WAT E R U T I L I T Y O P E R AT I O N S

Comparison of mineral acid pretreatments for sulfide removal Both sulfuric and carbonic acid improved sulfide removal by lowering pH, but only carbonic acid preserved alkalinity and proved less corrosive. Steven J. Duranceau, Robert K. Anderson, and Robert D. Teegarden

R

emoving sulfur species, particularly hydrogen sulfide (H2S) and sulfate (SO4), from a public water supply is an important aspect of potable water treatment. Sulfides are most often found in groundwater and at the bottom of water impoundments where anaerobic conditions prevail. Left untreated, sulfides affect finished water quality and corrosivity, create undesirable taste and odor, and To determine water quality effects on groundwater treatment for oxidize to form visible hydrogen sulfide, the authors compared sulfuric acid (H2SO4) and turbidity and color.1,2 carbonic acid for pH adjustment pretreatment in a pilot-scale (40This article presents gpm [2.5-L/s]) randomly oriented packed tower. Pretreatment the results of pilot-scale with either H2SO4 or carbon dioxide (CO2) to pH 6.0 resulted in air-stripping demonstra> 95 percent sulfide removal for tower feedwater sulfide tions using sulfuric acid concentrations of 2.5 mg/L. However, utilization of H2SO4 for pH (H 2 SO 4 ) and carbonic adjustment resulted in a loss of alkalinity in the finished water acid (H 2 CO 3 ) pretreatand an increase in sulfur (as sulfate), whereas CO2 pretreatment ment for sulfide removal preserved alkalinity in the finished water and did not increase sulfur (as sulfate). For executive summary, see page 167.

Copyright (C) 1999 American Water Works Association MAY 1999

S.J. DURANCEAU ET AL

85

In 1998, Orange County’s new Eastern Regional Water Supply Facility was placed in service. The design of the 20-mgd plant was based on information developed during the pilotand demonstration-scale studies.

α

adjustment of pH has been shown to be best accomplished using mineral acid addition. 8–11 H2SO4 is commonly used because of its effectiveness, availability, and relatively low cost. Conversely, a disadvantage of H2SO4 is that it is difficult to handle and presents safety challenges. Furthermore, when H2SO4 is used for air-stripping prefrom a groundwater supply. Sulfide was measured treatment, a portion of the groundwater’s natural at varying bed depths in the pilot air-stripping unit in alkalinity is destroyed in the chemical reaction, and order to compare sulfide mass transfer parameters additional treatment may be required to replenish using H2SO4 and H2CO3 pretreatment. Additional lost alkalinity in order to produce a nonaggressive, data were collected to evaluate secondary effects on well-buffered finished water. water quality using H2SO4 or H2CO3 pretreatment. Pilot-plant study. Orange County Utilities in Florida used H2SO4 pretreatment for air-stripping Mineral acids and pH adjustment H2S at the Econ Water Treatment Plant and received Previous researchers have described air-stripping regular red water complaints from customers. Addiof volatile organic contaminants.3–6 Less information tionally, the county found H2SO4 a difficult material is available about the removal of H2S. An excellent to handle. In deciding to explore H2CO3 as an alteroverview of the equilibrium and gas transfer of H2S native to H2SO4, it was reasoned that H2CO3 preis available.7 In water and wastewater treatment, treatment would be safer to handle and might provide the same level of sulfide removal as H2SO4 while maintaining the water’s natural alkalinity. FIGURE 1 Distribution of sulfur species over pH range 3–14 The authors did not find any previ– –2 ous documentation comparing H2S mass H2S HS S transfer parameters or finished drinking H2S HS 1.0 water quality effects for air-stripping 0.9 when different mineral acids are used for pretreatment. A pilot-plant study was 0.8 initiated to investigate and compare min0.7 eral acids for pretreatment effectiveness 0.6 in air-stripping processes. 0.5 The pilot-scale demonstrations were 0.4 conducted at the Econ Water Treatment 0.3 Plant. The facility consists of six forceddraft air-stripper towers with structured .02 –2 S media atop a 150,000-gal (57 X 104-L) 0.1 clearwell and supplies approximately 0 9.5 mgd (36 X 103 m3/d) average daily 2 4 6 8 10 12 14 flow to some 88,000 customers in pH Orange County’s Eastern Region Water System. Copyright (C) 1999 American Water Works Association 86

VOLUME 91, ISSUE 5

JOURNAL AWWA

H2S(gas) = H+ + HS–(aq)

FIGURE 2

Distribution of carbonate species over pH range 3–14 –

H2CO3* gas

1.0

–2

HCO3 aqueous

CO3

aqueous

0.6 0.8 0.7 0.6 α

Equilibrium chemistry in air-stripping processes. Some chemical reactions between a gas and water (termed hydrolysis) form soluble ionic species. These reactions are often reversible, i.e., the soluble (aqueous) species will revert to the gas form. In central Florida groundwater treatment, the gas–water equilibrium of H2S is often of primary concern. H2S dissociates in water according to the following equation:12–16

0.5 0.4 0.3

pK1 = 7

(1)

0.2 0.1

HS–

(aq)

=

H+

+

S2–

(aq)

pK2 = 14

0

(2)

2

K1 and K2 are the ionization constants for weak acids (hydrosulfuric) per Eq 1 and 2, respectively. The amount of each sulfide species in the water is pH-dependent. At a neutral pH of 7.0, the raw water contains 50 percent of the total sulfide as H2S gas and 50 percent as the bisulfide ion (HS–). Figure 1 shows distribution of sulfur species over a pH range of 2 to 14. The ordinate  in Figure 1 is the ratio of the molar

FIGURE 3

4

6

8 pH

10

12

14

concentration of a single species (H2S, HS–, or S2–) to the sum of the molar concentrations of the species (H2S + HS– + S2–).14 The H2S gas species is volatile and can be removed from solution to a great extent by mass transfer to the

Pilot-plant flow diagram Blower air to atmosphere Feed sample port

Feed nozzles

View port

Flange

9 ft (2.7m) Air-stripping packed tower 7 ft (2.1 m)

Sulfuric acid feed

Ball valve 5 ft (1.5 m) Packing material

Raw water sample port

3 ft (0.9 m) Plate

12 ft (3.6 m)

Carbon dioxide feed

2-in.- (5.1-cm-) diameter packs

11 ft (3.3 m) Sample port

Flowmeter 1 ft (0.3 m)

Pump

Globe valve

Totalizer meter

Clearwell

Sheave

Blower fan

Treated water to stormwater outfall Well

Treated water sample port

Copyright (C) 1999 American Water Works Association MAY 1999

S.J. DURANCEAU ET AL

87

Figure 2 is the ratio of the molar concentration of a single species (H2CO3, HCO3–, or CO32–) to the sum of the molar concentrations of the species (H2CO3 + HCO3– + CO32–).13 Mass transfer principles applied in project. Mass transfer theory development for the air-stripping process is not shown in detail in this article because it is addressed in the literature.3–7,17–20 Full-scale airstripping process design for volatile gases for most water treatment applications can be performed using a known or assumed overall liquid-phase mass transfer coefficient (MTC), often seen in the literature as Kla. Kla is the product of the liquid-phase MTC Kl

I

n water and wastewater treatment, adjustment of pH has been shown to be best accomplished using mineral acid addition.

The pilot equipment used for this research was located at Econ Water Treatment Plant in Orange County, Fla. Flow, pH, and sulfide were monitored daily in the field.

gas phase at a gas–water interface in accordance with Henry’s law.2,9 When sulfide is stripped from water, only that portion of the total sulfide that is H2S gas can be stripped. Sulfide-removal efficiency is improved by pH adjustment to convert the aqueous HS– to H2S gas to improve removal efficiency. As the pH is lowered to the range of 6.0–6.5, approximately 95–75 percent of the sulfide in the water is available for stripping. Carbon dioxide (CO2) is present in most groundwater. A volatile gas, CO2 exists in equilibrium with other carbonate species according to the following equations:12–16 H2CO3*(aq) = H+ + HCO3–

pK1 = 6.3

(3)

HCO3–(aq) = H+ + CO32–

pK2 = 10.3

(4)

In basic solutions, hydrated CO2 ionizes to give hydrated protons (H+), bicarbonate ion (HCO3–), and carbonate ion (CO32–). Figure 2 illustrates the effect of pH on the carbonate species. The ordinate  in

and the specific interfacial area a, which is the mass transfer interfacial area in the system volume. The Kla is also known as the volumetric MTC. The Kla is unique for each type and size of airstripping packing material and is sometimes available from packing manufacturers but is frequently determined by pilot studies or predicted using models such as the Onda correlations.19,20 The Kla is used to calculate the height of a transfer unit (HTU); the HTU multiplied by the number of transfer units (NTU) provides the required height of mass transfer packing needed to achieve a desired removal efficiency. For dilute, volatile gas air-stripping in which the contaminant’s liquid- and gas-phase mole fractions, along with the liquid and gas volumetric flow rates and Henry’s constant, are known, NTUs can be calculated from NTU = (R/(R–1)) ln[(Cin/Cout)*(1–1/R) + 1/R]

(5)

in which R is the stripping factor, Cin is the clearwell in, Cout is the clearwell out. R is calculated from R = H(Qgas/Qliquid)

(6)

in which H is Henry’s constant and Qgas and Qliquid are the gas and liquid molar flow rates.3 Howe and Lawler found the overall liquid-phase MTC Kla to be independent of pH for the H2S species and not for total sulfide.7 Howe and Lawler showed that optimal packed-tower air-stripping design can be achieved using conventional design parameters without considering chemical kinetics. They note, however, that pH dramatically affects the total sulfide removal that can be obtained.

Copyright (C) 1999 American Water Works Association 88

VOLUME 91, ISSUE 5

JOURNAL AWWA

FIGURE 4

pH versus packing depth at liquid flow rate of 30 gpm/sq ft (20.4 mm/s)

8.5

107 cfm/sq ft (32.6 m/min) 179 cfm/sq ft (54.6 m/min) Packing Depth—m 0 0.61 1.22 1.83 2.44 3.05 3.66

8.0 7.5 pH

The research described in this article was intended to show whether the type of pH pretreatment chemical used influences packed-tower air-stripping of H2S or the finished water quality that exits the packed tower. The overall liquidphase MTC Kla was calculated using data obtained from the pilot studies using H 2 SO 4 pretreatment and H 2 CO 3 pretreatment.

7.0

Materials and methods

6.5 6.0 5.5 0

2

4

6

8

10

12

Packing Depth—ft

5

Total sulfide versus packing depth CO2, pH 6.3 CO2, pH 6.0 H2SO4, pH 6.3 H2SO4, pH 6.0 Packing Depth—m 0.00 2.50

1.52

3.05

4.57

30 gpm/sq ft – 179 cfm/sq ft (20 mm/s –54.6 m/min)

2.00 Total Sulfide—mg/L

Pilot-plant description. A pilot-scale facility located at the Econ water treatment facility was designed and constructed for the study. Figure 3 shows a schematic of the pilot-scale air-stripping packed tower. The tower consists of two 8-ft- (2.4-m-) long sections of 16-in.(41-cm-) diameter polyvinyl chloride FIGURE (PVC) pipe flanged together, positioned vertically and mounted on a concrete slab. The unit houses 12 ft (3.7 m) of random loosed dumped packing, a mist eliminator in the top, and a clearwell in the bottom. Six vertically aligned sample ports allow liquid samples to be withdrawn at specific intervals along the 12ft (3.7-m) depth of the packing. Raw water is taken directly from wells located on site and is sent through a 3in.- (7.6-cm-) diameter PVC pipe where the pH is adjusted before the water enters the tower. A gate valve, ball valve, flow totalizer, and rotameter are located in the 3-in. (7.6-cm) piping to allow for adjustment of flow and monitoring of water use. Two chemical injection ports for addition of pH-adjustment chemicals (CO 2 and H 2 SO 4 ) are located downstream of the rotameter. A raw water sample port is located on the 3-in. (7.6cm) pipe just before the chemical injection points, and a feedwater sample port is located between the chemical injection points and the spray nozzles for monitoring feedwater pH. The water flows through the pipe, enters the top of the tower, and is sprayed over the top of the packing through nozzles. The water then flows through the 12-ft (3.7-m) depth of random loosed dumped packing by gravity to a clearwell located in the bottom of the tower. In this case, the packing consists of 2-in.- (5.1-cm-) diameter packs.* Water exits the clearwell through a 3-in.- (7.6-cm-) diameter pipe, continues to a stand pipe at which finished water samples are taken, and ultimately is discharged to the treatment plant’s stormwater runoff and finished water storage tank overflow system. Air is introduced into the bottom of the tower by a centrifugal fan. The air flows upward through the tower packing countercurrent to the water. Air flow

1.50

1.00

0.50

0.00 0.00

5.00

10.00

15.00

Packing Depth—ft

is regulated by adjustment of a sheave located on the fan discharge duct. The fan was calibrated at seven sheave settings by converting velocity (measured using a pitot tube in a temporary pipe connected to the fan’s intake) to flow. During the demonstrations, CO2 and H2SO4 were used to adjust the pH of the raw water. CO2 was fed as a gas and was stored in 150-lb (68-kg) gas containers. CO2 feed pressure was kept constant at 80 psig (652 kPa) by a pressure regulator located on the storage container discharge piping. A 40-scf air/h (1.13 m3 air/h) flow control rotameter† was used to monitor and regulate the flow of CO2 gas into the raw water. H2SO4 was stored in a 55-gal (208-L) plastic barrel and was pumped into the raw water feed pipe by an *Jaeger Tri-Packs, Jaeger Products Inc., Spring, Texas †Wallace & Tiernan, Belleville, Ill.

Copyright (C) 1999 American Water Works Association MAY 1999

S.J. DURANCEAU ET AL

89

FIGURE 6

Sulfide removal efficiency versus water loading, with and without pH adjustment

Sulfide Removal Efficiency—percent

pH 6.3, no forced air pH 7.8, no forced air pH 6.3, 107 cfm/sq ft (32.6 m/min) pH 7.8, 107 cfm/sq ft (32.6 m/min) pH 6.3, 179 cfm/sq ft (54.6 m/min) pH 7.8, 179 cfm/sq ft (54.6 m/min) Water Loading—gpm/sq m 10 14 17 20 24 27 31 100 90 80 70 60 50 40 30 20 10 0 15

20

25

30

35

40

45

Water Loading—gpm/sq ft

FIGURE 7

Sulfide Removal Efficiency—percent

100

Sulfide removal efficiency versus stripping factor No pH adjustment CO2, pH 6.3 CO2, pH 6.0

H2SO4, pH 6.3 H2SO4, pH 6.0

90

sured in the field.‡ Sulfide samples were analyzed immediately after collection to enhance accuracy by limiting the loss of H2S gas to the atmosphere. DO was measured in the field using the iodometric method with an azide modification.21 Both laboratory and field chemical analysis were performed in support of the pilot demonstrations. Field testing included pH, total sulfide measurements, DO and corrosivity of the water stream, and specific CO2 and H2S measurements of the pilot exhaust air stream. Samples collected in the field were packed in ice and submitted to the Orange County Utilities Laboratory for laboratory measurement of TDS, alkalinity, sulfate, calcium hardness, and total hardness. Pilot-plant operation. During the testing period, feedwater pH was adjusted from approximately 7.9 to either 6.0, 6.3, or 6.5 using H2SO4 or CO2 gas. Liquid loading rates of 20, 25, 30, 35, and 40 gpm/sq ft (13.6, 17.0, 20.4, 24.0, and 27.2 mm/s) and air loading rates of 107 and 179 cfm/sq ft (32.6 and 54.6 m/min) (150 and 250 cfm [4.25 and 7.08 m3/min)] were varied at adjusted pH values of 6.0, 6.3, and 6.5. The performance of 2-in.- (5.1- cm-) diameter packs were evaluated for each pH and loading rate.

Results and discussion

Evaluating sulfide removal. Packedtower air-stripping of a nondissociative 70 gas from water is a first-order reaction 60 and exhibits nonlinear behavior because of the reduction in driving force in the 50 mass transfer zone as the gas is stripped.18 40 A second order is added to the reaction No feedwater pH adjustment 30 when the effect of pH on acid-base dissociation is included.7 Air-stripping H2S 20 from groundwater is influenced not only 10 by the effect of pH on the sulfide species dissociation but also by the carbonate 0 0 5 10 15 20 25 30 species dissociation because CO2 is also Stripping Factor—R stripped in the tower. Figure 1 shows the influence of pH on sulfide species dissociation to be relatively linear in the pH range of concern (i.e., 6–8). Figure 2 electronic metering pump.* The H2SO4 feed rate was shows the influence of pH on carbonate species discontrolled by monitoring the pH of the feedwater. sociation to be relatively linear from pH 6 to 7 and Chemical methods. Water quality parameters nonlinear from pH 7 to 8. When the rate expression derived in Howe and evaluated for mass transfer comparisons included total sulfides, alkalinity, hardness, and pH. Alkalinity, Lawler7 is used, the rate of sulfide removal can be hardness, calcium–magnesium hardness, turbidity, shown to be a reaction that is first order in driving CO2, dissolved oxygen (DO), corrosivity, and total force and first order in . The nonlinear behavior of dissolved solids (TDS) were measured to provide *Pulsafeeder Pulsatron Series C, Rochester, N.Y. insight on secondary water quality effects. †Model 608 Digital pH Meter, Jenco, San Diego, Calif. Water pH was measured in the field using a digital ‡LaMotte Pomeroy Method Sulfide Test Kit, LaMotte Co., Chestertown, Md. pH meter.† Total sulfide concentration was also mea80

Mineral acid feedwater pH adjustment

Copyright (C) 1999 American Water Works Association 90

VOLUME 91, ISSUE 5

JOURNAL AWWA

Average K1a—1/min

Average K1a—1/min

sulfide removal through a packed tower FIGURE 8 Average K1a versus liquid loading for both pretreatment should be equally influenced by mass acids transfer driving force and pH change. Spacing sample ports at 2-ft (0.61-m) CO2 H2SO4 intervals along the packing depth proLiquid Loading—mm/s 6 10 14 17 20 24 27 31 vided the means of showing mass trans1.4 fer rates and water chemistry changes at those intervals through the tower. Fig1.2 ure 4 shows pH versus packing depth for two air loading rates at one water flow 1.0 condition at a feedwater pH of 6.0 and illustrates the increase in pH as water 0.8 moves through the pilot tower. Feedwater pH started at 6.0 or 6.3 pH units 0.6 entering the tower top and increased to 7.9 and 8.0, respectively, upon exiting 0.4 the tower. Figure 4 shows that the change in pH 0.2 is greatest in the first 12 in. (0.3 m) of packing and gradually decreases until the 0 10 15 20 25 30 35 40 45 pH reaches approximately 7. Near pH 7, Liquid Loading—gpm/sq ft the rate of pH change appears to increase slightly. In the bottom half of packing at pH values above 7, the rate of pH change FIGURE 9 Average K1a versus liquid loading for feedwater pH 6.3 slowly decreases and becomes nearly linand 6.0 ear. The rate of pH change follows closely with the rate of change in CO 2 gas pH 6.3 pH 6.0 (H 2 CO 3gas ) available for stripping obLiquid Loading—mm/s served in Figure 2 over the pH range 6–8. 6 10 14 17 20 24 27 31 1.4 Figure 5 shows representative raw data as total sulfide concentration as a 1.2 function of tower packing depth. Higher 1.0 sulfide removals are observed at lower feedwater pH values. As the water flows 0.8 down through the packing, CO2 removal results in an increase in water pH. There0.6 fore, the amount of sulfide available for 0.4 stripping is reduced as the water flows through the tower. 0.2 For example, at the top of the tower for a feedwater pH of 6.0, more than 90 0 10 15 20 25 30 35 40 45 percent of the total sulfide is available Liquid Loading—gpm/sq ft for stripping, whereas halfway through the tower at pH of 7.2, approximately 50 percent of the total sulfide is available for stripping. CO2 removal that occurs through the tower results in an increase in pH removal because both methods reduce the pH to a and thereby reduces H2S available for stripping. The value at which 95–100 percent of the total sulfide content is initially available for stripping, and there less H2S available for stripping, the less driving force are no significant differences in sulfide removal is available for mass transfer at the liquid–gas interface and the slower mass transfer occurs. The com- through the tower. Figure 6 depicts sulfide-removal efficiency as a bination of pH change resulting from CO2 stripping function of water loading for various air loading rates and decrease in H2S driving force resulting from stripping of hydrogen explains the nonlinear behav- under conditions with and without pH adjustment. Air loading rates include 72 cfm/sq ft, 179 cfm/sq ft, ior shown in Figure 5. Approximately 50 percent of H2S removal occurs in the first 12 in. (0.3 m) of and zero flow. Zero-flow conditions would resemble induced velocity draft aeration in the stripping packed packing depth. Figure 5 shows that sulfide removal using CO2 tower. In general, as air loading increases for a given feedwater pH, the efficiency increases for the condiis essentially the same as it is using H2SO4 for pH adjustment. Using either CO2 or H2SO4 should pro- tions tested. The greatest removal efficiency occurred vide an efficient pretreatment for total sulfide at the highest air loading rate at the lowest hydraulic Copyright (C) 1999 American Water Works Association MAY 1999

S.J. DURANCEAU ET AL

91

Total Sulfide—mg/L

CO2—mg/L

loading rate. When pH adFIGURE 10 Liquid-phase CO2 and total sulfide concentrations justment was used, the inversus packing depth crease in efficiency was significant. Efficiencies can be Packing Depth—m 0 0.61 1.22 1.83 2.44 3.05 3.66 improved from < 40 percent 140 2.50 to  85 percent by suppressFeedwater pH adjusted to 6.3 with CO2 ing pH to 6.3. 120 2.00 The increase in efficiency 100 attributable to air loading appeared to be even more sig1.50 80 nificant. Sulfide removal for Measured CO2 60 feedwater of pH 6.3 using no 1.00 forced air was < 40 percent, Total sulfide 40 whereas the pH 7.8, 179 0.50 cfm/sq ft (54.6 m/min) air 20 Calculated CO2 loading data fell between 60 0 0.00 and 70 percent. The pH 7.8, 0 2 4 6 8 10 12 179-cfm/sq ft (54.6 m/min) Packing Depth—ft air loading data were interesting because at pH 7.8 only 17 percent of the sulfide was initially available to strip, yet TABLE 1 Average water quality comparison removal efficiencies between H2SO4 CO2 60 and 70 percent were obRaw tained. With no forced air at Parameter Water Feed Finished Feed Finished pH 7.8, only 10–25 percent pH 7.8 6.0 7.5 6.0 7.9 removal efficiencies were Sulfide 2.3 2.3 0.2 2.3 0.1 obtained, which falls more in Sulfate 10 90 90 10 10 line with the 17 percent avail120 50 50 120 120 Alkalinity—mg/L as CaCO3 106 106 105 106 105 Hardness—mg/L as CaCO3 able sulfide. Total dissolved solids 130 137 137 130 135 One possible explanation for this is that the sulfide was being oxidized in the tower. Howe and Lawler7 attempted to account for oxidation in one of the computer mod- ciple to proceed. It is also possible that a combinaeling programs they developed to predict sulfide tion of oxidation and equilibrium shift accounted for removal in an existing facility at which sulfide oxi- the higher-than-expected removals at the higher air dation had been observed. Another possible expla- loading rate. nation is that as available sulfide was stripped in the A common air-stripping parameter used for comupper portion of the tower, the sulfide species equi- parison of removal efficiencies and for design is the librium shifted in accordance with Le Chatelier’s prin- stripping factor, R, determined using Eq 6. In Figure ciple and made more sulfide available to strip in lower 7, sulfide removal efficiency is shown as a function portions of the tower. At pH 7.8, only 3 percent of the of the stripping factor R for H2CO3, H2SO4, and no acid pretreatment. The data show that it was difficult to obtain sulfide removal efficiencies > 70 percent witheft untreated, sulfides affect finished out pH pretreatment. The data show that to achieve water quality and corrosivity, create sulfide removal > 95 perundesirable taste and odor, and oxidize cent for a feedwater with a pH of 6.0, stripping factors to form visible turbidity and color. > 12 and 15 were required for H 2 CO 3 and H 2 SO 4 , carbonate species was available to strip as CO2 gas, respectively. For a feedwater pH of 6.3, stripping facand CO2 mass transfer appeared to be limited because tors > 15 and 20 were required to achieve removal little pH change was observed through the tower. efficiencies > 95 percent for H 2 CO 3 and H 2 SO 4 , With little CO2 mass transfer occurring in the mass respectively. transfer zone, perhaps the increased air loading carDetermining mass transfer information. The ried the H2S gas at the liquid–gas interface away from overall liquid–phase MTC Kla has been reported to be the interface fast enough to allow Le Chatelier’s prinindependent of pH for the H2S species and not for total

L

Copyright (C) 1999 American Water Works Association 92

VOLUME 91, ISSUE 5

JOURNAL AWWA

sulfide.7 The mass transfer TABLE 2 Representative corrosion indexes calculations rate expression r T = gKla(CTs–CT), in which CT is Sample Raw CO2 Feed H2SO4 Feed the total concentration of Location Water Finished Water Finished Water solute in solution and CTs is Sample characteristics the concentration of solute in 25 25 25 Temperature—oC Total dissolved solids—mg/L 132 137 135 the gas phase, shows the rate 77.6 78.9 78 Calcium—mg/L as CaCO3 of sulfide removal was to be 27.2 27.2 27 Magnesium—mg/L as CaCO3 first order in driving force 104.8 106.1 105 Hardness—mg/L as CaCO3 134 138 40 Alkalinity—mg/L as CaCO3 (CTs–CT) and first order in g.7 pH 7.9 7.9 7.9 The nonlinear behavior of Water quality parameters sulfide removal through the Bicarbonate ion 132.76 136.72 39.6 Carbonate 1.19 1.23 0.36 tower (shown in Figure 5) is Carbon dioxide 3.52 3.62 1.05 the result of both changing Langelier saturation index 0.17 0.19 –0.36 pH of stabilization 7.73 7.71 8.26 driving force and changing Calcium carbonate precipitation 2.92 3.35 –1.64 pH through the tower. potential In Figure 8, average valAggressiveness index 10.23 10.24 9.97 Equilibrium conditions ues of Kla are plotted versus pH 7.76 7.74 8.24 liquid loading for CO2 and 74.68 75.55 79.64 Calcium—mg/L as CaCO3 H2SO4 pretreatment condi27.2 27.2 27.0 Magnesium—mg/L as CaCO3 131.08 134.65 41.64 Alkalinity—mg/L as CaCO3 tions. Average values of Kla were similar for both pretreatment mineral acids for liquid loadings of 20, 25, and alkalinityT 30 gpm/sq ft (13.6, 17.0, and 20.4 mm/s) with CO2 [H+] +  = [HCO3–] + 2[CO32–] + [OH–] (7) 50,000 Kla values slightly higher than those of H2SO4. In Figure 9, average values of Kla are plotted verFrom Eq 4, an equilibrium equation can be sus liquid loading for experiments run at feedwater pH derived for the second ionization of H2CO3: 6.3 and 6.0. The data show higher average Kla values [H+] [CO32–]/[HCO3–] = K2

P

retreatment with CO2 for pH adjustment resulted in no loss of alkalinity through the packed tower.

in the experiments run at feedwater pH 6.0 than at pH 6.3, confirming that total sulfide mass transfer increases with decreasing feedwater pH. CO 2 profiles. To ascertain efficiency of the removal of elevated CO2 concentrations in the water following H2CO3 pretreatment, water phase CO2 concentrations were measured at each sample port through the packed-tower pilot plant. Because of the limitations of the test kit’s sensitivity, concentrations of CO2 < 20 mg/L were calculated using pH and alkalinity values at each sample port. CO2 concentrations were measured through the tower at those levels that could be detected by the test kit. Calculated CO2 concentrations can be obtained from known pH and total alkalinity measurements. If total alkalinity is a measure of the equivalent concentration of cations associated with the alkalinityproducing anions (to maintain electroneutrality), then

(8)

K2 is the second ionization constant for the carbonate system. Given a measured pH value and thus [H+], the following equilibrium equation for water can be used to determine [OH–]: Kw [OH–] =  [H+]

(9)

in which KW is the ionization constant of water at 25oC, which is 10–14. Simultaneous solution of Eqs 7 and 8 yields HCO3 alkalinity (Eq 10) and carbonate alkalinity (Eq 11), both expressed as mg/L as calcium carbonate (CaCO3), at 25oC:13 HCO3 alkalinity = 50,000 [(alkalinityT /50,000) + [H+] – (10–14/[H+])]  (10) 1 + (2K2/[H+])

Carbonate alkalinity = 50,000 [(alkalinityT /50,000) + [H+] – (10–14/[H+])]  (11) 1+ ([H+]/2K2)

Assuming that no hydroxide alkalinity [OH–] was present within the test conditions, the actual CO2 concentrations present can be calculated using Eq 11 and converting from mg/L as CaCO3 to mg/L as CO2.

Copyright (C) 1999 American Water Works Association MAY 1999

S.J. DURANCEAU ET AL

93

Based on information developed from the pilot-plant study, Orange County implemented a 12-mgd demonstration project for carbonic acid pretreatment at the Econ Water Treatment Plant.

Dissolved Oxygen—mg/L

Figure 10 shows measured CO2 concentration as a function of packing depth for an adjusted feedwater pH of 6.3. The figure also depicts the calculated concentrations of CO2 for comparison. Figure 10 shows a high initial CO2 concentration in the feedwater stream following CO2 injection. As the water falls through the packing, CO2 is stripped and eventually dispersed into the atmosphere through the top of the tower. Measured CO 2 values showed less removal in the first 3 ft (0.9 m) of packing than was shown by the calculated CO2 values. Total sulfide removal data for the same feedwater conditions were plotted using a secondary axis for comparison. Total sulfide data had a slope similar to the calculated CO2 data at the top of the tower, indicating that the rate of change of removal was similar under the given conditions. DO profiles. Figure 11 shows DO concentration as a function of packing depth for air-to-water ratios of 27 and 45 for an adjusted pH of 6.5. This data verified that the air-to-water ratio has minimal effect on the amount of oxygen absorbed into the water over the FIGURE conditions tested. The rate of oxygen absorption into the water was rapid at the top of the tower and decreased near the bottom of the tower as the water became more oxygen-saturated. This nonlinear behavior is similar to the stripping of H2S and CO2 in that the rate of mass transfer was greatest at the top of the tower and decreased toward the bottom of the tower. Alkalinity and choice of mineral acid. The average water quality parameters measured during this study are shown in Table 1 for both H2SO4 and H2CO3 pretreatment. Water quality values shown under the “Feed” columns were determined from water samples following acid addition and prior to the water entering the pilot tower; values shown under

the “Finished” columns were determined from samples taken from pilot-plant discharge. As expected, the H2SO4-finished water had a higher sulfate concentration and < 50 percent of the alkalinity of the H2CO3 finished water. Screening corrosion effects. To gain insight into what was occurring inside the tower packing with regard to corrosion potential, the authors evaluated general corrosion potential using water quality indexes and linear polarization methods. The feedwater after CO2 injection, the water at each sample port through the entire tower height, and the finished water for the pilot tower were evaluated relative to corrosivity. One of the more common water quality indexes that indicates water could be corrosive is the Langelier Saturation Index (LSI). The LSI is a comparison of the measured pH of the water and a calculated pH of calcium saturation based on equilibrium chemistry. The pH of calcium saturation is often referred to as pHs and is the pH at which CaCO3 neither dissolves nor precipitates in solution. Although the LSI has historically proven useful, it presents limitations when it is used to assess copper corrosion control rates in distribution systems.22,23 Other representative corrosion indexes include CaCO3 precipitation potential and aggressiveness. Table 2 shows the representative corrosion indexes for the raw water, H2SO4-pretreated finished water, and H2CO3-pretreated finished water. A computer pro-

11

7.0

Dissolved oxygen versus packing depth for air-to-water ratios of 27 and 45 Air-to-water ratio—27 Air-to-water ratio—45 Packing Depth—m 0 0.61 1.22 1.83 2.44 3.05

3.66

pH 6.5

6.0 5.0 4.0 3.0 2.0 1.0 0.0 0

2

4

6 8 Packing Depth—ft

10

12

Copyright (C) 1999 American Water Works Association 94

VOLUME 91, ISSUE 5

JOURNAL AWWA

General Corrosion—mils/yr

gram* was used to process water quality FIGURE 12 Linear polarization probe corrosion evaluation for data. Values shown at the bottom of Table H2CO3-pretreated water 2 under “Equilibrium Conditions” were calculated by the program and represent stable equilibrium conditions. Copper Mild steel General corrosion potential can be 8.0 measured using an electrochemical 7.0 method or a linear polarization method. 6.0 Electrochemical methods have the ad5.0 vantage of providing instantaneous mea4.0 3.0 sure of the corrosion rate. These meth2.0 ods do not provide a cumulative 1.0 measure, nor do they yield a long-term 0.0 corrosion rate. Electrochemical methRaw Feed 11 9 7 5 3 1 Clearwell ods are relatively easy to perform and Sample Location are cost-effective. Figure 12 shows average general corAir-to-water ratio—27, pH—6.5, sulfide—2.3 mg/L rosion rate data for copper and mild steel as a function of packing depth. The air-to-water ratio was maintained at 27, the adjusted feedwater pH was 6.5, and • Pretreating with either H2SO4 or CO2 to pH 6.0 the sulfide loading was 2.3 mg/L during this evalresults in > 95 percent efficiency in the removal of uation. H2CO3 was used to provide pH adjustment. For the raw water, the general corrosion rate was H 2 S from groundwater for water loading rates 5.0 mils/yr for mild steel and 6.5 mils/yr for copper. between 20 and 35 gpm/sq ft (13.6 and 24.0 mm/s). Following CO2 injection, the general corrosion rate of Sulfide concentrations were reduced from an averthe feedwater was 7.3 mils/yr for mild steel and 4.9 age of 2.5 mg/L in the raw water to < 0.10 mg/L in mils/yr for copper. The general corrosion rate for mild the pilot discharge effluent for all hydraulic flows steel and copper dropped to 0.5 mils/yr and 0.9 evaluated, when the highest air flow setting (179 mils/yr, respectively, at the bottom of the tower. This cfm/sq ft [54.6 m/min]) was used. In addition, sulmeasured decrease in corrosion potential corrosivity fide removal efficiency increased from < 70 percent with no pH adjustment to > 95 percent when the pH was adjusted to 6.0. • Pretreatment with CO2 for pH adjustment resulted orrosivity indexes indicate that in no loss of alkalinity H2SO4-pretreated finished water is more through the packed tower; corrosive than H2CO3-pretreated finished accordingly, the buffering capacity of the water was water following packed-tower maintained during the aeration process and no loss of air-stripping. HCO3 by conversion to CO2 was realized. of the water indicates the benefit of having CO2 and • H2SO4, when used for pH adjustment, resulted H2S stripped from the water. in a loss of alkalinity. HCO3– was converted to CO2, The measured and calculated corrosion indexes which was lost by the aeration process. In addition, indicate that the H2SO4-pretreated finished water is as expected, sulfate concentrations increased from more corrosive than the H2CO3-pretreated finished 10 mg/L in the raw water to > 70 mg/L following water. The data and calculations indicate that the H2SO4 addition. H2SO4-pretreated finished water will need further • Changing from H2SO4 to H2CO3 pretreatment treatment in order to stabilize it and provide buffer- resulted in no significant differences in total suling capacity to equal the H2CO3-pretreated finished fide mass transfer in the pilot air-stripping packed water. Conversely, use of H2CO3 as a pretreatment tower. chemical significantly retains the finished water • Increasing air loading rate had a dramatic effect buffering capacity in a packed-tower air-stripping on total sulfide removal. Over the range of liquid process. loadings tested (20–40 gpm/sq ft [13.6–27.2 mm/s]), average removal efficiencies increased 46 percent Conclusions for no feedwater pH adjustment and 64 percent for Results of the packed-tower pilot study led to sev*Hydranautics, ROCHEM, Oceanside, Calif. eral conclusions.

C

Copyright (C) 1999 American Water Works Association MAY 1999

S.J. DURANCEAU ET AL

95

feedwater pH adjusted to 6.3 for the condition of no air loading and air loading of 179 cfm/sq ft (54.6 m/min), respectively. • Increasing the air-to-water ratio resulted in a slight increase in the overall liquid-phase sulfide MTC K l a. A more predominant increase in sulfide K l a occurred when pH was adjusted from 6.3 to 6.0. • For raw water, the general corrosion rate was 5.0 mils/yr for copper and 6.5 mils/yr for mild steel. However, the general corrosion rates measured in the finished water were reduced to 0.5 mils/yr for copper and 0.9 mils/yr for mild steel after CO2 and H2S were removed from the raw water supply. Corrosivity indexes indicate that H2SO4-pretreated finished water is more corrosive than H 2 CO 3 -pretreated finished water following packed-tower air-stripping.

Acknowledgment The authors thank Orange County Utilities for project funding and support and the field crews, supervisors, and managers who participated and assisted in the project, especially Robert Dehler, Pat DiVecchio, Harry Schumacher, Kim Kunihiro, and Jackie Torbert. The authors also thank Wayne Mather and Jackie Foster of Boyle Engineering Corporation for their advice and assistance, Paul Jacobs and John Verscharen of The Jacobs Group for their efforts, and Chuck Mattson, Stephanie Dobbs, and Tamara Richardson for their input and assistance. Portions of this work were presented at the 1994 Water Quality Technology Conference in San Francisco, Calif.

References 1. LYN, T.L. & TAYLOR, J.S. Assessing Sulfur Turbidity Formation Following Chlorination of Hydrogen Sulfide in Groundwater. Jour. AWWA, 84:9:103 (Sept. 1992). 2. WELLS, S.W. Hydrogen Sulfide Problems of Small Water Systems. Jour. AWWA, 46:2:160 (Feb. 1954). 3. KAVANAUGH, M.C. & TRUSSELL, R.R. Design of Aeration Towers to Strip Volatile Contaminants From Drinking Water. Jour. AWWA, 72:684 (1980). 4. AMY, G.L. & COOPER, W.J. Air-stripping of Volatile Organic Compounds Using Structured Media. Jour. Envir. Engrg., 112:4:56 (1986). 5. LAMARCHE, P. & DRISTE, R.L. Air-stripping Mass Transfer Correlations for Volatile Organics. Jour. AWWA, 81:1:78 (Jan. 1989). 6. FREIBURGER, E.J.; JACOBS, T.L.; & BALL, W.P. Probabilistic Evaluation of Packed-tower Aeration Designs for VOC Removal. Jour. AWWA, 85:10:73 (Oct. 1993). 7. HOWE, K.J. & LAWLER, D.F. Acid–Base Reactions in Gas Transfer: A Mathematical Approach. Jour. AWWA, 81:1:61 (Jan. 1989). 8. ROE, F.C. Aeration of Water by Air Diffusion. Jour. AWWA, 27:7:897 (July 1935).

9. POWELL, S.T.& VON LOSSBERG, L.G. Removal of Hydrogen Sulfide From Well Water. Jour. AWWA, 40:12:1277 (Dec. 1948). 10. MATSON, M.D. Mill Effluent Breathes Easier With Carbon Dioxide. Water Engrg. & Mgmt., 135:38 (1988). 11. FLENTJE, M.E. Aeration. Jour. AWWA, 29:6:872 (June 1937). 12. RUBIN, A.J. Chemistry of Water Supply Treatment and Distribution. Ann Arbor Science Publ., Inc., Ann Arbor, Mich. (1974). 13. SAWYER, C.N. & MCCARTY, P.L. Chemistry for Environmental Engineering. McGraw–Hill, New York (1978). 14. SNOEYINK, V.L. & JENKINS, D. Water Chemistry. John Wiley & Sons, New York (1980). 15. STUMM, W. & MORGAN, J.J. Aquatic Chemistry. John Wiley & Sons, New York (1981). 16. WEBER, W.J. JR. Physiochemical Processes for Water Quality Control. John Wiley & Sons, New York (1972). 17. TREYBAL, R.E. Mass Transfer Operations. McGraw– Hill, New York (1980). 18. MCCABE, W.L. & SMITH, J.C. Unit Operations of Chemical Engineering. McGraw–Hill, New York (1976). 19. ONDA, K.; SADA, E.; & MURASE, Y. Liquid Side Mass Transfer Coefficients in Packed Towers. Jour. AIChE, 5:2:235 (1959). 20. ONDA, K.; TAKEUCHI, H.; & OKUMOTO, Y. Mass Transfer Coefficients Between Gas and Liquid Phases in Packed Columns. Jour. Chem. Engrg. Japan, 1:1:56 (1968). 21. Standard Methods for the Examination of Water and Wastewater. APHA, AWWA, and WEF, Washington (18th ed., 1992). 22. POWELL R.M.; YOUSET, Y.A.; & MULFORD, L.A. Evaluation and Implementation of Corrosion Control by a Public Utility. Proc. 1992 AWWA Ann. Conf., Vancouver, B.C. 23. EDWARDS, M.; SCHOCK, M.R.; & MEYER, T.E. Alkalinity, pH, and Copper Corrosion By-product Release. Jour. AWWA, 88:3:81 (Mar. 1996). About the authors: Steven J. Duranceau is a senior engineer at Boyle Engineering Corp., 320 E. South St., Orlando, FL 32801. He is a graduate of Florida State University in Tallahassee with a BS in chemistry and of the University of Central Florida in Orlando with an MS in industrial chemistry and a PhD in environmental engineering. A founding member of the Southeast Desalting Association, Duranceau is past chair of AWWA’s Membrane Processes Committee and the 1995 and 1997 AWWA membrane technologies conferences. Robert K. Anderson is an associate engineer with Boyle Engineering Corp. Robert D. Teegarden is an assistant managing engineer in utilities engineering at Orange County Utilities, 8100 Presidents Dr., Orlando, FL 32801.

Copyright (C) 1999 American Water Works Association 96

VOLUME 91, ISSUE 5

JOURNAL AWWA