atoms in Na2SiO3: non-bridging oxygens (NBOs) and bridging oxygens (BOs). When Na2O and SiO2 react, Si-O bonds are broken, creating two NBOs. Each one of these NBOs bonds to one Si4+ atom and six Na+ atoms. The connection that the NBO provides between these two atoms is fairly weak and is not highly influential in the structure of the tetrahedral network (Stebbins and Xu, 1997). Each BO bonds to two Na+ atoms and two Si4+ atoms to link up and form chains of SiO4; these bonds are much stronger than the NBO bonds. The number of NBOs 29  per SiO4 tetrahedron (NBO/T) has a strong influence on the properties of melts. A higher NBO/T value means that each silica tetrahedron has a large number of NBOs compared to BOs, and thus fewer strong bonds. The values of NBO/T of each melt composition are calculated from Giordano (2001) and shown in Table 9.    Table 9. Calculated values of NBO/T for each melt composition.   The calculated values in Table 9 show a decrease in wetting angle with an increase in NBO/T (Figure 14). This suggests that a melt with more NBOs/T will have weaker cohesive forces between the molecules, which governs the liquid surface tension and, in turn, dictates the wetting angle.   Figure 14. Results show a non-linear, inverse relationship for calculated NBO/T vs. 𝜃. Sample Label SiO2 Na2O CO2 NBO/T Average θCarb 5 0.00 58.90 41.10 Infinite 3CS4 11.47 64.65 23.88 22.296 2.5CS2 36.91 50.96 12.13 4.471 8.5CS3 54.80 45.20 0.00 1.6 8.2CS6 62.30 37.70 0.00 1.17 5.25CS7 71.00 29.00 0.00 0.794 22.9CS8 80.00 20.00 0.00 0.494 -Recalculated Compositions (wt%)30  Furthermore, the addition of BOs can drastically change the cohesion between molecules. This is because the introduction of SiO2 allows oxygens to form bridging bonds rather than non-bridging bonds. Each Si atom in a Na2Si2O5 molecule is bonded to three BOs and one NBO (Ching and Murray, 1983). Each Na2Si2O5 molecule has one more BO than a Na2SiO3 molecule; this oxygen is unique in that it links SiO4 tetrahedral chains into sheets. The structures of both of these molecules are show in Figure 15 and Figure 16.  Droplets of 71 wt% SiO2 have a stoichiometric composition of approximately Na1.8Si2.2O5, which is structurally very similar to a Na2Si2O5 molecule, and produce a significantly larger wetting angle. This suggests that melt structure for compositions between Na2SiO3 and Na1.8Si2.2O5 has evolved via polymerization to the point where there is a major shift in wetting angle of the melt and, thus, its surface tension. The formation of Na2Si2O5 sheets may be the reason for larger wetting angles of the Na2Si2O5 droplets. If this is the case, then the melt of 71 wt% SiO2 may have reached the composition of Na2Si2O5 required to significantly affect its surface tension.  Surface tension values for pure liquid silicon at melting temperature are reported to be 749-827mN/m, which is significantly higher than surface tension values of Na2CO3 melts (Millot et al., 2008). For compositions with considerably larger amounts of SiO2 than Na2O, the SiO4 unit is maintained as the structural backbone of the melt (Ching and Murray, 1983). Therefore, the surface tension increase with an increase in SiO2 wt% may also be attributed to the silica component.  31    Figure 15. Na2SiO3 molecule forming chains. BO indicates represents bridging oxygens and NBO represents non-bridging oxygens. Adapted from Ching and Murray (1983).   32    Figure 16. Na2Si2O5 molecule. One bridging oxygen per molecule binds tetrahedron chains into sheets. BO indicates represents bridging oxygens and NBO represents non-bridging oxygens. Adapted from Ching and Murray (1983).     33  Geological Implications  Kimberlite Pyroclastic Deposits  Kimberlite volcanism produces both pyroclastic deposits and pelletal lapilli. Pelletal lapilli are concentric, rounded crystal or crystalline melt fragments surrounded by a rim of juvenile melt (Gernon et al., 2012). A commonly accepted model, proposed by Gernon et al. (2012), attributes their formation to particles being incorporated into liquid spheres during the rise of kimberlite magma. They suggest that pelletal lapilli form during fluidized spray granulation, which occurs as a pulse of kimberlite magma is injected into unconsolidated pyroclastic deposits within the kimberlite pipe. The physical conditions required for uplift include a sudden pressure drop, gas exsolution, and flow velocities high enough to support the weight of the particles without ejecting them from the system. These properties are all characteristic of kimberlite eruptions, making this process highly plausible for kimberlite melts. For most studied kimberlite pyroclastic deposits, there is a positive correlation between the surface area of the coating and the surface area of the inner particle, which suggests a uniform coating process. The formation of rounded pyroclasts is due to relaxation of a liquid on a solid surface, which is controlled by liquid surface tension (Moss and Russell, 2011). Therefore, the elliptical geometry observed in pelletal lapilli strongly indicates that their formation is dependent on surface tension (Gernon et al., 2012).  Analogous to pelletal lapilli are kimberlite pyroclastic deposits, which can be comprised of olivine crystals with full or partial kimberlitic melt coats, olivine crystals without kimberlitic coats, and kimberlite pyroclasts without olivine crystals (Moss and Russell, 2011). Differences between these three deposits can be attributed to properties of eruption processes as well as properties of the melt.  High liquid-solid interfacial energies cause the kimberlite melt to wet olivine crystals. Thin coats of kimberlite melts on olivine crystals with low contact angles between the coat and melt suggest that the work of adhesion between the melt and the crystal (𝑊𝑎𝑑ℎ) exceeds the work of cohesion of the melt (i.e. liquid surface tension of the melt) (Figure 17).  34    Figure 17. Large olivine grains coated by thin selvages of kimberlitic melt in a kimberlite pyroclastic deposit. The contact angle is small (𝜃 <30º) indicating a high interfacial tension between the melt and the grain (Moss and Russell, 2011).   High liquid-solid interfacial energies also reduce the ability of the melt to separate from crystals during eruption.  For a gas to be able to separate a melt from a solid, its pressure applied must overcome the work of adhesion (𝑊𝑎𝑑ℎ) between the melt and the solid, defined as:  𝑊𝑎𝑑ℎ = 𝛾(1 + cos(𝜃))         (7)   where 𝛾 is the liquid surface tension and 𝜃 is the contact angle between solid and liquid (Moss and Russell, 2011). Therefore, a liquid with high surface tension requires more work to be stripped off of a solid grain.  The spherical geometry of pyroclasts and the contact angle between kimberlitic melt coats and olivine crystals are likely controlled by the surface energy between the melt and the gas, as well as the interfacial energy between the melt and the crystals. Quantifying the surface tensions of melts can help interpret kimberlite pyroclastic deposits to better understand kimberlite eruptions.   Interconnectivity of Carbonate Melt in the Upper Mantle   Low viscosities of carbonatitic and kimberlitic melts allow for high mobility in the mantle at low melt fractions (Watson et al., 1990). However, this mobility is only possible if the melt can form an interconnected network between solid grain boundaries of the mantle rock. The 35  distribution and interconnectivity of a melt depends on the relative solid-melt and solid-solid interfacial energies (Gaetani and Grove, 1999). These energies determine the dihedral angle (𝜗) which is the angle the melt forms where it meets two grain boundaries, shown in Equation (8):  𝛾𝑠−𝑠 = 2𝛾𝑠−𝑓 cos (𝜗2)      (8)  where 𝛾𝑠−𝑠 is the solid-solid interfacial energy, 𝛾𝑠−𝑓 is the solid-fluid interfacial energy (Minarik and Watson, 1995). These interfacial energies are highly controlled by the surface tension of the melt. Smaller dihedral angles increase the tendency for the melt to wet the grains at the junction. If 𝜗 is less than 60º, the melt is able to wet the grains to form an interconnected network, thus allowing for melt percolation through the rock (Gaetani and Grove, 1999). Previous experiments have determined the dihedral angle for Na2CO3-rich carbonate melts in an olivine matrix to be 25-30º (Hunter and McKenzie, 1989; Watson et al., 1990; Minarik and Watson, 1995). The interconnected melt network can be seen in Figure 18. The small dihedral angle between carbonate melts and olivine grains allows the melt to ascend through the upper mantle, where olivine is the dominant mineral. If the change in surface tension as carbonate melts ascend and become more kimberlitic can be quantified, then the mobility of melts through the upper mantle can be better understood.                         Figure 18. Image of dolomitic melt forming an interconnected melt network between olivine grains. The drawing image emphasizes small channels which are not clear in the image (Hunter and McKenzie, 1989). Scale bar 20um.   36  Recent studies have found carbonate melts to have conductivities of up to 3 orders of magnitude higher than molten silica (Gaillard et al., 2008; Ni et al., 2011). These findings indicate that electrically conductive regions in the mantle are due to small amounts of carbonate melt in peridotite, rather than silica. This suggests that the melts do indeed form interconnected liquid networks at olivine grain boundaries, even at low melt fractions. The presence of carbonate melts may also explain the CO2 content in parts of the upper mantle, which may improve the understanding of the carbon cycle (Gaillard et al., 2008). Carbonate melts are also important transporters of soluble elements and carbon-rich volatiles and transport properties influence their effectiveness as metasomatizing agents (Lee et al., 2000; Minarik and Watson, 1995). Therefore, constraining melt mobility can better the understanding of magmatic processes that involve carbonate melts.              37  CONCLUSIONS   I used the sessile drop method to measure the wetting angles of borosilicate and soda lime droplets as a proxy for surface tension. Results show that borosilicate drops have higher wetting angles than soda lime drops, which correlates to their respective surface tensions (i.e. borosilicate melts have higher surface tensions than soda lime melts). Temperature, drop size, and substrate all affect the wetting angle and my results are in agreement with previous studies. An experimental method that constrains these factors and keeps them constant was developed. This was applied to experiments with carbonate-silica melts to accurately determine the relationship between melt composition and wetting angle of the drops.  Sintering Na2CO3 powder with SiO2 powder produced melts of varying Na2O-SiO2-CO2 weight percentages, depending on the weight percentages of the starting materials. The SiO2-Na2CO3 phase diagram indicates that liquid and solid Na2CO3, solid Na2SiO3, liquid Na2Si2O5, and solid SiO2 are the phases that can be produced (with exsolved CO2 gas).  Wetting angles were measured and non-bridging oxygens/SiO4 tetrahedral (NBO/T). Wetting angles did not change drastically for droplets of melt compositions up to 62 wt% SiO2. Wetting angles increased for droplets of 72 wt% SiO2. NBO/T values were calculated to decrease with increasing SiO2 wt%.  I propose that Na2CO3 is the controlling component of the surface tension of the melts until the melt composition contains over 62 wt% SiO2. Na2CO3 molecules have weak Na+ bonds of and a high NBO/T value. The molecular structure and weak interactions likely contribute to the low surface tension, and thus small wetting angles, of Na2CO3-rich melts.  Once the melt composition exceeds 62 wt% SiO2, Na2Si2O5 likely controls the surface tension of the melt via the introduction of a bridging oxygen that links SiO4 tetrahedral into chains. This change in interactions between the liquid molecules is likely what increases liquid surface tension and the wetting angle.        38  REFERENCES   Amirfazli, A., Chatain, D., and Neumann, A. W. (1998). Drop size dependence of contact angles for liquid tin on silica surface: Line tension and its correlation with solid–liquid interfacial tension. Colloids and Surfaces A: Physicochemical and Engineering Aspects, 142(2–3), 183-188. Amirfazli, A., and Neumann, A. W. (2004). Status of the three-phase line tension: A review. Advances in Colloid and Interface Science, 110(3), 121-141. Ching, W. Y., and Murray, R. A. (1983). Comparative studies of electronic structures of sodium metasilicate and α and β phases of sodium disilicate. Physical Review B, 28(8), 4724-4735. doi:10.1103/PhysRevB.28.4724 Dasgupta, R., Mallik, A., Tsuno, K., Withers, A. C., Hirth, G., and Hirschmann, M. M. (2013). Carbon-dioxide-rich silicate melt in the earth's upper mantle. Nature, 493(7431), 211. David, R., and Neumann, A. W. (2014). A theory for the surface tensions and contact angles of hydrogen-bonding liquids. Langmuir : The ACS Journal of Surfaces and Colloids, 30(39), 11634. Fatollahi, A. H., and Hajirahimi, M. (2013). Making sessile drops easier. Fujii, H., Matsumoto, T., Izutani, S., Kiguchi, S., and Nogi, K. (2006). Surface tension of molten silicon measured by microgravity oscillating drop method and improved sessile drop method. Acta Materialia, 54(5), 1221-1225. doi:10.1016/j.actamat.2005.10.058 Gaetani, G. A., and Grove, T. L. (1999). Wetting of mantle olivine by sulfide melt: Implications for Re/Os ratios in mantle peridotite and late-stage core formation. Earth and Planetary Science Letters, 169(1–2), 147-163. Gaillard, F., Malki, M., Iacono-Marziano, G., Pichavant, M., and Scaillet, B. (2008). Carbonatite melts and electrical conductivity in the asthenosphere. Science, 322(5906), 1363-1365. doi:10.1126/science.1164446 Genge, M., Price, G., and Jones, A. (1995). Molecular-dynamics simulations of caco3 melts to mantle pressures and temperatures - implications for carbonatite magmas. Earth and Planetary Science Letters,131(3-4), 225-238. doi:10.1016/0012-821X(95)00020-D Goleus, V. I., Belyi, A. Y., Sardak, M., and Belyi, Y. I. (1996). Calculation of the surface tension of molten borosilicate glasses. Glass and Ceramics, 53(8), 226-228. doi:10.1007/BF01213775 Hamuyuni, J., Taskinen, P., Akdogan, G., and Bradshaw, S. M. (2012). Measurement of surface tension of molten matte phases by an improved sessile drop method. Mineral Processing and Extractive Metallurgy,121(3), 173-177. doi:10.1179/1743285512Y.0000000013 HRMA, P. (1985). Reaction between sodium carbonate and silica sand at 874°C < T < 1022°C. Journal of the American Ceramic Society, 68(6), 337-341. doi:10.1111/j.1151-2916.1985.tb15236.x Hsieh, P., and Selman, J. R. (2010). A corresponding states method for the prediction of surface tension of molten carbonate mixtures. Journal of the Electrochemical Society, 157(6), F61. doi:10.1149/1.3371415 Huang, T. S., Rahaman, M. N., Eldred, B. T., and Ownby, P. D. (2001). Wetting of mullite by Y2O3–Al2O3–SiO2 and B2O3–SiO2 glasses. Journal of Materials Research, 16(11), 3223-3228. doi:10.1557/JMR.2001.0444 39  Hunter, R. H., and McKenzie, D. (1989). The equilibrium geometry of carbonate melts in rocks of mantle composition. Earth and Planetary Science Letters, 92(3), 347-356. doi:10.1016/0012-821X(89)90059-9 Igarashi, K., Tajiri, K., Asashina, T., Kosaka, M., Iwadate, Y., and Mochinaga, J. (1992). Surface tension around eutectic compositions of molten alkali carbonate mixtures. Zeitschrift Für Naturforschung A, 47(5), 675-677. doi:10.1515/zna-1992-0506 Jones, A. R., Winter, R., Florian, P., and Massiot, D. (2005). Tracing the reactive melting of glass-forming silicate batches by in situ 23Na NMR. The Journal of Physical Chemistry.B, 109(10), 4324. Jones, A. R., Winter, R., Greaves, G. N., and Smith, I. H. (2005). 23Na, 29Si, and 13C MAS NMR investigation of glass-forming reactions between Na2CO3 and SiO2. The Journal of Physical Chemistry.B, 109(49), 23154. Jones, A.P., Genge, M., and Carmondy, L. (2013). Carbonate melts and carbonatites. Reviews in Mineralogy and Geochemistry, 75, 289-322. Kimura, S., and Terashima, K. (1997). A review of measurement of thermophysical properties of silicon melt. Journal of Crystal Growth, 180(3-4), 323-333. doi:10.1016/S0022-0248(97)00263-7 Lee, W. J., Huang, W. L., and Wyllie, P. (2000). Melts in the mantle modeled in the system CaO-MgO-SiO2-CO2 at 2.7 GPa. Contributions to Mineralogy and Petrology, 138(3), 199-213. doi:10.1007/s004100050557 Li, J., Yuan, Z., Qiao, Z., Fan, J., Xu, Y., and Ke, J. (2006). Measurement and calculation of surface tension of molten Sn–Bi alloy. Journal of Colloid and Interface Science, 297(1), 261-265. Millot, F., Sarou-Kanian, V., Rifflet, J., and Vinet, B. (2008). The surface tension of liquid silicon at high temperature. Materials Science and Engineering: A, 495(1–2), 8-13. Minarik, W. G., and Watson, E. B. (1995). Interconnectivity of carbonate melt at low melt fraction. Earth and Planetary Science Letters, 133(3), 423-437. doi:10.1016/0012-821X(95)00085-Q Moss, S,. and Russell, J.K. (2011). Fragmentation in kimberlite: products and intensity of explosive eruption. Bull Volcanology, 79, 983-1003. Ni, H., Keppler, H., and Behrens, H. (2011). Electrical conductivity of hydrous basaltic melts: Implications for partial melting in the upper mantle. Contributions to Mineralogy and Petrology, 162(3), 637-650. doi:10.1007/s00410-011-0617-4 Parikh, N. M. (1958). Effect of atmosphere on surface tension of glass. American Ceramic Society, Journal, 41(1), 18-22. doi:10.1111/j.1151-2916.1958.tb13497.x Shishkin, A., and Basin, A. (2004). Surface tension of liquid silicon. Theoretical Foundations of Chemical Engineering, 38(6), 660-660. doi:10.1007/s11236-005-0043-2 Stebbins, J. F., and Xu, Z. (1997). NMR evidence for excess non-bridging oxygen in an aluminosilicate glass. Nature, 389(6655), 60. Watson, E.B., Brenan, J.M., and Baker, D.R. (1990). Distrubution of fluids in the continental lithospheric mantle, in M.A. Menzies (ed.), The Continental Lithospheric Mantle, Oxford University Press, Oxford, 1745-1758.  Wing, D. R. (2003). Factors influencing the density and surface tension of soda lime silica melts containing multivalent ions. ProQuest, UMI Dissertations Publishing. 40  Wolff, J. A. (1994). Physical properties of carbonatite magmas inferred from molten salt data, and application to extraction patterns from carbonatite–silicate magma chambers. Geological Magazine, 131(2), 145-153. doi:10.1017/S0016756800010682 Wyllie, P. J. (1980). The origin of kimberlite. Journal of Geophysical Research, 85(B12), 6902-6910. doi:10.1029/JB085iB12p06902 Yuan, Z. F., Mukai, K., Takagi, K., Ohtaka, M., Huang, W. L., and Liu, Q. S. (2002). Surface tension and its temperature coefficient of molten tin determined with the sessile drop method at different oxygen partial pressures. Journal of Colloid and Interface Science, 254(2), 338-345. Zhou, Z., Mukherjee, S., and Rhim, W. (2003). Measurement of thermophysical properties of molten silicon using an upgraded electrostatic levitator. Journal of Crystal Growth, 257(3), 350-358. doi:10.1016/S0022-0248(03)01430-1 41  APPENDIX   Appendix 1. Size and mass measurements of cores and wetting angles θ of droplets for borosilicate and soda lime experiments.      Experiment No. KH_14_15 KH_14_7 KH_14_19Sample No. Bo15 Bo7 Bo19Substrate Alumina Disk Alumina Disk Alumina DiskComposition Borosilicate Borosilicate BorosilicateAvg. Core Height (cm) - 0.52 0.50Avg. Core Diameter (cm) - 0.50 0.51Avg. Core Weight (g) 0.033 0.210 0.240Temperature (°C) 1000 1050 1100Time (hr) 5 5 5θ From Photos 140.5+-1 102.3+-0.5139.25+-1.5 104.1+-.7140.5+-1 100.6+-.4141.75+-1.26 98.7+-1.2139.67+-.58 99.5+-.7140+-.82 99.4+-2139.5+-.58 101.7+-.14141.5+-1.29 101.9+-.6Avg. θ Did not relax 140.33 101.00S.D. 0.91 2.0042   Experiment No. KH_14_2 KH_14_1 KH_14_3Sample No. SL2 SL1 SL3Substrate Alumina Disk Alumina Disk Alumina DiskComposition Soda Lime Soda Lime Soda LimeAvg. Core Height (cm) 0.51 0.48 0.52Avg. Core Diameter (cm) 0.48 0.49 0.50Avg. Core Weight (g) 0.230 0.220 0.230Temperature (°C) 1000 1050 1100Time (hr) 5 5 5θ From Photos 55+-1.63 37.25+-1.71 4.5+-0.455.75+-1.26 37.75+-.96 5.8+-0.548.75+-.96 39+-.82 5.3+-0.749.25+-.96 38.5+-.58 3.7+-0.350.25+-.58 38.25+-.5 4.5+-0.351.5+-1.29 41.5+-.58 5.4+-0.650.5+-1.73 40.5+-1.29 5.5+-0.752+-1.63 5.3+-0.2Avg. θ 51.63 38.96 5.00S.D. 2.55 2.00 0.6543   Experiment No. KH_14_10 KH_14_6 KH_14_9Sample No. Bo10 Bo6 SL9Substrate Alumina Disk Alumina Disk Alumina DiskComposition Borosilicate Borosilicate Soda LimeAvg. Core Height (cm) 0.50 0.49 0.49Avg. Core Diameter (cm) 0.49 0.53 0.48Avg. Core Weight (g) 0.195 0.220 0.219Temperature (°C) 1050 1050 1050Time (hr) 3 3 3θ From Optical Tensiometer 32.901+-.03133.493+-.03332.977+-.05233.723+-.03833.922+-.03133.315+-.058Avg. θ 33.39S.D. 0.40θ From Photos 139.33+-3.21140.9+-.3 138.67+-.58140.7+-.7 140+-.3139.4+-.8 141.5+-.71144.7+-.36 137.1+-.45140.76+-.7 140+-1.73139.9+-2141+-1.7Avg. θ 141+-0.9 139.30S.D. 2.00 1.6644   Experiment No. KH_14_11 KH_4_4 KH_14_12Sample No. Bo11 Bo4 SL12Substrate Alumina Disk Alumina Disk Alumina DiskComposition Borosilicate Borosilicate Soda LimeAvg. Core Height (cm) - 0.97 -Avg. Core Diameter (cm) - 1.00 -Avg. Core Weight (g) 0.022 1.707 0.027Temperature (°C) 1050 1050Time (hr) 5 3θ From Optical Tensiometer 149.078+-0.0415 34.934+-.0814148.227+-.0437 34.316+-.303132.3+-.12 35.257+-1.816129.2+-1.29 35.262+-1.196146.899+-.041 32.909+-.0404145.987+-.0413 33.674+-.0398134.3+-1.45139.29+-.15Avg. θ 147.55 34.3905S.D. 1.37 0.951θ From Photos 137.00 140+-2.4135.00 141.7+-0.6148.00 143.2+-4146.20 137.5+-2.7136.33+-0.6136.8+-0.4141.55 142.6+-0.66.50 141.9+-0.7Avg. θ 140.00S.D. 4.0045   Experiment No. KH_14_5 KH_14_13 KH_14_17Sample No. SL5 Bo13 Bo17Substrate Alumina Disk Dunite DuniteComposition Soda Lime Borosilicate BorosilicateAvg. Core Height (cm) 0.98 - -Avg. Core Diameter (cm) 1.00 - -Avg. Core Weight (g) 1.95 0.03 0.03Temperature (°C) 1050 1050 1050Time (hr) 3 5 5θ From Optical Tensiometer 64.12+-.138 81.890+-1.31 78.140+-.42955.01+-.169 81.890+-.473 77.756+-1.48460.99+-.165 90.123+-1.091 70.291+-.28961.14+-.033 87.666+-.4837 91.653-.07063.88+-.568 83.414+-.568 92.170+-.05257.5+-.0285 82.021+-.858 88.390+-.08164.89+-.12 88.390+-.08161.5+-.01456.31+-.12Avg. θ 60.04 84.50 83.83S.D. 3.41 3.23 8.42θ From PhotosAvg. θS.D.46   Samples with no height or diameter measurements are glass pieces weighed to a mass corresponding to an approximate size of 0.25x0.25mm. Either the optical tensiometer or photograph method was used to measure contact angle.           Experiment No. KH_14_14 KH_14_16 KH_14_18Sample No. SL14 SL16 SL18Substrate Dunite Dunite DuniteComposition Soda Lime Soda Lime Soda LimeAvg. Core Height (cm) - - -Avg. Core Diameter (cm) - - -Avg. Core Weight (g) 0.03 0.04 0.03Temperature (°C) 1050 1050 1050Time (hr) 5 5 5θ From Optical Tensiometer 18.756+-.642 18.362+-.397 18.346+-.38520.155+-.109 19.392+-.886 19.336+-.08817.317+/.828 18.044+-.030 22.582+-.42219.196+-.763 20.540+-035 20.529+-.07318.574+-.028 18.564+-.099 19.992+-1.01318.428+-.026 18.564+-.099 18.935+-.11119.468+-.036 18.335+-.946 16.919+-.59916.957+-.037 19.932+-.414 16.670+-.20417.859+-.567 19.02+-.949 19.563+-.76918.739+-.148Avg. θ 18.57 19.02 19.16S.D. 1.03 0.95 1.72θ FromPhotosAvg. θS.D.47  Appendix 2. Mass measurements for carbonate-silica droplets before and after melting and wetting angle measurements.           Experiment No. KH_14_28 KH_14_31 KH_14_27Melt Carb5 CS4 CS2Substrate Alumina Alumina AluminaTemperature (°C) 1000 1000 1000Time 1min 45sec 1min 3hr% SiO20.00 11.47 36.91%Na2O 58.90 64.65 50.96%CO241.10 23.88 12.13Mass Droplet 0.040 0.043 0.031Mass After Melting 0.039 0.042 0.028Mass Loss 0.001 0.001 0.003% Loss 1.763 0.011 9.936θ Measurements 4.8+-.6 3.2+-.6 7.1+-1.12.9+-.4 2.6+-1 10.5+/-2.13.2+-1 3.4+-.6 7.8+-.92.6+-2.7 2.8+-.9 6.9+-.62.7+-.4 2.9+-.8 9.9+-13.6+-.2 3.1+-.4 6.8+-1.23.5+-.2 2+-.5 7.6+-.42.7+-.3 3.7+-.4 11.1+-.6Average θ 3 2.9 8.5S.D. 1 0.6 1.7548     Experiment No. KH_14_33 KH_14_47 KH_14_50Melt CS3 CS3 CS3Substrate Alumina Alumina AluminaTemperature (°C) 1000 1000 1000Time 1hr 1hr 1hr% SiO254.80 54.80 54.80%Na2O 45.20 45.20 45.20%CO20.00 0.00 0.00Mass Droplet 0.035 0.035 0.034Mass After Melting 0.068 0.035 0.032Mass Loss -0.033 0.001 0.002% Loss -0.306 0.007 0.014θ Measurements 9.85+-.5 18.1+-.4 4+-.29.7+-.9 21.8+-.6 5+-.469.1+-.6 18.9+-.33 5+-19.3+-.3 14.7+-.87 5.5+-.559.9+-6 18+-.34 5.8+-.2310.7+-.9 17.9+-1.3 5.3+-.5510.1+-.9 19.1+-1.6 5.+-.610.2+-.6 18.2+-1.4 4.7+-.54Average θ 9.8 18.3 5.06S.D. 0.5 1.9 0.649    Experiment No. KH_14_51 KH_14_36 KH_14_52Melt CS3 CS6 CS6Substrate Alumina Alumina AluminaTemperature (°C) 1000 1000 1000Time 1hr 1hr 1hr% SiO254.80 62.30 62.30%Na2O 45.20 37.70 37.70%CO20.00 0.00 0.00Mass Droplet 0.039 0.031 0.025Mass After Melting 0.040 0.031 0.023Mass Loss -0.002 0.000 0.002% Loss -0.014 0.002 0.018θ Measurements 7.8+-.9 6.3+-.5 4.2+-.338.8+-.6 8.7+-2.4 3.5+-.47+-.3 6+-.7 4.1+-.25.2+-.62 5.9+-.8 3.4+-.36.9+-.7 5.8+-.6 4.9+-.57.4+-.8 6.5+-.2 4.4+-.46.4+-.5 6.2+-1.2 4+-.47.8+-.9 6.5+-.4 3.5+-.3Average θ 7.16 6.5 4S.D. 1.2 0.9 0.5250           Experiment No. KH_14_39 KH_14_49 KH_14_54Melt CS7 CS7 CS7Substrate Alumina Alumina AluminaTemperature (°C) 1000 1000 1000Time 1hr 1hr 1hr% SiO271.00 71.00 71.00%Na2O 29.00 29.00 29.00%CO20.00 0.00 0.00Mass Droplet 0.035 0.030 0.037Mass After Melting 0.102 0.031 0.036Mass Loss -0.067 -0.001 0.001% Loss -0.624 -0.009 0.010θ Measurements 20.63+-.95 27.4+-.8 24.8+-.2419.7+-0.6 26+-1.5 24.1+-.7819.8+-.9 26.5+-.3 21.4+-.2719.3+-1.2 27+-.4 23.7+-0920.4+-.7 25.2+-.9 21.1+-.0919.85+-1 24+-.1 20.2+-.6519.53+-1 26.5+-1.2 21.1+-.5726.1+-.6 23.3+-.34Average θ 19.8 26.1 22.46S.D. 0.5 1.1 1.7

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.52966.1-0053618/manifest

Comment

Related Items