Chapter 14. Gas-liquid Transfer

Transcription

1 Chapter 14 Ga-liquid Tranfer 14.1 GAS TRANSFER By ga tranfer i meant the ma movement of ga to or from olution, which occur when a liquid urface i in contact with a ga phae with which it i not in equilibrium. In chemical engineering practice the uptake of ga by a liquid i generally termed aborption while the removal of ga from a liquid i uually termed deorption or tripping. In water and watewater engineering practice the ga phae i uually air and hence ga tranfer i generally termed aeration. The gae of interet to the water and watewater engineer are O 2, CO 2, CH 4, H 2 S, NH 3, and Cl 2. The driving force that activate the tranfer of thee gae to and from water i the difference between their concentration in olution and their olubility under the prevailing condition GAS SOLUBILITY The main factor that influence ga olubility in water are: the water temperature, the partial preure of the ga in the ga phae, the diolved olid concentration in the water phae and the chemical nature of the ga. The olubility of gae, unlike that of olid, decreae with increae in temperature. For partial preure up to 1 atm and for gae that do not react with water to any great extent, the equilibrium concentration of ga in olution at a given temperature i proportional to the partial preure of the ga in contact with the water in accordance with Henry Law: c = Hp (14.1) where c (mg l -1 ) i the aturation or equilibrium concentration of the ga in olution, p (atm) i the partial preure of the ga in the ga phae in contact with the water and H i the olubility coefficient. Henry Law applie to mot of the gae of interet in water and watewater engineering including O 2, CH 4, CO 2, H 2 S. The latter two gae undergo reaction in olution. Diolved CO 2 react with water a follow: CO 2 + H 2 O H 2 CO 3 (14.2) H 2 CO 3 H + + HCO 3 - (14.3) HCO 3 - H + + CO 3 2- (14.4) For the condition normally encountered in water engineering the concentration of H 2 CO 3 will not exceed 1% of the CO 2 concentration (refer Chapter 1 for dicuion of the role of the carbonate ytem in water). Hydrogen ulphide react in olution a follow: H 2 S H + + HS - (14.5) HS - H + + S 2- (14.6) It i apparent from equation (14.5) and (14.6) that the diolved form of H 2 S depend on the ph of the olution. 146

2 Ammonia (NH 3 ) and chlorine (Cl 2 ) are highly oluble gae that readily react with water. Their preure-olubility relationhip deviate from Henry Law. Water olubility data for the above gae are preented in Table 1.7, 1.8 and 1.9 in Chapter 1. It i clear from the tabulated data that the gae which react with water, including CO 2, H 2 S, NH 3 and Cl 2, are coniderably more oluble than the remaining non-reactive gae MECHANISM AND RATE OF GAS TRANSFER When a water urface i expoed to a poorly oluble ga with which it i not in equilibrium, it i aumed that the water at the interface become intantaneouly aturated with the ga and that the ga i tranported into the body of the liquid by the proce of molecular diffuion: m = D c (14.7) t x where m/ t i the rate of ga tranport acro unit are of interface, D i the coefficient of molecular diffuion and c/ x i the concentration gradient normal to the interface. The implet phyical model of interface condition i that propoed by Whitman and Lewi (1924), hown in Fig They aumed that the reitance to ga tranfer reided in fixed ga and liquid film at the ga-liquid interface. The movement of ga acro the ga film implie the exitence of a preure gradient in the ga film and hence the ga preure P i at the interface i le than the bulk ga preure P g. However, for ga of low olubility the rate of tranfer of ga to the liquid phae i low and hence the preure gradient in the ga phae i negligibly mall. bulk ga well mixed (preure gradient negligible) interface fixed ga film fixed liquid film C L P g P i C i = C ga conc. in olution bulk liquid well mixed (conc. gradient negligible) Gradient acro film Fig 14.1 Two-film ga tranfer model Whitman and Lewi aumed a linear concentration gradient acro the liquid film: c C = x L C h (14.8) where C i the aturation or interface concentration of the ga in olution, and C L i the ga concentration in the bulk liquid (aumed to be well-mixed). The rate of tranfer acro a ga-liquid interface of area A i calculated from equation (14.7) and (14.8) to be Dividing by the liquid volume V: A m AD ( C L C ) t = h (14.9) A V m AD t Vh C L = ( C ) 147

3 Hence or dt dt ( C ) = a D h C (14.10) L ( C ) = k a C (14.11) L L where a = A/V, k L i called the liquid film coefficient (alo called the aborption coefficient ) and the product k L a i the overall tranfer coefficient. Although it can be verified experimentally that an equation uch a (14.11) accurately repreent ga-tranfer kinetic in ytem of the O 2 /water and CO 2 /water type, it i unlikely that the imple liquid film model, on which the equation i baed, correctly repreent actual condition at the interface. In the aeration ytem ued in water and watewater treatment procee, large area of air-water interface are created and continually renewed. When an element of interface i created, the initial rate of ga tranfer to or from a non-equilibrated liquid i high, but i reduced a the concentration gradient reduce (equation 14.7). Thu the more rapidly an element of interface i renewed the higher will be the mean rate of ga tranfer through it. Alo, ince ga concentration gradient in the bulk water are generally very mall, indicating little reitance to ga tranport within the bulk liquid phae, it i apparent that reitance to ga tranfer reide in the organied water at the interface. It i probable that the effective thickne of thi highreitance film i influenced by the level of turbulence in the bulk liquid. Thu the overall tranfer coefficient, which i a contant for a given ytem, i dependent on the total area of air-water interface created, the rate of interface renewal, the level of turbulence in the liquid and temperature OXYGEN TRANSFER The main function of aeration in watewater treatment i the olution of oxygen in the mixed liquor of the activated ludge proce. In the activated ludge proce, a in other imilar biochemical procee, the rate of oxygen input to the ytem mut balance the rate of oxygen uptake, ince, owing to it low olubility, the torage of oxygen in the ytem i negligible. Aerator are alo ometime ued in polluted river and lake to prevent exceive oxygen depletion. Aerator for oxygen tranfer are rated in term of oxygenation capacity (OC), which i defined a the rate at which an aerator tranfer oxygen to clean water under tandard condition, i.e. at a temperature of 20 o C (or 10 o C), an atmopheric preure of 1 atm and a zero concentration of oxygen in olution. Under thee condition, equation (14.11) become: dt = k L a C*(20) (14.12) where C *(20) i the equilibrium aturation concentration of oxygen in the water being aerated, at a temperature of 20 o C. It hould be noted that C *(20) may differ from C (20), the aturation concentration for a water in equilibrium with atmopheric air. Thi i particularly the cae for dipered air ytem with ubmerged diffuer, where C *(20) typically exceed C (20). With urface aeration ytem, C *(20) may be taken a equal to C (20) for practical computational purpoe. Since then OC = V (14.13) dt OC = k L a C*(20) V (14.14) where V i the liquid volume. The condition under which aerator operate in the activated ludge proce are quite different from the tandard condition to which their OC-value relate. The more ignificant difference are the preence 148

4 of olid in upenion, the preence of urfactant in olution and the temperature difference. Thee factor may influence both the value of k L a and C. Their effect on thee parameter are taken into account by the introduction of the empirical coefficient α and β, which are defined a follow: k La in upenion α = k a in tapwater L C in upenion β = C in tapwater (14.15) (14.16) The concentration of upended olid in the mixed liquor of activated ludge plant i uually within the range mg l -1. Since upended olid may alter the bulk and urface propertie of the upending medium, they are likely to affect ga tranfer kinetic (Karmo, 1972). Supended olid may effect an increae in vicoity, which tend to reduce the level of turbulence and rate of urface renewal. Particle that have a preferential affinity to adorb on the water urface may influence urface hydrodynamic. On the credit ide, air bubble may become attached to the olid in upenion and be carried down into the bulk liquid. Surface-active agent are preent in mot watewater treated by the activated ludge proce and for thi reaon their influence on oxygen tranfer ha been widely tudied. By reducing urface tenion, they effect an increae in the diameter of pherical bubble and a correponding increae in the airwater interfacial area per unit volume of air. On the negative ide, however, they tend to attach to the bubble urface, where they contitute a barrier to oxygen tranfer. The latter effect i more marked in diffued air ytem, where the riing bubble have a ufficiently long reidence time for thi layer to become etablihed. Tet reult how that the operational α-value in activated ludge procee i typically le than unity. It i well etablihed (Barnhart, 1969) that urface-active agent in olution reduce the oxygen tranfer rate in diffued air aeration ytem, reulting in a typical operational α-value range of The correponding typical α-value range for urface aerator i It hould be noted that an α-value in exce of unity infer an enhanced rate of oxygen tranfer under proce condition. Thi may arie where there i a rapid renewal of the air-water interface and hence inufficient time for the formation of a urfactant interfacial barrier to be formed. The value of the β-factor under proce condition i alo typically le than unity and clearly depend on the diolved olid concentration of the aqueou upenion being aerated. In activated ludge procee, the value of β i generally between 0.9 and INFLUENCE OF TEMPERATURE ON OXYGEN TRANSFER The oxygen aborption coefficient k L increae with increaing temperature, due mainly to the temperature-dependent nature of thoe water propertie that affect ga tranfer, namely diffuivity, urface tenion and vicoity. Oxygen diffuivity in water increae with increaing temperature. Surface tenion and vicoity, both of which affect the urface renewal rate, decreae with increaing temperature. Thee latter propertie alo influence the energy required to create air-water interface. The total area of air-water interface created per unit energy input hould increae with increaing temperature, a urface tenion and vicoity correpondingly decreae. Thi infer that the overall tranfer coefficient k L a hould be affected to a greater extent by temperature than i the aborption coefficient k L. In rating aeration ytem, it i uual practice to relate performance to a tandard temperature of 20 o C. The value at any temperature T may be related to it value at 20 o C by a temperature coefficient f, a follow: (k L a) 20 = (k L a) T f (20-T) (14.17) 149

5 where T i the temperature ( o C) at which k L a i meaured. Barnhart (1969) ha reviewed tudie on the temperature dependence of oxygen tranfer in diffued air aeration ytem, for which he reported an average value of 1.02 for f. The ASCE Standard for the Meaurement of Oxygen Tranfer in Clean Water (1992) recommend an f-value of It i alo worthy of note that the rate of oxygen tranfer by aeration ytem, at zero oxygen concentration, and within the temperature range generally encountered in water and watewater engineering practice, i not greatly affected by temperature ince the increae in k L a with riing temperature i offet by a correponding decreae in the olubility of oxygen. Uing the oxygen olubility value given in Table 1.7 and an f-value of 1.02, calculation how that, while there i an increae of 48.6% in the value of k L a due a temperature rie from 0 o C to 20 o C, the actual rate of oxygen tranfer, at zero oxygen concentration, by an aeration ytem at 20 o C i le than it value at 0 o C, but only by about 8% AERATION SYSTEMS USED IN THE ACTIVATED SLUDGE PROCESS The three main type of aeration ytem ued in the activated ludge (AS) proce are a follow: (a) (b) (c) dipered air ytem mechanical urface aeration ytem combination of (a) and (b) Dipered air ytem In dipered air ytem, air i introduced to the mixed liquor through ubmerged diffuer. Dipered air ytem are broadly claified according to the bubble ize generated by the diffuer, which may range from fine to coare. Fine bubble diffuer typically produce bubble in the 2-5mm ize range in clean water, while coare bubble diffuer produce bubble in the ize range 6-10mm. Coare bubble diffuer typically conit of drilled hole or lot in a ubmerged air ditribution ytem. The produced bubble may be maller than the orifice ize through being heared off by the water preure. The oxygen tranfer efficiency (OTE) of coare bubble ytem i generally lower than fine bubble ytem (OTE i defined a the percent of oxygen tranferred per metre ubmergence, at zero diolved oxygen concentration). Fine bubble diffuer (USEPA, 1989) are contructed from a range of material including ceramic, porou platic and perforated membrane. Ceramic and porou platic have interconnected pore tructure through which the air flow to be dicharged through the top urface a a bubble tream. Their pore tructure i rigid and hence ubject to fouling either due to impuritie in the air or precipitation from the water ide. To prevent clogging of porou diffuer, it ha been found neceary (Paveer and Sweeri, 1965) to reduce the dut content of the air ued to a level not greater than mg m -3. Thi i accomplihed by filtration. The perforated membrane are made from either thermoplatic or elatomeric heet material (platiized PVC, EPDM rubber, neoprene rubber) of thickne 1-2mm. They are mechanically perforated with a pattern of mall hole or lit, which expand a the air flow increae, hence they are commonly known a flexible membrane diffuer. Oxygen tranfer from an individual air bubble may be conidered to take place in three conecutive tage: bubble formation, bubble acent and bubble ecape at the urface. The oxygen tranfer rate i high during the bubble formation phae, ince a new air-water interface area i being created and the interfacial oxygen concentration gradient i therefore high. The extent of oxygen tranfer during the acent depend on the rate of interface renewal, bubble urface area and bubble contact time, factor that are dependent on bubble ize and ubequent depth of the diffuer. Barnhart (1969) found that for given volume of air and water, the overall tranfer coefficient had a maximum value at a bubble ize 150

6 of 2.2mm. Downing (1960) found that the average value of the overall tranfer coefficient decreaed with increaing depth of the diffuer. The total contact time, which effectively equal the time of acent, i greatly reduced if the water, into which the air i releaed, ha itelf an upward velocity, a in the cae of piral-flow tank, where diffuer are located on one ide of the tank only. The final phae of oxygen tranfer i due to urface diturbance generated by bubble ecape. Experimental evidence ugget that in mot cae thi i the leat ignificant of the three phae. While the overall oxygen tranfer coefficient k L a for diffued air aeration ytem i readily determined experimentally (refer ection 14.7), the determination of the aborption coefficient k L i more problematic. Reported value of k L vary from 3.67 mm min-1 for a ingle large bubble in a narrow tube ( Adeney and Becker, 1918, 1919, 1920) to mm min-1 for bubble in a highly turbulent aeration tank (Holroyd, 1952). Fig 14.2 how a chematic arrangement of a typical AS proce diffued air aeration ytem. air main typical tank depth in range 3-6m fine bubble dic diffuer fine bubble tubular diffuer Fig 14.2 AS proce diffued air aeration ytem layout Mechanical urface aeration ytem Mechanical urface aeration ytem effect oxygen tranfer by creating and rapidly renewing large area of air-water interface. Such interface reult from a combination of entrainment of air, praying of water into the air and the generation of a high level of urface turbulence. Mechanical urface aerator may be broadly claified in the following categorie: (a) (b) Impeller-type device, mounted on a vertical haft Rotor-type device, mounted on a horizontal haft. Vertical haft device are either of the plate or turbine type. Plate aerator (Downing et al., 1960) conit of a horizontal circular plate, mounted on a vertical haft, to the underide of which i attached a et of vertical radial or curved blade. In operation, the top of the plate i located at or lightly below the till water urface level. The rotation of the dic caue liquid to be dicharged radially, creating a circular hydraulic jump in which air i entrained. Air may alo be ucked down through opening in the plate to the low-preure zone created behind each vane a the plate rotate. A plate aerator on it own may not provide adequate mixing in an aeration tank. Supplemental mixing can be provided by extending the plate haft downward and attaching a mixing impeller at a lower level. Turbine aerator effect oxygen tranfer primarily by a pumping action. The impeller, which i located at the liquid urface, i deigned to pump large quantitie of liquid at a low head. Becaue the impeller i located at the liquid urface a large amount of air i entrained in the pumped liquid, which i thrown upward and outward in a low trajectory, creating coniderable turbulence a it trike the liquid urface. In deep tank, turbine aerator may be ued in conjunction with a draft tube, which extend 151

7 downward to within a hort ditance of the tank bottom, a hown in Fig Thi arrangement enure good mixing of the tank content. variable peed geared motor upport tructure AERATION TANK draft tube Fig 14.3 Typical arrangement of a vertical haft urface aeration ytem Rotor-type or bruh aerator conit of a horizontal revolving haft, to which are attached flat or angular blade, projecting radially outward. The overall diameter of bruh rotor i uually between m; their immerion depth may vary between 100mm and 300mm, depending on diameter. Bruh aerator (Caey, 1971) effect oxygen tranfer by entraining air in the liquid pray which i thrown upward and outward by the revolving blade. Coniderable urface turbulence i alo generated. Bruh aerator are generally ued in cloed-loop aeration tank of the oxidation ditch type, in which the mixing requirement i met by generating a horizontal flow velocity of ufficient magnitude to keep the mixed liquor olid in upenion. It hould be noted be noted that AS proce aeration ytem are required to provide adequate mixing a well a tranfer the required amount of oxygen into olution. Hence, the geometry of the aeration tank mut be matched to the mixing characteritic of the aeration device. In ome ituation it may be neceary to upplement a diffued air ytem by a mechanical mixing device to enure an adequate ditribution of mixing energy throughout the aeration tank EXPERIMENTAL DETERMINATION OF OXYGENATION CAPACITY The tandard oxygen tranfer rate (SOTR) of an aeration ytem (oxygenation capacity) i defined a the rate at which the aeration ytem tranfer oxygen into olution in clean water at a temperature of 20 o C, a zero oxygen concentration in olution and a prevailing atmopheric preure of 1 atm. In general, the prevailing tet condition will not be tandard and hence temperature and preure correction have to be applied to the tet reult. Combining equation (14.14) and (14.17), the reultant expreion for SOTR become: (20 T) SOTR = k a f C V P L (T) *(20) P b (14.18) where C *(20) i the aturation or equilibrium concentration of oxygen in olution (kg m -3 ) under tet condition at a water temperature of 20 o C, V i the aerated water volume (m 3 ), P b i the prevailing atmopheric preure (atm) at the tet ite and P i the tandard barometric preure of 1 atm. 152

8 For urface aeration ytem, the value of C *(20) can be taken a being equal to C (20), the aturation concentration correponding to an air preure of 1 atm, which, a given in Table 1.7, i 9.08 g m -3. For diffued air ytem, however, the value of C *(20) exceed C (20) by an amount which depend on the diffuer ubmergence and, hence, ha to be determined experimentally. A hown in ection 14.3, the rate of oxygen tranfer during an aeration proce i repreented by the expreion: or dt ( C) = k a C ( C C) * L * ( ) = k a dt L Integrating thi equation between an initial concentration of oxygen in olution C i at time zero and a concentration C t after an aeration interval of t: Thi give Ci t = k La(dt) Ct ( C C ) * L 0 ( * i ) ( C C ) ln C C * t k La t = (14.19) or ( ) ( ) ln C C = ln C C - k a t (14.20) * t * i L If the natural log of oxygen deficit (C * -C t ) i plotted a a function of aeration time t, then a traight-line graph i obtained in accordance with equation (14.20), a illutrated in Fig The overall tranfer coefficient k L a i the lope of thi line and ha the unit time Oxygen deficit (% aturation) rotor immerion depth (mm) Aeration time (min) Fig 14.5 Reult of an oxygenation capacity tet on a bladed rotor; rotor diameter 700 mm. 153

9 The determination of the oxygenation capacity of an aeration ytem require meaurement of the overall oxygen tranfer coefficient k L a (equation (14.18)). The unteady-tate re-aeration method i the uual experimental method for k L a determination (ASCE, 1992). The water i firt de-oxygenated by ga tripping or by the addition of odium ulphite (Na 2 SO 3 ). To accelerate the chemical reduction of oxygen, a catalyt uch a Co 2+ i added at a concentration of about 0.1 mg l -1. If an independent mixing ytem i not available it may be neceary to leave the aeration ytem in operation during the chemical de-oxygenation proce to enure diperion of the ulphite throughout the tet tank volume. Where the latter applie an amount of ulphite in exce of the toichiometric requirement hould be added to allow for the exce conumption by the aeration proce. The rate of re-oxygenation i then monitored, uually by DO enor at a number of location in the tet tank. For reliable k L a determination, it i deirable that the re-oxygenation meaurement hould extend over a wide value range, e.g. 10%-90% aturation. Since aeration tank are normally well mixed, the patial variation in oxygen concentration at a given intant in time tend to be mall. The meaured re-oxygenation rate data are ued to compute the value of k L a for the ytem, uing the emi-log graphical method or the three-parameter fit method a decribed in the following paragraph. The implet method of k L a determination i the emi-log graphical method, in which the oxygen aturation deficit i plotted on emi-log paper a a function of time, reulting in a traight line plot, a illutrated in Fig The value of k L a i calculated a the lope of the emi-log plot and i corrected for temperature in accordance with equation ( The SOTR value for the aeration ytem i computed by inertion of the appropriate parameter value in equation (14.8). Thi method i uitable for aeration ytem where the aturation concentration can be aumed to be the ame a the concentration in equilibrium with atmopheric air. To illutrate the three-parameter fit method, equation (14.20) i written in the form: ( ) ( ) C = C C C exp k a t (14.21) t * * i L Non-linear regreion i ued to find the bet-fitting value of the three parameter C *, C i and k L a (ASCE, 1992). An example of the output from uch a computation i preented in Fig Detail of the computational procedure, together with an example are given in Appendix C. Thi i the preferred method of k L a determination a it i baed on a bet-fit model of the experimental meaurement. It ha the advantage of not requiring an aumed value for the aturation concentration C *, which i of particular ignificance for diffued air ytem where the aturation concentration i a function of the immerion depth of the diffuer. 120 Oxygen deficit D (%C ) meaured value Bet-fit parameter value C i = 39.5 C * = k L a = min -1 Fitted equation: D = exp( t) %C Aeration time t (min) Fig 14.6 Experimental data for SOTR tet on diffued air aeration ytem Diffuer ubmergence 4m; DO enor calibrated to read 100% in air-equilibrated water 154

10 The teady-tate oxygen tranfer rate (OTR) by an activated ludge aeration ytem, operating at a fixed DO concentration C (mg l -1 ) and at a temperature T ( o C), i related to the ytem SOTR value a follow: OTR = α SOTR f ( ) ( β C C T 20 *(T) ) C *(20) P P b (14.22) The OTR value for an activated ludge diffued air aeration ytem can alo be meaured under field operating condition by monitoring the oxygen concentration in the off-ga from the aeration tank and thu quantifying the drop in the oxygen content of the air flow. Ideally, carbon dioxide and water vapour hould be tripped from both tream to eliminate partial preure effect. A et of guideline for carrying out off-ga tet ha been drawn up by the American Society of Civil Engineer (ASCE, 1996) The oxygen content of the influent and emergent air tream i conveniently done uing an aircalibrated DO meter. The procedure can be ued a a teady-tate method for meauring oxygen tranfer in activated ludge procee operating at a contant loading rate and hence at fixed DO level ENERGY CONSUMPTION AND OXYGENATION EFFICIENCY The tandard aeration energy utiliation efficiency (SAE) i uually defined in term of the energy expended per unit ma of oxygen tranferred under SOTR meaurement condition: SAE = SOTR P kg O 2 kwh 1 (14.23) where P i the wire-to-water power input (kw). The power conumption of air blower unit ued to upply air to diffued air aeration ytem i related to air flow and delivery preure a follow: P Q T a(m) 1 p 2 = η p kw (14.24) where Q a(m) i the intake air flow (kg -1 ), T 1 i the abolute temperature (K) of the intake air, η i the combined motor/blower efficiency, p 1 and p 2 are the abolute preure of the intake and compreed air tream, repectively. The SAE of diffued air ytem i a function of the combined efficiencie of the air delivery and oxygen tranfer procee. The SAE value for fine bubble diffued air ytem typically fall within the range kg O 2 kwh -1. The SAE of urface aeration ytem i typically in the range kg O 2 kwh -1. The energy utiliation efficiency of aeration ytem under field operating condition generally fall within a much narrower band than given by the foregoing SAE value range. Thi i due to the fact that the performance of diffued air ytem i more adverely affected (reduced α- and/or β-factor) by contaminant uch a urfactant than i the performance of urface aeration ytem. While fine bubble aeration ytem are generally more efficient than urface aeration ytem, they have the diadvantage of being uceptible to clogging and reduction in performance over time. The tandard oxygen tranfer efficiency (SOTE) of diffued air aeration ytem i the fraction of oxygen in the compreed air flow that i taken into olution under SOTR condition: SOTR SOTE = (14.25) 0.228Q a(m) where Q a(m) i the air ma flow rate with an oxygen ma fraction of

11 14.8 GAS STRIPPING Removal of carbon dioxide A may be een from the data in Table 1.9 relating to the olubility of atmopheric gae in water, the aturation concentration of carbon dioxide (CO 2 ) in a water body in contact with free air i le than 1 mg l-1 at ambient temperature. Water become uperaturated with CO 2 in a variety of circumtance. For example, many groundwater are found to have high concentration of CO 2, taken into olution a rainwater percolate through oil in which CO 2 i produced by microbial repiration. High CO 2 concentration are produced in the treatment of high-alkalinity water through the addition of trivalent aluminium or iron alt in the chemical coagulation proce ued for the removal of colour and turbidity in the production of drinking water from urface water. A dicued in Chapter 9, water that are uperaturated with CO 2 may be chemically untable and corroive. The baic ga tranfer principle already outlined for the diolution of oxygen in water are alo applicable to the tranfer of CO 2 out of olution. For ga removal it i important to keep the partial preure of the ga in the ga phae a low a poible - thi can be achieved by good ventilation with a high ratio of air/water. Spray nozzle, cacade, forced draft multiple tray aerator and packed column are ued for CO 2 tripping Removal of hydrogen ulphide Hydrogen ulphide (H 2 S) i found in ome groundwater and in effluent from anaerobic treatment. Even at low concentration it give rie to odour and corroion problem and ha a high chlorine demand. It removal by aeration i complicated by it diociation in olution (ee equation (14.5) and (14.6)). To control the extent of formation of the ionic pecie HS - and S 2- it may be neceary to reduce the ph by acid addition. Added oxygen may alo react with H 2 S to form free ulphur. While the latter reaction reduce the H 2 S content it may lead to further problem in chlorination and ditribution (growth of ulphur bacteria in the ditribution ytem). Although the general aeration principle already outlined are valid, becaue of the additional complicating factor involved in H 2 S removal, pilot plant tudie are adviable prior to plant deign Removal of ammonia by aeration Ammonia i produced in mot watewater, including dometic ewage, through the microbial degradation of organic compound containing nitrogen. Since it i an important nutrient, it removal from effluent i often neceary to control eutrophication in receiving water. Thi can be achieved biochemically by nitrification (converion to NO 3 - ) followed by ubequent denitrification to N 2, in which form it i no longer an available ource of nitrogen for mot organim. Ammonia can alo be removed by aeration. For efficient air-tripping a high ph i required to convert ammonium ion to NH 3 : NH 4+ + OH - NH 4 OH NH 3 + H 2 O (14.26) Becaue ammonia i a highly oluble ga (ee Table 1.8), large volume of air are required to effect it removal. The efficiency of the proce increae with increaing temperature owing to the reulting reduction in ga olubility GAS TRANSFER IN PACKED COLUMNS Inter-phae ga tranfer i achieved in packed column by the creation and renewal of the ga-water interface area a water trickle down through a bed of dicrete packing of the type hown in Fig The downward cacading water i brought into cloe contact with the ga contained in the void of the packing. Depending on the requirement of the application the ga phae may be maintained at a fixed preure (a in aturator ued in conjunction with the flotation proce) or may be in the form of a counter-current flow (a in ga-tripping application). 156

12 C i packing h dh C h Fig 14.7 Schematic repreentation of packed column Uing Fig 14.7 a a proce definition diagram, the ga tranfer characteritic of packed column can be derived by writing equation (14.7) in the form: = k L a dt(c C) Applied to a packed column, thi relationhip can be written a follow: dh dh = k a( C C) dt L dh ( ) = k a C C t L (14.27) h where t i the hold-up time in the column and h i the column height. It i aumed that dt/dh = t/h. i.e. there i no variation in the liquid reidence time over the column height. The ratio h/(k L a t) ha the dimenion of length and ha been found to be approximately contant for a given ga tranfer proce. Thi packed column characteritic i the tranfer unit height or HTU. Thu equation (14.27) can be written a 1 = dh HTU C C ( ) (14.28) Integration yield: ln ( C C i ) ( C C ) h h = HTU (14.29) where h i the packing height, C i i the inflow diolved ga concentration and C h i the outflow diolved ga concentration. Typical HTU value (air to water tranfer) HTU (mm) 25 mm and 38 mm phere mm tellerette mm tellerette mm Rachig ring

13 REFERENCES Adeney, W. E. and Becker, H. G. (1918; 1919, 1920) Sci. Proc. R. Dublin Soc., 15, 31 (1918); 15, 44 (1919); 16,13 (1920). ASCE (1993) A Standard for the Meaurement of Oxygen Tranfer in Clean Water, ANSI/ASCE 2-91, 345 Eat 47 th. St., New York. ASCE (1996) Standard Guideline for In-proce Oxygen Teting, ASCE-18-96, 345 Eat 47 th. St., New York. Barnhart, E. L. (1969) Proc. Am. Civ. Eng., San. Eng. Div., 95, SA3, 645 Caey, T. J. (1971) PhD Thei, Dept. Civ. Eng., Univerity College Dublin. Downing, A. L. (1960) J. Int. Public Health Engr., 59, 80. Downing, A. L., Bailey, R. W. and Boon, A. G. (1960) J. Int. Sew. Purif., Pt. 3, 231. Holroyd, A. (1952) Water Sanit. Eng., 3, 301 Karmo, O. T. (1972) M. Eng. Sc. Thei, Dept. Civ. Eng., Univerity College Dublin. Lewi, W. K. and Whitman, W. G. (1924) Ind. Eng. Chem., 16, Paveer, A. and Sweeri, S. (1965) J. Wat. Poll. Cont. Fed., 37, USEPA (1989) Deign Manual: Fine Pore Aeration Sytem, Centre for Environmental Reearch Information, Cincinnati. 158