Experimental methods for studying salt nucleation and growth from supercritical water

254

The Journal

of Supercritical

Fluids,

1991,4,

254-264

Experimental Methods for Studying Salt Nucleation and Growth from Supercritical Water? Fred J. Armellini

and Jefferson

W. Tester*

Chemical Engineering Department and Energy Laboratory, Massachusetts Institute of Technology, Room E40-455, Cambridge, MA 02139 Received May 20, 1991; accepted in revised form October 22, 1991

Experimental techniques have been developed and tested for examining phase behavior and precipitation in salt-water systems near and above the critical point of pure water (374 “C and 221 bar). These methods were conceived to address important issues relating to salt formation during Supercritical Water Oxidation (SCWO), which is an emerging waste treatment process. An experimental apparatus featuring an optically accessible cell was designed and constructed for operation up to 600 “C and 400 bar. The cell has been used to observe phase transformations and locate phase boundaries under isobaric conditions in the NaCl-H20 system at a pressureof 250 bar. A flow system has recently been incorporated into the cell for studying the rapid or “shock-like” precipitation of salts from supercritical water. Preliminary results from shock crystallization experiments on the NaCl-Hz0 and Na$S04-H20 systems at 250 bar suggest the importance equilibrium phase relationships have in determining the morphology and size of particles formed during the SCWO process. Keywords:

supercritical water, precipitation, salts, sodium chloride, sodium sulfate

MOTIVATION AND APPROACH Oxidation in a supercritical water medium is an efficient and clean way of treating many aqueous organic wastes,t-3 and it is also currently being considered as a means of waste disposal and water recycle on future long term space missions. 4~5 Supercritical water oxidation (SCWO) is defined as oxidation that takes place above the critical temperature and pressure of pure water, 374 “C and 221 bar (3205 psia), respectively. At typical process operating conditions of 500 to 600 “C and 250 bar, supercritical water has a low dielectric constant (<2) and dissociation constant (
is heated by a supercritical water recycle stream. This rapid or “shock-like” precipitation is a result of the extreme drop in salt solubility that occursas the feed stream becomes supercritical. For example, the approximate solubility of some common salts in supercritical water is 100 ppm for NaCl, 700 ppb for CaC12, and 2 ppb for CaS04 at 550 “C and 250 bar.6 The density of pure water at those conditions is 0.0786 g/cm3, and its dielectric constant is 1.3 (compared to a value of 79 at room temperature and pressure).7 Salts are contained as dissolved species in metabolic wastes and frequently in industrial aqueous wastes. They are also formed during the oxidation and subsequent neutralization of organic compounds containing chloride, fluoride, sulfur, and other heteroatoms. Examples of the latter include methylene chloride, dimeth-yl sulfoxide, and any chlorofluorocarbon (CFC). This paper describes experimental methods designed to obtain fundamental quantitative data on this unique precipitation phenomenon. In an earlier paper, we described a parallel modeling effort to characterize the thermodynamic properties of the NaCl-HZ0 system and also our initial experimental efforts8 Related research carried out by our group has dealt with the kinetics of oxidation reactions in supercritical water and is reported elsewhere.3s9-13

0 1991 PRA Press

of Supercritical

The Journal

Fluids,

Vol. 4, No. 4, 1991

Salt Nucleation and Growth

255

A) Type I system Initial flow experiments have also been performed to characterize salt formation at temperatures, pressures, and supersaturation values similar to those found in the actual SCWO process. In the flow experiments, shock crystallization is simulated by injecting a cool aqueous salt solution into a coaxially flowing supercritical-water stream at a constant pressure above the critical pressure of pure water (221 bar). These experiments have been used to identify important mechanisms of salt growth and estimate particle sizes of the solids formed in the process.

t

/ /, / / ..--.

2 5 $ fz

pure water (w) ,/‘.---*’ cp / **.= .e-tp ..=-v + s, ..-i/ E -pure salt (s)

Temperature B) Type II system

, , ,*.‘“. .’ q

/ --*

*\

\

-. I*

\

\

,

‘+**V +L : \ : ‘CP :

: :

J tP pure salt (s)

Temperature Figure 1. Pressure-temperature projections of phase diagrams for two types of binary salt-water systems (adapted from Morey, reference 14). (w), water; (s), salt; (cp), critical end point; (tp), triple point; (V), vapor phase; (L), liquid phase; (S), solid phase; (E), eutectic point; (- - -), critical curve (V=L); (***** ), three phase curve (V+L+S,); (p), lower critical point; (q), upper critical end point. For this study, an apparatus has been constructed for high-temperature and high-pressure operation in salt-water systems near and above the critical point of pure water. The main component of the apparatus is an optically accessible cell, which has been used to examine phase behavior and precipitation phenomena. This initial study has involved the binary systems of water and sodium chloride and water and sodium sulfate as these salts are commonly found in waste streams. The effect of other components, such as a second salt, organics, or oxidants, will be considered in future experiments. An experimental procedure has been developed for examining phase relationships in salt-water systems at constant pressure. This is desirable, since the oxidation and salt separation steps in the SCWO process are essentially isobaric, and testing in this mode is warranted. Furthermore, salt formation mechanisms may depend on phase behavior. To evaluate this technique, static isobaric experiments have been performed on the NaCl-HZ0 system at the typical process pressure of 250 bar.

PHASE RELATIONSHIPS AND PROPOSED PRECIPITATION MECHANISMS Phase behavior encountered in salt-water systems at high temperatures and pressures is complicated by twophase, vapor-liquid regions, three-phase equilibrium points, and critical solution points. The conditions required for different phase equilibria are particularly important in the SCWO process near the reactor inlet where extreme temperature gradients exist. Phase equilibrium relationships in binary salt-water systems at high-temperatures and pressures can be divided into two general types.14g15Two-dimensional pressure (P)-temperature (7’) projections of the three-dimensional (P, T, composition) phase space for the two-system types are shown in Figure 1. In the Type I system, salt solubility continuously increases with temperature along the solid-vapor-liquid saturation curve until the melting point of the salt is reached. Type I systems also exhibit a continuous critical locus as shown on the pressure-temperature projection of the phase diagram in Figure la. This curve extends from the critical point of pure water to the critical point of the pure salt. The critical point of NaCl, which is a Type I salt, has been estimated to occur at 3900 “C and 260 bar.16 Other common Type I salts include LiCl, KCl, and KF.15 In Type II binary salt-water systems, the salt composition in the saturated liquid-phase (in equilibrium with solid salt and a vapor phase) falls to negligible values as the critical point of water is approached. In these systems, the critical curve intersects the saturation curve at two points 0, and 9 in Figure lb). Type II salts include Na$04, K2S04, and Na2C03.15 For the Na2S04-HZ0 system, the lower critical curve end point, p, is very close to the critical point of pure water, while the end point of the upper critical curve, q, is approximately 440 “C and 1140 bar.” In this study, the Type I salt-water system of interest is sodium chloride-water. This system is very important to the SCWO process, and there is an extensive thermodynamic data base available for it. An isobaric section of the phase diagram of the NaCl-HZ0 system at 2.50 bar is shown in Figure 2 (note that the NaCl concentration axis has a logarithmic scale). The experimental points shown in the figure were interpolated from the isothermal data of the referenced researchers. 19,20,22,24,25 A key feature of the

256

Armellini

and Tester

vapor-solkl

The Journal

of Supercritical

Fluids,

Vol. 4, No. 4, 1991

regbn

f L400

fluid-solid salt region

\ IL l-

bn one phase flukl ragbn

#/

200

100 0.01

0.1

1 wt% NaCl

10

at this pressure

one phase fluid region \

1

100

Figure 2. Temperature-composition NaCl-H,O phase diagram at 250 bar. (----), compilation of Bischoff and Pitzer (1989)‘*; (0), Sourirajan and Kennedy (1962)r9; (W), Bischoff et al., (1986)20; (O), estimated from Linke, (1958)2’; (A), Parisod and Plattner (1981)22; (a), prediction of Pitzer and Pabalan (1986)23; (V), alander and Liander ( 1950)24; (+), Khaibullin and Borisov (1964)2s. phase diagram

4oo /

is the existence

of a two-

phase, vapor-liquid region at temperatures above 374 “C, the critical point of pure water. A critical solution point, where the vapor- and liquid-phase compositions are identical, occurs at approximately 387 “C (at the tip of the vapor-liquid dome). At a temperature of 450 “C, threephases are in equilibrium (vapor, liquid, and pure solid salt) corresponding to a point on the univariant saturation pressure curve at 250 bar. The saturated vapor has a salt concentration of approximately 0.03 wt % NaCl and a density of 0.11 g/cm3 as predicted by the Pitzer-Tanger equation-of-state. 26 The saturated-liquid phase has a density of approximately 1.1 g/cm3 and a salt content of about 50 wt %,27 Between 450 and 700 “C, only very dilute salt concentrations are found in the supercritical fluid phase at equilibrium. The phase behavior of the Type II sodium sulfate-water system, also of major importance to the SCWO process, is not as well characterized as the sodium chloridewater system. The extensive study of Ravich and Borovaya17 on this system has established that, at a pressure of 250 bar and at temperatures of interest in the SCWO process, only one fluid phase exists in equilibrium with solid salt. Thus, this system exhibits much different phase behavior than the NaCl-HZ0 system at the process conditions. An isobaric slice of the phase diagram for the NaS04-H20 system at 250 bar is shown in Figure 3. A key feature of the phase diagram is the rapid falloff in solubility of Na2S04 as temperatures approach and pass the critical temperature of pure water. As Figure 3 shows, there is a large discrepancy in the reported solubility of Na$04 in water at 500 “C. The interpolated data points from Morey and Hesselgesser2s and Martynova6%29differ

wt?/. N%SO, Figure 3. Temperature-composition diagram at 250 bar.

Na,SO,-H,O

phase

by over three orders of magnitude. At temperatures above 700 “C, a vapor-liquid equilibrium region should appear.17 This region is not shown in Figure 3, since these conditions are not typically reached in the SCWO process. As a first approximation, binary salt-water phase relationships can be used to devise possible mechanisms for solid salt formation during the SCWO of a waste stream containing a single dissolved salt. Using the NaCl-HZ0 system as a simplified model for a waste stream containing sodium chloride, a possible mechanism for shock crystallization is that a highly concentrated aqueous brine serves as a precursor to solid precipitation in the process. As a cool NaCl-H20 aqueous feed solution is rapidly mixed with a hot supercritical water stream, it will be simultaneously heated and diluted. Even if one assumes a practical dilution factor of ten, a feed stream initially containing a few weight percent of NaCl will still penetrate the vapor-liquid region shown in Figure 2 as it is isobaritally heated at 250 bar. Assuming equilibrium, this will create concentrated liquid droplets, which will become more concentrated as the solution gets hotter. During heating, the volume of the liquid phase will diminish due to its increasing salt composition. Above 450 “C, the liquid droplets will become unstable resulting in the formation of solid salt in the presence of a dilute, low density vapor phase. For a model waste stream containing sodium sulfate, the mechanism of solid formation should be distinctly different from that described for sodium chloride. As a cool Na2S04-H20 feed is rapidly mixed with a supercritical water stream, solid precipitation can occur directly from the homogeneous waste stream without first forming a concentrated liquid brine phase. This difference in phase behavior between the sodium chloride-water and sodium sulfate-water systems may affect the morphology and size

Salt Nucleation and Growth

The Journal of Supercritical Fluids, Vol. 4, No. 4, 1991

257

Flow Tuba

Copper Gasket metering

S-hire

Window

back

relief

vaivs

12.7cm. (5.09

l-

HdNe Laser

m

Window Holder Retainer

ceramic insulation

Figure 5. Schematic of apparatus for static isobaric equilibrium experiments.

FlowTube Retainer

Figure 4. Cross-sectional view of high-temperature high-pressure optically accessible cell.

lnconel625~ Optical cell

and

of the formed solids as well as precipitation rates. Salt formation may also be affected by transport rates, intrinsic nucleation and growth kinetics, and metastable regions. In the actual SCWO process, components such as dissolved gases and organic wastes form a significant percentage of the reactor feed medium. The mechanisms of solid salt formation in these mixtures may also depend on their specific phase relationships. OPTICALLY ACCESSIBLE HIGH-TEMPERATURE AND HIGH-PRESSURE CELL An optical cell was designed and constructed in cooperation with Harwood Engineering (Walpole, MA) for direct in situ measurements of phase equilibria and solid salt formation in supercritical water. The cell has been operated successfully up to temperatures of 600 “C and pressures of 300 bar. The cell’s body is constructed out of a corrosion resistant, high strength alloy, Inconel 625, and the windows are made out of optically clear synthetic sapphire (a-A120s). Similar view cells have been used in the past by Franck’s group at the University of Karlsruhe in Germany.30x31 A cross-sectional diagram of the optical cell is shown in Figure 4. The cell is cubic in shape with an outer length of 12.7 cm (5 in.), and an inner volume of approximately 25 cm3. All six faces of the cell contain threaded ports, in which either Inconel 625 window holders, plugs, or flow tubes can be inserted. The sapphire windows have a diameter of 1.91 cm (0.75 in.), and are 0.953 cm (0.375 in.) thick. The actual diameter of the circular viewing area is 1.27 cm (0.5 in.). Clear circular glass plates are attached to the outer face of the window holder retainers to reduce heat losses from the window ports. Besides being used for solution delivery to and from the cell, the flow tubes are also used for internal pressure and temperature measurements. Copper gaskets are used to make the high-pressure seals at all the ports. During early testing of the cell, un

phase

desirable corrosion of the gaskets occurred, and this was eliminated by plating the copper rings with a thin layer of nickel and/or gold. The cell’s internal pressure seals the windows by mating the optically flat surface of the window with a similar surface in the window holder. Following suggested procedures used by Franck and his colleagues at Karlsruhe, a gold-foil spacer (0.025 mm thick) was compressed between the two flat surfaces to reduce the effects of any imperfections in the surfaces. The two thermowells shown in Figure 4 are used to measure the block temperature at both the top and bottom of the cell. In addition, thermocouples inserted through a flow tube can be placed directly in contact with the fluid contained in the cell. STATIC ISOBARIC EXPERIMENTS Apparatus and Procedures. A schematic of the experimental apparatus for examining, phase behavior in salt-water systems at extreme temperatures and pressures is shown in Figure 5. For the static experiments, only one flow tube is inserted in the top port of the cell, and it is used for temperature measurement and solution bleedoff during heating. The flow tube has an inner diameter of 0.318 cm and the thermocouple which passes through it has an outer diameter of 0.159 cm. Inconel 625 plugs are placed in the bottom and two side ports of the cell, while window holders are placed in the remaining two side ports (front and back). The cell is heated by twelve electrical resistance strip heaters (two attached to each outer face), and it is encased with approximately four inches of high temperature ceramic insulation (see Figure 5). A high performance liquid chromatography (HPLC) pump with a self-flushing head is used to pressurize the cell with the salt solution, and a micrometering valve is used for fine pressure adjustments. Pressure is measured with a transducer, which is rated to 517 bar (7500 psi) with an accuracy of f0.5 bar. As shown in Figure 5, optical equipment for light extinction measurements have been incorporated into the apparatus, which rests on an optical table to damp out any undesirable vibrations.

The Journal of Supercritical Fluids, Vol. 4, No. 4, 1991

Armellini and Tester

(4

(b)

Figure 6. View inside optical cell during an isobaric experiment at 250 bar with an initial NaCl concentration of 3.0 wt %. (a) Initial nucleation of vapor phase (average temperature = 390 “C). (b) Two-phase, vapor-liquid equilibrium (average temperature = 392 “C). In a typical static isobaric experiment, the cell is first filled at room temperature with a homogeneous solution of known salt concentration. The cell is then slowly heated at constant pressure. A phase transformation is identified visually or by light extinction. The temperature of the inner solution and of the cell body is recorded when the nucleation of the new phase is first identified. Since the salt solution expands as it is heated, the pressure can be kept constant by varying the bleed-off rate with temperature. A pressure control loop and a solenoid micrometering valve could possibly be used for this, but due to the small inner volume of our cell (25 cm3) and the nonlinear density properties of water, the required bleed-off rates would be too small and wide varying to make this method practicable. For example, in order to heat 25 cm3 of pure Hz0 isobarically at 250 bar and a rate of 1.5 “C/min, the calculated required average bleed-off rate changes from 0.022 g/min between 50 to 100 “C to a high of 0.51 g/min between 375 to 400 “C, and finally to 0.006 g/min between 550 to 600 “C. A simpler technique has been developed for conducting the isobaric experiments. During an experiment, the high-pressure pump is used to flow pure water through a straight length of tubing, while the pressure is held constant using a back pressure regulator. The optical cell is connected to the system using a high-pressure cross (as shown in Figure 5), and as the cell is heated the bleed-off combines with the pumped pure water. Since the flow rate of the pump is kept appreciably above the required bleed-off rate for isobaric heating, the total pressure in the system remains nearly constant. For a typical isobaric run with a heating rate between 1.0 and 2.0 “C/min, the pressure varies by only 1 to 2 bar if the flow rate of the

pure water is kept at 2.0 g/min.

Results and Discussion. Solutions of known sodium chloride concentration were heated isobarically at a rate of approximately 1 “C/min at a pressure of 250 bar. The initial concentration of the solutions ranged from 0.1 to 20.0 wt % NaCl. As each solution was heated from room temperature, it passed from the one-phase fluid region into the vapor-liquid region of the phase diagram (see Figure 2). Video taping was routinely carried out to document the experiments. In the isobaric runs with initial concentrations of 0.1 and 0.3 wt % NaCl, a sharp phase transformation was not visually observed in the cell. For these initial concentrations, a new phase with liquid-like density should have nucleated as the solutions were heated. Nucleation of the droplets most likely occurred on the inner walls of the cell, and then flowed downward out of the view provided by the windows in the cell. In the 0.3 wt % run, a haze slowly developed in the solution, but no distinct initial nucleation temperature could be obtained. For solutions of concentration of 1.0 wt % or greater, a well-defined phase transformation was identified visually, and appeared as a burst of very small vapor bubbles nucleated from the homogeneous fluid. The vapor bubbles seemed to emerge from the inner walls of the cell, which were hotter than the bulk solution because ttie cell was heated externally. Neglecting any significant penetration into the metastable region, the temperature at which the bubbles first appeared locates the vapor-liquid boundary in the phase diagram at the concentration of the initial solution. The initial nucleation of the vapor bubbles from a 3.0 wt % NaCl solution can be seen in Figure 6a. The dark region in the photos is caused by the scattering

The Journal of Supercritical Fluids, Vol. 4, No. 4, 1991

Salt Nucleation and Growth

259

set of nucleation. The average of these two experimental temperatures is plotted as the solid triangular points. Since nucleation occurred consistently on the inner walls of the cell, it is believed that the bulk solutiofl temperature and the block temperature provide an upper and lower bound for the actual phase nucleation temperature.

400

0.01

0.1

1 wt%Nacl

10

100

Figure 7. Results of isobaric experiments shown on the temperature-composition NaCI-H,O phase diagram at 250 bar. (v), present study; (--), compilation of Bischoff and Pitzer (1989)‘*; (0), Sourirajan and Kennedy (1962)19; (m), Bischoff et al., (1986)20; (@), estimated from Linke, (1958)*‘; (A), Parisod and Plattner (1981)**; (+), prediction of Pitzer and Pabalan (1986)23; (V), ijlander and Liander (1 950)24; (+), Khaibullin and Borisov ( 1964)25. of light from the small bubbles. After continued heating, the coexistence of the low-density vapor phase, which rose to the top of the cell, and the high-density liquid phase, which collected at the bottom of the cell, was observed. An example of this is shown in Figure 6b. The liquid phase appears dark because of the continuous nucleation of vapor bubbles which occurred as the solution was heated. The numbers in the photos are the run times, which were recorded by the video camera. Figure 7 shows the results of the isobaric runs at a pressure of 250 bar, and Table I lists the recorded nucleation temperatures. Our data are plotted as solid triangles with error bars. The bottom temperature limit corresponds to the measured bulk solution temperature as the initial vapor bubbles nucleated, and the top temperature limit corresponds to the cell block temperature at the on

Results

of Static

FLOW/SHOCK CRYSTALLIZATION EXPERIMENTS Apparatus and Procedures. The dynamic experiments are designed to simulate the rapid mixing of the supercritical water recycle stream and waste stream in the SCWO process. The two feeds in the experiments are a pure supercritical water (SCW) stream and a cool salt solution. Figure 8 shows the experimental apptiatus for the flow experiments. Two large flow tubes (0.912-cm i.d.) are placed in the top and bottom ports of the optical cell, and two window holders are placed in the front and back side ports to allow for visual observation and video recording. The other two side ports of the cell are closed with Inconel625 plugs. An electrical resistance heater is used to raise the temperature of the pure water feed to supercritical (upper right hand comer of Figure 8). The SCW stream then enters the optical cell through an annular space in the top flow tube. A low-pulse HPLC pump is used to deliver the salt solution to the cell. The solution is injected coaxially into the SCW stream as a jet using a nozzle, which passes down the center of the top flow nipple. The nozzle is cooled with either nitrogen or water to keep the solution’s temperature sub-critical and thus prevent solid salt from precipitating in the nozzle. The internal coolant flow for the nozzle is finely controlled to provide ample insulation of the salt solution jet, while not removing a significant amount of heat from the SCW feed stream. The inner diameter of the nozzle is 0.084 cm. Both nozzles with flat and conical shaped tips have been used in the flow experiments. Three primary temperature measurements are made

TABLE I Isobaric Experiments

at 250 bar

Nucleation Temperatures Initial NaCl Cont. (wt%)

Bulk Solution (“C)

0.1

0.3 1.0

3.0 5.0 10.0

20.0 * No distinct nucleation

temperature

was observed

Cell Block m

Average w>

*

*

*

* 384 385 385 388 395

* 389 395 392 396 399

* 386.5 390.0 388.5 392.0 397.0

260

Armellini and Tester

The Journal of Supercritical

Fluids, Vol. 4, No. 4, 1991

Salt Solution

SCW feed -

w/ 0.912 cm ID flow tub

-Nozzle Temp. thermocouple location

Window View

_ drain

t

Figure 8. Schematic of apparatus for flow/shock crystallization experiments. during a flow experiment. The temperature of the supercritical water feed is measured by a thermocouple in contact with the feed at the high-pressure cross above the cell. An estimate of the exit temperature of the salt solution jet is obtained using a small OD thermocouple (0.051 cm) which passes down the center of the nozzle to its tip. The tip of the small ID thermocouple is situated approximately 2 mm before the exit of the nozzle. Finally, an 0.159-cm OD thermocouple passes up through the bottom flow tube of the cell to measure the temperature of the mixed stream before it exits the inner chamber of the cell. All three thermocouples are Type K (ChromelAlumel), and have Inconel 600 sheathing. As in the static experiments, the block temperature of the optical cell is also monitored. Once the mixed stream exits the cell, it passes through a water-cooled heat exchanger, a pressure transducer, and a back pressure regulator. A micrometering valve is used to make fine pressure adjustments during a run. To begin a flow experiment, the two feed pumps are turned on, and the pressure of the system is raised using the back pressure regulator. The pure water feed heater and the cell’s strip heaters are then turned on to slowly heat the system until the desired temperatures are reached. During the heating period, pure water is fed though the nozzle. After steady state has been obtained, the nozzle feed is switched to the salt solution for a short time. At the end of an experiment, the pump feeds are stopped and the system pressure is slowly lowered using the micrometering valve. During cooling, the system is flushed with nitrogen to remove water vapor from the system, which

Mixed Temp. thermocouple

1 1.27 cm. (i/2 in.)

Figure 9. Schematic of cell interior during the flow/shock crystallization experiments. Nozzle tip is shown positioned near the center of the optical cell. TABLE II Experimental Conditions for Flow Runs Nozzle Tip at Top of Window Pressure: Nozzle: Jet Feed: SCW Feed: Mixed Stream:

with

248-25 1 bar 0.084-cm ID, conical tip flow = 0.5 g/min, T = 100-105 “C flow = 10.1 g/min, T = 550 “C T = 520-540 T I

when condensed would dissolve any formed solids. After reaching ambient temperature, the cell is opened, and the solid salt is collected for scanning electron microscopy (SEM) analysis. Results: Initially Mixed Region. This section describes the results of several exploratory shock crystallization runs conducted at a system pressure of 250 + 2 bar and a constant cell block temperature of 600 ‘C. In the first series of experiments, a nozzle with a conical shaped tip was used to inject salt solutions of various compositions into a coaxially flowing supercritical water stream. A schematic of the inside of the cell during these flow/shock crystallization experiments is shown in Figure 9. The supercritical water feed flowed along the outer surface of the nozzle, until it mixed with the nozzle feed. The placement of the nozzle tip at the top of the view provided by the windows in the optical cell enabled the observation of the initial mixing of the two streams. The conditions for the first set of flow experiments

The Journal of Supercritical Fluids, Vol. 4, No. 4, 1991

(4

Salt Nucleation and Growth

261

(b)

(4 Figure 10. View of various jets during a flow/shock crystallization experiment (see Table II for experimental conditions). (a) Pure H,O; (b) 3.0 wt % NaCl; (c) 3.0 wt % Na,SO,; (d) 10.0 wt % NaCl. are listed in Table II. With a SCW-to-jet feed ratio of 2O:l and additional heating by the cell body, the mixed stream temperature remained well above 500 “C. Figure 10 shows photos of the different jets tested. The photos were taken using 35-mm film and a camera shutter speed of l/500 set, with white light entering through the rear window of the cell for illumination. As shown in Figure 10, the pure water jet appeared fairly turbulent due to the density differences in the two feeds, while the 3.0 wt % NaCl jet had a similar apperirance. The 3.0 wt % Na$Od

‘jet and surrounding solution, on the other hand, appeared as turbulent, but much darker. Possibly, this was caused by small particles which nucleated homogeneously from the solution, scattering the light. Since the solution around the jet also appeared cloudy, there was some recirculation occurring during the experiment. The final jet shown in Figure 10d had an initial concentration of 10.0 wt % NaCl. This jet appeared much narrower and more focussed than the pure water jet, and can be seen breaking up in the photo. The difference in appearance between the

262

Armellini and Tester

The Journal of Supercritical Fluids, Vol. 4, No. 4, 1991

TABLE III Experimental Conditions for Flow Runs with Nozzle Tip in Top Flow Tube Pressure: Nozzle: Jet Feed: SCW Feed: Mixed Stream:

249-25 1 bar 0.084-cm ID, flat tip flow = 1.0 g/min, T = 105-140 “C flow = 10.1 g/min, T = 530 “C

T=540-550'C

sodium chloride and sodium sulfate jets is believed to be primarily a result of the different phase behavior exhibited by the two salts in supercritical water at a pressure of 250 bar. This hypothesis will be discussed further at the end of this section.

Results: Fully Mixed Region and Particle In the second set of flow experiments, Morphologies. the nozzle tip was located inside the top flow tube at a distance of 5.7 cm above the center of the cell. The object of these experiments was to collect solid salt samples for SEM analysis and, also, to observe the fully mixed region of the two feed streams. The conditions for this set of experiments are listed in Table III. In addition to the new nozzle location, the flow rate of the salt solution feed was increased to 1.0 g/min. Nozzle feeds of 3.0 wt % sodium chloride and 3.0 wt % sodium sulfate were studied. In the first run, there was no visible change in the appearance or the flow pattern of the mixed stream as the nozzle feed was switched from pure water to the 3.0 wt % NaCl solution. The mixed stream was clear except for some slight density fluctuation waves. The cell was then cooled while flowing nitrogen through the system. Upon opening the cell, white particles were found covering the bottom surface. Since the mixed temperature of the feeds did not fall below 550 “C, it is likely that the NaCl particles were formed from the bulk solution, and did not form heterogeneously at the bottom of the cell. At the conditions of the mixed stream, NaCl has a solubility in water of approximately 150 ppm. Therefore, a large supersaturation was quickly created as the injected stream was rapidly heated. In the second run, the mixed stream became extremely clouded when the nozzle feed was switched to the 3.0 wt % Na2S04 salt solution. After opening the cell, the bottom surface was again covered with white particles. One of the particles was fairly large. Since salt particles at these conditions tend to be “sticky,“3 the small particles which caused the turbidity most likely agglomerated together at the bottom of the cell. As with the NaCl runs, extreme levels of supersaturation were achieved during the experiment. Figure 11 shows SEM photomicrographs of the solids collected after the two runs described above. The NaCl solids shown in Figure 1la were grouped in clusters of highly amorphous kernel shaped particles with lengths

(4

(b) Figure 11. SEM photomicrographs of particles collected during shock crystallization experiments in supercritical water (see Table III for experimental conditions). (a) NaCl cluster; (b) Na,SO, solid.

between 10 and 100 ,nm. At higher magnification, the particles appeared to have many hollow inner regions. The Na2S04 solid shown in Figure 1 lb comprised many small particles which were fused together. Most of the primary particles were between 1 and 2 pm long. Interpretation of Flow Results. As stated earlier, the phase behavior of NaCl-HZ0 and Na2S04-HZ0 systems at 250 bar provides a possible explanation for the different visual appearance of the mixed streams and the morphology of the solids formed in the experiments described above. Entering the vapor-liquid region in the NaCl-HZ0 system most likely leads to the formation of concentrated liquid droplets, which become unstable once the three-phase temperature is reached. Unstable solid growth from these droplets could account for the amorphous morphology and relative large size of the NaCl

The Journal

of Supercritical

Fluids,

Vol. 4, No. 4, 1991

solids. These hypotheses based on the phase behavior shown in Figures 2 and 7 seem reasonable, but further testing is required before they can be confirmed. Conversely, the absence of a vapor-liquid region in the Na2S04-Hz0 system at the experimental conditions most likely resulted in direct homogeneous nucleation of many small particles from the injected jet. Nucleation would commence once the solubility curve was crossed, and continue as the jet was heated to the final mixed temperature, since the NazS04 solubility continuously decreases with temperature. Scattering of the incident light by the collected primary particles of diameter of approximately 1 pm would account for the extreme level of turbidity observed during the experiments. Since the salt mass flow rate was equal in both the NaCl and Na2S04 experiments and the average size of the Na2S04 particles was over an order of magnitude less than the NaCl particles, the number of NazS04 particles formed was much greater than the number of NaCl particles. This would explain why the mixed stream in the NaCl experiments appeared clear with a turbidity too low to be visually detected. CONCLUSIONS AND RECOMMENDATIONS This paper discussed phase relationships in the NaClHZ0 and Na2S04-H20 systems and proposed possible mechanisms of solid salt formation at conditions encountered in the SCWO of metabolic and industrial waste. An experimental apparatus employing a high-temperature and high-pressure optical cell was designed, constructed, and tested. Experimental methods were also developed for examining phase relationships and rapid precipitation in salt-water systems at high temperatures and pressures. A method for observing phase transformations and locating phase boundaries under isobaric conditions was tested on the NaCl-HZ0 system at a pressure of 250 bar. The method not only provides direct observation of phase changes and location of phase boundaries by visual means, it can also be used to verify interpolated data from traditional isothermal experiments which use sampling techniques. Future tests will be conducted on ternary systems with two salts to examine how the second salt (such as NazSO,) affects the isobaric phase diagram for the NaCl-Hz0 system. The flow experiments have for the first time simulated the shock crystallization of salts in supercritical water at temperatures, pressures, and salt concentrations encountered during the SCWO of industrial wastes. These experiments suggested possible mechanisms of salt formation in the SCWO process and, also, provided data on morphology and sizes of solids formed by rapid heating of a salt solution. Preliminary results identified the importance phase relationships in salt-water systems may have on determining the morphology and size of particles formed during the SCWO process. In future experiments, a wider range of conditions for precipitation will be exam-

Salt Nucleation and Growth

263

ined, and the use of in situ light scattering techniques for quantitatively determining particle size distributions will be explored. Theoretical methods for predicting temperature profiles in the jet would aid in the understanding of the flow experiments. These are currently being investigated. This task is extremely difficult due to the unstable nature of the jet caused by the density gradients between it and the supercritical water stream. ACKNOWLEDGMENTS The authors would like to acknowledge the Office of Aeronautics and Exploration Technology and NASAJohnson Space Center (Grant #NAG9-252) for their financial support of this work under the guidance of Donald Price. In addition, we would like to thank William Killilea, Glenn Hong, and David Ordway at MODAR, Inc (Natick, MA) for their assistance and insights, and also William Newhall and his colleagues at Harwood Engineering (Walpole, MA) and E. Ulrich Franck at the University of Karlsruhe (Germany) for their help with the design and construction of the optical cell. Professors Adel Sarofim, Janos Beer, and William Deen at MIT have also provided thoughtful comments and discussions. REFERENCES

(1) (2) (3)

(4)

(51

(6)

(7) Go

(9) (10)

Thomason, T. B.; Modell, M. Haz. Waste 1984, I, 453. Modell, M. In Standard Handbook of Hazardous Waste Treatment and Disposal; Freeman, H. M., Ed.; McGrawHill: New York, 1989; pp 8.153-8.168. Tester, J. W.; Holgate, H. R.; Armellini, F. J.; Webley, P. A.; Killilea, W. R.; Hong, G. T.; Barner, H. E. Presented at the Third Annual Symposium on Emerging Technologies for Hazardous Waste Management, Atlanta, GA, October 1991; paper 113.2, in press. Hong, G. T.; Fowler, P. K.; Killilea, W. R.; Swallow, K. C. SAE Technical Paper Series No. 871444; 17th Intersociety Conference on Environmental Systems, Seattle, WA, July 13-15, 1987. Hong, G. T.; Killilea, W. R.; Thomason, T. B. In Engineering, Construction, and Operations in Space: Proceedings of Space ‘88; Johnson, S. W.; Wetzel, J. P., Eds.; American Society of Civil Engineers: New York, 1988, pp 987-998. Martynova, 0.1. In High Temperature and High Pressure Electrochemistry in Aqueous Solutions; Jones, D. de G.; Staehle, R. W., Eds.; National Association of Corrosion Engineers: Houston, TX, 1976, pp 131-138. Uematsu, M.; Franck, E. U. J. Phys. Chem. Ref. Datu 1980, 9, 1291. Armellini, F. J.; Tester, J. W. SAE Technical Paper Series No. 901313; 20th Intersociety Conference on Environmental Systems, Williamsburg, VA, July 912, 1990. Helling, R. K.; Tester, J. W. Energy Fuels 1987, I, 417. Webley, P. A.; Tester, J. W. In Supercritical Fluid Science and Technology; Johnston, K. P.; Penninger, J. M. L., Eds.; ACS Symposium Series 406; American Chemical Society: Washington, DC, 1989, pp 259275.

264

Armellini and Tester

(11)

Webley, P. A.; Tester, J. W. Energy Fuels 1991, 5, 411. Webley, P. A.; Tester, J. W.; Holgate, H. R. Ind. Eng. Chem. Res. 1991,30, 1745. Holgate, H. R.; Tester, J. W. Presented at the 2nd International Symposium on Supercritical Fluids, Boston, MA, May 20-22, 1991, submitted for publication in Comb. Sci. Tech. Morey, G. W. Econ. Geol. 1957,52, 225. Ravich, M. I. Russ. .I. Inorg. Chem. 1970, 15, 1041. Pitzer, K. S. Chem. Phys. Lett. 1984, 105, 484. Ravich, M. I.; Borovaya, F. E. Russ. J. Inorg. Chem. 1964, 9, 520. Bischoff, J. L.; Pitzer, K. S. Am. .I. Sci. 1989, 289, 217. Sourirajan, S.; Kennedy, G. C. Am. J. Sci. 1962,260, 115. Bischoff, J. L.; Rosenbauer, R. J.; Pitzer, K. S. Geochim. Cosmochim. Acta 1986,50, 1437. Linke., W. F. Solubilities - inorganic and metal organic compounds; American Chemical Society: Washington, DC, 1958.

(12) (13)

(14) (15)

(16) (17) (18) (19)

(20) (21)

The Journal

of Supercritical

Fluids,

Vol. 4, No. 4, 1991

(22) Parisod, C. J.; Plattner, E. .I. Chem. Eng. Data. 1981,

26, 1620. Pitzer, K. S.; Pabalan, R. T. Geochim. Cosmochim. Acta 1986,50, 1445. (24) Glander, A.; Liander, H. Acta. Chim. Stand. 1950,4, 1437. I. Kh.; Borisov, N. M. High Temperature (25) Khaibullin, 1966,4, 489. (26) Tanger IV, J. C.; Pitzer, K. S. Geochim. Cosmochim. Acta 1989,53, 973. (27) Urusova, M. A. Russ. J. Inorg. Chem. 1975, 20, 1717. C-W Morey, G. W.; Hesselgesser, J. M. Econ. Geol. 1951, 46, 821. (29) Martynova, 0. I.; Samoilov, Yu. F. Russ. .I. Inorg. Chem. 1962, 7, 372. (30) Japas, M. L.; Franck, E. U. Ber. Bunsenges. Phys. Chem. 1985,R9, 793. (31) Schilling, W.; Franck, E. U. Ber. Bunsenges. Phys. Chem. 1988,92, 631. (23)