<<

Enthalpy of of Hypersaline Brine from 230 to 280 bar

A thesis presented to

the faculty of the Russ College of Engineering and Technology of Ohio University

In partial fulfillment

of the requirements for the degree

Master of Science

David D. Ogden

May 2018

© 2018 David D. Ogden. All Rights Reserved.

2

This thesis titled

Enthalpy of Vaporization of Hypersaline Brine from 230 to 280 bar

by

DAVID D. OGDEN

has been approved for

the Department of Mechanical Engineering

and the Russ College of Engineering and Technology by

Jason P. Trembly

Associate Professor of Mechanical Engineering

Dennis Irwin

Dean, Russ College of Engineering and Technology 3

ABSTRACT

OGDEN, DAVID D., M.S., May 2018, Mechanical Engineering

Enthalpy of Vaporization of Hypersaline Brine from 230 to 280 bar

Director of Thesis: Jason P. Trembly

There is a need for thermodynamic data of mixed brine solutions in order to properly treat brines generated from oil/ wells and CO2 sequestration. Thermodynamic properties of high concentration, multicomponent brine solutions are unknown, and limited to estimations based on single component solution data. Experimental data for mixed brine solutions at elevated and does not exist due to the extreme operating conditions above the supercritical point of pure . This study combines experimental results for mixed brine solutions, with thermodynamic models previously created for single component NaCl solutions to identify deviations resulting from additional dissolved species. Experiments were conducted using a heated desalinator at pressures of 230 to 280 bar and temperatures of 387 to 406 ºC.

4

ACKNOWLEDGMENTS

I would like to thank my advisor, Dr. Jason Trembly for his support of my research as well as assisting in the development of this thesis. I would also like to acknowledge the financial support of the National Energy Technology (NETL) and the Ohio

Water Development Authority. Further, the efforts of Dr. Wen Fan, Mr. Eli Fox, Ms.

Rachel Schack and Mr. Dominik Steinberg for assistance in operation of the experimental apparatus and conducting analytical analyses are to be recognized, as well as Mr. Colton

Nissen and Mr. Stuart Burkett for assistance in construction of the experimental apparatus. 5

TABLE OF CONTENTS

Page

Abstract ...... 3 Acknowledgments...... 4 List of Tables ...... 6 List of Figures ...... 7 Chapter 1: Introduction ...... 9 1.1 Background ...... 9 1.2 Objectives ...... 11 Chapter 2: Literature Review ...... 13 2.1 Supercritical Water Overview ...... 13 2.2 Supercritical Brine Thermodynamics ...... 15 2.3 of Dissolved in Supercritical Water ...... 17 Chapter 3: Methods ...... 24 3.1 Materials ...... 24 3.2 Desalination System ...... 26 3.3 Data Acquisition and Controls ...... 27 3.4 Sample Analysis ...... 29 3.5 Water Recovery ...... 29 3.6 Thermodynamic Modeling ...... 30 3.6.1 Specific ...... 30 3.6.2 Enthalpy of Vaporization ...... 34 Chapter 4: Results ...... 37 4.1 Objective 1 ...... 37 4.2 Objective 2 ...... 43 4.3 Objective 3 ...... 45 4.4 Corrosion ...... 50 Chapter 5: Conclusions ...... 52 Chapter 6: Recommendations ...... 53 References ...... 55 6

LIST OF TABLES

Page

Table 1. The test brine concentrations were selected based on a review of water produced by oil/gas and CO2 injection wells [32]–[36]. All concentrations are presented in mg∙L-1 or g∙L-1 measured at normal and (293.15 K and 101.3 kPa)...... 25 Table 2. Summary of experimental results for 50 and 180 g∙L-1 NaCl brines at 250 bar. 40 Table 3. Summary of experimental results for 50 and 180 g∙L-1 multicomponent brines. Data presented are three trial averages with standard deviations. Pseudocritical temperature of pure water is included for reference...... 41 Table 4. product compositions for 50 and 180 g∙L-1 multicomponent brine experimental trials. Concentrations are presented in mg∙L-1, and vapor fractionation is presented as a percentage range from all trials conducted. Note that was below the detection limit and is presented as BDL...... 43

7

LIST OF FIGURES Page

Figure 1: [A] Specific heat and [B] of pure water at critical and supercritical pressures [29]...... 14 Figure 2: Subcritical and supercritical water divided into 4 distinctions regions with a common intersection at the supercritical point [29]. © 2011 Igor Pioro and Sarah Mokry. Originally published in - Theoretical Analysis, Experimental Investigations and Industrial Systems under Creative Commons Attribution 3.0 Unported license. Available from: 10.5772/13790 ...... 14 Figure 3: Driesner results compared to other state equations. [19] ...... 17

Figure 4: diagram of NaCl + H2O system at a constant temperature of 375 C.[18] ...... 18 Figure 5: for subcritical and supercritical brine at a constant pressure of 250 bar.[9] ...... 19

Figure 6: P-T-x surface of NaCl-H2O system as described by Benchoff and Pitzer.[25] 20 Figure 7: NaCl concentrations presented against pressure on constant temperature lines from Bischoff and Pitzer results. [25] ...... 21 Figure 8: Solubility of , and chloride and sulfate species at 250 Bar [30], pseudocritical line presented in green for reference. [29] ...... 23 Figure 9: Dielectric constant and ion dissociation constant for water at 250 bar. [30] .... 23 Figure 10. (a) Process and instrumentation diagram for the prototype Joule-heating desalination system. (b) Cross section view of the prototype Joule-heating brine desalinator, a close up view of the vapor equilibrium is shown...... 27 Figure 11: for water and NaCl aqueous brines at (a) 230 bar, (b) 250 bar and (c) 280 bar...... 31 Figure 12: Constant concentration lines plotted on the temperature enthalpy diagram at a constant pressure of 230 bar...... 32 Figure 13. Control volume analysis (a) upper desalinator control volume used to estimate TVLE, (b) lower desalinator control volume used to calculate the enthalpy change associated with liquid and vapor product streams, (c) combined control volume used to determine the enthalpy of vaporization...... 35 Figure 14. Representative prototype system data for 180 g∙L-1 NaCl brine. a) Product flowrates and desalinator resistance, b) Inlet/outlet stream temperatures and desalinator power, and c) inlet/outlet stream compositions...... 38 Figure 15. Comparison of vapor TDS concentrations from 180 g∙L-1 NaCl brine and 50 and 180 g∙L-1 multicomponent brine study results with Bischoff and Pitzer data [25]. 8

Bischoff and Pitzer results are shown as smoothed grey lines of constant vapor temperature...... 42

Figure 16. Experimental TVLE results with vapor concentration for 230, 250 and 280 bar. Provided lines of pseudocritical temperature derived from Driesner model [19]. Results from the 50 g∙L-1 trials are presented as open markers, while 180 g∙L-1 results are presented as filled markers. Typical single trial standard deviations were below 50 mg∙L-1 with three trials resulting in standard deviations near 150 mg∙L-1...... 44 Figure 17: Test results from the 50 g∙L-1 electrochemical reactions test. The heating power loss is calculated based on the sloped of the low and high power linear regions. . 46 Figure 18. Water recovery from experimental trials based upon desalinator power. Water recovery defined by Equation 1. All trials completed with an inlet flowrate of 100 -1 -1 mL∙min . Results corrected for desalinator heat loss. 50 g∙L multicomponent data -1 represented by hollow markers, 180 g∙L multicomponent data represented by solid filled markers...... 47 Figure 19. Enthalpy of vaporization for 180 g∙L-1 NaCl brine and 50 and 180 g∙L-1 multicomponent brines at evaluated pressures...... 48 Figure 20. Process lines plotted on the temperature enthalpy diagram with lines of constant NaCl concentration. The process lines symbolize the enthalpy change of from the inlet to the lower outlet of the desalinator. It should be noted that the 50 g∙L-1 trial shows significantly different thermodynamics than the 180 g∙L-1 trial...... 50 Figure 21: Corrosion of inner electrode removed after 50 hours of high power operation...... 51 Figure 22: Corrosion of outer electrode after 250 hours of high powered operation. Pitting on right side of the figure can be compared to the left side of the figure where electrical current is minimal...... 51

9

CHAPTER 1: INTRODUCTION

1.1 Background

The ability to find adequate freshwater resources for energy production is a growing concern in the United States and around the world. A 2015 study completed by Konikow

[1] showed that nearly all US aquifers have shown significant depletion since 1900, with a distinct increase in depletion rate during the 1940’s and after the year 2000 [1]. While water resources are scarce in some locations, they are being contaminated in others.

Hydraulic fracturing is a water intensive process that also generates produced water, also known as brine. A study conducted by Mielke [2] et al. showed that a single lateral can consume up to 5.6 million gallons of fresh water during the hydraulic fracturing process. Further, as carbon emissions from power production come under increased scrutiny interest in sequestration of captured CO2 via deep saline aquifers will increase.

3 Projections indicate injection of one metric ton of CO2 can generate 1 m of saline water with total dissolved solids (TDS) concentrations ranging from 6 to 210 g∙L-1[3], [4].

There have been many methods developed for decontaminating or disposing of hypersaline brines. These methods range from injection into underground reservoirs to treatment at municipal water plants [5]. Membrane technology including forward and reverse osmosis are being pursued, but membrane fouling and pH sensitivity are still issues along with limited water recovery [6], [7]. Simple distillation of the brines could be completed, but is highly energy intensive for large scale desalination of high salinity waste streams [7]. pits are a simple technology that have been successfully used in desalination of low concentration brines, mainly from membrane desalination 10 permeate, but require a large footprint and do not recover the desalinated water [7].

Mechanical vapor compression and multi-effect desalination systems are currently in use for desalination of produced water, but have low recovery ratios for high concentration brine feed water [8].

Ohio University (OHIO) is developing a supercritical water based treatment process for hypersaline brines generated from industrial activities including oil/gas wells and CO2 injection wells. Like many , as pure water approaches its critical point, its enthalpy of vaporization (∆hvap) diminishes dramatically, until the critical point is reached where a single phase is formed (i.e. no change of state occurs with heat addition). Utilizing supercritical water properties is a known means to desalinate brine waste generated by industrial processes [9]–[13]. However, most supercritical water reactors utilize externally heated designs, which possess high thermal lag, internal scaling issues, and high manufacturing costs. Due to these issues, such systems are unable to be cost-effectively scaled. To address this issue, new reactor designs which utilize alternative heating designs are being considered [14], [15]. In this paper, Joule-heating, a result of the power dissipated as current travels through a non-ideal conductor, is evaluated for directly heating the brine.

This novel design allows for direct control of power applied to the brine using the electrical conductivity of the brine when applying an alternating current (AC), resulting in a lower

TDS content vapor product and a concentrated TDS liquid product. These streams will be referred to as vapor and liquid phases throughout the remainder of the document.

Early research into the properties of brines was driven by the need for thermodynamic data above standard conditions (temperature and pressure) for use in 11 geological modeling including groundwater [16], [17] and sea floor hydrology [18], [19].

The properties of high concentration brines under elevated pressure have been studied as early as 1931 [20], while elevated temperature and pressure tests began as early as 1942

[16]. To date, numerous experimental studies have been completed for a variety of brines ranging in temperatures and pressures up to 646.2 ºC and 4,137 bar [20], [16], [21]–[24],

[9]. Several reviews of these and other experimental results have been conducted to develop models for the thermodynamic properties of brines at elevated temperature and pressure [17]–[19], [23], [25]–[28]. The results from these studies show that brine dissolved solids content has a distinguishable effect on brine density, enthalpy and specific heat.

1.2 Objectives

Minimal experimental data was found regarding Joule heating of or the ∆hvap for multicomponent hypersaline brines. The primary purpose of this research is to evaluate the ability to utilize direct electrical Joule-heating of hypersaline brines and develop information regarding thermodynamic properties associated with hypersaline brines with multicomponent dissolved solids (salt) compositions. Such results will allow for better understanding of the energy requirements for supercritical water treatment.

 Objective 1– Experimentally determine the composition of vapor generated by

Joule-heated multicomponent brines at 230, 250 and 280 bar with brine

containing 50 and 180 g∙L-1 TDS. 12

 Objective 2 – Experimentally determine the vapor-liquid equilibrium (VLE)

temperature of multicomponent brines at 230, 250 and 280 bar with brine

containing 50 and 180 g∙L-1 TDS.

 Objective 3 – Experimentally determine the enthalpy of vaporization of

multicomponent brines at 230, 250 and 280 bar with brine containing 50 and 180

g∙L-1 TDS.

13

CHAPTER 2: LITERATURE REVIEW

2.1 Supercritical Water Overview

As pure water nears the critical point, 373.95 ºC and 220.6 bar, the thermodynamic properties undergo significant changes including the density, bond strength and heat capacity [9], [10], [29]. The critical point is defined by rapid changes in specific heat, reaching a peak value at the supercritical temperature and pressure, as shown in Figure 1.

Above the critical pressure water is defined in two distinct phases, compressed fluid with liquid like properties and with a gas like behavior, divided by the pseudocritical line, see Figure 2. Pseudocritical points, which make up the pseudocritical line, are the temperature and pressure combinations at which the specific heat peaks above the critical point [29]. As water transitions from the low temperature compressed fluid phase to the high temperature supercritical phase, it undergoes a large decrease in dielectric constant, transitioning the water from its classical polar state to a non-polar state, significantly decreasing the solubility of inorganic salts [10]. 14

[A] [B]

Figure 1: [A] Specific heat and [B] density of pure water at critical and supercritical pressures [29]. © 2011 Igor Pioro and Sarah Mokry. Originally published in Heat Transfer - Theoretical Analysis, Experimental Investigations and Industrial Systems under Creative Commons Attribution 3.0 Unported license. Available from: 10.5772/13790

Figure 2: Subcritical and supercritical water divided into 4 distinctions regions with a common intersection at the supercritical point [29]. © 2011 Igor Pioro and Sarah Mokry. Originally published in Heat Transfer - Theoretical Analysis, Experimental Investigations and Industrial Systems under Creative Commons Attribution 3.0 Unported license. Available from: 10.5772/13790 15

2.2 Supercritical Brine Thermodynamics

Archer [26] used a modified version of a formula first proposed by Pitzer to model the thermodynamic properties of NaCl solutions. The resulting equations predicted the specific heat as well as the specific volume of the solution, but used extensive formulas and parameters in the calculation.

Driesner [19] reviewed a number of studies that had investigated the thermodynamic properties including density, heat capacity and enthalpy of NaCl brines at elevated temperatures and pressures. Driesner states that although many equations of state that had been previously formed were reliable for specific areas of the pressure, temperature and concentration region in question, there were none that covered the entire range that has been investigated. Driesner proposed a model for the entire range by developing a temperature correction coefficient that could then be evaluated with the thermodynamic properties of enthalpy, specific heat and density at the corrected temperature. The Driesner model is compared to many different equations of state from previous research studies in

Figure 3. The results show acceptable correlation to the previous equations over the entire range of the study.

The addition of dissolved solids to water introduces an additional degree of freedom, changing thermodynamic properties and phase change behavior in comparison to pure water. At a constant temperature and pressure, properties such as enthalpy, density and specific heat vary based on the salinity of the brine. In systems with pressures above the critical point of pure water, and at constant temperatures, high salinity gradients can 16 induce the presence of both vapor phase and a compressed liquid phase due to the shift in pseudocritical temperature based on concentration. As heat is added to the system, a low concentration vapor phase is produced. The remaining dissolved solids concentrate in the compressed liquid phase, further increasing the pseudocritical temperature. The amount of heat that is required to produce a low concentration vapor phase, ∆hvap, is a result the concentration change. This is similar to the two phase region of subcritical water where quality changes at a constant temperature except instead of a change in quality, the salt concentration changes.

17

Figure 3: Driesner results compared to other state equations. [19]

2.3 Solubility of Dissolved Solids in Supercritical Water

Driesner [18] shows that brine concentration has a significant effect on the phase diagram shown in Figure 4. It is clear that there is a distinct vapor liquid zone on the left hand side of the figure near the critical point of water, 220.64 bar and 373.95 ºC [29]. 18

Figure 4: Phase diagram of NaCl + H2O system at a constant temperature of 375 C.[18]

To aid in the development of a zero discharge desalinator, Odu et al.[9] conducted experiments in capillary tubes to visually evaluate the phases in subcritical and supercritical NaCl brines. The results of the study show a clear visual distinction between the phases, and are plotted on the phase diagram in Figure 4. It should also be noted that all of their experiments showed on the line between one phase fluid and

V-L. This results in the water splitting into two separate phases that have concentrations defined by their temperature and pressure.

19

550

500 Vapor-Solid C) ° 450 Vapor-Liquid 400 Temperature ( 350 Liquid Liquid-Solid 300 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 NaCl Concentration (wt. %)

Figure 5: Phase diagram for subcritical and supercritical brine at a constant pressure of 250 bar.[9]

Bischoff and Pitzer[25] created a thermodynamic model by creating a database of experimental data at temperatures between 300 and 500 ºC, as well as conducted their own experiments to create a more complete database. The paper described the existence of supercritical brine on a pressure, temperature and concentration surface (P-T-x), shown in

Figure 6. The results from their study are shown in Figure 7, with NaCl concentration plotted against pressure with lines of constant temperature. The results of this study align nicely with previous research and are an accurate way to predict the concentration NaCl in both the liquid and vapor phases. 20

Figure 6: P-T-x surface of NaCl-H2O system as described by Benchoff and Pitzer.[25] 21

Figure 7: NaCl concentrations presented against pressure on constant temperature lines from Bischoff and Pitzer results. [25]

Leusbrock et al.[11]–[13] published a number of articles analyzing the solubility of various ions in subcritical and supercritical water. The articles covered experimental analysis as well as modeling of the ions in question. The experimental setup was an open loop system where a feedstock was preheated and then flowed through a furnace to reach its final temperature, followed by a filtration. The outlet concentrations of the system were 22 measured and used as the solubility for the ion. The system has the capabilities of operating up to 250 bar and 450 °C. Luesbrock [10] followed these articles with an additional article that summarized the methods and results for various constituents with the intent of showing supercritical water can be a useful method for removing these ions.

Tester [30], began researching supercritical water to evaluate its ability as an oxidation medium. The results of the study found that supercritical water is an excellent choice for oxidation due to its high solubility of inorganic compounds, as well as its low solubility of organics. The inorganic salt solubility results of studies that Tester reviewed are shown in

Figure 8. Tester concludes that the low solubility of inorganic salts is a result of the drastically reduced dielectric constant above the supercritical conditions of water, shown in Figure 9.

23

100000

10000

1000

100

10

1

Solubility Solubility (ppm) NaCl 0.1 KCl CaCl2 0.01 Na2SO4 CaSO4 0.001 350 375 400 425 450 Temperature °C Figure 8: Solubility of sodium, potassium and calcium chloride and sulfate species at 250 Bar [30], pseudocritical line presented in green for reference. [29]

Figure 9: Dielectric constant and ion dissociation constant for water at 250 bar. [30]

24

CHAPTER 3: METHODS

3.1 Materials

All source materials were purchased from Fisher Scientific, with the exception of colloidal silica and ammonium hydroxide, purchased from Sigma-Aldrich and barium chloride, purchased from Reagents. The concentration and source materials for all constituents in the test brines are presented in the multicomponent brine was prepared at a concentration of 180 g∙L-1 in a 25 gallon tank circulating for a minimum of 12 hours using a 3250 GPH Hydor Koralia Magnum pump. After mixing was complete, precipitates were filtered using a 0.35 micrometer pleated cartridge filter. The 50 g∙L-1 brine was prepared by diluting the premixed 180 g∙L-1 brine by a factor of 3.6 using DI water. Density of the

50 and 180 g∙L-1 test brines were calculated using NaCl data from EES [31], and found to be 1,032 and 1,115 kg∙m-3 with salt content of 4.9 wt. % and 16.1 wt. %, respectively.

25

Table 1. The test brine concentrations were selected based on a review of water produced by oil/gas and CO2 injection wells [32]–[36]. All concentrations are presented in mg∙L-1 or g∙L-1 measured at normal temperature and pressure (293.15 K and 101.3 kPa). Brines Ions Ion Source(s) 50 (g∙L-1) 180 (g∙L-1) NaCl (> 99.0%), NaHCO3 (> 99.7%), + Na 14,956 53,429 Na2SO4 (> 99.0%) 2+ Ca 4,261 15,222 CaCl2 (> 95%) 2+ Ba 27 97 BaCl2 (> 95%) 2+ Sr 109 389 SrCl2 (> 99%) K+ 54 194 KCl (> 99.0%) NaCl (> 99.0%), CaCl2 (> 95%), - Cl 30,671 109,572 BaCl2 (> 95%), SrCl2, KCl (> 99.0%) - HCO3 82 292 NaHCO3 (> 99.7%) 2- SO4 109 389 Na2SO4 (> 99.0%) + NH4 109 389 Ammonium Hydroxide (28 wt. %)

SiO2 10 34 Colloidal Silica (40 wt. %) -1 TDS (mg∙L ) 50,387 180,008 Density (kg∙m-3) 1,032 1,115 Note: Sodium added as sodium carbonate and sodium sulfate, followed by sodium chloride to complete the sodium balance. Chloride added as calcium chloride, barium chloride and chloride, follow by sodium chloride to complete the chloride balance.

The multicomponent brine was prepared at a concentration of 180 g∙L-1 in a 25 gallon tank circulating for a minimum of 12 hours using a 3250 GPH Hydor Koralia

Magnum pump. After mixing was complete, precipitates were filtered using a 0.35 micrometer pleated cartridge filter. The 50 g∙L-1 brine was prepared by diluting the premixed 180 g∙L-1 brine by a factor of 3.6 using DI water. Density of the 50 and 180 g∙L-

1 test brines were calculated using NaCl data from EES [31], and found to be 1,032 and

1,115 kg∙m-3 with salt content of 4.9 wt. % and 16.1 wt. %, respectively. 26

3.2 Desalination System

A supercritical water test system using Joule-heating for treatment of brines was developed with the aim of investigating the proposed water management process. A process and instrumentation diagram of the prototype test system and a cross section view of the desalinator are shown in Figure 10. The system utilizes a high pressure liquid chromatography pump (P-100) to supply a flow rate up to 300 ml per minute. The brine was pumped through a high pressure tube-in-tube heat exchanger (HX-100) to recover heat from the processed fluid, followed by a preheater (HX-101) to control the desalinator inlet temperature. The desalinator has a radial electrode configuration with an inner electrode diameter of 0.25 inches, and is constructed from Hastelloy C-276. The 1 inch Hastelloy C-

276 tubular body of the desalinator is used as the outer electrode with an inner diameter of

0.76 inches, resulting in a 0.255 inch gap between the electrodes. The inner electrode has

24 inches of exposed surface along the length of the desalinator; however, the inner electrode is not entirely immersed in the highly concentrated brine during steady state operation. The potential for electrochemical reactions in the desalinator was analyzed through experimental recording of the current as the electrode voltage was slowly ramped, the effective power from electrochemical reactions was estimated and used in the final results calculations.

The desalinator was operated at three pressures, 230, 250 and 280 bar, and two inlet

TDS concentrations (50 and 180 g∙L-1). Triplicate trials were conducted at each condition with a minimum of 90 minutes of steady state operation per trial. The desalinator produces a low TDS vapor stream exiting at the top of the desalinator and a concentrated liquid 27 stream extracted from the bottom. After the desalinator, the high concentration brine was cooled in the heat exchanger (HX-100) before passing through a secondary (HX-

104) to cool the brine to room temperature. The vapor flow was condensed (HX-105) before its pressure was reduced and the effluent was sampled. Liquid and vapor effluent samples (50 ml) were collected at five minute intervals.

a) b)

BPR-103 BPR-101

PT PT To Fume 107 106 Hood TE SP-100 SP-101 107 Liquid Vapor BPR-102 Sample Port Sample Port BPR-100 V-102 HX-105 Liquid/Gas Sample TE Separator Condenser 105

PT MI-100 MI-101 105 Liquid Vapor Balance Balance TE 106 TE 201 HX-104 Condenser TE 202 HX-103 TE Upper 203 PT Desalinator TE 104 Heat Tape 204 TE TE 104 205 V-101 TE Desalination TE PT TE PT HX-101 PT TE 206 Vessel 100 100 101 101 Pre-Heater 102 102 TE 207

HX-102 TE Inlet Heat HX-100 208 Tape V-100 P-100 Tube in Tube TE Test High Pressure Pump Heat Exchanger 209 Solution TE Tank 210 TE PT 103 103

Figure 10. (a) Process and instrumentation diagram for the prototype Joule-heating desalination system. (b) Cross section view of the prototype Joule-heating brine desalinator, a close up view of the vapor liquid equilibrium is shown.

3.3 Data Acquisition and Controls

The system was controlled by a National Instruments CRio controller running embedded LabView software. Temperature is measured and recorded at essential points in the process using Type-K thermocouples which directly contact the process fluid.

Temperature measurements for thermodynamic analysis are located at the inlet to the desalinator (TE 102), as well as the vapor outlet (TE 104) and the liquid outlet (TE 130). 28

Ten additional Type-K thermocouples were installed on the outer wall of the desalinator to develop a temperature profile. System pressure is monitored at the inlet and outlet of the desalinator as well as six other high pressure locations in the process; see Figure 10 for all essential pressure and temperature measurement locations. Datum were logged at an interval of five seconds for the duration of the experimental trials.

Desalinator voltage was controlled using a variable transformer with a step down voltage transformer, both rated at 10 kVA. The transformer combination allows for operation up to 833 A at 12 Vac or 416 A at 24 Vac. All trials were conducted at a frequency of 60 Hz, allowing for direct usage of utility power without need for frequency modification. The voltage and amperage waveforms were measured in real time, allowing for calculation of the true power, phase angle, and desalinator resistance.

To achieve steady state operation, it was necessary to control the pressure as well as the flowrate of both the vapor and liquid effluents from the system. The vapor effluent was used to control system pressure on a feedback control loop using a pressure transducer mounted downstream of the desalinator (PT-103). Liquid level in the desalinator was determined measuring the resistance of the desalinator. Once a steady state concentration was achieved in the desalinator, the resistance could be used to control the desalinator level via a process control valve (BPR-101). This setup allowed for a dynamic control of the desalinator liquid level. If power is increased, the vapor fraction in the desalinator is increased, lowering desalinator liquid level and in turn lowering liquid effluent flowrate.

The liquid and condensed vapor product flowrates were measured using Mettler Toledo 29

ICS435 and Mettler Toledo PG6002-S balances, respectively. The mass of each effluent stream was recorded using the CRio controller.

3.4 Sample Analysis

Samples were analyzed using Inductively Coupled Optical Emission

Spectrometer (ICP-OES) (iCAP 6000, Thermo Scientific). A tailored ICP standard solution series was used for elemental calibration to allow for single dilution for each sample.

Dilution factors ranged from 1 to 400 depending on the expected concentration of the sample, targeting the mid-range of the calibration.

3.5 Water Recovery

TDS content of the vapor effluent was determined by ICP for the major cation constituents, sodium, calcium, strontium and potassium. Water recovery rates were tested in the range of 50 to 80 % for 50 g∙L-1 and 20 to 40% for 180 g∙L-1 inlet brines, respectively.

The water recovery percentage was limited to avoid supersaturation within the desalinator, preventing salt precipitation within the vessel. All trials were completed at a constant inlet flowrate of 100 ml∙min-1. Water recovery percentage was calculated using Equation 1.

ṁ Water Recovery (%)= v × 100 (1) ṁ i

-1 -1 where ṁ i (g∙min ) is the inlet the mass flowrate and ṁ v (g∙min ) is the vapor stream mass flowrate.

30

3.6 Thermodynamic Modeling

3.6.1 Specific Heat Capacity

Driesner [19] reviewed a number of studies that had investigated the thermodynamic properties including density, heat capacity and enthalpy of NaCl brines at elevated temperatures and pressures. Driesner states that although many equations of state that had been previously formed were reliable for specific areas of the pressure, temperature and concentration region in question, there were none that covered the entire range that has been investigated. Driesner proposed a model for the entire range by developing a temperature correction coefficient that could then be evaluated using existing thermodynamic data for pure water. The correlations presented by Driesner [19] have been selected for use in the thermodynamic analysis of the experimental results produced from the Joule-heated desalinator. The selected model employs pure water enthalpy with a corrected temperature function to calculate the enthalpy of the NaCl brine at a given temperature, pressure and concentration (Equation 3). The value of the corrected temperature T* is calculated based on a linear fit to experimental data obtained by Driesner

[19]. Taking the derivative of the enthalpy corrected temperature function with respect to temperature yields the corrected specific heat function (Equation 3).

* hsolution(T, P, XNaCl)=hH2O (T h, P) (2)

C (T ,P, X )=q Cp (T* , P) psolution NaCl 2 퐻2O h (3)

The results of the Cp modeling are presented in Figure 11. It can be seen the concentration of NaCl has a significant effect on the temperature at which the Cp spike occurs, increasing the pseudocritical temperature of the fluid. It should also be noted that 31 the maximum value of the Cp spike lowers as the concentration increases. Figure 12 displays the enthalpy results using the Driesner model, showing the same increase in temperature that was seen in the Cp data. It is important to note that at a constant temperature and pressure, as the salt concentration increases, the enthalpy of that solution decreases, this energy will need to be accounted for in the modeling of the enthalpy of vaporization of the brine as concentration gradients will be present in the desalinator system.

Figure 11: Specific heat capacity for water and NaCl aqueous brines at (a) 230 bar, (b) 250 bar and (c) 280 bar.

32

Figure 12: Constant concentration lines plotted on the temperature enthalpy diagram at a constant pressure of 230 bar.

A MATLAB program using the Driesner [19] model was developed to evaluate the corrected properties of the large data sets. After calculating the value of the corrected temperature for the desired property the program interpolates the value from a pure water thermodynamic database with a resolution of 1 °C and 1 bar [31]. Increasing NaCl content lowers the brine specific heat capacity by roughly 1 % per 1 wt. % brine. Decrease of heat capacity is accompanied by an increase in brine temperature at which the maximum specific heat occurs and can be directly related to an increase in the solution pseudocritical temperature when pressure is held constant. It may also be noted that as the NaCl content increases, the enthalpy at a constant temperature and pressure decreases.

The addition of dissolved solids to water introduces an additional degree of freedom, changing thermodynamic properties and phase change behavior in comparison to pure water. At a constant temperature and pressure, properties such as enthalpy, density and specific heat vary based on the salinity of the brine. In systems with pressures above the critical point of pure water, and at constant temperatures, high salinity gradients can induce the presence of both vapor phase and a compressed liquid phase due to the shift in 33 pseudocritical temperature based on concentration. As heat is added to the system, a low concentration vapor phase is produced. The remaining dissolved solids concentrate in the compressed liquid phase, further increasing the pseudocritical temperature. The amount of heat that is required to produce a low concentration vapor phase, ∆hvap, is a result the concentration change. This is similar to the two phase region of subcritical water where quality changes at a constant temperature except instead of a change in quality, the concentration is changing.

For thermodynamic analysis, the system was assumed to be a pure NaCl brine.

Experimental evaluation comparing the effects of various salt species shows there is no significant difference in specific heat when comparing KCl to NaCl [22]. A comparison of specific heat capacity at normal temperature and pressure between the Driesner model and the experimental CaCl2 results from Toner and Catling [37] show less than a 2.5% deviation in the concentration range of the test matrix. The other salts were assumed to follow the same trends as NaCl, although at the low concentrations that are present should have minimal effects on the thermodynamics. In addition two experimental trials were

-1 conducted using brines containing 50 and 180 g∙L NaCl to allow for comparison with multicomponent brine results.

34

3.6.2 Enthalpy of Vaporization

-1 To determine ∆hvap (kJ∙kg ) a three step procedure was used, summarized in Figure

13, initiated by calculation of the TVLE, followed by determination of energy absorbed by the desalinator vapor and liquid streams, and finally an energy balance on the desalinator.

First, TVLE was estimated through an energy balance on a control volume over the upper

-1 portion of the desalinator (Figure 13a) using the vapor product mass flowrate ṁ v (kg∙s ) and temperature (TV (ºC)) of the vapor outlet in combination with the heat transfer (Q̇ Upper

(kW)). Q̇ Upper (kW) was determined (Equation 4) by summing the power applied to the heat tape (Q̇ tape (kW)), HX-103, to achieve desired vapor outlet temperature and an estimated heat loss function (Q̇ loss,Up (kW)) based on the desalinator wall temperature. Due to the rapid changes in specific heat near the pseudocritical point, an integration of the specific heat was completed with respect to temperature at constant pressure and constant brine concentration (Equation 5). TVLE was determined using an integral bound via a minimizing function in Engineering Equation Solver software [31].

Q̇ Upper=Q̇ tape-Q̇ loss,Up (4)

TV Q̇ Upper=ṁ v ∫ CpdT (5) TVLE

Next, enthalpy change between the vapor (∆HVapor (kJ) and liquid (∆HLiquid (kJ)) outlet streams was determined using an energy balance on two control volumes for each product stream (Figure 13b). Due to the large change in brine concentration in the lower portion of the desalinator, as well as specific heat dependence on concentration, it is necessary to assume a concentration profile in the liquid phase of the desalinator. Due to 35 limited knowledge of the concentration gradient in the lower desalinator, a linear concentration gradient was assumed. Enthalpy changes for both outlets was completed using a similar integral of specific heat with relation to temperature at constant pressure, but with the assumed brine concentration profile (Equations 6 and 7).

TVLE ∆HVapor=ṁ v ∫ Cp,vapdT (6) TInlet

TL ∆HLiquid = ṁ L ∫ Cp,liqdT (7) TInlet

-1 -1 -1 Where ṁ L (kg∙s ) is the liquid product flowrate, Cp,vap (J∙kg ∙K ) is the specific heat of

-1 -1 at the specified system pressure and Cp,liq (J∙kg ∙K ) is the specific heat of the liquid product.

Figure 13. Control volume analysis (a) upper desalinator control volume used to estimate TVLE, (b) lower desalinator control volume used to calculate the enthalpy change associated with liquid and vapor product streams, (c) combined control volume used to determine the enthalpy of vaporization.

Finally, the ∆hvap is determined by first completing an energy balance (Equation 8) on a control volume over the lower portion of the desalinator using the electrical power 36 input through the electrode (Q̇ Elect (kW)), heat loss around the lower portion of the desalinator (Q̇ loss,low (kW)), ∆HVapor, and ∆HLiquid. Then, ∆hVap (Equation 9) is determined using the result of Equation 8 and ṁ v.

∆HVap=Q̇ Elect-Q̇ loss,low-∆HVapor-∆HLiquid (8)

∆HVap ∆hVap= (9) ṁ v

37

CHAPTER 4: RESULTS

As published in Desalination 424(2017), 149-158

4.1 Objective 1

A series of trials were conducted evaluating the ability of the prototype Joule- heating desalinator to produce a clean water product from multicomponent brines

-1 containing 50 to 180 g∙L TDS, modeled after brines generated by oil/gas and CO2 injection wells. As limited information exists for the H2O/NaCl brine system at supercritical conditions, trials were completed with 50 and 180 g∙L-1 NaCl brines to compare with previous experimental and modeling results. Figure 14 presents prototype system operating data from a single trial conducted at 23 MPa with 100 mL∙min-1 of 180 g∙L-1 NaCl brine. As shown in Figure 14a, approximately 50 min is needed to establish steady state conditions when heated from room temperature. Once operating conditions are achieved, the prototype system exhibits stable operation as demonstrated by the liquid/vapor flowrates and desalinator power (Figure 14a) and inlet/product temperatures

(Figure 14b). Average desalinator power (Figure 14b) was approximately 875 W throughout with slight fluctuations due to minor changes in liquid conductivity and liquid product flowrate. Vapor TDS (Figure 14c) decreases to a steady state concentration of approximately 1,000 mg∙L-1 within the first 25 minutes of operation. This phenomena was associated with the purging of brine flowing through the system before fluid transition to pseudocritical state.

38

a) 140 Vapor Flowrate 0.1

Liquid Flowrate 0.09 )

120 Electrode Resistance Ω 0.08 ) 1 - 100 0.07

80 0.06 0.05 60 0.04 40 0.03 Flowrate(g·min 0.02

20 ( Resistance Desalinator 0.01 0 0 50 70 90 110 130 150 170 Time (min)

b) 410 Vapor Temperature 1,200 Liquid Temperature 1,150 400 Inlet Temperature Desalinator Power 1,100 390 1,050

380 1,000 950 370 900 850

Temperature(ºC) 360

800 (W) Power Desalinator 350 750 340 700 50 70 90 110 130 150 170 Time (min)

c) 100 Na Liquid Outlet 2,500 Na Inlet 90

Na+) Na Vapor Outlet 1 - 80 Vapor Outlet TDS 2,000 Na+) 1 - )

70 1 - 60 1,500 50 40 1,000 30 20 500 Vapor TDS (mg·L VaporTDS 10 Liquid Concentration (g·L Concentration Liquid 0 0 50 100 150 (mg·L VaporConcentration Time (min)

Figure 14. Representative prototype system data for 180 g∙L-1 NaCl brine. a) Product flowrates and desalinator resistance, b) Inlet/outlet stream temperatures and desalinator power, and c) inlet/outlet stream compositions.

39

Results from the 50 and 180 g∙L-1 NaCl brine trials are shown in Table 2, while results from the 50 and 180 g∙L-1 multicomponent brine trials are provided in

40

Table 3. As negligible information regarding properties of supercritical fluids generated from multicomponent brines exists, data available for H2O/NaCl systems was selected for comparison. The Bischoff and Pitzer study was selected for comparison, as the authors reported a wide range of H2O/NaCl brine results and NaCl is the primary component of the brines investigated in this study[25]. Figure 15 presents the Joule-heated desalinator vapor product results with the Bischoff and Pitzer data.

Table 2. Summary of experimental results for 50 and 180 g∙L-1 NaCl brines at 250 bar. Parameters 50 g∙L-1 180 g∙L-1 Pressure (bar) 250.00 ± 0.10 249.99 ± 0.31 Pseudocritical Temp (ºC) 384.9 384.9 Inlet Temp (ºC) 351.58 ± 0.19 352.11 ± 0.15 Vapor Outlet Temp (ºC) 390.036 ± 0.51 391.06 ± 0.78 Liquid Outlet Temp (ºC) 392.11 ± 0.40 400.73 ± 0.31

TVLE (ºC) 383.1 386.6 -1 ṁ inlet (g∙min ) 104.5 ± 7.4 112.3 ± 8.6 Inlet TDS (ppm) 60,692.8 209,774.5 Vapor TDS (ppm) 2,010.81 ± 60.0 1,128.44 ± 23.1

41

Table 3. Summary of experimental results for 50 and 180 g∙L-1 multicomponent brines. Data presented are three trial averages with standard deviations. Pseudocritical temperature of pure water is included for reference. 50 g∙L-1 Operating Pressure (bar) Temperatures 230.01 ± 0.45 249.99 ± 0.97 280.00 ± 1.00 Pseudocritical (ºC) 377.5 384.9 395.4 Inlet (ºC) 343.5 ± 2.05 349.7 ± 2.02 362.8 ± 8.43 Vapor Outlet (ºC) 387.2 ± 0.64 391.0 ± 0.32 406.4 ± 0.06 Liquid Outlet (ºC) 369.5 ± 4.63 373.7 ± 10.91 388.2 ± 7.03

TVLE (ºC) 379.1 ± 0.47 386.3 ± 0.53 397.3 ± 2.65 -1 ṁ inlet (g∙min ) 104.3 ± 1.30 104.4 ± 1.98 105.4 ± 1.98 Inlet TDS (ppm) 41,639.4 ± 298.3 44,261.9 ± 252.7 59,130.1 ± 311.3 Vapor TDS (ppm) 655.1 ± 158.5 1,240.0 ± 237.7 2,836.1 ± 97.5 180 g∙L-1 Operating Pressure (bar) Temperatures 230.00 ± 0.31 249.99 ± 0.89 280.00 ± 0.17 Pseudocritical (ºC) 377.5 384.9 395.4 Inlet (ºC) 346.0 ± 0.94 351.5 ± 0.74 360.4 ± 0.67 Vapor Outlet (ºC) 378.7 ± 0.87 390.3 ± 0.74 402.7 ± 0.63 Liquid Outlet (ºC) 381.9 ± 1.16 392.0 ± 0.74 404.2 ± 0.76

TVLE (ºC) 380.1 ± 1.43 387.8 ± 1.19 398.4 ± 1.55 -1 ṁ inlet (g∙min ) 103.9 ± 1.11 105.4 ± 0.78 111.1 ± 0.48 146,501.7 ± 154,033.3 ± 157,645.9 ± Inlet TDS (ppm) 11,129.2 2,128.25 1,334.82 Vapor TDS (ppm) 589.4 ± 40.9 1,095.4 ± 75.2 2,607.5 ± 263.3

42

320 Multicomponent Brine NaCl Brine 300

280 410 ºC

260 400 ºC

240 390 ºC Pressure(bar) 380 ºC 220

200 0.01 0.1 1 10 100 Vapor Concentration (wt. %)

Figure 15. Comparison of vapor TDS concentrations from 180 g∙L-1 NaCl brine and 50 and 180 g∙L-1 multicomponent brine study results with Bischoff and Pitzer data [25]. Bischoff and Pitzer results are shown as smoothed grey lines of constant vapor temperature.

Product vapor TDS ranges for the investigated pressures were 622.3 ± 163.7,

1,167.7 ± 249.3, and 2,721.8 ± 280.8 mg∙L-1 at 230, 250, and 280 bar, respectively.

Individual cation concentrations as well as vapor fractionation, defined by the ratio of vapor concentration to inlet concentration, are shown in Table 4. This phenomena of increasing vapor TDS with pressure is a result of increased supercritical water vapor density. of the supercritical water produced in the prototype system were determined to be 227.0, 274.4, and 295.6 kg∙m-3 using vapor/liquid interface temperatures at 230, 250, and 280 bar, respectively [31]. As water density increases, its ability to dissolve inorganic halide salts increases. Similar results have been reported for single component halide salts in supercritical water including NaCl, KCl, CaCl2, and MgCl2 [11]–[13], [38],

[39]. Further, there is no significant difference between vapor TDS generated from the 50 43 and 180 g∙L-1 multicomponent brines at each operating pressure, further supporting the theory that vapor phase density controls composition. The removal efficiencies for the major cations were found to be 97.98-99.59% and 99.81-99.98% for sodium and calcium, respectively. Overall TDS removal range was 95.2-99.6 % over the evaluated conditions.

Table 4. Vapor product compositions for 50 and 180 g∙L-1 multicomponent brine experimental trials. Concentrations are presented in mg∙L-1, and vapor fractionation is presented as a percentage range from all trials conducted. Note that barium was below the detection limit and is presented as BDL. 230 250 280 Constituent (mg∙L-1) (mg∙L-1) (mg∙L-1) Ca2+ 15.03 ± 18.81 13.60 ± 11.53 24.46 ± 8.19 K+ 7.71 ± 11.26 4.97 ± 0.75 10.09 ± 2.49 Na+ 232.08 ± 29.82 440.37 ± 66.78 1,038.59 ± 77.76 Sr2+ 0.49 ± 0.55 0.62 ± 0.39 0.81 ± 0.34 Ba2+ 0.30 ± 0.57 0.11 ± 0.23 1.73 ± 2.37

4.2 Objective 2

Experimental data generated by the prototype Joule-heated desalinator was used to calculate TVLE for each operating condition using the method described in Section 3. Figure

-1 -1 16 presents the calculated TVLE values for the 180 g∙L NaCl and 50 and 180 g∙L multicomponent brines with pseudocritical temperature calculated using the Driesner model [19] for NaCl brines. It can be seen that as the NaCl concentration increases, the

Driesner model predicts a slight increase in pseudocritical point. These results are similar to those shown in Figure 15, as the model also predicts increased vapor TDS concentration with operating pressure due to increased supercritical fluid density. 44

-1 The 180 g∙L NaCl trial TVLE value is in good agreement with the Driesner model.

TVLE values determined from multicomponent brine experiments generally are slightly higher (2-5 °C) in comparison to the predicted NaCl model values, as shown in Figure 16.

This could be a result of thermocouple error (± 2.2 °C) and additional brine components, most notably CaCl2. No discernable difference is seen between TVLE values from the 50 and 180 g∙L-1 multicomponent brines at each operating pressure, correlating with data in

Table 4, which showed no discernable difference in vapor composition generated from either fluid at each pressure. These data further suggest thermodynamic properties of multicomponent brines, with NaCl as the dominant component, may be suitably predicted using experimental data and models developed for H2O/NaCl brines [19], [25].

405

400

395 280

(ºC) 390 VLE T 230 bar Multicomponent 385 250 250 bar Multicomponent 380 280 bar Multicomponent 230 250 bar NaCl 375 0 1000 2000 3000 Vapor TDS Concentration (mg·L-1)

Figure 16. Experimental TVLE results with vapor concentration for 230, 250 and 280 bar. Provided lines of pseudocritical temperature derived from Driesner model [19]. Results from the 50 g∙L-1 trials are presented as open markers, while 180 g∙L-1 results are presented as solid filled markers. Typical single trial standard deviations were below 50 mg∙L-1 with three trials resulting in standard deviations near 150 mg∙L-1. 45

4.3 Objective 3

A benefit of the Joule-heating desalinator is that electrical power is directly converted into thermal energy within the liquid brine via resistive heating, removing conduction and/or convection heat transfer associated with external heating mechanisms, while potentially reducing scaling/plugging within the system.

Due to the Joule-Heating method requiring an electrical current to be passed through brine in the desalinator, it was necessary to consider the effects of electrochemical reactions, including water splitting. A test was completed for each inlet concentration by slowly ramping the voltage while monitoring desalinator current. Two distinct slopes were found in the voltage and current data, the first region below 1.25 V ac and second above

2.25 Vac. Data between the two regions was not linear, therefore it was not included in the distinct linear ranges. It was assumed current while operating below 1.25 V ac was purely resistive heating of the fluid, while current above 2.25 V ac was a combination of resistive heating and electrochemical reactions. The slope of the two linear regions was used to calculate the power that was associated with heating of the process fluid. This test resulted in a power loss of up to 20% and 40% for the 50 and 180 g∙L-1 solution respectively, Figure

17 displays the results of the 50 g∙L-1 test.

46

120 Low Voltage 50 High Voltage 100 40 Electrochemical 80 Power 30 60 20 40 Consumption (%) Consumption

Desalinator Current (A) DesalinatorCurrent 10 Power Electrochemical 20

0 0 0 2 4 6 Desalinator Voltage (V ac)

Figure 17: Test results from the 50 g∙L-1 electrochemical reactions test. The heating power loss is calculated based on the sloped of the low and high power linear regions.

Figure 18 presents clean water recovery from the generation of vapor within the desalinator based upon supplied power. Typical operating voltages for the trials were 5.9-

10.0 Vac with a current range of 124-173 A depending upon brine composition and targeted water recovery. A clear linear relationship between desalinator power and water recovery is seen regardless of operating pressure or inlet composition. To avoid precipitation of salts within the desalinator, the system may be operated at a water recovery level which prevents an ion product greater than its associated equilibrium constant. Further water recovery may be achieved via flashing the desalinator liquid product as described previously [9], [15]. 47

100 230 bar 90 250 bar 280 bar 80 70 60 50 40 30 Water Recovery (%) WaterRecovery 20 10 0 600 800 1,000 1,200 1,400 1,600 Desalinator Power (W) Figure 18. Water recovery from experimental trials based upon desalinator power. Water recovery defined by Equation 1. All trials completed with an inlet flowrate of 100 -1 -1 mL∙min . Results corrected for desalinator heat loss. 50 g∙L multicomponent data -1 represented by hollow markers, 180 g∙L multicomponent data represented by solid filled markers.

To understand the energetics associated with the proposed Joule-heated brine treatment process, the energy necessary to produce a clean water product needs to be understood. ∆hvap represents the energy necessary to generate a quantity of low-TDS vapor from the hypersaline brine within the desalination system. Experimental data from the prototype system were used to derive of vaporization values using the methodology described in Section 3. Figure 19 presents the results from the ∆hvap analyses. Energy required to generate low-TDS vapor from the 50 g∙L-1 multicomponent brine was determined to be 495 ± 43.1, 389 ± 47.4, and 392 ± 40.4 kJ∙kg-1, while values for the 180 g∙L-1 multicomponent brine were determined to be 421.7 ± 37.3, 289.3 ± 23.6, and 178.3 ± 69.9 kJ∙kg-1 at 230, 250 and 280 bar respectively. Operating pressure was 48 found have a significant effect on ∆hVap, with the highest value for both multicomponent brines at 230 bar.

600 50 g·L Multicomponent Brine 180 g·L Multicomponent Brine 180 g·L NaCl Brine 500 ) 1 - 400

300 (kJ·kg v ap h

Δ 200

100

0 220 240 260 280 300 Pressure (bar)

Figure 19. Enthalpy of vaporization for 180 g∙L-1 NaCl brine and 50 and 180 g∙L-1 multicomponent brines at evaluated pressures.

It was found that the higher inlet concentration brine resulted in a lower ∆hvap, similar to the increase in efficiency of mechanical vapor compression and multi-effect desalination as the inlet concentration is increased [8]. Comparing

Figure 18 and Figure 19, the linear relationship between desalinator power and water recovery is counterintuitive when compared to the ∆hvap results. As ∆hvap for the

180 g∙L-1 multicomponent brine was 18-120% less than the 50 g∙L-1, one would expect power required to generate vapor from the lower strength brine to be greater. However, the linear relationship between water recovery and desalinator power can be attributed to 49 the increased ratio of electrochemical reactions in the 180 g∙L-1. Further evaluation of the thermodynamic of the lower portion of the desalinator revealed a difference in liquid enthalpy change due to increasing salt concentration within the lower portion of the reactor.

This phenomena can be further explained through Figure 20, which shows the T-h diagram for the multicomponent brines (50 and 180 g∙L-1) entering the desalinator and the resulting liquid product (310 g∙L-1) exiting the lower outlet. Process paths indicating the transition of the brines entering the desalinator to the liquid product are also shown. The specific enthalpy of the resulting lower liquid product from the 50 g∙L-1 brine is lower than its corresponding inlet brine specific enthalpy; however, in the 180 g∙L-1 case the liquid product has a greater specific enthalpy than the corresponding inlet brine. This specific enthalpy change in the 50 g∙L-1 case, results in an energy release from the liquid in the bottom of the desalinator, while showing an increase in the 180 g∙L-1 case, further complicating the relationship between applied power and ∆hvap. 50

-1 0 g·L-1 50 g·L 420 -1 -1 180 g·L 310 g·L 50 g·L-1 Inlet 180 g·L-1 Inlet 400 C)

380

360 50 g∙L-1 Inlet

Temperature( 340 180 g∙L-1 Inlet 320

300 1250 1300 1350 1400 1450 1500 h (kJ·kg-1)

Figure 20. Process lines plotted on the temperature enthalpy diagram with lines of constant NaCl concentration. The process lines symbolize the enthalpy change of fluid from the inlet to the lower outlet of the desalinator. It should be noted that the 50 g∙L-1 trial shows significantly different thermodynamics than the 180 g∙L-1 trial.

4.4 Corrosion

Corrosion was encountered in the lower portion of the desalinator where liquid brine underwent Joule-heating. Throughout the trials a potential range of 5.9-10.0 V ac was applied across the electrodes in the lower portion of the desalinator. Images of the inner and outer electrodes after 50 and 250 hrs of operation are shown in Figure 21Figure 22, respectively. The inner electrode experienced more significant corrosion, likely due to the higher current density at this location. The 0.5-in diameter inner electrode experienced significant corrosion, forming large pits along its length exposed to brine during joule- heating tests. Corrosion at the outer electrode was found to be significantly lower, with pitting no deeper than 100 µm found after 250 hours of operation with brine. 51

Figure 21: Corrosion of inner electrode removed after 50 hours of high power operation.

Figure 22: Corrosion of outer electrode after 250 hours of high powered operation. Pitting on right side of the figure can be compared to the left side of the figure where electrical current is minimal.

52

CHAPTER 5: CONCLUSIONS

The reported results indicate a Joule-heated desalinator is capable of treating

-1 multicomponent hypersaline brines with TDS content ranging from 50 to 180 g∙L at supercritical water pressures. The prototype system demonstrated stable operation and the ability to treat hypersaline brines that are problematic for other water treatment/management techniques. Clean water recovery ranges of 20-80 % were demonstrated over the evaluated operating envelope. TDS content of the clean water product was found to be controlled by operating pressure, with a TDS range of 600-2,800 mg∙L-1 found from 230 to 280 bar. Results from multicomponent hypersaline brine tests were found to be similar to those reported by Bischoff and Pitzer [20] and Driesner [13] for single component NaCl hypersaline brines. TDS content of the hypersaline brine was shown to decrease the associated heat of vaporization, however due to change in liquid product enthalpy, a clear linear relationship between electrical power input and water recovery was established.

53

CHAPTER 6: RECOMMENDATIONS

Although the Joule-heating method has been proven as an effective process to treat hypersaline brines, there is still research that needs to be completed to develop this technology into a commercially viable process. One of the largest concerns with this technology is the high concentration brine stream that is produced. This brine effluent will still need to be treated or disposed of after leaving the desalinator. One method to reduce or eliminate the high concentration effluent would be to introduce a flash separator into the process post desalinator. This would allow for the highly concentrated, high temperature brine to be separated into a solid salt and a clean vapor, as well as a small amount of saturated brine that is based on the quality of the flashed brine. Due to the size and the flowrate of the test system used in this investigation, it is not feasible to flash salts as the small orifice that would be required would be easily plugged. It is my recommendation that a separate test system is constructed to allow for a large volume batch flash tests to be conducted under the conditions that exist at the liquid outlet of the desalinator. A simple experiment could be conducted with a heated pressure vessel, a high temperature ball valve and a flash tank. This would allow for the products of the flash to be evaluated as well as offer insight into the thermodynamics of the flash system, aiding in the design of a commercial scale system.

Due to the large power loss resulting from electrolysis in the desalinator, it is imperative to gain a better understanding of the power losses associated with electrochemical reactions. In this investigation the power associated with electrochemical reactions was assumed to be in line with the trials that were conducted using 50 and 180 54 g∙L-1 inlet solutions, having a constant concentration throughout the desalinator. In experimental operation there will not be a constant concentration, rather a concentration gradient in the lower portion of the desalinator, significantly effecting the electrochemical reactions. To gain a better understanding of the actual effects of electrochemical reactions,

I recommend that a series of experimental trials be conducted using an external heating method to supplement the electrode heating. By operating the system at various levels of external heating, a comparison of electrode power to external power will yield important data regarding the amount of electrochemical reactions taking place inside of the desalinator. The data from these trials can be analyzed to increase the precision of the electrochemical reaction correction for a more accurate calculation of ∆hvap.

In addition to the previous improvements, I recommend that the desalination system be tested using a field derived produced water. By doing this the system can be tested for its ability to treat that are not ideally produced in a lab. Experiments conducted with water from hydraulic fracturing wells will not only prove that the technology has the ability to treat this water, but will also give insight on how different constituents and their concentrations effect desalinator operation.

55

REFERENCES

[1] L. F. Konikow, “Long-Term Groundwater Depletion in the United States,” Groundwater, vol. 53, no. 1, pp. 2–9, 2015. [2] E. Mielke, L. Diaz Anadon, and V. Narayanamurti, “Water Consumption of Energy Resource Extraction, Processing, and Conversion,” Energy Technol. Innov. Policy Res. , no. October, 2010. [3] P. E. S. Bergmo, A.-A. Grimstad, and E. Lindeberg, “Simultaneous CO2 injection and water production to optimise aquifer storage capacity,” Int. J. Greenh. Gas Control, vol. 5, no. 3, pp. 555–564, May 2011. [4] S. A. Dastgheib, C. Knutson, and Y. Yang, “Produced Water from CO2-EOR in the Illinois Basin,” Energy Procedia, vol. 63, pp. 6878–6886, 2014. [5] M. Zoback, S. Kitasei, and B. Copithorne, “Addressing the environmental risks from shale gas development,” 2010. [6] B. D. Coday and T. Y. Cath, “Forward Osmosis: Novel desalination of produced water and fracturing flowback,” J. - Am. Water Works Assoc., vol. 106, pp. E55– E66, Feb. 2014. [7] R. Kaplan, D. Mamrosh, H. H. Salih, and S. A. Dastgheib, “Assessment of desalination technologies for treatment of a highly saline brine from a potential CO2 storage site,” Desalination, vol. 404, pp. 87–101, Feb. 2017. [8] G. P. Thiel, E. W. Tow, L. D. Banchik, H. W. Chung, and J. H. Lienhard, “Energy consumption in desalinating produced water from shale oil and gas extraction,” Desalination, p. 94, 2015. [9] S. O. Odu, A. G. J. van der Ham, S. Metz, and S. R. A. Kersten, “Design of a process for supercritical water desalination with zero liquid discharge,” Ind. Eng. Chem. Res., vol. 54, pp. 5527–5535, 27 2015. [10] I. Leusbrock, “Removal of inorganic compounds via supercritical water,” University of Groningen, 2011. [11] I. Leusbrock, S. J. Metz, G. Rexwinkel, and G. F. Versteeg, “ of inorganic compounds in supercritical water.” [12] I. Leusbrock, S. J. Metz, G. Rexwinkel, and G. F. Versteeg, “The solubility of chloride and calcium chloride in near-critical and supercritical water,” J. Supercrit. Fluids, vol. 53, no. 1–3, pp. 17–24, Jun. 2010. [13] I. Leusbrock, S. J. Metz, G. Rexwinkel, and G. F. Versteeg, “Solubility of 1:1 Alkali Nitrates and Chlorides in Near-Critical and Supercritical Water,” J. Chem. Eng. Data, vol. 54, no. 12, pp. 3215–3223, Dec. 2009. [14] X. Dong, J. Trembly, and D. Bayless, “Techno-economic analysis of hydraulic fracking flowback and produced water treatment in supercritical water reactor,” Energy, vol. 133, pp. 777–783, Aug. 2017. [15] D. E. López and J. P. Trembly, “Desalination of hypersaline brines with joule- heating and chemical pre-treatment: Conceptual design and economics,” Desalination, vol. 415, pp. 49–57, Aug. 2017. [16] N. B. Keevil, “Vapor pressures of aqueous solutions at high temperatures,” J. Am. Chem. Soc., vol. 64, no. 4, pp. 841–850, 1942. 56

[17] Christopher Palliser, “A model for deep geothermal brines: state space description and thermodymanic properties,” Massey University, 1998. [18] T. Driesner and C. A. Heinrich, “The system H2O–NaCl. Part I: Correlation formulae for phase relations in temperature–pressure–composition space from 0 to 1000°C, 0 to 5000bar, and 0 to 1 XNaCl,” Geochim. Cosmochim. Acta, vol. 71, pp. 4880–4901, Jan. 2007. [19] T. Driesner, “The system H2O–NaCl. Part II: Correlations for molar volume, enthalpy, and isobaric heat capacity from 0 to 1000°C, 1 to 5000bar, and 0 to 1 XNaCl,” Geochim. Cosmochim. Acta, vol. 71, pp. 4902–4919, Jan. 2007. [20] L. H. Adams, “Equilibrium in binary systems under pressure. I. An experimental and thermodynamic investigation of the system, NaCl-H2O, at 25,” J. Am. Chem. Soc., vol. 53, no. 10, pp. 3769–3813, 1931. [21] C. W. Blount and F. W. Dickson, “The solubility of anhydrite (CaSO4) in NaCl- H2O from 100 to 450°C and 1 to 1000 bars,” Geochim. Cosmochim. Acta, vol. 33, pp. 227–245, 01 1969. [22] J. K. Hovey, K. S. Pitzer, J. C. Tanger, J. L. Bischoff, and R. J. Rosenbauer, “Vapor-liquid phase equilibria of potassium chloride-water mixtures: equation-of- state representation for potassium chloride-water and sodium chloride-water,” J. Phys. Chem., vol. 94, no. 3, pp. 1175–1179, 1990. [23] F. Armellini, “Phase equilibria and precipitation phenomena of sodium chloride and sodium sulfate in sub- and supercritical water,” Massachusetts Institute of Technology, 1993. [24] L. Hnědkovský, V. Hynek, V. Majer, and R. H. Wood, “A new version of differential flow heat capacity calorimeter; tests of heat loss corrections and heat capacities of aqueous NaCl from T= 300 K to T= 623 K,” J. Chem. Thermodyn., vol. 34, no. 6, pp. 755–782, Jun. 2002. [25] J. L. Bischoff and K. S. Pitzer, “Liquid-vapor relations for the system NaCl-H2O: summary of the P-T-x surface from 300 degrees to 500 degrees Celsius,” Am. J. Sci., vol. 289, pp. 217–248, 1989. [26] D. Archer, “Thermodynamic properties of NaCl + H2O system II. Thermodynamic properties of NaCl(aq), NaCl-2H2O(cr), and phase equilibria,” J. Phys. Chem. Ref. Data, vol. 21, no. 4, pp. 793–829, 1992. [27] D. Archer, “Thermodynamic properties of NaCl + H2O system I. Thermodynamic properties of NaCl(cr),” J. Phys. Chem. Ref. Data, vol. 21, no. 1, p. 121, 1992. [28] C. Palliser and R. McKibbin, “A model for deep geothermal brine, II: thermodynamic properties - density,” Transp. Porous Media, vol. 33, pp. 129–154, 1998. [29] I. Pioro and S. Mokry, “Thermophysical properties at critical and supercritical conditions,” Heat Transf. Theor. Anal. Exp. Investig. Ind. Syst. InTech Rij., 2011. [30] J. W. Tester et al., “Supercritical water oxidation technology: process development and fundamental research,” ACS Publications, 1993. [31] Engineering Equation Solver. 2017. 57

[32] K. Guerra, K. Dahm, and S. Dundorf, “Oil and gas producted water management and beneficial use in the western United States,” U.S. Department of Interior Bureau of Reclamation, 157, Sep. 2011. [33] H. L. Barnes and S. B. Romberger, “Chemical aspects of acid mine drainage,” Water Polution Control Fed., vol. 40, no. 3, pp. 371–384, Mar. 1968. [34] L. O. Haluszczak, A. W. Rose, and L. R. Kump, “Geochemical evaluation of flowback brine from Marcellus gas wells in Pennsylvania, U.S.A.,” Appl. Geochem., vol. 28, pp. 55–61, Jan. 2013. [35] L. S. Land and D. R. Prezbindowski, “The origin and evolution of saline formation water, lower cretaceous carbonates, south-central Texas, U.S.A.,” J. Hydrol., vol. 54, no. 1–3, pp. 51–74, 1981. [36] P. F. Ziemkiewicz and Y. Thomas He, “Evolution of water during Marcellus Shale gas development: A case study in West Virginia,” Chemosphere, vol. 134, pp. 224–231, Sep. 2015. [37] J. D. Toner and D. C. Catling, “A Low-Temperature Thermodynamic Model for the Na-K-Ca-Mg-Cl System Incorporating New Experimental Heat Capacities in KCl, MgCl 2 , and CaCl 2 Solutions,” J. Chem. Eng. Data, vol. 62, no. 3, pp. 995–1010, Mar. 2017. [38] M. M. DiPippo, K. Sako, and J. W. Tester, “Ternary phase equilibria for the sodium chloride–sodium sulfate–water system at 200 and 250 bar up to 400 C,” Fluid Phase Equilibria, vol. 157, no. 2, pp. 229–255, 1999. [39] Phillip E. Savage, Sudhama Gopalan, Thamid I. Mizan, Christopher J. Martino, and Eric E. Brock, “Reactions at supercritical conditions: Applications and fundamentals,” AIChE J., vol. 41, no. 7, pp. 1723–1778, Jul. 1995.

! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

Thesis and Dissertation Services ! !