UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Fouling by calcium oxalate in aqueous solution Lencar, Diana R. 2001

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

Item Metadata

Download

Media
831-ubc_2001-0448.pdf [ 7.57MB ]
Metadata
JSON: 831-1.0058675.json
JSON-LD: 831-1.0058675-ld.json
RDF/XML (Pretty): 831-1.0058675-rdf.xml
RDF/JSON: 831-1.0058675-rdf.json
Turtle: 831-1.0058675-turtle.txt
N-Triples: 831-1.0058675-rdf-ntriples.txt
Original Record: 831-1.0058675-source.json
Full Text
831-1.0058675-fulltext.txt
Citation
831-1.0058675.ris

Full Text

FOULING BY CALCIUM OXALATE IN AQUEOUS SOLUTION by Diana R. Lencar B.Sc. (1992) Polytechnic Institute of Bucharest, Romania A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE STUDIES Department of Chemical and Biological Engineering We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April, 2001 © Diana Lencar, 20,0^ >o In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of CHBMJcfiL ANb /5lOLOGilC4L 6ti GINEERIN Gt The University of British Columbia Vancouver, Canada Date f~1c\yj 7 \ 2o0 / DE-6 (2/88) Abstract Calcium oxalate, a normal solubility salt, gives rise to deposits in the human kidneys, and in process equipment in the food and pulp and paper industries. Although crystallization kinetics are available for medical conditions, and solubility data have recently been developed for bleach plant situations, no deposition studies under industrial processing conditions have been reported. In this work, the solubility-temperature relationship for calcium oxalate was first confirmed in beaker tests. A heat transfer loop originally designed for heating experiments was completely re-built to permit thermal fouling measurements under cooling conditions. Supersaturated solutions of calcium oxalate at controlled pH and 75°C formed by rnixing calcium nitrate and sodium oxalate solutions from a 175 L tank were recirculated over periods of about 6 days through a double pipe heat exchanger cooled with water at 5-10°C. The decline in heat transfer coefficient as the deposit formed was monitored with time. A continuous bleed of fresh chemicals was necessary to sustain fouling. The effects on fouling rate and deposition morphology of the pH, initial supersaturation, calcium/oxalate ratio and flow rate were explored. Interpretation of results was complicated by the presence of suspended crystals at high supersaturation. The fouling thermal resistance increased linearly with time so long as fresh oxalate species were added. Fouling ceased when the particles concentration in the solution decreased. In the 6-day runs, the overall heat transfer coefficient declined by 7-58% depending on operating conditions. The pH was varied between 2.2 and 6.3. The initial fouling rate was found to rapidly increase with pH reaching a maximum at pH = 3 and ii then rapidly decrease and remain constant for pH>4.5. The initial fouling rate also increased with Reynolds number from 8,000, reached a maximum at about 15,000, and then decreased for Re up to 22,000. The maximum initial fouling rate was obtained for a Ca/Ox ratio of 2/1. The initial fouling rate was found to decrease as the initial relative supersaturation increased from 1 to 6 and then remained fairly constant as the supersaturation was further increased to 14.5. Deposits were confirmed to be calcium oxalate monohydrate through thermogravimetric and wet chemical analyses. ii i TABLE OF CONTENTS ABSTRACT ii LIST OF TABLES vi LIST OF FIGURES vii ACKNOWLEDGEMENT ix 1 INTRODUCTION 1 1.1 Categories of fouling 2 1.2 Stages of the fouling process and factors that affect them 4 1.3 Measurement of heat exchanger fouling 6 2 LITERATURE REVIEW 9 2.1 Chemistry of calcium oxalate 10 2.2 Studies regarding the solubility of calcium oxalate 14 2.3 Crystallization forms of calcium oxalate 18 2.4. Calcium oxalate crystallization 20 2.5 Calcium oxalate scaling control in the pulp and paper industry .24 3 OBJECTIVES 27 4 EXPERIMENTAL APPARATUS AND PROCEDURE 28 4.1 Test Section 28 4.2 Hot water flow loop 31 4.3 Cooling water circuit 32 4.4 Fresh chemicals automatic feed system 32 iv 4.5 Reagents used 34 4.6 Step by step procedure for a scaling run 34 4.7 Cleaning the experimental apparatus 36 5 SOLUBILITY TESTS 38 5.1 Introduction 38 5.2 Experimental Setup 41 5.3 Experimental Procedure 43 5.4 Analytical Techniques 44 5.4.1 Soluble Ca^ determination 44 5.4.2 Soluble C 2 0 4 = determination 44 5.5 Results and discussion 45 5.6 Conclusions 50 6 RESULTS AND DISCUSSION 51 6.1 General Overview 51 6.2 Preliminary experiments 52 6.3 Calculation of fouling resistance versus time from raw data 59 6.3.1 Fouling resistance 59 6.3.2 Fouling rate 60 6.3.3 Data presentation 61 6.3.4 Correction factor for fluctuations in the hot water flow rate 61 6.4 Reproducibility of the results 62 6.5 Deposit morphology 65 6.6 Flow rate effect on calcium oxalate fouling 66 6.7 pH effect on the fouling process 71 6.8 Initial relative supersaturation effect on the fouling process 75 6.9 Ca/Ox ratio influence on initial fouling rate 7 9 6.10 Deposit analyses 84 6.10.1 Thennogravimetric analysis of deposit 84 6.10.2 SEM and EDX analyses of the deposit 87 6.10.3 Quantitative analysis of the total calcium and oxalate content of deposits...,...., 89 7 CONCLUSIONS 92 8 RECOMMENDATIONS FOR FUTURE STUDY 94 NOMENCLATURE _ 95. REFERENCES 99 APPENDIX I.... 103 APPENDIX II 106 APPENDIX III 114 List of Tables TABLE 5.5.1 S U M M A R Y OF T H E MAIN PARAMETERS A N D RESULTS FOR SOLUBILITY TESTS PERFORMED A T PH = 4 49 TABLE 5.5.2 S U M M A R Y OF T H E MAIN PARAMETERS A N D RESULTS FOR SOLUBILITY TESTS PERFORMED AT PH = 3.5 49 TABLE 6.1.1: SOLUBILITY PRODUCT OF C A L C I U M O X A L A T E IN AQUEOUS SOLUTION FOR DIFFERENT TEMPERATURES A N D PH V A L U E S 52 TABLE 6.6.1: F L O W R A T E E F F E C T O N C A L C I U M O X A L A T E FOULING 67 TABLE 6.7.1: PH E F F E C T O N T H E C A L C I U M O X A L A T E FOULING .73 TABLE 6.8.1: INITIAL R E L A T I V E SUPERSATURATION (SIO) EFFECT O N T H E C A L C I U M O X A L A T E FOULING 79 TABLE 6.9.1: C A / O X EFFECT O N T H E C A L C I U M O X A L A T E FOULING 80 TABLE 6.10.1.1: T G A RESULTS FOR DEPOSIT A N A L Y S E S A N D COMPARISON WITH C A L C U L A T E D V A L U E S 85 TABLE III-l: CALCULATED CRYSTAL CONCENTRATIONS FOR SUSPERSATURATED CONDITIONS 114 y i List of Figures FIGURE 1.1.1: TYPES. OF FOULING CURVES 3 FIGURE 1.3.1:TEMPERATURE PROFILES AND THERMAL RESISTANCES ...8 FIGURE 2.1.1: OXALIC ACID IONIZATION AS A FUNCTION OF PH 13 FIGURE 2.2.1: SOLUBILITY OF CALCIUM OXALATE IN PULP MILL STREAM FILTRATES (CONTINUOUS LINES) AND PURE WATER (THE DASHED LINE) 15 FIGURE 2.2.2: THE OBSERVED SOLUBILITY OF CALCIUM OXALATE AS A FUNCTION OF PH AT CONSTANT IONIC STRENGTH 16 FIGURE 2.2.3: THE OBSERVED SOLUBILITY OF CALCIUM OXALATE AS A FUNCTION OF CA/OX RATIO FOR PH=3 AND PH = 5 17 FIGURE 2.2.4: THE OBSERVED SOLUBILITY OF CALCIUM OXALATE AS A FUNCTION OF PH 17 FIGURE 2.3.1: VARIATION OF ACTIVITY PRODUCT WITH TIME 19 FIGURE 2.5.1: SOLUBILITY OF CALCIUM AS A FUNCTION OF ALUMINUM ADDITION 25 FIGURE 4.1: PICTURE OF THE EXPERIMENTAL APPARATUS USED FOR CALCIUM OXALATE FOULING RUNS 29 FIGURE 4.2: SCHEMATIC REPRESENTATION OF THE EXPERIMENTAL APPARATUS USED FOR CALCIUM OXALATE FOULING RUNS 30 FIGURE43: TEST SECTION 33 FIGURE 5.1.1: CALCIUM OXALATE SOLUBILITY IN PURE SOLUTIONS AS A FUNCTION OF PH 39 FIGURE 5.1.2: SOLUBILITY VALUES REPORTED BY ULMGREN AND RADESTROM (1997) FOR THE CONDITIONS OF INTEREST FOR THE PRESENT STUDY 40 FIGURE 5.2.1: SOLUBILITY TESTS EXPERIMENTAL SETUP 42 FIGURE 5.4.1: TITRATION CURVE FOR C 2 0 4 = 46 FIGURE 5.5.1: RESULTS OF SOLUBILITY TESTS AT PH = 3.5, AND CA/OX MOLAR RATIO = 2/1 : 4 7 FIGURE 5.5.2: RESULTS OF SOLUBILITY TESTS AT PH = 4, AND CA/OX MOLAR RATIO = 2/1 48 FIGURE 6.2.1: VARIATION OF THE INVERSE OF OVERALL HEAT TRANSFER COEFFICIENT AND CONCENTRATION OF OXALATE IONS IN SOLUTION IN EXPERIMENT F4 54 FIGURE 6.2 2: VARIATION OF THE INVERSE OF OVERALL HEAT TRANSFER COEFFICIENT AND CONCENTRATION OF OXALATE IONS IN SOLUTION IN EXPERIMENT F5 55 FIGURE 6.2.3: VARIATION OF THE INVERSE OF OVERALL HEAT TRANSFER COEFFICIENT AND CONCENTRATION OF OXALATE IONS IN SOLUTION IN EXPERIMENT F10 56 FIGURE 6.2.4: VARIATION OF THE INVERSE OF OVERALL HEAT TRANSFER COEFFICIENT AND CONCENTRATION OF OXALATE IONS IN SOLUTION TN EXPERIMENT F12 57 FIGURE 6.4.1: FOULING RESISTANCE VERSUS TIME EST EXPERIMENTS F18B AND F18, PRESENTED AS MOVING AVERAGE 63 FIGURE 6.4.2: FOULING RESISTANCE EVOLUTION VERSUS TIME FOR F12 AND F24, PRESENTED AS MOVING AVERAGE OVER 21 DATA POINTS 64 FIGURE 6.4.3: OXALATE CONCENTRATION PROFILE DURING F12 AND F24 64 FIGURE 6.6.1: DEPENDENCE OF FOULING RESISTANCE ON THE SOLUTION FLOW RATE (RESULTS PRESENTED AS MOVING AVERAGE OVER 21 DATA POINTS) 68 FIGURE 6.6.2: DEPENDENCE OF INITIAL FOULING RATE ON THE SOLUTION FLOW RATE .. 68 FIGURE 6.6.3: DEPOSITS MORPHOLOGY AS A FUNCTION OF FLOW RATE 70 FIGURE 6.7.1. DEPENDENCE OF FOULING RESISTANCE ON THE PH (RESULTS PRESENTED AS MOVING AVERAGE OVE 21 POINT 72 FIGURE 6.7.2: DEPENDENCE OF INTTIAL FOULING RATE ON PH 72 FIGURE 6.7.3. DEPOSITS MORPHOLOGY AS A FUNCTION OF SOLUTION PH 74 vii FIGURE 6.8.1: DEPENDENCE OF THE FOULING RESISTANCE ON THE INITIAL RELATIVE SUPERSATURATION (RESULTS PRESENTED AS MOVING AVERAGE OVER 21 DATA POINTS) 76 FIGURE 6.8.2: DEPENDENCE OF THE FOULING RESISTANCE ON THE INITIAL RELATIVE SUPERS ATURATION S I O 76 FIGURE 6.8.3: DEPOSITS MORPHOLOGY AS A FUNCTION OF THE INITIAL RELATIVE SUPERSATURATION SK>....:.: 7 8 FIGURE 6.9.1: DEPENDENCE OF THE FOULING RESISTANCE ON CA/OX INITIAL RATIO (RESULTS PRESENTED AS MOVING AVERAGE OVER 21 DATA POINTS) 81 FIGURE 6.9.2: DEPENDENCE OF THE FOULING RESISTANCE ON CA/OX INITIAL RATIO 81 FIGURE 6.9.3: DEPOSIT MORPHOLOGY AS A FUNCTION OF CA/OX MOLAR RATIO 83 FIGURE 6.10.2.1: SEM ANALYSIS, CRYSTALS FROM THE TEST TUBE SURFACE 88 FIGURE 6.10.2.2: EDX RESULT FOR F 28 89 FIGURE 1-1: CALIBRATION CURVE FOR MAGNETIC FLOW METER 103 FIGURE 1-2: CALIBRATION CURVE FOR ROTAMETER 104 FIGURE 1-3: WILSON PLOT 105 FIGURE HI-1: EXPERIMENT F15 - OVERALL HEAT TRANSFER COEFFICIENT, TOTAL CALCIUM AND TOTAL OXALATE CONCENTRATION VS. TIME 115 FIGURE III-2 EXPERIMENT F16 - OVERALL HEAT TRANSFER COEFFICIENT, TOTAL CALCIUM AND TOTAL OXALATE CONCENTRATION VS. TIME 116 FIGURE m-3 EXPERIMENT F18 - OVERALL HEAT TRANSFER COEFFICIENT, TOTAL CALCIUM AND TOTAL OXALATE CONCENTRATION VS. TIME 117 FIGURE m-4:EXPERIMENT F19 - OVERALL HEAT TRANSFER COEFFICIENT, TOTAL CALCIUM AND TOTAL OXALATE CONCENTRATION VS. TIME 118 FIGURE III-5 EXPERIMENT F20 - OVERALL HEAT TRANSFER COEFFICIENT, TOTAL CALCIUM AND TOTAL OXALATE CONCENTRATION VS. TIME 119 FIGURE ELI- 6.EXPERIMENT F21 - OVERALL HEAT TRANSFER COEFFICIENT, TOTAL CALCIUM AND TOTAL OXALATE CONCENTRATION VS. TIME 120 FIGURE III-7EXPERIMENT F22 - OVERALL HEAT TRANSFER COEFFICIENT, TOTAL CALCIUM AND TOTAL OXALATE CONCENTRATION VS. TIME 121 FIGURE ffi-8 EXPERIMENT F24 - OVERALL HEAT TRANSFER COEFFICIENT, TOTAL CALCIUM AND TOTAL OXALATE CONCENTRATION VS. TIME 122 FIGURE III-9 EXPERIMENT F26 - OVERALL HEAT TRANSFER COEFFICIENT, TOTAL CALCIUM AND TOTAL OXALATE CONCENTRATION VS. TIME 123 FIGURE ni-lOEXPERIMENT F27 - OVERALL HEAT TRANSFER COEFFICIENT, TOTAL CALCIUM AND TOTAL OXALATE CONCENTRATION VS. TIME 124 FIGURE m-11 EXPERIMENT F28 - OVERALL HEAT TRANSFER COEFFICIENT, TOTAL CALCIUM AND TOTAL OXALATE CONCENTRATION VS. TIME 125 FIGURE III-12EXPERIMENT F29 - OVERALL HEAT TRANSFER COEFFICIENT, TOTAL CALCIUM AND TOTAL OXALATE CONCENTRATION VS. TIME 126 FIGURE m-13 EXPERIMENT F30 - OVERALL HEAT TRANSFER COEFFICIENT, TOTAL CALCIUM AND TOTAL OXALATE CONCENTRATION VS. TIME 127 FIGURE IH-14EXPERIMENT F17 - OVERALL HEAT TRANSFER COEFFICIENT, TOTAL CALCIUM AND TOTAL OXALATE CONCENTRATION VS. TIME 128 Acknowledgement I wish to express my sincere thanks to Dr. A.P. Watkinson, my supervisor, for his beneficial guidance, utmost patience and support throughout the duration of this work. I am also thankful to the workshop staff and the electrical workshop for their help in installation and maintenance jobs. Special thanks are also due to Mr. Horace Lam of the stores for his help in ordering and procuring chemicals and parts. Financial support provided by NSERC is also gratefully acknowledged. This volume is dedicated to my parents, Viorica and Ilarie Hinoveanu. Chapter I - Introduction 1 Introduction Defined by Epstein (1983) as the accumulation of undesirable solid materials on heat transfer surfaces, fouling is a phenomenon that causes important supplementary costs and production downtime in many industrial applications. These deposits can significantly reduce the heat transfer rate and increase the pressure drop, which results in important energy losses. In order to compensate for this decrease in heat transfer, the heat exchangers are initially designed with an excess surface, which further increases the costs. Despite the high costs associated with fouling, for decades only a very limited attention was given to this subject. Fouling resistance values are recommended by TEMA (Tubular Exchangers Manufacturers Association) standards for certain conditions and are largely used for heat exchangers design. However, these tables can only be applied for shell and tube heat exchangers and have the major disadvantage of considering a constant fouling resistance value from the beginning of the heat exchanger operation. In recent years, more research effort has been put into rigorous study of fouling. The first step is to understand the mechanism governing the fouling process and then to develop models that can describe with reasonable accuracy the fouling risk in the specific conditions of each system. 1 Chapter 1 - Introduction 1.1 Categories of fouling A generally accepted classification of fouling categories includes: • Crystallization fouling — this category includes both precipitation and solidification fouling according to Epstein (1983); precipitation fouling, also known as scaling, refers to the crystallization of dissolved salts onto the heat transfer surface; solidification fouling is somehow different, it involves the freezing onto the heat transfer surface of a liquid substance from a mixture due to its higher solidification temperature; some scientists have considered solidification fouling as a different category. • Particulate fouling — implies the migration and deposition of suspended particles from the solution onto the heat transfer surfaces; when this is due to gravity, the process is also called sedimentation. • Chemical reaction fouling - in this case the deposit is formed as a result of a chemical reactions taking place within the flowing fluid; the heat transfer surface is not a reactant in this situation but sometimes it may act as a catalyst. • Corrosion fouling - is due to the accumulation of corrosion products resulting from the reaction between the heat exchanger surface and the flowing fluid. • Biological fouling - in this case the deposit is formed by micro-organisms and their products, as well as macro-organisms that may attach to the surface to use the micro-organisms as a source of food. The extent of fouling is characterized by its thermal resistance, which for a 2 Chapter 1 - Introduction thin deposit is given by: Rf = x/Xf, where x is the thickness of the deposit and Xf is its thermal conductivity. Monitoring the fouling resistance evolution in time, four types of fouling curves were identified: - Linear: when the fouling resistance increases in time at a constant rate; - Falling rate: when the fouling resistance increases in time at a declining rate but the fouling rate never reaches zero; - Asymptotic: the fouling rate decreases over time to zero, at which time the fouling resistance approaches an asymptotic value; - Saw tooth: in this case, due to the weak deposit consistency the fouling resistance increase is periodically interrupted by decreases. Overall, linear or falling rate behaviour is observed. Figure 1.1.1; Types of Fouling Curves May time, t d Time 3 Chapter 1 - Introduction 1.2 Stages of the fouling process and factors that affect them In all the fouling categories presented in the previous paragraph the following sequence of events usually occur: - Initiation: a delay period before measurable fouling occurs; for crystallization fouling it is related to nucleation time whereas for chemical reaction fouling the buildup time for fouling precursors might be involved; the heat transfer surface characteristics can also play an important role in this stage of fouling (roughness and high energy metal surfaces can decrease the initiation time); other factors that can reduce the delay time are the increased temperature and supersaturation. - Transport: is the transfer of foulant, precursors or particles from the bulk solution to the heat transfer surface. The deposition flux (ma) is proportional to the driving force, which in this case is the difference in the concentration of transferred spiecies in the bulk and at the heat transfer surface (ct,-cs). The proportionality factor is the transport coefficient (k t). md = kt(cb-cs) (1.2.1) When the transferred species are ions or molecules, one deals with diffusional transport, which means that the transport coefficient is equal to the mass transfer coefficient (km). The mass transfer coefficient can be determined from correlations available in literature when the diffusivity and flow characteristics are known. - Attachment: in the case of crystals that nucleate and grow on the heat transfer surface, the attachment is referred to surface integration and its rate is given by: md = kr(crcsat)n (1.2.2) where k r is a function of surface temperature: k r = A ' exp(-E a/RT s) 4 Chapter 1 - Introduction From Eq. 1.2.1-1.2.2 it follows: (1.2.3) 1 1 When the species transported to the heat transfer surface are particles, their surface concentration can be considered zero and Eq. 1.2.1 replaced by: where: S p is in the probability that a particle that reaches the heat transfer surface will remain there. - Removal: in some situations it is possible that parts of the deposit are removed from the deposit; in these situations the accumulation of the deposit on the tube will be equal to the difference between the deposition rate and the removal rate; Kern and Seaton (1959) proposed a model where the net deposition rate is expressed as the difference between a deposition term, and a removal term that is a function of fluid shear stress (x), the deposit mass per unit area (m), and the strength of the deposit (vj/): Integrating Eq. 1.2.5. with m and t as variable and initial condition m = 0 at t = 0, it follows: mj = k,SpCb (1.2.4) Bzm (1.2.5) m = m*(l-emc) (1.2.6) where: m* - asymptotic mass per unit area tc - time constant, tc - m/mr = m*/mj = y//Brs 5 Chapter I - Introduction - Aging: is the change of chemical or physical properties of the deposit that may occur in time; it may introduce complications if fouling is measured by thermal effects, because aging can cause the change of thermal conductivity of the deposit. 1.3 Measurement of heat exchanger fouling Fouling can be studied either on site, or in a laboratory using small scale model heat exchangers. The latter technique allows for a more in depth study of the mechanisms involved in the fouling processes, since process conditions can be better controlled. Several techniques can be used to determine the extent of fouling in laboratory equipment and the most common ones are presented below: - Direct weighing of the deposit: this method offers a very simple solution but cannot be applied in many practical situations. The following relation can be used to calculate the fouling resistance once the deposit weight per unit area (mf) and the physical characteristics (density pf and thermal conductivity A,f) of the deposit are known: dR/=dm//( pjXf) - Deposit thickness measurement: this technique can relatively easily be applied for deposits on the outside of the probe that have a hard consistency. However, things are more difficult when the deposit is inside the tube, or when dealing with deformable deposits especially since often the thickness of the deposit is very small. Bott (1997) describes a method based on electrical conductivity measurements that can be employed to determine the thickness of this kind of deposits. - Heat transfer measurements: this method implies the determination of changes that occur in the heat transfer during fouling due to the increased thermal resistance 6 Chapter I - Introduction caused by the deposit layers. In this case, the fouling resistance is calculated as the difference between the total thermal resistance before fouling occurred and the total thermal resistance after deposits have started to build up. Figure 1.3.1 shows the temperature profiles and the thermal resistances that exist in a system before and after the fouling occurred. Based on this figure, the heat flow and overall heat transfer coefficients U c (clean) and U f (fouled) based on the area A i and a mean temperature difference can be expressed as: Qc^UcAiCTbi-Tbdm (1.2.3) Qf=UfA1CrbJ-Tb2)m (1.2.4) ± = R I + R W + A R 2 (1.2.5) -]- = /?, +RN +RW + A* /2 +4-^ 2 (1-2-6) For plate heat exchangers the two areas A i and A2 are equal but for others (i.e. the double pipe heat exchangers used in the present study) they are different. From Eq 1.2.3-1.2.4 it follows that: R = J L = R +AR (1.2.5) In laboratory systems, one would prefer to design conditions such that fouling occurred on only one side of the exchanger surface. In APPENDIX II the calculation for the overall heat transfer coefficients as well as for the heat transferred between the two fluids is presented. 7 Chapter 1 - Introduction Figure 1.3.1; Temperature profiles and thermal resistances for clean (a) and fouled (b) heat transfer surface. Wall l b , l Hot fluid *-K M lw,2 l b , 2 Cold fluid b.l Wall Deposit lb.2 Cold fluid (a) (b) 8 Chapter 2 - Literature Review 2 Literature Review Calcium oxalate is an important source of scale formation in the food industry and in the pulp and paper industry. The deposits of calcium oxalate are hard, smooth and very difficult to remove causing downtime and additional production costs. In sulfite mills calcium oxalate scale can be found in weak liquor storage tanks, evaporators, heat exchangers, pipes, and liquor nozzles. Also, calcium oxalates, calcium carbonate and barium sulfate are the main species that cause scale problems in bleach plants (Houdlette 1985, Ulmgren and Radestrom 1997). In order to achieve the demands of new and more severe environmental regulations, the tendency is to increase the closure level of the pulp mills by recycling streams to approach zero effluent discharge. Besides the important advantages of closed processes, one of the related disadvantages is the increased risk of scale formation due to the increased filtrate concentration. Calcium oxalate forms when both Ca44" and oxalate anion are in solution. The source for oxalate anion is oxalic acid. Oxalic acid comes with wood, especially with bark, but is also formed along the cooking and bleaching stages. Krasowski and Morton (1982, 1983) have shown that in the bleaching process, oxalic acid results mainly from lignin oxidation. They concluded that the amount of oxalic acid formed in the bleaching stage is proportional to the initial lignin content of the pulp. Nilvebrant and Reiman (1996) have found that the source of oxalic acid in the ozone bleaching process is not necessarily the lignin. Based on the experiments they performed, they concluded that a hexonuronic acid containing xylan is the main source when pulp is bleached with industrially used charges of ozone and lignin becomes the main source when much higher ozone charges are used. Calcium comes mainly with the wood and with the fresh water. 9 Chapter 2 - Literature Review Softwood contains less calcium than hardwood and for all the wood species, most of the calcium is in the bark. In an article published in 1994, Ester explained that the common bleaching stages where oxalate can be found are chlorine dioxide, sodium hypochlorite and ozone stages. Calcium oxalate deposits usually occur between pH of 3 and 9, but at high levels of oxygen in the peroxide and oxygen bleaching stages, calcium oxalate scale can even be found at pH=l l . 2.1 Chemistry of calcium oxalate Calcium oxalate is the salt of oxalic acid, which is a dibasic acid. The following relationships should be taken into account in calcium oxalate solutions: - Oxalate species and water equilibrium: H2C2O4 -c->HC204" + H + K i (2.1.1) HC2O4" <->C204= + H + K 2 (2.1.2) - Ion-pair equilibrium: CaC 2 0 4 <->Ca++ + C 2 0 4 = K 3 (2.1.3) K i , K 2, K 3 have the following values at 298 K, quoted by Tomazic andNancollas (1980): Kx = -OSiSi £L_ = (18450)-' mol/L (2.1.1 .a) aH2C204 K2 = — = (17.861)-' mol/L (2.1.1.b) aHC204-10 Chapter 2 - Literature Review K^ =_c2o1 ^ _ = ( i537i)- ' m o / /z , aCaC2Ot (2.1.1.C) Using equations 2.1.1 and 2.1.2 and assuming the activity coefficients to be equal with unity, one can estimate the distribution of oxalate species vs. pH once the initial concentration of oxalic acid is known: Make the following notations: pH = p=> [H¥] = 10p Total oxalate added initially = q = [HC204~] + [C2Of] + [H2C204] (2.1.4) From Eq. 2.1.2 it follows: [HC204] = K2 10p [C204=] (2.1.5) From Eq. 2.1.1 and 2.1.5 it follows: [H2C204] = Kj K2 10'2p [C204=] (2.1.6) Replacing in Eq. 2.1.4: q =K2 Wp [C204=]+ [C2Or] + Ki K2 10-2p [C204=] (2.1.7) Whence: [C 20 4 =] = q/(l + K2 10* + K, K2 l(T2p) (2.1.8) FromEq. 2.1.8 and 2.1.5: [HC204-J = K2 10p q/(l + K2 l(fp + K, K2 W2p) (2.1.9) From Eq. 2.1.8 and 2.1.6 [H2C204] = K, K2 ia2p q/(l + K2 Wp + K, K2 ia2p) (2.1.10) 11 Chapter 2 - Literature Review Based on Eq. 2.1.8 - 2.1.10, and considering q = 0.0008 mol/L, the oxalate species distribution diagram from Figure 2.1.1 was calculated at 298 K. The species distribution at 343 K (the temperature of interest for the fouling experiments), was calculated by evaluating K i and K 2 at 348 K, using the following relation (Martell and Smith, 1979): logKT= logK298 + AH(T-298) /(2.303R(298)2)=> logKT= logK298 + AH(T-298)0.00246 (2.1.11) Enthalpy values at 298 K were obtained from the same source (Martell and Smith, 1979): AHi = 1.6 kcal/mole, AH 2 = 0.9 kcal/mole Eq. (2.1.11) is based on the assumption that AC P = 0. This assumption is considered acceptable for the purposes of this work. Where more concentrated electrolytes are used, appropriate activity coefficients should be calculated for use in Eq. 2.1.1.a-c. 12 O N o o VO i n xt- C N I—1 O o o o o o o o o O o o o o o o o o o o o o o o o o o o o o o o o o o " o o [q/JOUl] U 0 U B J J U 3 3 U 0 3 13 Chapter 2 - Literature Review 2.2 Studies regarding the solubility of calcium oxalate Calcium oxalates are components of urinary stones and because of that, much of the research on calcium oxalates is related to medical situations. In the last few years, the problems related with calcium oxalate fouling in pulp and paper industry have earned an increased attention and some solubility studies were carried out for calcium oxalates at higher temperatures and also in mill streams. Ulmgren and Radestrom (1997) have studied the solubility of calcium oxalate in pure supersaturated solutions and also in filtrates from a kraft bleached pulp mill. In such mills with a high degree of closure, calcium oxalates deposits occur not only in the bleaching stages but also when the acidic filtrates are stored at high temperatures. The experimental data were used together with a computer program (SOLGASWATER), to elaborate a steady state model that can be used to estimate the conditions under which calcium oxalate deposits form in a bleach plant. The experiments were carried out in polytetrafluoroethylene containers immersed in a water bath at constant temperature with continuous shaking or stirring. The calcium oxalate pure supersaturated solutions were obtained by adding calcium and oxalate ions in excess to the hot experimental solution. The experiments were carried out at 50, 70 and 90°C. The ionic strength was adjusted with sodium chloride and the pH with hydrogen chloride or sodium hydroxide. They found that the solubility of calcium oxalate defined by the solubility product L s (L s = [Ca 2 +] s a t[Ox 2"] s at) is much higher in bleach plant filtrates than in pure solution (Figure 2.2.1). The authors explained this aspect as a result of the possible formation of colloids of calcium oxalates particles and the formation of complexes between calcium 14 Chapter 2 - Literature Review ions and carboxylic groups of the organic acids present in the filtrates. As the pH is raised from 2 to 6, the solubility decreases markedly. Figure 2.2.1: Solubility of calcium oxalate in pulp mill stream filtrates (continuous lines) and pure water (the dashed line). Note: COD is the chemical oxygen demand and is a measure of the organic mater content. (Ulmgren and Radestrom, 1997) They also found that in both pure solution and plant filtrate the solubility of calcium oxalate increases with temperature and with ionic strength of the solution (Figure 2.2.4, 2.2.1 and 2.2.2). The solubility of calcium oxalate was also studied as a function of Ca/Ox molar ratio and these results are presented in Figure 2.2.3. The solubility has also 15 Chapter 2 - Literature Review decreased with pH increase when pH<4 and for pH>4 the solubility does not depend upon the pH variations (Figure 2.2.4). The main component of deposits for both types of solutions was calcium oxalate monohydrate. The degree of supersaturation needed to obtain a spontaneous precipitation was low. Once started, the precipitation speed of calcium oxalate was high. Figure 2-2; The observed solubility of calcium oxalate as a function of pH at constant ionic strength (0.03, 0.1, 0.3 and 1 mol/L). Log L a (mol/L)2 PH Note: The temperature was 70°C, Ca/Ox ratio was lmole/mole. (Ulmgren and Radestrom, 1999) 16 Chapter 2 - Literature Review Figure 2.2.3: The observed solubility of calcium oxalate as a function of Ca/Ox ratio for pH=3 and pH = 5. log L s , <mol/L)' -6.0 h -6.5 h 1 2 3 Ca/Oxf mol/mol Note: Temperature was 70°C and ionic strength 0.1 mole/L. Continuous line - in presence of 5 mmol/L magnesium ions, dashed line - in pure aqueous solution. (Ulmgren and Radestrom, 1999) Figure 2.2.4: The observed solubility of calcium oxalate as a function of pH at 30, 50,70 and 90°C. log Lg.CmoVty -5 -6 -7 O o 30°C 50'C a A 90°C A a. M i; .... i • 2 4 6 8 10 PH Note: The Ca/Ox ratio was 1 mole/mole, and the ionic strength 0.1 mol/L. (Ulmgren and Radestrom, 1999) 17 Chapter 2 - Literature Review 2.3 Crystallization forms of calcium oxalate Calcium oxalate forms three hydrates, the monoclinic monohydrate (whevellite), the tetragonal dihydrate (weddelite), and triclinic trihydrate. Monohydrate is the thermodynamically stable form. Brecevic and Skrtic (1985) have performed batch precipitation experiments in order to study the kinetics of di- and tri-hydrate calcium oxalates transformation in the thermodynamically stable monohydrate. Figure 2.3.1 presents their results. Observing the increase in the activity product of the solution as the metastable hydrates dissolve under the influence of undersaturation, followed by the fall of the activity product as the growth of monohydrate dominates, the authors concluded that the transformation mechanism seems to be solution mediated with a first stage consisting in the dissolution of di- and tri-hydrates followed by the growth of monohydrate form. They also found that the type of mixing influenced the initial stage of precipitation distribution. Thus, when a propeller was used for mixing, all three hydrates precipitated (COT in the highest proportion and COD and C O M in about the same amounts) whereas when a magnetic stirrer was used, COT and C O M only precipitated with a higher COT proportion. The diameter and rotation speed were similar for propeller and magnetic stirrer. Whence, the difference in precipitation distribution was probably caused by the magnetic field rather than by a difference in mixing intensity. 18 Chapter 2 - Literature Review Figure 2-5: Var i a t ion of activity product wi th time. 6 f 0_ < 2 L * flL <\i_ 2> ° U < ^ S O . COD 3 0.2 M 2D (a) Unstirred 40 T ime |h] KSO.COM , 0.2 M 100 < JL •SO, C O I 0.2 M Kso.cor;,0.2M Kso. C O M . 0.2 M 20 (b] Magnetic stiiiei 40- 60 Time [h] 90 Note: The dashed lines indicate the thermodynamic solubility products of the three hydrates. (Brecevic and Skrtic, 1985) 19 Chapter 2 - Literature Review 2.4 Calcium oxalate crystallization Nancollas and Gardner (1973) have studied the kinetics of crystal growth for calcium oxalate monohydrate in supersaturated suspensions, in a temperature range between 15-45°C and pH = 6.5. A l l experiments used calcium oxalate seeds to initiate the nucleation process. The supersaturation was in the range 0.2 to 4 and was defined as: S = ([Ca2+],-[Ca2+]Q)/[Ca2+]0 (2.4.1) Where: [Ca2+]i - total concentration of Ca ions added [Ca 2 +] 0 - ionic solubility under experimental conditions For Ca/Ox = 1/1, they found a initial fast reaction followed by a period when the rate of crystallization was described by an equation of form: d[Ca2+] = kgsaCa2+]-[Ca2+]oY (2-4.2) dt Where: s - surface of the added seeds crystals kg - the observed rate constant for crystal growth The authors do not specify the value of seed crystals surface. They mention though that the seeds were of prismatic shape and the edge length was in the range 1-2 pm and 4 - 5 pm . kg was found to decrease with the increase in seeds concentration (1.7-44 mg seeds/100 ml) in the range 6.1 - 0.4 [L mol"1 min'^mg seeds/100 ml"1)]. For uneven Ca/Ox ratios, ranging from 3.5 to 0.28, the rate of crystallization was expressed as: dt g = k sN2 (2.4.3) 20 Chapter 2 - Literature Review Where N is the number of mole/L calcium oxalate monohydrate deposited from the supersaturated solution before equilibrium is reached. N was calculated from the following equation: U =/2([Ca2+J - NXfOx2-] -N)=21CT9 mof/L2 (2.4.4.a) Where f2 is the activity coefficient of the divalent cation and anion, assumed to be equal and calculated with the extended form of Debye-Huckel equation (cited by authors): -log f2 = 2.04 z 2 [I 1 / 2/(l + IM) - 0.3 I) (2.4.4.b) Where I is the ionic strength of the solution [mole/L] The rate constant (kg) ranged from 1.1 to 8.7 [L mol"1 min"'(mg seeds/100 ml"1], being higher when Ca/Ox > 1. From the experimental results they concluded that the crystallization mechanism of calcium oxalate monohydrate is surface reaction controlled rather than diffusion controlled. Thus, plots of log(k) vs. 1/T were linear (T between 15-45 °C) with a slope corresponding to an activation energy Ea = 11.7+/- 1 kcal/mole. The magnitude of Ea, and the negligible change in rate constant with mode and rate of stirring eliminated the idea of bulk diffusion as the limiting step. The authors suggest that assuming the crystal surface is surrounded by a hydrated monolayer, the crystals growth occurs through the simultaneous dehydration of pairs of calcium and oxalate ions at the active growth sites. In an experimental study made in 1975, Gardner has found that the crystallization rate for calcium oxalate trihydrate is proportional to the square of its supersaturation (expressed as in Eq. 2.4.1) in solution but is independent of the stirring rate and the temperature. Brown and others (1989) have studied the factors favoring the precipitation 21 Chapter 2 - Literature Review of calcium oxalate dihydrate. They have found that the precipitation of COD has occurred only in the presence of citrate and was more likely at 22°C than at 37°C, being favored by high calcium-to-oxalate ratios (Ca/Ox range studied was 3.5 - 9) and low relative supersaturation (<25, where the relative supersaturation is defined by authors as AP/LS). Nancollas and Tomazic (1980) have studied the crystallization kinetics for all the calcium oxalate hydrates, mono-, di- and tri-hydrates. The experiments have been carried out in supersaturated seeded calcium oxalate solution with sodium chloride at 37°C. They found that the rate of crystallization for the three hydrates decreases in the following order. kg(COT)>kg(COM)>kg(COD) (kg is the rate constant [Lmol 'min 'm 2]) and is proportional with the square of the relative supersaturation of each of them. Supersaturation with respect to each hydrate was expressed in terms of free energy change: -AG = RT ln(AP/LJ (2.4.4) Where AP is the activity product and is L s the thermodynamic solubility product. The range of supersaturation expressed in Eq. (2.4.4) was: 6.87 - 5.46 for COM, 4.81 - 3.4 for COD and 3.82 - 2.4 for COT (kJ/mole). The dissolution of calcium oxalate hydrates has been found to be diffusion controlled with a first order reaction rate that decreases in the order: ka(COT) >kd(COD)>kd(COM), with kd expressed as [min"' (mg seeds/100 ml)], by Tomazic and Nancollas in a previous study (1978). Brecevic and Skrtic (1985) have studied the dissolution kinetics of calcium oxalate dihydrate and trihydrate, and the consequent growth of calcium oxalate monohydrate. Experimental conditions were: 298 K, pH = 6.5, ionic strength of 0.2 22 Chapter 2 - Literature Review mol/L NaCl, unseeded and with Ca/Ox = 8.83. Three series of experiments were performed, first without stirring, the second one with mechanical stirring, and the third one with magnetic stirring. They obtained the following equations for dihydrate and trihydrate dissolution kinetics: COD: V-l,3^- = 2A5xl05(Ls-AP)06n (2.4.5) COT: V'in ^ = 1.12x10'(L s - APf1915 (2.4.6) For COM growth rate they obtained the following relationship: V~ui — = 9.3xlO-3S2 (2.4.7) dt Where: V - crystal volume [u.m /cm solution], and S - supersaturation (S = ln(AF7Ls)). Eq. 2.4.7 shows a good agreement with the relationship Nancollas and Gardner found for COM crystallization rate (Eq. 2.4.2). Babic and others (1985), have studied the influence of the initial concentration of calcium and oxalate anion in an unseeded solution containing also sodium chloride on the morphology and composition of calcium oxalates precipitates (at 298 K, pH = 6.5 and molar Ca/Ox =1). The analyses have shown that calcium oxalate has precipitated only as monohydrate or dihydrate. The dihydrate form precipitation was favored by excess concentration of calcium or oxalate ions. At supersaturation less than 48 (corresponds to Sb = 5.92), (where the relative supersaturation is defined by authors as AP/LS) they concluded that the crystallization rate is surface reaction controlled, whereas for higher supersaturations the diffusion of ions through the bulk becomes the limiting step. 23 Chapter 2 - Literature Review 2.5 Calcium oxalate scaling control in the pulp and paper industry Brackenbury and Seccombe (1997) have performed laboratory simulations to determine the usefulness of enhanced alkaline hydrogen peroxide bleaching (EAHPB) and hot acid hydrolyses (HAH) for improving the prospects of mill closure in a kraft pulp mill using filtrate recycling. HAH is a new metals control strategy in which the metals in chemical pulp are removed under neutral-to-alkaline conditions to prevent build up in the water circuit. Besides other advantages, they reported that the use of HAH resulted in reduced calcium oxalate scaling. Hultman and Nilsson (1981) have developed a method to avoid the deposition of calcium oxalate in a sulfite mill, using aluminum ions added as aluminum sulfate. The variations of calcium oxalate solubility in the sulfite liquor, with and without aluminum ions added, were calculated using a computer program (SOLGASWATER). The experiments have shown a minimum of the calcium oxalate solubility at pH: 3.5-4, that was slightly different than the calculated solubility which attended a minimum at pH=3. The temperature at which the experiments were carried out was 80°C. The pH and ion concentrations (Ca"1-1" and C 2 C V ) were varied. Figure 2.5.1 shows how the solubility of calcium oxalate has been found to increase linearly with increased amounts of aluminum ions added. Laboratory tests shown that addition of aluminum ions up to 50 mg/L, has no significant effect on paper quality. However, authors found that as little as 6 mg/L aluminum is sufficient to inhibit calcium oxalate scale formation. Besides the positive effect of aluminum addition, authors also mention its negative aspects, the formation of deposits containing aluminum in the pH range of 6-8. These deposits were found in the top of the reaction tower (deposits containing silica and 24 Chapter 2 - Literature Review Figure 2-6: Solubility of calcium as a function of aluminum addition. C a 2 * , mg/ l 4 Note: Oxalate addition was 660 mg/L, temperature 80°C and pH=4. (Hultman and Nilsson, 1981) aluminum) as well as in the heat exchangers following the reaction tower (aluminum hydroxide). High pressure water was used to wash the deposits in the reaction tower and sodium hydroxide to dissolve the aluminum hydroxide scale formed in the heat exchangers. The authors also outline the low costs associated with aluminum inhibitor addition. Perez and Zidovec (1995) conducted scale inhibition tests with a new environmentally friendly non-phosphorus (NPA) scale inhibitor. This inhibitor was 25 Chapter 2 - Literature Review developed and patented by Betz Laboratories and is based upon polyepoxysuccinic acid. According to authors, this scale inhibitor was proved efficient in preventing calcium carbonate, barium sulfate and calcium oxalate deposition under conditions typical to the paper industry. Calcium oxalate tests were conducted by using the static beaker test. In this test, the treatment was added to a solution containing calcium oxalate ions at the required temperature and pH in the range 4-9 .5 . The beakers were incubated at 60°C for 1.5 hours. A measured portion of the solution was then filtered and the calcium oxalate concentration measured by ICP. The percent inhibition was calculated from the analyses of the treated, stock and control solutions. Although authors do "not give many details about the testing procedure and quantities of inhibitor used, they concluded that at pH=4 NPA outperformed other commonly used inhibitors, its inhibiting effect being about five times higher. At pH=95, NPA inhibition effect was still 20% higher that that of the most efficient inhibitor used by authors for comparison. 26 Chapter 3 - Objectives 3 Objectives The literature review has shown that some data are available on solubility of calcium oxalates, and on crystallization kinetics. However, no studies have been reported on the fouling process in spite of the impact of calcium oxalate deposition in the pulp and paper industry. This study intends to determine the main parameters that influence the fouling behaviour of calcium oxalate as well as the extent of their influence. It also aims to shed some light on the mechanism involved in the fouling process. Therefore, the specific objectives of this work are: 1. To perform solubility tests for calcium oxalate in the range of temperature, pH and Ca/Ox molar ratio that are of interest for the fouling experiments, and to compare the results with the available literature data. 2. To build an experimental apparatus that meets the specific demands for studying the calcium oxalate fouling. 3. To determine the influence of the following parameters on the initial fouling rate as well as on the morphology and composition of the deposits: flow rate, pH, supersaturation level, and Ca/Ox molar ratio in the initial solution. 27 Chapter 4 - Experimental Apparatus and Procedure 4 Experimental Apparatus and Procedure The apparatus used for fouling experiments is presented in Figure 4.1 and Figure 4.2 An experimental unit previously used to study fouling in a heating system was rebuilt in order to meet the demands of the present research that involved a cooling system. The detailed description of this system is further presented in this chapter. Because of the increased solubility of calcium oxalate at high temperatures and low pH, it was necessary to build a system that allows high temperatures (~80°C) and is corrosion resistant. Stainless steel was whence used for all the equipment parts that came in contact with the experimental solution. 4.1 T e s t S e c t i o n The test section is a vertical double pipe heat exchanger (Figure 4.3). The shell consists of a 1.4 m long stainless steel tube having 3.81 cm internal diameter. At both ends the tube has Pyrex glass sections to allow for the visual observation of the fouling process during the experiments. The glass sections have the same internal diameter as the stainless steel shell has. The inner tube is also stainless steel and has a 1.905 cm external diameter and a 1.58 cm internal diameter. The cooling water and the hot solution containing calcium oxalate in counter current flow. The water flows downwards through the inner tube and the solution flows upwards through the annular section. Fouling is expected to occur on the outer surface of the inner tube. 28 Figure 4.1: Picture of the experimental apparatus for calcium oxalate fouling runs 29 Chapter 4 - Experimental Apparatus and Procedure Figure 4.4-1: Schematic representation of Ae experimental apparatus used for calcium oxalate fouling runs. Steam in Steam Oiit ^ Tap water Drain P C Data Logger A V " Steam T e m p . Controller Magnetic flowmeter Chemicals Supply Tank 7 r i i i Drain 30 Chapter 4 - Experimental Apparatus and Procedure 4.2 Hot water flow loop Because calcium oxalate is a sparingly soluble salt, the experimental solution cannot be readily prepared by dissolving the appropriate amount of calcium oxalate in water. Calcium oxalate is rather obtained by adding the corresponding quantities of calcium nitrate and sodium oxalate in the water contained in the supply tank. The water is heated up to the desired experimental temperature prior to adding the chemicals. The supply tank is a 220 L stainless steel tank. The solution in the tank is heated up with steam that flows through a stainless steel coil inside the tank. A control valve installed on the steam supply pipe allows variation in steam flow to maintain a constant temperature in the tank during the experiment. The actuator of the automatic valve is linked to an Omega CN temperature controller that also monitors the temperature in the tank through a copper constantan thermocouple installed in the supply tank at the lower side. The solution containing chemicals is pumped from the tank with a Cole-Palmer pump. The flow rate is monitored by a magnetic flow meter before entering the annulus of the test section, at its lower end. The magnetic flow meter is linked to the computer data acquisition system and the flow values measured can be recorded. The temperature of the solution is measured both at the entrance and the exit of the test section using copper-constantan thermocouples. These temperatures can also be recorded on the computer data logger. A back up system also allows the reading of temperatures and hot water flow rate on digital displays on the control panel. A l l the pipes and fittings that are in contact with the solution are stainless steel 316 L. In order to minimize the heat loss, 31 Chapter 4 - Experimental Apparatus and Procedure the entire hot solution loop has been insulated with fiberglass insulation (1" thickness), including the supply tank and the test section. A stainless steel filter was installed before the magnetic flow meter on a bypass line. 4.3 Cooling water circuit The cooling water is supplied from a 220 L plastic tank and it is directed into the inner tube of the test section using a Cole-Palmer pump. The flow rate is measured with a Brooks rotameter. The temperature of the cooling water is measured both at the inlet and outlet of the test tube using copper-constantan thermocouples. These temperatures are recorded on the computer data logger and can also be read on the digital display of the control panel. 4.4 Fresh chemicals automatic feed system Since a continuous addition of chemicals to the tank during the experiment proved to be necessary, two small pumps have been added to the initial setup. One of the pumps is an IDEX MICROPUMP 415A, which is used to transfer a calcium nitrate solution from a 25 L plastic container into the supply tank at a 13.8 ml/min. rate. The second pump is a VARIAN PCR-1 micropump and it is used to pump a sodium oxalate solution from an 8 L plastic container into the supply tank at a 5 ml/min. rate. 32 Chapter 4 - Experimental Apparatus and Procedure Figure 4.3: Test Section cold water inlet internal diameter = 2.5 cm hot solution outlet internal diameter = 2.5 cm e o LO glass section internal diameter = 3.81 cm stainless steel shell tube internal diameter= 3.81 cm E o o CO inner tube internal diameter = 1.37 cm outer diameter = 1.91 cm E o LO glass section internal diameter = 3.81 cm E o hot solution inlet internal diameter = 2.5 cm cold water outlet internal diameter = 2.5 cm 33 Chapter 4 - Experimental Apparatus and Procedure 4.5 Reagents used Calcium oxalate is a sparingly soluble salt with a normal solubility that increases with both temperature and acidity of the solution. As mentioned previously, calcium oxalate in solution was obtained by mixing solutions of highly soluble chemicals. Calcium nitrate and sodium oxalate were chosen because of their good solubility in water, as well as because at low pH the reaction product is nitric acid which is less corrosive than other acids (i.e. hydrochloric acid). The calcium nitrate was certified A.C.S. grade from Fisher Scientific with molecular formula Ca(NC>3)2*4H20 and M.W. = 236 g/mole. Sodium oxalate was bought from AVOCADO Research Chemicals Limited, with a purity of 99 wt.%. The molecular formula is Na2C204, and the M.W. is 134 g/mole. pH was adjusted by means of addition of nitric acid 20 wt.%, prepared from nitric acid 78 wt.% from Fisher Scientific. 4.6 Step by step procedure for a scaling run 1. Fi l l the supply tank with tap water to the level corresponding to 175 L. 2. Turn on the temperature controller for the supply tank and set temperature at 75°C. 3. Turn on the controlled steam valve actuator. 4. Open the steam valve and wait about 2 hours to heat up the water in the supply tank to 75°C. 5. Turn on the magnetic flow-meter and the temperature reading. 6. Turn on the computer and set the sampling time for temperatures and hot solution flow rate at every 10 minutes. 7. Turn on the pump for the hot solution loop and set the flow rate at the desired value. 8. Open the tap water valve and fil l up the cold-water tank. 34 Chapter 4 - Experimental Apparatus and Procedure 9. Turn on the pump for the cold-water loop. 10. Set the cold water flow rate at the desired value using the rotameter. 11. Adjust the opening of the tap water valve such as to maintain a constant level in the cold-water tank. 12. Turn on the computer data logger and set the sampling interval at 10 minutes for recording the hot water flow rate and the inlet and outlet temperatures of the cold and hot water. 13. Start the data acquisition on the computer to monitor the experimental data in clean conditions (without any chemicals added) for several hours (usually 5 to 12 hours). 14. Prepare separately in two 4 L plastic containers the calcium nitrate and sodium oxalate solutions. 15. Add first the sodium oxalate solution to the supply tank. 16. After about 15 minutes add calcium nitrate solution in the supply tank, slowly and using four different holes existing in the supply tank lid to avoid creating local supersaturation sites. 17. After another 15 minutes take sample to analyze for total calcium and respectively oxalate content of the solution and to check the pH; adjust the pH at the desired value by adding nitric acid. 18. Prepare the fresh makeup chemicals stock for 24 hours. 19. Turn on the two small pumps to start the automatic fresh chemicals feed at one hour after the calcium nitrate solution was added. 20. Take samples for calcium and oxalate analyses, check pH and remake the fresh makeup chemicals stock at every 24 h. 35 Chapter 4 - Experimental Apparatus and Procedure 21. Continue the run for 5-7 days. 22. During the run, feed the saved data to the excel program to calculate the overall heat transfer coefficient in order to monitor the fouling process. Make visual observation of the scale build up through the glass windows of the test section. 23. Stop the data saving and turn off the control switches for temperature and flow measurements when the run is finished. 24. Turn off the two pumps that fed fresh makeup chemicals in the supply tank. 25. Close the steam valve and turn off the actuator of the automatic steam valve. 26. Evacuate the hot water by opening both the exit valve at the bottom of the supply tank and the exit valve placed right above the supply tank. 27. Turn off the two pumps on the hot water and respectively cold water loop. 28. Carefully remove the inner tube of the test section and let it dray on a horizontal surface. 29. Let the system cool down. 30. Examine the dried deposit on the experimental tube and take pictures. 31. Scrape the deposit from the tube using a stainless steel spate and collect it in a plastic vial for further analyses. 4.7 Cleaning the experimental apparatus. 1. Open the exit valve at the bottom of the tank and let tap water flow through the supply tank to wash the crystals deposited at the bottom. 2. Closed the exit valve and fil l up the supply tank with tap water. 3. Turn on the steam and the automatic valve actuator to heat up the water in the tank at 75°C. 36 Chapter 4 - Experimental Apparatus and Procedure 4. Add nitric acid in the supply tank to lower the pH at about 3.5. 5. Replace the inner tube in the test section. 5. Turn on the pump on the hot solution loop. 6. Let the acidified hot water flow at high flow rate (about 0.7 kg/s) overnight. 7. Evacuate the water from the supply tank. 8. Repeat steps 2 to 5. Let the hot water run through the apparatus for a shorter time, around 3-4 hours. 9. Drain the water and turn off the steam and the pump. 10. Wash the bottom of the tank with tap water and i f crystals are still present repeat the cleaning procedure. 37 Chapter 5 - Solubility Tests 5 Solubility Tests 5.1 Introduction When studying the fouling behaviour of sparingly soluble salts the solubility data are of critical importance. This is because the difference between the solubility of the studied salt at bulk temperature and at the heat transfer surface temperature creates the driving force for the mass transfer of the species from the bulk to the surface which consequently leads to scale formation. Many solubility data can be found in the literature for calcium oxalate at room temperature and at the human body temperature, because most of the existing studies for this salt are related to the medical field. As described in Chapter 2 Ulmgren and Radestrom (1997, 1999, 2000) have performed a detailed study of calcium oxalate solubility at higher temperature, ranging between 50 and 90°C. Their results confirmed that calcium oxalate is a normal solubility salt, which means its solubility is increasing with the increase in temperature. Figure 5.1.1 is the graphical presentation of their 1997 findings. As mentioned by authors, the curves were generated with an equilibrium model using a computer program called SOLGASWATER. They also found that calcium oxalate solubility is strongly dependent on the pH for these higher temperatures. However, one can see in Figure 5.1.2 that as the temperature decreases, calcium oxalate solubility dependency on pH decreases. In their 1999 work (Figure 2.2.4), this effect is less noticeable. Because this was the only source of data in the range of interest, it was considered useful to conduct a set of solubility tests for calcium oxalate to compare with the data published by Ulmgren and Radestrom. 38 Chapter 5 - Solubility Tests Figure 5.1.1: Calcium oxalate solubility in pure solutions as a function of pH at 50, 70 and 90°C. Log L a (mol/L)2 _ § I i I . I i I i I i I . J I 0 2 4 6 8 10 12 pH Note: The ionic strength was 0.03 mol/L, and the ratio Ca/Ox = 3 moVmo\.(Ulmgren and Radestrom, 1997) The solubility of calcium oxalate was studied in pure aqueous solution for different temperatures and pH. The temperatures and pH values considered for experiments were chosen from the range of interest for the future fouling experiments. These values were 50°C, 60°C and 70°C for temperature and 3.5, 4 for pH. The molar ratio C a + + / C 2 0 4 ~ was in the range 2-2.2/1 in most of the experiments. The strategy adopted in this work to measure the solubility of calcium oxalate at different temperatures was to obtain a solution saturated in calcium oxalate at 75°C and 39 40 Chapter 5 - Solubility Tests then to decrease its temperature. Since calcium oxalate is a normal solubility salt, the decrease in temperature will make the solution become supersaturated and the excess calcium oxalate will then crystallize. If the solution is filtered and the crystals removed, the filtrate wil l be saturated in calcium oxalate at the lower temperature. Therefore, by measuring the concentration of the filtrate one can determine the solubility of calcium oxalate at the given conditions. Calcium oxalate was whence obtained by mixing H2C2C»4*H20 and Ca(NC»3)2*4H20. Both chemicals used were technical grade from Fisher Scientific. Note that in subsequent scaling work, the same calcium nitrate was used but oxalic acid was replaced with sodium oxalate. 5.2 Experimental Setup Figure 5.2.1 shows the schematic of the experimental setup used for the solubility tests. The main instruments were: • VWR Scientific controlled temperature water bath; this bath was able to maintain the set temperature with +/- 0.05°C accuracy; • SYBRON Thermoiine stir-plate heater Model Nuovo 7; • Millipore vacuum filter; a coil of copper tubing (with a 4 mm internal diameter and 1 mm wall thickness) was wound around the filter funnel and connected to a second controlled temperature water bath to make possible to heat up the funnel during filtration; 0.45 \im Millipore filter paper was used for filtration; • Constant temperature water bath from COLORA, model K4508; • ACCUMET pH-meter from Fisher Scientific. 41 co XI i CD co •o CD c o o CD ® o 5 CO c c ^ •5 CD CD i_ Q . Q . fl) O . P o 2 I I I m co c o CL CD CO *C0 c CD E *l— CD C L X CD co t o CD O C/5 CM LO* i_ D O L L c o o CO "co «4-» c CD E l_ CD C L X CD CD CD — CD CO CO CD x: <= _c: S 8 © CO x> _ CD CO CD co . CD ~ C L i co CM CO 42 Chapter 5 - Solubility Tests 5.3 Experimental Procedure A 550 ml solution was first prepared by dissolving a measured amount of oxalic acid in distilled water at the room temperature. This solution was mixed using the magnetic stirrer until the oxalic acid was completely dissolved. The pH of the resulting solution was then measured, and adjusted to the desired value by adding NaOH IN. A 50 ml sample was taken from the solution in order to determine the initial concentration of C2O4" ions in solution. The solution was then heated up to 75°C in a temperature controlled water bath. When the solution temperature reached 75°C, a measured amount of calcium nitrate was added. The solution was again mixed with the magnetic stirrer until the salt was completely dissolved. Two samples were then taken for analysis of the initial concentration of Ca"^ in solution. The solution was kept for about 30 min. at 75°C to allow for calcium oxalate formation reaction to take place. The temperature was then decreased to 70°C, 60°C and 50°C respectively. The solution was maintained at each temperature for 30-45 min. in the water bath. This time interval was found to be long enough for the equilibrium to be attained in solution (by analyzing the C 2 0 4 = and Ca 4 4" content of solution at different instants of time). For each above mentioned temperature, samples were taken for the analyses of C 2 C V and Ca44^ content of the solution. Prior to analyses, the samples were each time filtered through a 0.45 urn pore size Millipore filter paper in order to eliminate all the calcium oxalate crystals already formed. This pore size was chosen based on the information found in literature (Ulmgren and Radestrom, 1997). Special precautions were taken to avoid the change of the solution 43 Chapter 5 - Solubility Tests temperature during the filtration. The funnel of the filter was heated and maintained at the desired temperature by the circulation of hot water from a controlled temperature water bath through the coil of copper wound around it. Also, during filtration, the flask of the filter was placed on the heated plate to avoid calcium oxalate crystallization in filtrate. 5.4 Analytical Techniques 5.4.1 Soluble Ca4* determination In order to determine the total C a 4 4 content of the solution, the filtrates were analyzed using a GBC 904 Atomic Absorption Spectrophotometer (from GBC Scientific Equipment Pty Ltd., Victoria, Australia). The flame used was air-acetylene with 427 nm wavelength. The optimum concentration range was 1-4 ppm. The samples were diluted with a HCI solution 0.0IN containing 4 ppm Sr (dilution solution was prepared with IN HCI and 1000 ppm Sr reference solution, both from Fisher Scientific). A calibration curve was determined each time by using four reference solutions with different Ca44^ concentrations, between 1 and 4 ppm. The reference solutions were prepared using a Ca reference solution from Fisher Scientific diluted with distilled water containing HCI and Sr as mentioned before. 5.4.2 Soluble CzO^ determination The total C2CV content of the solution was determined by titration with a 0. IN KMnOu solution following a method found in literature (Kekedy, 1986). The titration reaction is: 2 MnCV + 5(COO) 2 = + 16H+ = 2MH 4 4 + 10CO2 + 8H 2 0 44 Chapter 5 - Solubility Tests Procedure: to 50 ml solution are added 8 ml H2SO4 20%; the solution is then heated up to 70-80°C (to avoid calcium oxalate precipitation) and titrated at this temperature with KMn04 0.1N (from Fisher Scientific) until the solution became light purple. Titration curve: to determine the titration curve, nine samples with known concentration of C2O4" have been titrated. The samples were prepared with sodium oxalate (from AVOCADO, analytical grade). The concentration of C204= given in the titration curve is based on the hypothesis that a 50 ml sample is titrated each time (Figure 5.4.1) 5.5 Results and discussion Two main groups of solubility experiments have been performed, whose characteristics are listed in the Table 5.5.1 and Table 5.5.2. The results of the Ca^ and C204= analyses have been used to calculate the solubility product values as follows: L s = [C204lsat[Ca++]sa t (5.1) L s values obtained from the performed tests have been plotted together with the values that correspond to the solubility levels reported by Ulmgren and Radestrom (1997) in the given experimental conditions. Figures 5.5.1 and 5.5.2 present the results of these experiments. The results of the experiments type A (Figure 5.5.2), at pH = 4 show a good agreement with values reported by Ulmgren and Radestrom at 50 and 60°C. At 70°C and 40°C however, the values differ by up to 15% compared with those found in literature. 45 Chapter 5 - Solubility Testa Figure 5.4.1: Titration curve for C 2 O 4 -0.12 1 0 0.2 0.4 0.6 0.8 1 1.2 1.4 K M n 0 4 [ml] 46 o oo 47 48 Chapter 5 - Solubility Tests Table 5.5.1 Summary of the main parameters and results for solubility tests performed at p H = 4. j Temp. Test A l Test A2 log(Ls) lit. [mol/L]2 [°C] [C 2(V] [mol/L] [Ca++] [mol/L] log(U) [mol/L]2 [C 2 0 4 = ] [mol/L] [mol/L] iog(Ls) [mol/L]2 70 0.00018 0.00048 -7.06349 0.0002 0.00049 -7.00877 -7.0798 60 0.00014 0.00038 -7.27409 0.00014 0.0004 -7.25181 -7.2428 50 0.00012 0.00033 -7.4023 0.00012 0.00037 -7.35262 -7.4058 40 0.0001 0.000258 -7.58838 0.00009 0.000245 -7.65659 -7.5688 Table f p H = 3. Temp. [°C] 5.5.2 Summary of the main parameters and results for solubility tests perl Formed at TestCl Test C2 iog(Ls) lit. [mol/L] 2 [C 2 0 4 = ] [mol/L] [CaT [mol/L] log(U [mol/L]2 [C 2 0 4 = ] [mol/L] [ C O [mol/L] log(Ls) [mol/L]2 70 0.00031 0.00071 -6.65738 0.00036 0.00073 -6.58037 -6.6667 60 0.00027 0.00056 -6.82045 0.00029 0.00055 -6.79724 -6.8667 50 0.0002 0.00042 -7.07572 0.0002 0.00041 -7.08619 -7.0667 The results of test CI and test C2 gave results close to those reported by Ulmgren and Radestrom at 50°C. However, for 60 and 70°C the solubility values obtained are higher (up to 18 %) than those found by Ulmgren and Radestrom. Equations fitted to the experimental data were: - for pH = 3.5: log L s = 0.0231 T(°C) - 8.2225 - for pH = 4 : log L s = 0.0187 T(°C) - 8.3552 49 Chapter 5 - Solubility Tests 5.6 Conclusions The solubility tests that were carried out for calcium oxalate helped in developing the analytical methods that will be further used in the fouling experiments for detennining the total calcium and oxalate ions concentration. The results obtained confirmed that calcium oxalate solubility increases with temperature increase and pH decrease. They also emphasized the strong variation of solubility with pH in the range 3 . 5 -4 (at 70°C the solubility based on oxalate analyses was almost double at pH = 3.5 compared with pH = 4). The solubility values obtained showed a good agreement with those reported by Ulmgren and Radestrom and differences were not higher than 18 %. 50 Chapter 6 - Results and Discussion 6 RESULTS AND DISCUSSION 6.1 General Overview The experimental solution was made up by adding measured amounts of calcium nitrate and sodium oxalate to 175 L tap water that was previously heated up at 75°C in the supply tank. The initial quantities of chemicals to be added were calculated as a function of the desired initial relative supersaturation and Ca++/C204'= for each experiment. According to Ulmgren and Radestrom (1997), the relative supersaturation of the initial solution may be calculated as: s _ ([Ca++MC2QA=] - 1 (6-1) where: L s - the solubility product of C aC204 [mol IL ] [Ca""], [C2O4*] - total molar concentration of calcium and oxalate ions in solution [mol/L] L s values corresponding to the temperature and pH used in each experiment have been taken from the experimental results published by Ulmgren and Radestrom in 1997, and are tabulated in Table 6.1. Since L s is known, one can calculate from the above formula the [Ca**] and [C2CV] values that give the desired Sj 0 and Ca/Ox. For high values of Sj0, inevitably some crystals will be found in suspension. During experiments samples were taken from the fluid returning to the supply tank in order to determine the total content of oxalate and calcium in suspension. The former was determined by titration with 0.1 N KM11O4 and the latter by Atomic Absorption. Samples were immediately diluted with acid for both type of analysis (20 wt. % H2SO4 for oxalate and 0.01 N HCI for calcium) to ensure that the calcium oxalate 51 Chapter 6 - Results and Discussion Table 6.1.1; Solubility product of calcium oxalate in aqueous solution for different temperatures and pH values. Temperature rci 90 70 50 Log(L s) [mol2/L2] pH>5 -7.15 -7.35 -7.65 pH = 3.5 -6.75 -7.1 -7.4 pH = 4 -6.3 -6.6 -7.1 Note: after Ulmgren and Radestrom, 1997 crystals present in solution were dissolved. The detailed description of these analyses is presented in Chapter 4. 6.2 Preliminary experiments When studying the fouling behaviour of a certain system in an experimental rig, one of the most important problems is that significant fouling has to be obtained during a short period of time. In industry the scale might be accumulated in months but for research studies is not possible to perform experiments over such a long period of time due to the high related-costs. Since no prior work on fouling of such solutions has been published, preliminary experiments were performed in order to establish conditions necessary to obtain significant fouling of calcium oxalate on the cooled tube surface in a reasonable amount of time. "Significant" means in this case a change in the overall heat transfer coefficient that would be well above the extent of scatter of the data for these experiments, which was found to be +1-3%. A "reasonable" amount of time was taken to mean five days of continuous operation. In terms of main process parameters, these conditions were found to be: 52 Chapter 6 - Results and Discussion • pH higher than 5; • temperature of the hot fluid as high as possible (~ 75°C) and low flow rate for the hot fluid (0.15 - 0.4 kg/s); • highest possible flow rate for the cold fluid (0.75 kg/s). Preliminary tests have also highlighted the importance of the way chemicals are introduced into the system. Once seed calcium oxalate crystals are formed in the solution, the crystallization of calcium oxalate becomes an almost instantaneous process, which results in the accumulation of large crystals at the bottom of the supply tank. The concentration of calcium oxalate in the hot fluid stream rapidly decreases after the crystals start forming in solution. In an attempt to overcome the tendency of crystals to settle, the agitation of the solution in the tank was improved. Indeed, when mixing was increased the crystals remained suspended in the flowing stream but surprisingly that did not result in scaling on the experimental tube surface. As one can see in Figure 6.2.1 where the thermal resistance 1/Uh is plotted over time, with Si 0 = 6, and crystals in suspension (see Appendix UI, Table m-1, for calculated suspended solids concentrations) no evident fouling occurred. Oxalate concentration decreased very slightly from 0.0017 to 0.0015 mol/L. At this acidity (pH=6.3), and Ca/Ox = 2mol/mol, the saturation condition is [Ox] s = 0.00018 mol/L, which is an order of magnitude lower. Further tests have shown that in order to obtain significant fouling, fresh makeup chemicals have to be added in a continuous way during each run. Initially, these chemicals were added intermittently throughout the experiment in order to overcome the decrease in the concentration such that the oxalate concentration was roughly constant with time, at 0.00045 mol/L (see Figure 6.2.2). Measurable fouling occurred in this experiment even i f 53 Chapter 6 - Results and Discussion Figure 6.2.1: Variation of the inverse of overall heat transfer coefficient and concentration of oxalate ions in solution in experiment F4 (Si 0 = 6, ma=0.3 kg/s, mc=0.75 kg/s, Th,in=750C, T,y n=7.5 0C > pH = 6, strong mixing, no fresh make up chemicals added during experiment). In the given experimental conditions, [Ox]sai = 0.00018 mol/L. 0 . 6 i 0.5 -1 § 0.2 0.1 -I — i 1 1 1 1 1 1 i i 0 200 400 600 800 1000 1200 1400 1600 1800 Time [min] 0.0025 ^ 0.002 c — 0.0015 C J B O C J 0.001 -\ 9, 0.0005 0 0 • • 1 1 1 1 1 1 r 1 1 200 400 600 800 1000 1200 1400 1600 1800 Time [min] 54 Chapter 6 - Results and Discussion Figure 6.2.2: Variation of the inverse of overall heat transfer coefficient and concentration of oxalate ions in solution in experiment F5 (S i o = 2, mj=0.45 kg/s, nv=0.75 kg/s, Th,i„=750C, Tc,in=10°C, pH = 6, no mixing, fresh make up chemicals added at every 3-4 hours during experiment). In the given experimental conditions, [Ox]** = 0.00018 mol/L. 0.7 -j 0.6 -0.5 A w 0.2-^ 0.1 -0 -I 1 1 1 1 > 0 1000 2000 3000 4000 5000 Time [min] 0.0008 A o JS c o o o u 0.0006 0.0004 0.0002 0 • • i i 1 { I 0 1000 2000 3000 Time [min] 4000 5000 Note: arrows indicate when fresh chemicals were added during experiment. 55 Chapter 6 - Results and Discussion Figure 6.2.3: Variation of the inverse of overall heat transfer coefficient and concentration of oxalate ions in solution in experiment F10 (S i o = 10, mt=0.15 kg/s, mc=0.75 kg/s, TVi„ =75°C, Tc,m=4.5°C, pH = 6, low mixing, fresh make up chemicals added at every 6 hours during experiment). In the given experimental conditions, [Ox]^ = 0.00018 mol/L. Q007 n „ 0.005 1 0.005 0.004 4 o I o 0.003 5 0.002 3 0.001-0 -0 1000 2000 3000 4000 5000 6000 7000 Time [min] Note: arrows indicate when fresh chemicals were added during experiment. 56 Chapter 6 - Results and Discussion Figure 6.2.4: Variation of the inverse of overall heat transfer coefficient and concentration of oxalate ions in solution in experiment F12 (S i o = 6, mh=0.15 kg/s, mc=0.75 kg/s, Th,i„=750C, Tc^.5°C , pH = 6.3, low mixing, fresh make up chemicals added continuously during experiment). In the given experimental conditions, [Ox]** -0.00018 mol/L. 1 -i 0.5 A 0.4 4— 1 1 • ' ' 0 2000 4000 6000 8000 10000 Time [min] 0.003 -V0.0025 | 0.002 i 0.0015 o o «^ 0.001 o <J 0.0005 -0 --0 2000 4000 6000 Time [min] 8000 10000 5 7 Chapter 6 - Results and Discussion the amount of overall heat transfer coefficient decrease was small (7%). It was found that even if the concentration of the solution increases above the initial values, fresh chemicals still have to be added in order to maintain the fouling process. In experiment F10 periodic additions of make up chemicals were made, every 6 hours, as shown in Figure 6.2.3. Significant fouling was obtained for the first time (up to 15% decrease in overall heat transfer coefficient), with a more than doubling of oxalate concentration from 0.0018 to 0.0042 mol/L. Hence, it was decided to continuously add fresh chemicals during the future experiments. This procedure partially simulates the once through flow of oxalate solution. Figure 6.2.4 presents the results for the first experiment performed with continuous addition of fresh make-up chemicals, and one can see that the thermal resistance increased from 0.67 to 0.92 m2K/kW (30% decrease in overall heat transfer coefficient) in one week. Oxalate concentration increased from 0.0012 to 0.0025 mol/L. Therefore, two metering-pumps have been added to the experimental apparatus to add the two solutions to the supply tank. For calcium nitrate an IDEX MICROPUMP was used to deliver 0.916 L/h of a 3.9 g/L Ca(N03)2«4H20 solution. A 3g/L sodium oxalate solution was fed in the supply tank at a 0.33 L/h flow rate, using a VARIAN PCR-1 pump. Since a total of 30 L was added in this way for every 24 h of run, each day it was necessary to discard from the tank the same volume in order to maintain the level of liquid. The liquid was discarded through a hose connected at the bottom of the tank. Since calcium oxalate crystals accumulate at the bottom, the water that was discarded contained a high concentration of crystals in suspension. The same amount of fresh make-up chemicals at the same flow rate was added in each experiment. This amount was found to be enough to sustain the fouling process. 58 Chapter 6 - Results and Discussion When a larger quantity was added, the result was an increased amount of crystals accumulated at the bottom of the tank without any change to the scaling process. From the results of the preliminary experiments it seemed that a competition exists between crystallization of calcium oxalate onto crystals suspended in solution and onto the cooled tube surface. This hypothesis was confirmed by subsequent experiments and will be further discussed in this chapter. The inverse overall heat transfer coefficient versus time as well as the total oxalate ion concentration in suspension versus time are presented for all experiments in APPENDIX III. One can see from these plots (e.g.: Figure 6.2.4) that increases in oxalate concentration correlates very well with the evolution of the overall thermal resistance in time. For cases where 1/Uh had linearly increased in time, the concentration of oxalate had also first increased and then remained constant. That means, during all the runs with linear increase of fouling resistance, crystals were present in solution at all times and with increasing concentration. In contrast, for experiments where the fouling resistance had reached a constant value (e.g.: Figure III-1, III-4, III-6), one can see that the oxalate ion concentration has decreased in a similar fashion over that time period. Visual observation indicated that when 1/Uh leveled out the solution became clear. 6.3 Calculation of fouling resistance versus time from raw data 6.3.1 Fouling resistance When the calcium oxalate deposit forms on the surface of the experimental tube, the thermal resistance of the wall increases due to the added layer. The increased thermal resistance will determine a decrease in the heat transferred between the hot and the cold fluids along the double pipe heat exchanger. This will consequently determine a decrease 59 Chapter 6 - Results and Discussion between the inlet and outlet temperatures for both fluids. Hence, the experiments are run with fixed inlet temperature and flow-rates but with decreasing heat flow over time. This change is reflected in the data saved on the computer data logger during the experiment and allows one to calculate the fouling resistance at any instant of time as well as the final fouling resistance. R f = - (6-2) U f U c Where: Uf - is the overall heat transfer coefficient at any time t after fouling occurred [kW/m2K] Uc- is the overall heat transfer coefficient before chemicals were added [kW/m2K] The overall heat transfer coefficients are calculated from the inlet and outlet temperatures for the hot and cold water as well as from their flow rates using Eq (II-4). Because of the flow rates selected, the temperature change along the experimental tube is much higher on the hot water side (typically 10°C versus 3°C on the cold water side). The overall heat transfer coefficient calculation was therefore based on the heat flux calculated for the hot water side. 6.3.2 Fouling rate The initial fouling rate for each run was determined. For experiments where variation of Rf versus time had a linear trend, the initial fouling rate was obtained by performing linear regression over the whole range of time. For experiments where falling rate or the asymptotic behaviour was evident a linear regression was performed over the time range where the increase was linear to give an initial rate. 60 Chapter 6 - Results and Discussion 6.3.3 Data presentation Because data were saved at every 10 minutes and the experiments were generally carried on for oneweek, a very large amount of data was obtained. For the sake of clarity, the results are presented as moving averages over 21 data points (210 minutes) in Figures 6.4.1, 6.4.2, 6.6.1, 6.7.1, 6.8.1, 6.9.1. The values of Uf for 21 consecutive data points are averaged and the average value is assigned to the time of the mid-interval (the eleventh point of the interval). The formula used to calculate the moving average set of data is: (6-3) 7=1 where: Xj - the experimental data points (1 < j < n, where n is the total number of experimental data points) Y; - the moving average values (11 < i < n ) 6.3.4 Correction factor for fluctuations in the hot water flow rate During the experiment, some small fluctuations in the hot water flow rate occurred and these fluctuations have an impact on the overall heat transfer coefficient because they change the hot film convective heat transfer coefficient. Since the change in the overall heat transfer coefficient was used as a measure of the fouling resistance, to eliminate other influences a correction factor was calculated for the overall heat transfer coefficient that accounts for small changes in hot water flow rate (up to 15%) that may occur during each experiment. Several experiments were carried out under clean conditions at constant cold-water flow rate and varying the hot water flow rate. A Wilson plot was drawn with the 61 Chapter 6 - Results and Discussion data obtained from these runs. This plot gives a linear variation of the convective heat transfer coefficient of the hot water film versus Re 0 ' 8. Using the slope of this plot we could calculate a correction factor for fluctuations in hot water flow rate. This calculation is presented in APPENDIX II. For high fluctuations in the flow rate, other phenomena may occur that cannot be accounted for, such as the partial removal of the deposit from the tube surface due to the turbulence created by the sudden flow rate change. 6.4 Reproducibility of the results Satisfactory reproducibility of the results has been confirmed in two experiments. Experiment F18 was repeated because of an accidental stop of the initial experiment due to a power break down. In this case the comparison of the results shows an identical evolution of the fouling resistance in the two experiments over the 3500 minutes that was the duration of the first one prior to the power break down. As one can see in Figure 6.4.1, there was no measurable difference between these two experiments over the common time interval. Experiment F12 has been deliberately repeated as F24 and thus covers the same time interval (10,000 minutes, which is roughly the equivalent of 6.5 days). The results, as shown in Figure 6.4.2 (fouling resistance) and 6.4.3 (concentration), are similar for these two experiments: the initial fouling rate is 10.6% higher in F12 than in F24, and the final fouling resistance is 6.3% higher for F12 (Table 6.6.1). The trend of Rf versus time 62 Chapter 6 - Results and Discussion Figure 6.4.1: Fouling resistance versus time in experiments F18b and F18, presented as moving average. 0.5 -i 0.45 -0.4 -0.35 -12000 Time [min.] was linear in both experiments and the concentration of oxalate ions in solution during the experiment was also similar. In both experiments the total concentration of oxalate ions in solutions has remained fairly constant and equal with the initial value of 0.0012 mol/L for the first 3000 minutes of the run and then increased at the same rate for another 3000 minutes reaching a value that was almost double the initial concentration. Then, the concentration of oxalate ions remained constant in both experiments with a slight decreasing tendency for F24 in the last 1000 minutes of the run. The average inlet temperature of the cold water was 3.5°C during F12 and about 12°C in F24 as seen in Table 6.6.1. This difference was due to the fact that the first experiment was performed during winter and the second one in summer. 63 Chapter 6 - Results and Discussion Figure 6.4.2: Fouling resistance evolution versus time for F12 and F24, presented as moving average over 21 data points. 0.5 n o.<H Figure 6.4.3: Oxalate concentration profile during F12 and F24 0.003 -i §0.0025 1, 0.002 §0.0015 o C 0.001 o rj0.0005 -| 0 0 O A O O A O O 2000 A O O A A circles: F12 triangles: F24 O A 4000 6000 T i m e [min] 8000 O 10000 64 Chapter 6 - Results and Discussion The slight difference between F12 and F24 could be attributed to the significant difference between the cold water temperatures in the two cases. The similarity of the results shows in this case the good repeatability of the results. The small difference in calculated surface temperature, estimated respectively as 24.7°C and as 29.6°C had relatively weak influence on the fouling process of calcium oxalate. 6.5 Deposit morphology After each run, the inner tube of the heat exchanger was left to dry over night and then was removed to examine the scale formed on its surface. Photos were then taken of the fouled tube. These photos are presented further in this chapter. For asymptotic fouling, deposits were fairly strong and coherent, and could be readily examined. However, for all the experiments that exhibited a linear fouling trend, the deposit was very weak and except for one case, it washed out from the tube once the flow through the heat exchanger was stopped and the solution was drained. Visual observations that were made during these experiments through the glass windows of the test section showed that the deposit looked similar in all cases and was fairly smooth. For one of these experiments, F30, the deposit was successfully recovered and a photo of this deposit is presented in Figure 6.6.3 (corresponding to 0.35 kg/s). They confirm the visual observations made for the rest of experiments that exhibited a linear fouling trend. In order to stabilize the deposit in run F30, the run was continued three days at the end without adding fresh make up chemicals. That was done to stop the linear build-up of the deposit to permit strengthening by the continuous flow of the solution, which is known as the aging of the deposit. 65 Chapter 6 - Results and Discussion 6.6 Flow rate effect on calcium oxalate fouling Flow rate generally has an important influence over the fouling process for sparingly soluble salts. This influence depends on the fouling mechanism involved. If the fouling process is controlled by mass transfer of reacting ionic species from the bulk solution to the heat transfer surface, then the fouling rate is expected to increase with the flow rate due to the increased turbulence. However, this increased turbulence would have a negative impact on fouling rate if adhesion or surface integration (crystallization) is the controlling step in the process. Thus, the results of the flow rate effect on the fouling rate studies can give important information regarding the fouling mechanisms governing that system. The flow rate effect on calcium oxalate fouling was studied in the turbulent regime with corresponding Reynolds numbers in the range 8,000-22,000. The lower limit of the range was established in order to remain in the turbulent regime. The upper limit was dictated by the fact that as the flow rate increases the temperature drop along the test section in the hot water side decreases. If this temperature drop becomes very small, the errors in calculated heat flux become much higher and whence the results less trustful. For Re=22,000, the bulk temperature change in the hot water side was already less than 8°C under clean conditions. Since during the experiment this temperature difference is expected to further decrease due to the added fouling resistance, the 0.4 kg/s flow rate (corresponding to Re=22,000) was chosen as the highest flow rate tested in this study. The main parameters of these experiments are presented in Table 6.6.1. Both linear and asymptotic fouling have been observed during these experiments and falling rate was also exhibited in one of the cases (F19). Figure 6.6.1. shows the fouling 66 Chapter 6 - Results and Discussion resistance evolution over time in all these experiments. In all runs there is an induction period of about 300 minutes after the addition of chemicals at 710 minutes. One can see that for the lowest flow rate (0.15 kg/s) the fouling trend is linear and becomes falling rate or asymptotic when the flow rate increases to the range 0.2-0.25kg/s. For the high flow rates however, the trend is becoming again increasingly linear. At 0.3 kg/s one can see from Figure 6.6.1 that the fouling resistance exhibits an almost linear increase and for 0.35 and 0.4 kg/s the trend is clearly linear. Table 6.6.1: Flow Rate Effect on Calcium Oxalate Fouling Run # m h [kg/s] Ren Th,in avg. [°C] T • 1 ('.III avg. ra Initial Fouling Rate [m2K/kJ] Range of Initial rate calc. [min.] Rf final [m2K/k\V] U„ final decrease [%] (**) Fouling Trend F12 0.15 8,270 75 4 0.471E-06(LR) 710-9,470 0.267 (9800 min) 29.1 L F24 0.15 8,270 76.05 9.93 0.42 E-06(LR) 710-10,000 0.253 (10000 min) 28.8 L F22 0.2 11,020 75.58 9.71 1.66 E-06(LR) 2,100-3,810 0.265 (6300 min)* 31.8 A F19 0.25 13,780 75.18 7.18 2.90 E-06(LR) 2,100-5,240 0.69 (9100 min) 57.5 F F21 0.3 16,530 74.44 8.86 3.12E-06(LR) 2,100-3,200 0.26 (6000 min)* 35.7 A F30 0.35 19,290 75.76 14 6.7175E-07(LR) 710-7,000 0.227 (7000 min) 35.6 L F20 0.4 22,040 73.26 7.39 1.17E-06(LR) 2,100-7,120 0.47 (7200 min) 52.7 L (*) - For asymptotic trend runs R f final is Rf* (**) - In these experiments the reduction in Uh was between 30 - 58% and the corresponding drop in heat flux was between 25 - 52 %. (***) - F stands for "falling rate fouling", A stands for "asymptotic fouling" and L stands for "linear fouling". Note: For all experiments, the chemicals where added at 710 min. after the beginning of the experiment. In all these experiments: S i o = 6, pH=6.3 and initial molar Ca/Ox = 2/1 67 Chapter 6 - Results and Discussion Figure 6.6.1: Dependence of fouling resistance on the solution flow rate (results presented as moving average over 21 data points) F19 (try, = 0.25 kg/h) 1000 2 0 0 0 3 0 0 0 4 0 0 0 5 0 0 0 6 0 0 0 7 0 0 0 8 0 0 0 9 0 O 3 10000 Time [min] Figure 6.6.2: Dependence of initial fouling rate on the solution flow rate (Sio - 6, Ca/Ox = 2/1, pH = 6.3) Flow kate [kg/s] Chapter 6 - Results and Discussion For the experiments with linear trend the induction time was very short and the slope of fouling resistance increase over time remained unchanged during the whole run. For the experiments that showed an asymptotic fouling trend, after an initial moderate increase of fouling resistance, the fouling rate exhibited an increase followed by a leveling out, resulting in an "S" shaped curve. The initial fouling rate for each run was detenriined through linear regression. For experiments F12, F24 and F20 that had a linear trend, the linear regression was performed over the time of the whole experiment. For the asymptotic evolutions, the linear regression was performed over the time range where the increase was linear, which was after a induction time of about 2100 minutes. For experiment F19 that had a falling rate, linear regression was performed over the range of about 50 h with maximum fouling rate. The results are also presented in Table 6.6.1. The initial fouling rate has been found to increase with the flow rate increase (Figure 6.6.2) and reach a maximum at mh = 0.3 kg/s (Re around 16,000). When the flow rate was further increased to 0.35 kg/s (Re=19,000), the initial fouling rate has steeply decreased. Figure 6.6.3 shows the photos of the deposits for the experiments with an asymptotic fouling trend. For the lowest flow rate, 0.2 kg/s, one can see that high ridges are present on the surface of the scale normal to the direction of the flow. For the higher flow rate of 0.3 kg Is, these ridges are still present and very well defined but they are much smaller than in the previous case. When the flow rate was further increased to 0.35 kg/s, high ridges almost totally disappeared. For 0.4 kg/s hot water flow rate, the fouling trend was linear, which corresponded to a smooth deposit surface. As stated earlier, the deposit for linear fouling experiment could not be recovered, and the information about 6 9 Inner tube diameter 70 Chapter 6 - Results and Discussion its appearance is based on the visual observations that were made during experiments through the glass windows of the test section. 6.7 p H effect on the fouling process The pH of the solution during the fouling runs is expected to have a significant influence over the scale formation process because it strongly influences the distribution of the ionic species in the solution. Figure 2.1.1 shows the calculated concentration of ionic species for oxalic acid as a function of the solution pH. One can see that for pH>6 most of the acid is completely dissociated and oxalate ions concentration is more than 99 mole %. By contrast, at pH=2, less than 20 mole % of oxalic acid is completely dissociated. Experiments were performed at four different pH values: 2.2, 3, 3.7, and 6.3. Using calcium nitrate and sodium oxalate, the pH of the solution is normally 6.3. To achieve lower values, nitric acid (20% wt) was added. For the experiments that were carried out at low pH values, nitric acid was added both in the initial solution and in the fresh make up chemicals that were added during the run. The pH was checked up regularly and was found to be stable once the desired value was achieved at the beginning of the experiment. In terms of fouling trend, one can see from Figure 6.7.1 that all the runs at pH lower than 6.3 exhibited asymptotic fouling, whereas for pH=6.3 the trend was linear. Fouling was heaviest at pH=3, where Rf reached a value of 0.35 m K/kW. At pH = 2.2, Rf* value decreased to 0.14 m 2K/kW. The initial fouling rate was found to increase as pH is raised from 2.2, pass through a maximum at pH=3, and decrease with increases of pH in the range 3-6.3 71 Chapter 6 - Results and Discussion Figure 6.7.1: Dependence of fouling resistance on the pH (results presented as moving average over 21 point 0.45 A 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 Time [min] Figure 6.7.2: Dependence of initial fouling rate on pH (Sio = 6, mh = 0.15 kg/s, Ca/Ox = 2/1) pH 72 Chapter 6 - Results and Discussion (Figure 6.7.2). Coincidentally, this follows the trend of concentration of HC2O4" shown in Figure 2.1.1. The morphology of these deposits, presented in Figure 6.7.3, also shows a strong variation with the pH of the solution. For the lowest pH, one can see that the deposit has a hard crusty appearance with big lumps irregularly distributed on the heat transfer surface. For pH=3 the scale has the same crusty consistency but the lumps are smaller than in the previous case. For even higher pH of 3.7 the deposit looks more like a powder; its surface still has a rough appearance but the particles are much smaller. Finally, at pH=6.3, the scale was formed linearly, had a powdery appearance and a smooth surface, and was too weak to remain on the tube for the photographs. Table 6.7.1: pH Effect on the Calcium Oxalate Fouling. Run # pH m h Th^n avg. avg. Initial Fouling Rate Range of Initial rate calc. Rf final Fouling Trend [kg/s] P C J rq [m2K/kJl [min.] [m2K/kW] (*) (**) F26 3 0.15 76.3 11.45 4.07 E-06 (LR) 1,260-2,500 0.34 (5000 min) A F27 3.7 0.15 76.15 11.74 1.54 E-06 (KS) 1,100-6,200 0.217 (6200 min) A F28 2.2 0.15 75.97 12.47 1.74 E-06 (LR) 1,000-1,800 0.163 (6000 min) A F12 6.3 0.15 75 4 0.471E-6(LR) 710-9,740 0.267 (9800 min) L F24 6.3 0.15 76.05 9.93 4.20 E-07 (LR) 800-10,000 0.253 (10000 min) L Note: In all experiments: S i o = 6, initial molar Ca/Ox = 2/1, Re = 8,270. (*) - For asymptotic trend runs Rf final is Rf* (**) - A stands for "asymptotic fouling" and L stands for "linear fouling". From previous precipitation studies (Kolthoff and Sandell, 1952) it was found that if calcium oxalate is precipitated from acidic solutions, the precipitate is more thermally 73 Inner tube diameter Chapter 6 - Results and Discussion stable than if the precipitation took place in neutral or basic solutions. That could explain the increasing coherence of the deposit as the pH decreased in the above-mentioned fouling experiments. 6.8 Initial relative supersaturation effect on the fouling process The initial concentration of the reactants governs the bulk concentration of the studied salt and as a result the driving force for the transport by diffusion of the reactants to the surface of heat transfer. At high supersaturations, it would affect the concentration of suspended crystals (Appendix III). An increase in the initial concentration of the reactants is thus expected to increase the initial fouling rate whether the mechanism is scaling or particulate deposition. However, experimental studies have suggested that the rate of crystallization of sparingly soluble salts is controlled by reaction kinetics, rather than by diffusion (Bott, 1995). Thus, in a study published in 1979, Nancollas and Gardner concluded that the rate of crystallization reaction of calcium oxalate monohydrate has a quadratic dependence upon the relative supersaturation of the solution. Four different relative supersaturations of the solution were studied during five experiments, all at a molar Ca/Ox = 2/1. The results are presented in Figure 6.8.1 and Table 6.8.1. The fouling behaviour of this system seems to be different for the low and high range of initial supersaturation. When the initial relative supersaturation of the solution was below 2, an asymptotic fouling trend was obtained with a very high initial fouling rate. Final Rt- values were 0.12-0.13 m2K/kW. For higher initial relative supersaturations, the fouling trend was linear and the initial fouling rates were much lower and close to each other. In these experiments the deposit was built up more slowly 75 Chapter 6 - Results and Discussion Figure 6.8.1: Dependence of the fouling resistance on the initial relative supersaturation (results presented as moving average over 21 data points) 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 Time [min] Figure 6.8.2: Dependence of the fouling resistance on the initial relative supersaturation S fo (Ca/Ox = 2/1, pH = 6.3, mh = 0.15 kg/s) 76 Chapter 6 - Results and Discussion but at an almost constant pace during the whole run and as a result the final fouling resistances were much higher (0.23-0.27 m2K/kW) than in the low supersaturation case. One can see that the highest fouling rate was obtained for the lowest initial supersaturation of 1.0 and it decreased sharply by about 50%, as the initial supersaturation has increased to 2.0. At higher Sj 0 values the rate increased slightly. However, in terms of final fouling resistance, the increase of the initial supersaturation clearly increased the amount of scale formed during a certain time interval. The result of these experiments is rather unexpected since an increase in supersaturation would be expected to increase the fouling rate. The explanation of this result might reside in the composition of the experimental solution during each of these runs. At the beginning of this chapter the hypothesis was mentioned that competition might exist in the calcium oxalate mixture between crystallization onto particles in suspension and crystallization on the tube surface. From this point of view it makes sense that a high supersaturation would actually enhance the crystallization in the suspension (by creating a larger number of crystal seeds in solution) being detrimental to the crystallization on the tube surface and determining a lower initial fouling rate. However, the linear fouling trend of the experiments with higher relative supersaturations suggests that suspended particles are also steadily deposited on the surface. When the particles concentration in solution was higher (for higher initial relative supersaturations), a higher final fouling resistance was therefore obtained. Figure 6.8.3 shows the pictures of the two types of deposits that occurred in these experiments. First one corresponds to experiment F15 that took place at low 77 Inner tube diameter 78 Chapter 6 - Results and Discussion supersaturation (S-,0 =1), and without visible particles in suspension. The second picture corresponds to the high supersaturation experiments, F16, F12 and F24 where a large amount of particles were present in solution during the runs and the fouling trend was linear. Table 6.8.1: Initial Relative Supersaturation (S io) Effect on the Calcium Oxalate Fouling. Run # Sio nii, Th,in avg. Tcjn avg. Initial Fouling Rate Range of Initial rate calc. R f final Fouling Trend [kg/s] [°C] [°C] [m 2K/kJ] [min.] [m 2K/kW] (*) (**) F15 1 0.15 75.4 5 1.12E-06(LR) 1,550-3000 0.12 (8000 min) A F29 1.8 0.15 76 14 6.63E-07(LR) 1,550-4000 0.132 (4000 min) L F12 6 0.15 75 4 4.71 E-07(LR) 710-9740 0.267 (9800 min) L F24 6 0.15 76 10 4.20E-07(LR) 710-10,000 0.253 (10000 min) L F16 14.5 0.15 75.5 5.8 5.83 E-07(LR) 710-6540 0.226 (6500 min) L Note: In all these runs: Ca/Ox = 2/1, Re = 8,270, pH = 6.3. (*) - For asymptotic trend runs Rf final is Rf* (**) - A stands for "asymptotic fouling" and L stands for "linear fouling". One can see that at low concentration the deposit consists of lumps growing directly on the tube surface, and incompletely covering it whereas for the high concentrations the deposit is uniformly distributed on the tube surface and has a smooth appearance. 6.9 Ca/Ox ratio influence on initial fouling rate The initial Ca/Ox ratio is expected to influence the fouling process because it influences the chemistry of the solution. From the study performed by Ulmgren and Radestrom (1999) it follows that the solubility of calcium oxalate (expressed as L s) differs in the same condition of temperature and pH if the initial Ca/Ox ratio is different. They found that for the same pH and temperature the solubility of calcium oxalate 79 Chapter 6 - Results and Discussion decreases as the Ca/Ox ratio increases from 0 to 1 and then increases with the increase of Ca/Ox ratio, as shown in Figure 2.2.3. In the present study, three experiments were performed in similar conditions but with different Ca/Ox ratios: 1/1,2/1 and 3/1. In these experiments the sodium oxalate was added in the same quantity, corresponding to the molar concentration of calcium oxalate for a initial relative supersaturation equal to unity in the given experimental conditions, for a 1/1 molar ratio. Table 6.9.1 gives a summary of the main parameters involved in these experiments as well as the results. Table 6.9.1: Ca/Ox Effect on the Calcium Oxalate Fouling. Run# Ca/Ox Sio Thjn T • Initial Range of R f final Fouling avg. avg. Fouling Initial rate Trend Rate calc. 1°C] [m 2K/kVV| 1°C] |m2K/kJ] [min.] (*) (") F18b 3/1 1.4 75.15 7 1.01 E-06(LR) 1,250-4,100 0.294(10000 min) F F15 2/1 1.8 75 4 1.12E-06(LR) 1,551-3,000 0.12 (8000 min) A F17 1/1 1 74.43 6.5 0.436 E-06 (LR) 1,000-3,600 0.041(9800 min) A Note: In all runs: mh = 0.15 kg/s, pH = 6.3 and Re = 8,270. (*) - For asymptotic trend runs Rf final is Rt* (**) - F stands for "falling rate fouling" and A stands for "asymptotic fouling" fouling". In all these experiments the initial oxalate concentration was the same and equal to the corresponding saturation concentration of calcium oxalate in the given conditions (0.00026 mol/L). Figure 6.9.1 presents the moving average values of Rf versus time for all these three experiments. Asymptotic fouling was obtained for Ca/Ox molar ratios of 1/1 and 2/1. For 3/1 the fouling resistance had a falling rate, which appeared to approach asymptotic behavior. From the 1/Uh versus time and oxalate concentration versus time presented in APPENDIX III for these experiments, one can see that the main difference between F17 and F15 on one hand and F18 on the other hand resides in the absence and 80 Chapter 6 - Results and Discussion Figure 6.9.1: Dependence of the fouling resistance on Ca/Ox initial ratio (results presented as moving average over 21 data points) F18 (Ca/Ox=3M) -0.05 J 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 Time [min] Figure 6.9.2: Dependence of the initial fouling rate on Ca/Ox initial ratio (Si„ = 0 - 2 mh = 0.15 kg/s, pH = 6.3) Chapter 6 - Results and Discussion respectively presence of particles in solution. During most of the time (8000 min.) in experiment F18 the concentration of oxalate ions in solution was above the saturation limit, which means that particles were present. For F15 the concentration of oxalate has rapidly (about 4000 min.) dropped below the saturation limit, which means that the solution was clear most of the time. For F17 the concentration of oxalate in solution decreased below the saturation limit of calcium oxalate even faster, after only 2000 minutes of run. From Figure 6.9.1 one can see that for the experiment with Ca/Ox = 3/1 the induction period was very short, almost zero whereas for the two other experiments the induction period was about 800 min. From Figure 6.9.1 one can see that the final fouling resistance over the same period of time (about 8000 min.) was much higher (almost double) for F18 (Ca/Ox = 3/1) than for F15 (Ca/Ox = 2/1). For a molar ratio of 1/1 (F17), the final fouling resistance was three times lower than for 2/1 molar ratio (F15) over a longer period of time (9800 minutes compared with 8000 minutes for F15). The initial fouling rate was calculated by performing linear regression over the range where the fouling resistance increase was linear for the asymptotic experiments. For the experiment with falling rate, the initial fouling rate was obtained by linear regression over the range were the fouling rate was highest. The highest initial fouling rate was obtained for a Ca/Ox ratio of 2/1 or greater. At 2/1 the final fouling resistance was also the highest. The initial fouling rate at Ca/Ox = 3/1 remained virtually as high as at 2/1. For a 1/1 ratio however, the final fouling resistance was very low, only 0.041 m K/kW, which corresponds to a decrease in overall heat transfer coefficient of 5.7%, and the rate was about 40% of the value at 2/1. 82 Inner tube diameter Chapter 6 - Results and Discussion From Figure 6.9.3 one can see that for 1/1 ratio the deposit was thin but well defined, with small crystal lumps randomly distributed on the tube surface. The scale looks similar for a 3/1 ratio but is much thicker and the lumps are bigger. For a 2/1 ratio however, one can see that ridges are formed normal to the flow direction. 6.10 Deposit analyses 6.10.1 Thermogravimetric analysis of deposit Thermogravimetry is an analysis technique in which the mass of a substance is measured as a function of temperature whilst the substance is subjected to a controlled temperature program (Hatakeyama and Zenhai, 1998). The deposits which resulted from the fouling runs were analyzed using a TGA-50H apparatus from Shimadzu. A small quantity of deposit, usually 5-8 mg, was loaded in the alumina-ceramic pan of the apparatus and heated up to 850°C under continuous high purity N2 flow of 50 ml/min. When a calcium oxalate monohydrate deposit is heated up, some decomposition reactions are expected to take place. According to "Introduction to Thermal Analysis -Shimadzu TGA-50 Guidebook", the corresponding decomposition reactions for calcium oxalate monohydrate are: ot one.® CaC 2 0 4 • H 2 0 > CaC 2 0 4 + H 2 0 - expected weight loss: 12.33% wt. at 500° C CaC 2 0 4 > CaC0 3 + C 0 2 - expected weight loss: 19.18%wt. 3.t 750^ C CaC0 3 > CaO + C 0 2 - expected weight loss: 30.14%wt. By comparing the weight loss of the sample during the TGA analysis with the expected weight loss for calcium oxalate monohydrate at the same temperatures, one can 84 Chapter 6 - Results and Discussion determine if the sample is indeed calcium oxalate monohydrate. The results of these analyses are presented in Table 6.10.1.1. Table 6.10.1.1: TGA results for deposit analyses and comparison with calculated values Run# Weight loss in the first step (hydration water) Weight loss in the second step (C02) Weight loss in the third step (C02) Overall weight loss [wt %] [wt %] [wt %] [wt%] TGA Calc. Diff. TGA Calc. Diff. TGA Calc. Diff. TGA Calc. Diff. result result result result (1) (2) (2)-(l) (3) (4) (4)-(3) (5) (6) (6)-(5) (7) (8) (6)-(5) F10 12.266 12.33 0.063 18.038 19.18 1.142 28.27 30.14 1.87 58.573 61.64 3.071 F12 12.353 12.33 -0.024 18.415 19.18 0.765 28.873 30.14 1.267 59.608 61.64 2.036 F15 11.903 12.33 0.426 19.098 19.18 0.082 29.149 30.14 0.991 59.750 61.64 1.894 F16 12.397 12.33 -0.068 18.588 19.18 0.592 28.916 30.14 1.224 59.900 61.64 1.744 F17 12.261 12.33 0.068 18.948 19.18 0.232 29.017 30.14 1.383 60.200 61.64 1.444 F18 12.399 12.33 -0.07 18.531 19.18 0.649 28.967 30.14 1.173 59.351 61.64 2.293 F19 12.528 12.33 -0.199 18.179 19.18 1.001 28.786 30.14 1.354 59.493 61.64 2.151 F20 12.09 12.33 0.239 18.482 19.18 0.698 28.418 30.14 1.722 58.989 61.64 2.655 F21 11.797 12.33 0.532 27.035 19.18 -7.855 25.307 30.14 4.833 64.025 61.64 -2.381 F22 11.591 12.33 0.738 24.826 19.18 -5.646 25.541 30.14 4.599 61.849 61.64 -0.205 F23 12.643 12.33 -0.314 18.630 19.18 0.55 27.589 30.14 2.551 58.962 61.64 2.682 F24 12.091 12.33 0.238 18.312 19.18 0.868 28.270 30.14 1.87 58.672 61.64 2.972 F26 11.87 12.33 0.459 20.394 19.18 -1.214 26.123 30.14 4.017 58.250 61.64 3.394 F27 12.568 12.33 -0.239 18.571 19.18 0.609 28.587 30.14 1.553 58.916 61.64 2.728 F28 12.277 12.33 0.052 18.494 19.18 0.686 28.820 30.14 1.32 59.045 61.64 2.599 F30 12.025 12.33 0.304 18.153 19.18 1.027 27.835 30.14 2.305 58.013 61.64 3.631 A V G 12.191 - - 18.63* - - 28.03 - - 59.600 - -STDEV 0.299 - - 0.583* - - 1.261 - - 1.483 - -* The average and the standard deviations for the second step were calculated excluding the values of the weight loss for F21 and F22 The TGA results show a very good agreement with the calculated values with respect to the hydration water loss (first step). The average weight loss was 12.191% compared to the expected value of 12.33%. The largest deviation was -0.74%. Since the calculated weight loss assumes that the deposits contain 100% calcium 85 Chapter 6 - Results and Discussion oxalatmonohydrate, the TGA analyses indicates that only the monohydrate form is present in these samples. For the decomposition of CaC2C>4 to CaC03 and CO2, the TGA results are also very close to the calculated values. Fourteen samples had an average weight loss of 18.63%, and a standard deviation of 0.583%, which compares with the expected value of 19.18%. However, two samples (F21, F22), showed weight losses of 27% and 25%, which are much higher than the average value of the other fourteen samples. There was no apparent reason for this deviation. The biggest deviations occur for the third decomposition step in which CaC03 decomposes to CaO and CO2 but these differences are still under 5%, with deposits giving 25.5-29.2% weight loss with an average of 28.03%, when for pure calcium oxalate monohydrate, 30.1% is expected. Samples F21 and F22 which gave the largest deviations for the second decomposition step, also were least accurate for the third step. The overall difference between the experimental values and the calculated weight loss of 61.6% are less than 3.4% for all the samples except F30 where the difference is slightly higher, at 3.6%. TGA is an indirect analysis method and which should be confirmed by direct analysis methods, such as spectroscopic and wet analytical methods. Whence, the results of other analyses performed for these deposits are further presented in this chapter. For crystal materials, morphological observations of crystals habit are also of use. 86 Chapter 6 - Results and Discussion 6.10.2 SEM and EDX analyses of the deposit Scanning Electron Microscopy (SEM) examination of deposits samples were performed and Figure 6.10.2.1 presents images obtained during these analyses. The SEM apparatus is a HITACHI S/2300 model, which operates at 8-20kV and allows for a broad range of magnifications. The samples shown are coated with gold to allow better quality pictures. For most runs though, carbon was used to coat samples for SEM because this way was also possible to perform EDX analysis on those samples. One can identify in these pictures the characteristic hexagonal shape of calcium oxalate monohydrate samples. A QUARTZ X-ONE Energy Dispersive X-ray microanalysis system that is attached to the SEM was used to examine the deposit samples. The samples are coated with carbon, which is used as a reference for these analyses. The EDX analysis is mostly a qualitative method but it also allows for a rough quantitative estimation of the elements present in the sample. Figure 6.10.2.2 shows the result of such an EDX examination of deposit from experiment F28. One can see that Ca is the major element present in the sample. The second peak that shows up in Figure 6.10.2.2 right after the Ca peak is also given by calcium. Oxygen also shows up but the peak is very small because EDX is not very suitable for identifying gases. Also, carbon doesn't show up because as mentioned above, it is used as reference for the EDX analysis and as coating for the sample. Whence, the EDX analysis supports the TGA analyses, which suggest the high purity of the deposits even though it cannot give the quantitative composition. It actually shows that the sample doesn't contain impurities in quantities higher than the detection limit of EDX, which is 0.5 wt. % 87 Figure 6.10.2.1: SEM analysis, crystals from the test tube surface (X 6k magnification) a. Deposit formed in 24 hours: b. Deposit formed in 6 days: Chapter 6 - Results and Discussion 6.10.3 Quantitative analysis of the total calcium and oxalate content of deposits The total content of calcium and oxalate in each deposit have been determined through the same methods used to determine the total calcium and oxalate ions in solution during the runs: titration with K M n 0 4 0. I M for oxalate and Atomic Absorption for calcium. Figure 6.10.2.2: EDX result for F 28 Counts 13410 10728 4 8046 4 5364 4 2682 4 0 o 0 4 8 keV F28-EDX-1 89 Chapter 6 - Results and Discussion 6.10.3 Quantitative analysis of the total calcium and oxalate content of deposits The total content of calcium and oxalate in each deposit have been determined through the same methods used to determine the total calcium and oxalate ions in solution during the runs: titration with KMnCu 0 .1M for oxalate and Atomic Absorption for calcium. Figure 6.10.2.2: EDX result for F 28 F28-EDX-1 89 Chapter 6 - Results and Discussion A small quantity of deposit, about 0.02 g was dissolved into 4.5 ml of I N HCI, from Fisher Scientific. The acid readily dissolved some of the samples but others had to be heated in order to completely dissolve. The samples were further diluted with distilled water up to 250 ml. Then, two 50 ml portions have been used to determine the oxalate content through titration with 0.1N KMn04. The rest of solution has been further diluted to obtain the 1-4 ppm Ca concentration needed for A A analysis. Table 6.10.3.1 summarizes the results of these analyses. The calculated values for calcium and oxalate mass-fractions in the deposit are based on the hypothesis that the deposit consists only of calcium oxalate monohydrate. The measured calcium mass fraction was 0.297 +/- 0.0204, compared to expected value of 0.274. The measured oxalate mass fraction 0.607 +/- 0.035, compared to the expected value of 0.603. The data resulted from these analyses show a good agreement with the calculated values, which confirms that calcium oxalate monohydrate is the main component of the deposit and that any other impurities are present in minor proportion 90 Chapter 6 - Results and Discussion Table 6.10.3.1: Calcium and oxalate content of the deposit resulted from duplicate analyses and the corresponding calculated values. Run# Ca-mass fraction -measured-Ca-mass fraction -calculated-Ox-mass fraction -measured-Ox-mass fraction -calculated-F30 0.266 0.274 0.549 0.603 F29 0.270 0.274 0.539 0.603 F28 - 0.274 0.595 0.603 F27 - 0.274 0.660 0.603 F26 - 0.274 0.627 0.603 F24 - 0.274 0.605 0.603 F22 - 0.274 0.621 0.603 F21 0.285 0.274 0.649 0.603 F20 0.281 0.274 0.594 0.603 F19 0.303 0.274 0.606 0.603 F18 0.322 0.274 0.587 0.603 F17 0.321 0.274 0.661 0.603 F16 0.313 0.274 0.595 0.603 F15 0.308 0.274 0.603 0.603 F12 0.303 0.274 0.616 0.603 Avg. 0.297 0.607 STDEV 0.0204 0.035 91 Chapter 7-Conclusions 7 Conclusions A new experimental apparatus was built for studying the fouling behavior of calcium oxalate from pure aqueous solutions in a cooling system. The influence of flow rate, pH, supersaturation level, and Ca/Ox molar ratio on the fouling process was determined at constant inlet temperature of 75°C. Fresh make up chemicals had to be continuously fed into the system. This procedure to some extent reflects the actual conditions in the pulp mill streams, where calcium oxalate is continuously formed from calcium and oxalic acid that enters the process with the wood. Experiments ran continuously for about 6 days, resulting in up to 58% decrease in the overall heat transfer coefficient U h . Depending on circumstances, linear, falling rate and asymptotic fouling behaviour have been observed. Measured oxalate concentrations were well above saturation, which indicated the presence of calcium oxalate crystals. Both crystallization and particulate fouling seem to be involved in calcium oxalate fouling. The deposits were found to consist of calcium oxalate monohydrate and had a high purity. The morphology of deposits was strongly influenced by the concentration of suspended particles. Thus, when calcium oxalate crystals were visibly present in bulk (suspension was cloudy), the resulting deposit was smooth and had a very weak consistency. When the particles concentration was low enough to make them undetectable by visual observation (solution was clear), the deposits were hard and randomly distributed lumps grew directly onto the heat transfer surface. 92 Chapter 7-Conclusions The flow rate was varied in the range 0.15 - 0.40 kg/s, which corresponds to Reynolds number between 8,270 - 22,000. The initial fouling rate was found to increase with flow rate and to reach a maximum for Reynolds number of about 16,5000 before declining sharply. The pH was found to have a strong influence on the initial fouling rate in the range 2.2-4, with a maximum at pH = 3. When pH was increased to 6.3 the initial fouling rate remained almost unchanged. The initial fouling rate dropped about three times as the initial relative supersaturation was increased from 1 to 6 and remained basically unchanged as the initial relative supersaturation was further increased to 14.5. For initial relative supersaturation (less than 2) the fouling trend was asymptotic whereas for the highest supersaturations the fouling trend became linear. Thus, despite the high initial fouling rate found for low initial relative supersaturations, the final fouling resistance accumulated during four day runs increased with initial relative supersaturation. Three experiments were performed at different Ca/Ox molar ratios of: 1/1, 2/1, and 3/1. The initial fouling rate was found to increase more than twice as the Ca/Ox ratio increased from 1/1 to 2/1. When the molar Ca/Ox ratio was increased to 3/1, the initial fouling rate remained almost the same. 93 Nomenclature 8 Recommendations for future study The results of the present work highlighted several paths that could be followed to further increase the understanding of the fouling behaviour of calcium oxalate. 1. Flow rate was found to have a strong influence on the initial fouling rate as well as on the general trend of the fouling process but the investigation was limited to one set of conditions with respect to initial relative supersaturation, Ca/Ox ratio and pH. It will be thus useful to see how the above parameters influence the fouling rate at different flow rates. Based on the previous results, the following flow rates could be considered for further experiments: 0.25 kg/s, 0.3 kg/s, 0.35kg/s. 2. A l l the previous experiments have been carried out at the same bulk temperature of 75°C. Calcium oxalate solubility has been reported to strongly increase with temperature increase (Ulmgren&Radestrom, 1997). The same authors also found that the rate of calcium oxalate formation increases with temperature in bleach plant filtrates. An increase in the bulk temperature would also increase the temperature difference between the two fluids, which is expected to enhance the fouling. A temperature of 90°C might be considered for further experiments since solubility data exist in literature for this temperature and is the maximum temperature allowed on the actual experimental apparatus. 3. Since the calcium oxalate solubility was found to be much higher in bleach plant filtrates (Ulmgren&Radestrom, 1997), calcium oxalate fouling should further be investigated in pulp mill streams. 94 Nomenclature Nomenclature a activity coefficient A area, m 2 A ' arrhenius pre-exponential factor AP 2 2 activity product, mol IL B constant inEq, 1.2.5 Cb bulk concentration, mol/L Cp specific heat of water, kJ/kg°C ACp heat capacity change of reaction, J/mol K C s concentration at the heat transfer surface, mol/L Csat saturation concentration, mol/L D e the inside diameter ox the exterior tube, m Di the inside diameter of the inner tube, m D 0 the outside diameter of the inner tube, m E a activation energy, J/mol f2 activity coefficient for divalent ion AG Gibbs free energy change AH enthalpy change, kJ/mol h convective heat transfer coefficient, kW/m K rate constant for crystals dissolution, min"1 (mg seeds/100 ml) k g rate constant for crystal growth, L mol"1 min" l(mg seeds/100 ml"1) k m mass transfer coefficient, m/s 95 Nomenclature k r attachment rate constant kt transport coefficient, m/s k x conductive heat transfer coefficient of the wall, kW/m K K s o solubility product in Figure 2.3.1 Kj, K2, K 3 thermodynamic equilibrium constants K T thermodynamic equilibrium constants at temperature T K thermodynamic equilibrium constants at 298 K L s solubility product m deposit mass per unit surface area, kg/m2 m* asymptotic value of m, kg/m2 ma deposition flux, kg/m2s m^  mass flow rate of cold fluid kg/s mf deposit weight, kg mh mass flow rate of hot fluid kg/s m r removal flux, kg/m2s N number of mole/L calcium oxalate monohydrate deposited from the supersaturated solution before equilibrium is reached p notation for pH in section 2.1 q notation for total oxalate added in section 2.1 R gas constant, 1.98 kcal/mol K Ri, R2 fluid thermal resistance, m K/kW Re 0 8 Reynolds number (pvD/p.) 96 Nomenclature Rf, Rn, R E fouling thermal resistance, m K/kW Rf* asymptotic value of Rf, m K/kW R w wall thermal resistance, m2K/kW S supersaturation relative supersaturation of the initial solution Sp sticking probability s surface of seed crystals, m2 T temperature, °C or K tc time constant, s td delay time, s U overall heat transfer coefficient, kW/m K V crystal volume, unrVcm3 solution Ax thickness of the tube wall, m x thickness of the deposit, m Xj the experimental data points in Eq. 6.3 Yi the moving average values in Eq. 6.3 Greek symbols thermal conductivity of the deposit, kW/m K P fluid density, kg/m3 Pf density of fouling deposit, kg/m fluid shear stress at surface, N/m strength of the deposit 97 Nomenclature Subscripts c cold h hot sat saturation s surface 98 References References Babic-Ivancic, V., H. Furedi-Milhofer, B. Purgaric, N. Braicevic, and Z. Despotoviv. "Precipitation of Calcium Oxalates from high Ionic Strength Solutions", J. Cryst. Growth, 71: 655, 1985. Bott, T.R., "Fouling of Heat Exchangers", Elsevier Science B.V., 1995. Brackenbury, K. and R. Seccombe. "Addressing mill closure issues with enhanced peroxide and hot acid hydrolysis stages", 51 s t Appita Annual general Conference, 1997. Proceedings (Appita), Paper Mo. 3A32: 397-403 (1997. Appita), 1997. Brecevic, L. and D. Skrtic. "Transformation of Calcium Oxalate Hydrates", J. Cryst. Growth, 74: 399,1986. Brown, P., D. Ackermann, and B. Finlayson. "Calcium Oxalate Dihydrate (Weddelite) Precipitation", J. Cryst. Growth, 98:285. 1989. Dahl, O., J. Niinimaki, and H. Kuopanportti. "Evaporation of acidic effluent from ECF bleaching: pilot tests", 1988 International Environmental Conference & Exhibit, TAPPI Proceedings, 1998 Epstein N. "Thinking about Heat Transfer Fouling: A 5 x 5 Matrix", Heat Transfer Engineering, 4: (1) 43-56, Jan-Mar, 1983. Epstein N. "Elements of Particle Deposition onto Nonporous Solid Surfaces Parallel to suspension Flows", Experimental Thermal and Fluid Science, 14: 323-324,1997. Eriksson, G. "An algorithm for the computation of aqueous multicomponent, multiphase equilibrium", Anal. Chim. Acta, 112: 375. 1979. 99 References Ester, D.R. "Reduction of bleach plant deposits yields better pulp, less downtime". Pulp&Paper, 68(9): 135-137, Sept. 1994. Gardner, G.L., "Nucleation and Crystal Growth of Calcium Oxalate Trihydrate", J. CrystGrowth, 30:158,1975. Golaszewski, R. and D. Keselica. "Organic and inorganic deposits control". TAPPI Short Course Notes. Chemical Processing Aids. 1991. Houdlette, G.H. "Controlling inorganic scale deposits increases bleach plant productivity", Pulp Paper 59. No. 6: 154-157. June 1985. Hultman, B., C. Nilsson and S. Sjoberg, "Inhibition of calcium oxalate deposits in a sulfite mill by addition of aluminum sulfate, Svensk Papperstid. 84: R163-168, Dec. 1981. Kern, D.G. and R.E. Seaton. "A theoretical Analysis of Thermal Surface Fouling", British Chemical Engineering, Vol. 4, p. 258, May 1959. Krasowski, J.A. and J. Marton, "The formation of oxalic acid during bleaching - a source of deposits", 1982 TAPPI Research and Development Division Conference, TAPPI Proceedings: 129-138,1982. Krasowski. J.A. and J. Marton, "The formation of oxalic acid during bleaching of Kraft pulp", Journal of Wood Chemistry and Technology, 3(4), 445-458 (1983). Kolthoff, M and E.B Sandell, "Textbook of quantitative inorganic analysis", 3 r d Edition, McMil lan Company, 1952 Martell, A.E. and R.M. Smith, "Critical Stability Constants", v3, Plenum Press, New York, 1979. 100 References Nancollas, G.H. and G.L. Gardner. "Kinetics of Crystal Growth of Calcium Oxalate Monohydrate", J. Cryst. Growth, 46: 355,1979. Nilvebrant, N. and A. Reimann. "Xylan as a source for oxalic acid during ozone bleaching", Fourth European Workshop on Lignocellulosics and Pulp: 485-491,1996. Perez, L.A. and D.F. Zidovec. "Scale control by using a new non-phosphorus, environmentally friendly scale inhibitor", Mineral Scale Formation, Plenum Press, New York, 1995. Perry, R.H. and D. Green, "Perry's Chemical Engineer's Handbook", 7 t h Edition, New York. McGraw Hil l , 1997. Sallis, J.D., W. Juckes, and M.E. Anderson. "Phosphocitrate-potential to influence deposition of scaling salts and corrosion", Mineral Scale Formation, Plenum Press, New York, 1995. Tomazic, B. and G.H. Nancollas. "The Kinetics of Dissolution of Calcium Oxalate Monohydrate", J. Cryst. Growth, 46:355. 1979. Tomazic, B. and G.H. Nancollas. "Crystal Growth of calcium Oxalate Hydrates: A Comparative Kinetics Study", J. colloid and interface science, 75(1): 149. 1980. Ulmgren P., R. Radestrom. "Calcium Oxalate in Bleach Plant Filtrates". Minimum Effluent Mil ls Symposium: Proceedings (TAPPI): 51-62. TAPPI Press. 1997. Ulmgren, P. and R. Radestrom. "An equilibrium model for calcium oxalate solubility in sodium chloride medium", Nordic Pulp and Paper Research, 14: (3), 214-220. 1999. 101 References Ulmgren, P. and R. Radestrom, "Solubility of calcium oxalate in the presence of magnesium ions, and solubility of magnesium oxalate in sodium chloride medium", Nordic Pulp and Paper Research Journal, 14(4):330-335, 1999. 102 Appendix I APPENDIX I Figure 1-1; Calibration curve for magnetic flow meter 103 Appendix I Figure 1^ 2: Calibration curve for rotameter 104 Appendix I Figure 1-3: Wilson plot 105 Appendix II APPENDIX II SAMPLE CALCULATIONS II. 1 Calculation for the rate of heat transfer The following formula was used to calculate the rate of heat transfer: Q = m x c p x A T ( H - l ) Where: m - is the mass flow rate of the fluid, either the mass flow rate of the hot fluid (mh), either the mass flow rate of the cold fluid (rn.) [kg/s] AT - is the temperature difference between the inlet and outlet of either the hot fluid (ATh) or cold fluid (AT.) [K] c p — is the mean specific heat of water averaged over the temperature range of interest [kJ/kg°C] The hot water flow rate was measured with a magnetic flow meter; the signal was acquired, on the computer data logger that gave the output in kg/s. The cold-water flow rate was measured with a rotameter using the calibration curve (Figure 1-2). The rate of heat transfer was calculated for both the hot and the cold fluid as follows: - the rate of heat transferred from the hot fluid: Q h = m h x c p x A T h (H-2) - the rate of heat transferred to the cold fluid: Q c = m c x c p x A T c (H-3) II.2. Calculation of the overall heat transfer coefficients The overall heat transfer coefficient can be calculated using the following general formula: 106 Appendix II U= ^ (H-4) A x A T l m Where: Q - is the heat transferred between the cold and the hot fluid along the test section [kW] A - the heat transfer surface: it can be either the inside or the outside surface of the inner tube [m2] A T . - the log mean temperature: A T - ^Th-' ~ T c - ° )~( T h -° (H-5) Th,o ~ TCjj T ^ . , T H O , T C L , T C O - the inlet and outlet temperatures for hot and cold water; they were measured with four type T (copper-constantan) thermocouples, and the signal was acquired on the computer data logger that gave the output in °C. Because on the hot water side the temperature difference between the inlet and the outlet is much larger than in the cold water side (since m^  < mc), the measurements were considered more reliable and the overall heat transfer coefficient was calculated based on U h = 9s (II-6) A o x A T l m Where A Q is the outside area of the inner tube and is given by: A ^ x D x L (H-7) II-3. The fouling resistance calculation The fouling resistance at any instant of time was calculated using the following formula: 107 Appendix II R f = J — L (II-8) u h u h 0 Where: U h o — the overall heat transfer coefficient based on the heat transferred from the hot water side in clean conditions; [kW/m2K] U h - the overall heat transfer coefficient based on the heat transferred from the hot water side in fouled conditions. [kW/m2K] II-4. Reynolds number calculation The following equations were used to calculate de Reynolds number: - on the hot water side: R e 4 x m h (II-9) h u hx J tx ( D e + D 0 ) - on the cold water side: R e _ 4 x t n c ( 1 1 - 1 0 ) U c X 7 t x D i Where: D o - is the outside diameter of the inner tube [m] Dj - is the inside diameter of the inner tube [m] D e - is the inside diameter ox the exterior tube [m] II-5. Calculation of convective heat transfer coefficient on the hot water side Wilson plot method was used to obtain the convective heat transfer coefficient of the hot water from the overall heat transfer coefficient. In clean conditions: - i -= -L + - L x ^> + ^L ( i i - i i ) U 0 h h h c A; k w Where: h h, h - convective heat transfer coefficients for hot and respectively cold water [kW/m2K] Ao, A i - the outside and respectively inside surface of the inner tube [m2] 108 Appendix II AX - the thickness of the inner tube wall [m2] k x - conductive heat transfer coefficient of the inner tube wall [kW/mK] Ao, A i , A x kx are constant. Several runs were conducted keeping constant the cold water flow rate (0.75 kg/s) and varying the hot water flow rate (0.45, 0.4, 0.35, 0.3, 0.25, 0.2, 0.15 kg/s). n 0.8 Because h was kept constant and given that ^ ~ h (II-12) a It follows that: _L * . + b (H-13) U h Re°h8 1/Uh was plotted versus 1/Reh0-8. One can see that the slope of the graph is equal with a and whence, 1^  can be determined for each Re number (Eq. II-11). From Figure I- 3: a = 587.67 rn^K/kW and b = 0.2272 ntK/kW H-6. Calculation of a correction factor to account for fluctuations in hot water flow rate Let ir^' be the actual value of the hot water mass flow rate at any instant of time and Re h ' and U h ' the corresponding Re and overall heat transfer coefficient. If is the setup value of the hot water flow rate in that experiment, then: m h/m' h = n (11-14) From Eq. (II-8) and (II-13) it follows: Reh/Reh'= n (II-15) Eq. (11-13) becomes: 1 = a . b (11-16) U h n° - 8 x(Re h r Rearranging Eq. (11-15) it follows: 109 Appendix II 1_ 1 u h n (Re h) 0.8 + b + 1 -1 ^ n a 8y (Re h) a u 1 • + b = — r 0.8 1 1 + f i - _ » |b Whence: _J_ = _ U h n08 U h ' T n08 Replacing n from Eq. (11-13) it follows: J _ c • \ m 0.8 0.8" i _ EJL b • (11-17) The above formula gives the corrected value for the overall heat transfer coefficient. One can see that when m^  = m h ' => U h = U h ' . II-7 Water properties calculation The density and specific heat of water at average bulk temperature were deteraiined using the following relationships by applying a polynomial regression to the available literature data {Perry's Chemical Eng. Handbook, 1997): - for cold water (2-14°C): o = -0. 0.0797 T + 1000.3 [kg/m?] He c.avg L & J c p c= -0.0032Tc a v g + 4.2251 [kJ/kg°C] ^ c= -0.00004T c a v g + 0.0017 [Ns/m2] Where Tc,avg is the average bulk temperature of the cold fluid. - for hot water (60-75°C): p h=-0.554T h a v g+1016.5 [kg/irf] c p = 0.0006T h a v g + 4.1483 [U/kgoC] (11-18) (11-19) (11-20) (11-21) (11-22) 110 Appendix II ^ = -0.000006Th a v g + 0.0008 [Ns/m2] (11-23) Where T h a v g is the average bulk temperature of the hot fluid. X _ T h J n + T h . o (H-24) A h.avg 2 T - T c - i n + T c - ° (11-25) 1 c.avg 2 II-8 Numerical Sample calculation for F12 Experiment F12 was carried out in the following conditions: m h = 0.15 kg/s, m c = 0.75 kg/s; pH = 6.3; S.Q = 6; Ca/Ox = 2/1 molar. Chemicals were added at time t = 710 min. At time t = 4000 min from the beginning of the experiment the temperatures reading were: T. . =74.9°C h,m T. =63.7°C h,o T . = 3.3 °C c, m T =5.7°C C,0 From Eq. (11-24) and (11-25): T^ a v g= 69.3 °C and T c a v g = 4.5 °C Water properties are then calculated using Eq. (11-18) to (11-23): p c= 999.94 kg/m3 c p = 4.211 kJ/kg°C „ = 0.00152 Ns/m2 p h= 978.1078 kg/m3 c p h= 4.1899 kJ/kgoC , ,= 0.0003842 Ns/m2 111 Appendix II Applying Eq. (II-9) and (11-10): R e = 4 x 0 - 1 5 =8702 h 0.0003842x7ix(0.0381 + 0.01905) R e = 4x0.75 =WR<) c 0.00152 X T C X 0.0158 Using Eq. (II-5): (74.9-5.7)-(63.7-3.3 ) : AT, - v ' V /=64QQor l m 74.9-5.7 In 63.7-3.3 Applying Eq. (II-2) and (II-3): Q h = 0.15 x 4.1899 x (74.9 - 63.7)= 7 - 0 3 9 k w Q c = 0.75 x 4.211 x (5.7 - 3.3) = 7 -579 kW The outside area of the inner tube was calculated using Eq. (II-7): A0 = 7t X 0.01905 x 1.4 = 0.0837 m 2 From Eq. (II-6) it follows: U = 7.039 = T 9QQ wW/m2K 0.0837x64.99 112 Appendix II Similarly was calculated the overall heat transfer coefficient, before chemicals were added (in clean conditions) and these values were averaged. The average value of the heat transfer coefficient in clean conditions for F12 was: U h o = 1.512 kW/m2K The fouling resistance at t = 4000 min. can then be calculated using Eq. (II-8): R f =_! L_= 0.1083 itfK/kW 1.299 1.512 If the mass flow fluctuation was +10%, m' h = 1.1(0.1) kg/s Then with U ' h = 1.299 kWrn^K, b = 0.2272 Eq. (11-17) gives: U h = 1.230 kW/itfK 113 Appendix III APPENDIX III Table III-1: Calculated crystal concentrations for suspersaturated conditions Exper. Na 2C 20 4 added initially [g] Ca(N03)24H2 O added initially fel [Ox]0 mmol/L [Ox]f* mmol/L mmol/L Suspended Solids** mg/L F12 21 75 6 0.92 2.52 0.18 108 F15 6.1 21.5 1 0.22 0.38 0.18 5.8 F16 48 172 14.5 2 1.5 0.18 265.7 F17 6 11 2 0.25 0.18 0.18 10.2 F18 6.1 32 6 0.24 0.6 0.18 8.8 F19 21 75 6 0.85 0.68 0.18 98 F20 21 75 6 0.89 0.48 0.18 103.7 F21 21 75 6 0.89 0.4 0.18 103.7 F22 21 75 6 0.85 0.38 0.18 98 F24 21 75 6 0.89 1.8 0.18 103.7 F26 50 180 1 1.96 0.3 0.45 220.5 F27 32.3 115.4 6 1.42 0.5 0.28 166.4 F28 168 593.25 6 5.8 0.8 1.45 635 F29 8.66 30.5 1.8 0.38 1.2 0.18 29 F30 21 75 6 0.9 0.65 0.18 105 * Oxalate concentration at the end of the range of initial fouling rate calculation ** Assumes solution to be saturated and crystals to be calcium oxalate monohydrate. 114 Appendix III Figure lll-l : Experiment F15 - overall heat transfer coefficient, total calcium and total oxalate concentration vs. time (S i o = 1, T h = 75°C, pH = 6.3, initial Ca/Ox = 2/1, m h =0.15 kg/s). In the given experimental condition [Ox]^ = 0.18 mmoI/L ^ 0.5-, E 0.4 4 0.3 I 0.2 0.1 o u o -:-2000 4000 6000 Time [min] 8000 10000 115 Appendix III Figure 111-2 : Experiment FI6 - overall heat transfer coefficient, total calcium and total oxalate concentration vs. time (Sj0 = 14.5, T|, = 75°C, pH = 6.3, initial Ca/Ox = 2/1, mi, = 0.15 kg/s). In the given experimental condition [Ox]^ = 0.18 mmol/L 3 i 1 ^  ^ 1000 2000 3000 4000 Time [min] 5000 6000 7000 1 Appendix III Figure 111-3 : Experiment F18 - overall heat transfer coefficient, total calcium and total oxalate concentration vs. time(S;0 = 2, T h = 75°C, pH = 6.3, initial Ca/Ox = 3/1, m h = 0.15 kg/s). In the given experimental condition [Ox]^ = 0.18 mmol/L 1 0.9 0.8 0.7 0.6 0.5 -I 0.4 2000 4000 6000 Time [min] 8000 10000 5 4 3 -2 -1 -0 2000 4000 6000 Time [min] 8000 10000 0.8 0.6 -0.4 -0.2 0 0 2000 4000 6000 Time [min] 8000 10000 117 Appendix III Figure IH-4 : Experiment F19- overall heat transfer coefficient, total calcium and total oxalate concentration vs. time(S io = 6, T h = 75°C, pH = 6.3, initial Ca/Ox = 2/1, nth = 0.15 kg/s). In the given experimental condition [Ox]^ = 0.18 mmol/L 1.4 H 0.4 H 1 , , , , 0 2000 4000 6000 8000 10000 Time [min] 2000 4000 6000 Time [min] 8000 10000 •4 o o u O U 1 0.8 - 0 . 6 e 0.4 0.2 0 2000 4000 6000 Time [min] 8000 10000 118 Appendix III Figure 111-5 : Experiment F20 - overall heat transfer coefficient, total calcium and total oxalate concentration vs. time(S;0 = 6, T h = 75°C, pH = 6.3, initial Ca/Ox = 2/1, ma = 0.4 kg/s). In the given experimental condition [Oxjsat = 0.18 mmol/L 1.1 -1 0.9 -"B 4 0.7 -r—( 0.5 -0.3 --0 2.5 i © 2 -c3 1.5 -c o o 1 -t u 0.5 -0.8 H £ a " 0.6 o o u 0.4 O U 0.2 • * * • **** 'Cr^yy «**v»v» Time [min] i i • i i i 1 1 — i 1000 2000 3000 4000 5000 6000 7000 8000 Time [min] 0 1000 2000 3000 4000 5000 6000 7000 8000 Time [min] 119 Appendix III Figure IIT-6 : Experiment F21- overall heat transfer coefficient, total calcium and total oxalate concentration vs. time(Si0 = 6, Th = 75°C, pH = 6.3, initial Ca/Ox = 2/1, mi, = 0.3 kg/s). In the given experimental condition [Ox]sat = 0.18 mmol/L 0.9 0.8 W 0.7 ^ 0.5 0.4 0.3 o c o o o a 0.8 0.6 -0.4 -0.2 0 1 1 1 1 1 i i 0 1000 2000 3000 4000 5000 6000 7000 Time [min] — i 1 1 i i i i 1000 2000 3000 4000 5000 6000 7000 Time [min] 1 1.7 | 1.5 H " 1.3 i i . i o ° 0.9 6 0/7 0.5 T 1 1 1 I I I 1000 2000 3000 4000 5000 6000 7000 Time [min.] 120 Appendix III Figure IIJ-7 : Experiment F22 - overall heat transfer coefficient, total calcium and total oxalate concentration vs. time(Si0 = 6, Th = 75°C, pH = 6.3, initial Ca/Ox = 2/1, ma = 0.2 kg/s). In the given experimental condition [Ox]^ = 0.18 mmol/L 0.95 -0.85 -0.75 -0.65 -T - ( 0.55 -0.45 -0.35 -g- 2 I 1.5 H o 1 1 c t ° 0.5 -1 a U 0 0 o 0.8 B o c 0.6 o o O 0.4 CJ 0.2 0 2000 4000 6000 Time [min] 8000 — i 1 1 1— 2000 4000 6000 8000 10000 Time [min.] 2000 4000 6000 Time [min] 8000 10000 12000 10000 121 Appendix III Figure III-8 : Experiment F24 - overall heat transfer coefficient, total calcium and total oxalate concentration vs. time(Si0 = 6, Th = 75°C, pH = 6.3, initial Ca/Ox = 2/1, ma = 0.15 kg/s). In the given experimental condition [Ox]^ = 0.18 mmol/L i f 1 - i 0.9 -0.8 -0.7 -0.6 0.5 0.4 o o B a d fi o o + U 6 4 2H 0 3 -, 2.5 B B 1.5 o C o o o a i 0.5 0 2000 2000 2000 4000 6000 Time [min] 8000 10000 4000 6000 Time [min] 8000 10000 4000 6000 Time [min] 8000 10000 122 Appendix III Figure HI-9 : Experiment F26 - overall heat transfer coefficient, total calcium and total oxalate concentration vs. time(S io = 1, T h = 75°C, pH = 3, initial Ca/Ox = 2/1, ma = 0.15 kg/s). In the given experimental condition [Ox]sat = 0.45 mmol/L 1.2 1.1 I 1 § 0.9 .•=- 0.8 I 0-7 i 0.6 0.5 0 JL 1-5 I • o 0.5 O U 0 0 ^ ' 5 1 1 4 -* 3 CJ 0 i i 1 1 1 1 1000 2000 3000 4000 5000 6000 Time [min] 1000 2000 3000 4000 5000 6000 Time [min] 1000 2000 3000 4000 Time [min.] 5000 6000 123 Appendix III Figure ITMO: Experiment F27 - overall heat transfer coefficient, total calcium and total oxalate concentration vs. time(Si0 = 6, T h = 75°C, pH = 3.7, initial Ca/Ox = 2/1, im, = 0.15 kg/s). In the given experimental condition [Ox]^ = 0.28 mmol/L 1.1 -I 1 -1 0.9 -0.8 -0.7 -0.6 -0.5 - i i i i i 1 1 1000 2000 3000 4000 5000 6000 7000 Time [min] 2 o 3 -i 1 2 -one 1 -o t a 0 -U o C o o O O 1.5 i 1 H 0.5 1000 2000 3000 4000 5000 6000 7000 Time [min.] • • 1000 2000 3000 4000 5000 6000 7000 Time [min] 124 Appendix III Figure HI-11 : Experiment F28 - overall heat transfer coefficient and total oxalate concentration vs. time(S io = 6, T h = 75°C, pH = 2.2, initial Ca/Ox = 2/1, n^ = 0.15 kg/s). In the given experimental condition [Ox]^ = 1.45 mmol/L 0 1000 2000 3000 4000 5000 6000 7000 Time [min] 6 o C l C o O U 3 2 1 0 1000 2000 3000 4000 5000 6000 7000 Time [min] 125 Appendix HI Figure ni-12:Experiment F29 - overall heat transfer coefficient, and total oxalate concentration vs. time (S i o = 1.8, T h = 75°C, p H = 6.3, initial Ca/Ox = 2/1, m h = 0.15 kg/s). In the given experimental conditions [Ox] = 0.18 mmoI/L. I 0.95 0.9 -\ £ 0.85 g 0.8 H ~E 0.75 ^ 0.65 0.6 0.55 -\ 0.5 0 500 1000 1500 2000 2500 3000 3500 4000 Time [mini 1.5 -. o E 1 0.5 O u 0 -*r-0 1000 2000 Time [min.] 3000 4000 126 Appendix III Figure IH-13:Experiment F30 - overall heat transfer coefficient, total calcium and total oxalate concentration vs. time (S;0 = 6, Th = 75°C, pH = 6.3, initial Ca/Ox = 2/1, mh = 0.35 kg/s). In the given experimental conditions [Ox] = 0.18 mmol/L. 0.85 0 75 0.65 0.55 i—< 0.45 0.35 3 - i o 2.5 -§ 2 -1.5 -o a 1 -K 0.5 -cj 0 -1000 2000 3000 4000 5000 6000 7000 Time [min] 8000 1000 2000 3000 4000 5000 6000 7000 Time [min.] 8000 o c o o O U 1.2 1 0.8 H 0.6 0.4 -\ 0.2 0 0 1000 2000 3000 4000 5000 6000 7000 8000 Time [min] 127 Appendix III Figure III-14:Experiment F17 - overall heat transfer coefficient vs. time (Sj0 = 1, T h = 75°C, pH = 6.3, initial Ca/Ox = 1/1, m h = 0.15 kg/s). In the given experimental conditions [Ox] = 0.26 mmol/L. 0.8 -| 0.75 -Time [min] 128 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0058675/manifest

Comment

Related Items