Air Stripping of Ammonia and Methanol in a Bubble-Cap Column

D.H. Jenkins,a D.A. McCallum,a R. Ruzbacky,b S. Saunders,c and A.K. Brenta a CSIRO Minerals, Victoria 3169, Australia; [email protected] (for correspondence) b Human Resources, Swinburne University, Victoria 3122, Australia c Victoria 3187, Australia
Published online 13 November 2007 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ep.10235

A twenty-tray single bubble-cap column was used to air strip ammonia and methanol from a synthetic, as well as a plant wastewater stream. When feeding synthetic solutions containing $700 mg/kg ammonia, stripping efciencies of more than 98% were achieved at gas to liquid mass ratios (G/L) ratios of 2.52.7. This produced stripper bottom solutions contains $10 mg/kg ammonia. Reducing the air ow by 25% increased the ammonia content in the stripper bottoms to $30 mg/kg. Using the graphical McCabeTheile method, the overall column efciency for the bubble cap column during ammonia stripping was between 9 and 26% because of the low contact efciency of this column. When treating the as received plant wastewater, the ammonia removal efciency was only 84% at the lower G/L ratio. This increased to 99% by raising the pH from 9.8 to 11.5 and increasing the G/L mass ratio. The ammonia volumetric mass transfer coefcient (KLa) for these tests was calculated to be between 0.38 and 0.58 h21. Both stripping and absorption of methanol occurred in the column during stripping of ammonia. Based upon solution analyses, the methanol removal efciency varied between 55 and 95%. 2007 American
Institute of Chemical Engineers Environ Prog, 26: 365374, 2007

INTRODUCTION

Keywords: ammonia stripping, bubble-cap column, methanol stripping, mass transfer coefcient

2007 American Institute of Chemical Engineers

The Australian Magnesium Corporation (AMC) operated a 1500 tonnes per annum Magnesium Demonstration Plant at Gladstone, Queensland between 1998 and 2002, to evaluate the various unit operations of the AM process [1, 2]. Distillation of magnesium chloride leach liquor and glycol was an important unit operation in this process. Distillation produced a 15% magnesium chlorideglycol solution containing very little water as a bottom product. This bottom product was the feed for the production of anhydrous magnesium chloride required in the electrolytic cell for the production of magnesium metal. A schematic diagram of the distillation section is shown in Figure 1. The distillate from the No.1 dehydration column contained mainly water but also contained ammonia, methanol, glycol, and 2-chloroethanol (2-CE). The distillate was treated using a biological nutrient removal activated sludge process and achieved the required efuent total nitrogen standard of 5 mg/L. During the latter stages of the Demonstration Plant operation, AMC was completing the engineering design for the Stanwell Magnesium Project. The proposed Stanwell plant had a production capacity of 90,000 tonnes per annum and was expected to be the largest magnesium metal producer in the world. Initially, the Stanwell Magnesium Project proposed to treat the dehydration column distillate wastewater in a similar manner to that used at the Gladstone Demonstration Plant as this would allow the water to be discharged to a nearby watercourse. However, after engineering design, the capital cost for this treatment process appeared expensive, so an alternative treatment option was proposed using air stripping of
December 2007 365

Environmental Progress (Vol.26, No.4) DOI 10.1002/ep

Figure 1. Schematic diagram of AMC demonstration plant distillation.

ammonia. It was assumed that some methanol stripping would occur but there would be minimal stripping of 2-CE and ethylene glycol. Gas stripping for the removal of dissolved gases and volatile organic compounds (VOC) is an accepted method for the removal of these materials from waste water streams. Pilot scale air stripping of ammonia from waste water was conducted at Lake Tahoe, California as early as 1966 [3]. Generally, packed towers operating in a counter-current mode have been used for the removal of VOCs because of their cost effectiveness when high removal efciencies are required [4]. The suitability of air stripping for treating the dehydration distillate wastewater which contained methanol, glycol, and 2-CE as well as ammonia would largely depend on the management of air emissions. Emission standards for these components were required and appropriate dispersion provided to achieve these standards. An advantage of air stripping, unlike biological processes, is that stripping is not subject to process upsets and can be started and stopped relatively easily. AMC possessed little experimental data to conrm the accuracy of their Pro IITM air stripping simulations for the No 1 dehydration column distillate wastewater when the ammonia content was less than 1000 mg/ kg. Their simulated results obtained from models utilizing the Pro IITM database and various thermodynamic methods were extremely different compared with the results obtained using the AMC BIPs and the NRTL thermodynamic method. To conrm the feasibility of air stripping, tests were completed on both synthetic and actual dehydration column distillate wastewater using a 20-tray bubblecap column. This column was not used to provide design data for air stripping of this solution but would conrm the feasibility or otherwise of air stripping as a treatment option for treating this distillate. The design
366 December 2007

values for the stripper efuent quality for ammonia and methanol were 60 and 386 mg/kg, respectively.
EXPERIMENTAL DETAILS

Equipment Counter current air stripping of ammonia and VOCs is usually carried out in packed towers or sieve trayed columns because of the high surface area that these types of packing can provide for gas-liquid contacting. In the work reported here, however, the stripping was carried out in a bubble-cap column, which had been used previously for steam stripping of ammonia in the Flamemag process [5]. This 9.60 m high column contained 20 single bubble-cap trays. These solid-skirt bubble-caps, which were fabricated from titanium as were the downcomers, had a diameter of 219 mm and were housed in a Halar-coated mild steel shell of 350 mm diameter. Tray numbers 4, 8, 12, and 16 had thermocouples and sampling points to allow temperature and stripping proles to be observed during column operation. The air entered the column under tray 20, at a point opposite where the steam entered the reboiler. The reboiler was a single tube condensing steam heat exchanger, which discharged into the top of the column sump and was used to heat the column bottoms to approximately 43458C. This temperature was the expected temperature of the No.1 dehydration column distillate. A schematic diagram of the bubblecap column is shown in Figure 2 and the specications for the column are listed in Table 1. The bubble-cap column was not insulated but the average daily temperatures expected during the test work were between 21.8 and 24.18C. Therefore, heat losses from the column were expected to be low. Nevertheless from Test 3 onwards, the feed to the
Environmental Progress (Vol.26, No.4) DOI 10.1002/ep

Table 1. Dimensions of 20-tray bubble-cap column.

Total height (m) Tray height (m) Column diameter (m) Bubble-caps per tray Bubble-cap diameter (m) Bubble-cap height (m) Slots per bubble-cap Skirt clearance above tray (m) Weir height (m) Tray spacing (m) Downcomer diameter (m)

9.60 8.00 0.35 1 0.22 0.13 0 0.03 0.05 0.40 0.04

preweighed, large sample sizes. Non-zero results for low NH3 contents were obtained by entering 1/100th of the sample weight into the instrument to give NH3 results to three signicant gures.
Methanol, Glycol, and 2-Chloroethanol

Figure 2. Schematic diagram of stripping column.

column was preheated to produce more uniform temperatures throughout the column and a higher average column temperature. Solutions Blackwell et al., [6] found that stripping tests using synthetic solutions of methanol in tap water performed in a similar manner to their actual test solutions. It was therefore assumed that a synthetic solution would be appropriate for predicting species behavior in No.1 distillate wastewater. Four tests were completed using synthetic solutions containing $700 mg/kg NH3 prepared by adding methanol, glycol, and aqueous ammonia solution to demineralized water. No 2-CE was added to the synthetic solutions. Following the synthetic solution tests, two tests were completed using actual dehydration column No 1 distillate supplied from the Demonstration Plant at Gladstone. A sample of the feed solution was taken for chemical analysis prior to the commencement of each test and these analyses are shown in Table 2. Analytical Procedures
Ammonia

Ethylene glycol, methanol, and 2-chloroethanol (2-CE) were determined at the Demonstration Plant analytical laboratory at Gladstone, using a Waters HPLC system (high performance liquid chromatograph) equipped with a refractive index detector and running Millennium32 software. The column used was a BioRad Aminex HPX-87H 300 3 7.8 mm fast acid column running at 508C, while 0.01 v/v % phosphoric acid was used as the eluent at a ow rate of 0.8 mL/min.
pH Measurement

pH measurements were made on both the bottoms and the individual tray samples. These measurements were performed using an Orion Research model 701A digital pH/mV meter, which was calibrated using pH 7 and pH 9 buffer solutions. Column Efciency In the situation where an existing packed tower or trayed column is used, the number of ideal stages or transfer units can be determined by the graphical McCabeThiele method [7]. In this method, the number of stages are stepped off, between the operating and equilibrium lines. The slope of the operating line is drawn using the terminal concentrations achieved during a test, while the equilibrium line is based upon the Henrys law constant of the VOC being removed. Having determined the number of theoretical stages, the efciency of the column can be determined from the actual number of trays, which the column contains. The column efciency is given by Eq. 1. g 100 Ntheoretical NA (1)

Ammonia analyses were determined using a Tecator 1035 Automatic Kjeltec Analyzer. During calibration, recovery gures between 99.5 and 100.2% were obtained when using AR grade dried NH4Cl. Because of the low NH3 content expected in the stripped solutions (down to 10 mg/kg NH3), the solutions were analyzed directly without dilution using
Environmental Progress (Vol.26, No.4) DOI 10.1002/ep

Experimental Method The air ow to the column was adjusted to 50, 75, or 100% of maximum ow giving air ow rates of 70, 109, and 143 kg/h, respectively. Once the liquid feed
December 2007 367

Table 2. Composition of solutions used in air stripping tests.

Test No. 1 2 3 4 5 6

Test type Synthetic Synthetic Synthetic Synthetic Plant Plant

NH3 (mg/kg) 717 637 666 670 640 631

MeOH (mg/kg) 792 694 701 643 811 755

Glycol (mg/kg) 161 155 160 161 234 255

2-CE (mg/kg) 229 nd

nd, not determined.

was started, triplicate measurements of the feed rate were taken. The ow rates of liquid into and out of the column, as well as the levels in both the feed tank and the bottoms collection tank, were measured at approximately hourly intervals. The column bottoms were also sampled hourly while the individual trays were sampled at two hourly intervals. All liquid samples were taken in 200 mL Schott bottles. The pH of each sample was measured and duplicate subsamples were taken for methanol, glycol and 2-CE analysis. These subsamples were taken in AgilentTM 1 mL clear vials with push-in caps. The ammonia content of each sample was determined, in duplicate, from either a 25 mL or a 50 mL weighed aliquot, depending on the ammonia concentration. Temperature readings on each tray as well as the column bottoms and the reboiler were recorded by a Data Historian, along with the top and bottom column pressures and the column level at 30-min intervals. These readings were subsequently transferred to an Excel spreadsheet. The gas and liquid ow rates, along with the ammonia content of the feed solution and the bottoms product allowed the column operating line to be calculated. The low ammonia levels in both the liquid and gaseous phases meant that both the operating and equilibrium lines could be plotted as straight lines. The slope of the ammonia equilibrium line, at log mean column temperatures, was calculated by using information on the Henrys law constant for ammonia from the paper by Clegg and Brimblecombe [8] as shown in Eq. 2. LnKH NH3 8:09694 3917:507 0:00314T T (2)

Table 3. Henrys constant (kPa) and stripping factors

for air stripping tests. Test No. 1 2 3 4 5 6 NH3 H 135.8 139.7 163.2 156.2 164.2 167.3 R 1.97 2.15 2.54 1.83 1.94 2.55 MeOH H 39.1 40.4 48.5 46.1 48.8 49.9 R 0.57 0.62 0.75 0.54 0.58 0.76 H 0.001 0.001 0.001 0.001 0.001 0.001 2.03 2.16 2.18 1.64 1.65 2.14 Glycol* R 3 3 3 3 3 3 1025 1025 1025 1025 1025 1025

*H and R values at 208C only.

Henrys law constants are measured and published in many different units, making conversion between units difcult, therefore the Converting Henrys Law Constants web site by Sander [12] was used to convert from one form to another.
RESULTS

The Henrys law constant for methanol was obtained from the paper by Staudinger and Roberts [9] and this is shown in Eq. 3. Log Hcc MeOH 3:444 2142 T (3)

The Henrys law constant for glycol at 208C, was obtained from Sander, [10] and a calculated value for the Henrys law constant of 2-CE at 258C, was obtained from the paper by Yaffe et al. [11].
368 December 2007

Stripping Tests Using Synthetic Solutions The Henrys law constant (H) and the stripping factor (R) for NH3, methanol and ethylene glycol shown in Table 3 indicate that stripping should occur for NH3 as the stripping factor is greater than 1. Some methanol stripping should occur as the stripping factor for this compound was close to 1. However, the very low stripping factor values for ethylene glycol indicate that negligible stripping of this compound was expected. Stripping of 2-CE, with a stripping factor of 0.050.06 at 258C, was also expected to be small. The stripping factors for both NH3 and methanol in the plant solutions (Test No. 5 and 6) were similar to the stripping factors for the synthetic solutions (Tests No. 14) shown in Table 3. The ammonia content of the feed and bottoms solutions from the tests on synthetic solutions are shown in Table 4, along with the gas and liquid rates, while the methanol and glycol content in these solutions are shown in Table 5. The airow rate is the most important variable for minimizing stripping plant costs, [13] so in the fourth test the gas rate was reduced by $25%. Table 4 indicates that when treating the synthetic feed, the column bottoms easily attained the design value of 60 mg/kg NH3. Ammonia removal varied
Environmental Progress (Vol.26, No.4) DOI 10.1002/ep

Table 4. NH3 content of feed and column bottoms from synthetic feed stripping tests.

Test No. 1 2 3 4

Type Feed Bottom Feed Bottom Feed Bottom Feed Bottom

Time (h) 7.42 10.25 10.0 10.0

pH 10.8 9.0 10.8 8.8 10.6 8.6 10.7 8.8

NH3 (mg/kg) 717 13 637 12 666 13 670 30

G (kg/h) 143 143 143 109

L (kg/h) 57.5 53.9 54.4 54.6

Table 5. Methanol and glycol content of feed and

column bottoms from synthetic feed stripping tests. Test No. 1 2 3 4 Time (h) 7.42 10.25 10.0 10.0 MeOH (mg/kg) 792 129 694 181 701 99 643 290 Glycol (mg/kg) 161 157 155 142 160 154 161 155

Type Feed Bottom Feed Bottom Feed Bottom Feed Bottom

Figure 4. Column prole for ammonia at end of Test

No. 4 at G/L 5 2.0.

Figure 3. Column temperature prole at end of Test

No. 4 at G/L 5 2.0.

from 98% at the high gas rate to 96% at the lower gas rate. The design value for methanol in the stripper bottoms was also achieved at these G/L ratios as shown in Table 5, although the stripping efciency was lower, varying between 86% and 55%. Chemical analysis indicates that some stripping of ethylene glycol occurred, although the low stripping factor for this compound would suggest otherwise. According to Narbaibz et al., [14] R values lower than 1 mean that removal is limited by the equilibrium relationship. The temperature prole attained throughout the column at the end of Test No. 4 and shown in Figure 3 indiEnvironmental Progress (Vol.26, No.4) DOI 10.1002/ep

cates that the liquid temperature decreased from the feed point, with a minimum occurring at tray 4 and gradually increasing down the column towards the reboiler. It should be noted that these results are for a preheated feed where the ambient temperature was lower. The proles for ammonia and methanol at the end of Test No. 4, are shown in Figures 4 and 5, respectively and these gures indicate that stripping of ammonia occurred throughout the column. Won [15] in his paper mentions that some components may show composition maxima in the middle of the tower, and this behavior was exhibited by the methanol prole shown in Figure 5. Methanol stripping occurred up to tray 4 and from tray 16 onwards, but between trays 4 and 16 methanol absorption occurred. Blackwell [16] also observed the occurrence of both absorption and stripping in his work on methanol stripping. Figure 6 gives the glycol prole through the column for Test No. 4 and here only a slight decrease in the glycol concentration occurred conrming the initial supposition that negligible glycol stripping would occur. The small increase in the reboiler is probably an experimental error. Stripping Tests Using Plant Solutions Only 1600 L of plant dehydration column No 1 distillate was available to conrm whether the results achieved using synthetic solutions could be replicated
December 2007 369

Figure 5. Column prole for methanol at end of Test

No. 4 at G/L 5 2.0.

Figure 6. Column prole for glycol at end of Test No.

4 at G/L 5 2.0.

with plant solutions. In Test No. 5, using plant solution with an air rate of 109 kg/h only 83% of the ammonia was removed and Table 6 shows that the bottoms NH3 content was greater than the specication level of 60 mg/kg. This result suggested that the ammonia in the plant solution was present in an ionic form. That is, the ammonia equilibrium shown in Eq. 4 was displaced to the left. NH , NH3 aq H 4 (4)

The equilibrium shown in Eq. 4 is the basis of ammonia stripping and the US EPA [17] suggests that for ammonia stripping, the pH needs to be within the range 10.8 11.5 to move the ammonia equilibrium to the right. The pH in the as received plant solution was a lot lower than the synthetic solution but the addition of 548 mg NaOH/kg of solution raised the pH of the remaining plant solution from 9.8 to 11.5. The NH3 removal then increased to over 99%, with the residual ammonia well below the design level of 60 mg/kg as Table 6 shows. The methanol and glycol concentrations in the feed and bottoms during Test No. 5 and 6 are shown in Table 7. A higher removal efciency was observed for methanol, because of the increased G/L ratio. However, no improvement in the removal efciency was observed for glycol. The temperature prole for Test No. 6 was similar to that shown in Test No. 4 (Figure 3) although it showed a higher minimum temperature at tray 4. For Test No. 6, the average outdoor temperature for the last 3 h of operation was 21.58C, while for Test No. 4 it was only 18.38C. Both the ammonia and glycol concentrations during Test No. 6 decreased down the column indicating that stripping of these species occurred from the plant solution. In a similar manner to Test No. 4, the methanol concentration-prole, exhibited both stripping and absorption characteristics. Equilibrium Stages The stripping efciencies of the various components in the wastewater were calculated using the liq370 December 2007

uid rates into and out of the column plus the ammonia and methanol content of the feed and bottoms. These rates and compositions, along with the gas rate into the column, were also used to calculate the composition and the rate of the gas leaving the column. It was assumed that the gas leaving the top of the column was saturated with water. The average pressure in the column over the nal 3 h of operation, plus the daily values of atmospheric pressure obtained from the Victorian Climate Services Centre, [18] allowed the partial pressure of each component to be determined. This time frame was selected because the column conditions were approximately at steady state for the nal 3 h of operation. The long times needed to achieve steady state were probably a consequence of both the hold-up within the column and lack of insulation on the column. The actual values for gas and liquid ow rates and both the gas and liquid compositions, which were determined during the air stripping tests, are shown in Table 8 for the synthetic and plant solutions. The values for the Henrys law constant for both ammonia (KNH3) and methanol (KMeOH), calculated at the log mean column temperature over the nal 3 h of operation, are also shown in these tables. Because of the low ammonia and methanol concentrations in both the gas and liquid phases, the operating line and equilibrium line were drawn as straight lines. The number of equilibrium stages required for each test was determined for ammonia stripping and these are also shown in Table 8. When treating the synthetic solutions, the number of theoretical trays varied between 3.5 and 5.0 giving an efciency for the 20-tray bubble-cap column of between 17 and 25% depending on the G/L ratio. Once the pH of the plant solution was increased, similar values were obtained when treating the plant solution. The methanol stripping efciency, based upon solution analysis was determined for each test but the number of theoretical trays could not be determined because of both absorption and stripping occurring
Environmental Progress (Vol.26, No.4) DOI 10.1002/ep

Table 6. NH3 content of feed and bottoms from stripping tests using plant solutions.

Test No. 5 6

Type Feed Bottom Feed Bottom

Time (h) 10.25 10.25

pH 9.8 8.3 11.5 9.5

NH3 (mg/kg) 640 106 631 2

G (kg/h) 109 143

L (kg/h) 54.8 55.1

Table 7. Methanol, glycol, and 2-chlorethanol in feed and column bottoms from stripping tests

using plant solutions. Test No. 5 6 Type Feed Bottom Feed Bottom Time (h) 10.25 10.25 MeOH (mg/kg) 811 306 755 37 Glycol (mg/kg) 234 183 255 194 2-CE (mg/kg) 229 226 nd 114

Table 8. Log mean column conditions for synthetic solution tests.

Stripping test Log mean column temp. (8C) Log mean column pressure (kPa) L0 liquid rate (kg/h) x0 composition (mol NH3/kg) a0 composition (mol MeOH/kg) xn composition (mol NH3/kg) an composition (mol MeOH/kg) Gn11 gas rate (kg/h) yn11 pp* NH3 (Pa) zn11 pp* MeOH (Pa) y1 pp NH3 (Pa) z1 pp MeOH (Pa) KNH3 (mol kg21 Pa21) KMeOH (mol kg21 Pa21){ NH3 theoretical trays h NH3 tray efciency (%)

1 33.5 106.6 57.5 0.042 0.025 6.46 3 1024 3.03 3 1023 143 0 0 48 25 4.09 3 1024 1.39 3 1023 5.0 25

2 34.1 107.0 53.9 0.037 0.022 7.05 3 1024 5.02 3 1023 143 0 0 40 19 3.97 3 1024 1.34 3 1023 4.5 22.5

3 37.6 105.0 54.4 0.039 0.022 8.22 3 1024 3.53 3 1023 143 0 0 41 20 3.40 3 1024 1.10 3 1023 3.5 17.5

4 36.6 106.2 54.6 0.039 0.020 17.6 3 1024 8.65 3 1023 109 0 0 54 17 3.55 3 1024 1.17 3 1023 4.0 20

5 37.8 104.9 54.8 0.038 0.025 62.2 3 1024 9.36 3 1023 109 0 0 45 24 3.38 3 1024 1.10 3 1023 1.8 9

6 38.2 105.7 55.1 0.037 0.024 1.17 3 1024 1.15 3 1023 143 0 0 40 24 3.32 3 1024 1.08 3 1023 5.2 26

*pp, partial pressure. KNH3, Henrys constant for ammonia at log mean column temperature. { KMeOH, Henrys constant for methanol at log mean column temperature.

in different parts of the column. This produced an operating line which crossed over the equilibrium line for all tests apart from Test No. 4. In this test, the methanol operating line and equilibrium line were superimposed at the top of the column producing a pinch point and giving an innite number of stages. The un-insulated column meant that each tray temperature was dependent on the ambient temperature. This can be seen in Table 8 where the log mean column temperature increased from $348C for Test No. 1 and 2,
Environmental Progress (Vol.26, No.4) DOI 10.1002/ep

to 388C for Test No. 3 and 378C for Test No. 4. The increased column temperature meant that the Henrys law coefcients decreased, which resulted in lower levels of ammonia and methanol in the column bottoms. Mass Transfer Coefcient Calculations One of the most important parameters in the design of air strippers, according to Djebbar and Narbaitz, [19] is the overall system-specic liquid phase
December 2007 371

Table 9. Ammonia mass transfer coefcient, KLa determined from Eq. 9.

Test No. 1 2 3 4 5 6

Type Synthetic Synthetic Synthetic Synthetic Plant Plant

qL (mol/m3) 55,203 55,192 55,124 55,144 55,120 55,112

LM (mol h21 m22) 33,142 31,061 31,311 31,437 31,539 31,706

GM (mol h21 m22) 51,516 51,516 51,516 39,267 39,267 51,516

KLa (h21) 0.518 0.454 0.383 0.380 0.176 0.578

mass transfer coefcient, KLa. The American Institute of Chemical Engineers [20] denes the overall liquid transfer unit by Eq. 5.   Z KL aL Z dx  NOL LM x x Also, 1  NOG NOL (6) (5)

Table 10. Gas/liquid ratios and pollutant removal of efciencies attained in tests.

Test No. 1 2 3 4 5 6

G/L (molar) 1.55 1.65 1.63 1.24 1.24 1.61

G/L (mass) 2.48 2.65 2.63 2.00 1.99 2.60

KNH3 (%) 98.2 98.1 98.0 95.5 83.4 99.7

KMeOH (%) 83.7 73.9 85.9 54.9 62.3 95.1

Diab and Maddox, [21] gave the relationship between the number of transfer units and the number of equilibrium stages as  NOG Ntheoretical ln  1  (7)

Kspecies

  Xin Xout 100 % Xin

(10)

Therefore, combining Eqs. 6 and 7 gives  NOL  Ntheoretical ln  1  (8)

Thus KLa can be calculated by combining Eqs. 8 and 5 to give  KL a  Ntheoretical    ln  LM 1 L Z (9)

The values of KLa, for these tests were calculated from the values of LM, GM, and L calculated at the log mean conditions, as well as the slope of the equilibrium line and the number of theoretical stages. These KLa values are shown in Table 9. The large difference between the KLa values for Test No. 5 and 6 is due to the lower pH of the solution used in Test No. 5 and the higher G/L ratio used in Test No. 6.

The ammonia removal efciencies (KNH3) appear slightly higher than those of Slechta and Culp [3] who were able to achieve 95% NH3 removal when using a G/L ratio of 400 ft3/gal, that is, at a mass ratio of $3.6 when using 24 ft of packing (7.3 m). According to these authors, Kuhn achieved 99% NH3 removal with a G/L ratio of 750 ft3/gal (i.e., using a 6.7 G/L mass ratio). Benneworth and Morris [22] estimated the optimal G/L (mass) ratio to be 1.81 for 90% ammonia removal and as Table 10 shows our higher G/L ratios have achieved NH3 removal efciencies greater than 90%. However, this was with higher G/L ratios. The methanol removal efciencies, KMeOH, achieved in these tests compare favorably with the values achieved by others. In the paper by Blackwell et al., [16] a K value of 88.7% was achieved for their Test No. 11. They were treating a solution containing $5500 mg/kg methanol in a 9.15 m packed column operating at an average column temperature of 65.78C. The G/L molar ratio used in their test was 3.07. The maximum KMeOH value attained by Veluga et al. [23] was 3.05% at 258C when using a crisscross ow packed column with a G/L molar ratio of 0.16. However, as well as a lower G/L ratio when compared with the tests reported here, the height of their column was only of the order of 1 m.
CONCLUSION

Gas to Liquid Ratios Table 10 shows the pollutant removal efciency, K dened in Eq. 10, along with G/L ratios used in this test work.
372 December 2007

A 20-tray bubble-cap column was used to air-strip ammonia and methanol from both synthetic solutions and Demonstration Plant dehydration column No 1 distillate using gas to liquid (G/L) mass ratios of 2.02.7.
Environmental Progress (Vol.26, No.4) DOI 10.1002/ep

Ammonia removal efciencies of more than 98% were achieved from synthetic solutions containing 700 mg/kg NH3 at G/L ratios of 2.52.7. This produced stripper bottoms containing $10 mg/kg NH3. Reducing the air ow by 25% to give a G/L of 2.0, increased the ammonia content in the stripper bottoms to 30 mg/kg. Using the graphical McCabeTheile method four to six theoretical stages were needed to achieve 10 mg/ kg of ammonia in the bottoms, which meant that the overall column efciency for stripping was between 17 and 26%. The ammonia removal efciency achieved using the as received Demonstration Plant solution was only 84%. However by increasing the pH from 9.8 to 11.5 with NaOH, 99% ammonia removal was achieved with the bottoms containing less than 5 mg/kg NH3. At a G/L mass ratio of 2.6, six theoretical stages were needed to achieve this removal efciency. Volumetric ammonia mass transfer coefcients of between 0.38 and 0.52 h21 were obtained when treating synthetic solutions. When the pH of the plant solution was raised to 11.5, a value of 0.58 h21 was obtained. Unlike ammonia, stripping factors for methanol were less than 1. Nevertheless, this species was also stripped achieving removal efciencies between 58 and 95%. The number of theoretical stages required for the removal of methanol could not be calculated as the equilibrium line was below the operating line in most instances. The methanol prole down the column indicated that both absorption and stripping of methanol occurred during the tests. Negligible glycol was stripped as indicated by the small difference between the glycol content in the feed and bottom solutions. Both ammonia and methanol were removed from dehydration column No 1 distillate by air stripping using a bubble-cap column. However, the newer column internals now available, such as sieve plates and structured packing provide greater interfacial areas to enhance mass transfer [24] and these types of packing have become accepted for removing VOCs from waste waters. This is especially so when the solution has a propensity for fouling [25].
NOMENCLATURE

NOG NOL Ntheoretical NTU R T X x x* Z

= = = = = = = = = =

over all gas transfer unit actual over all liquid transfer unit theoretical number of stages number of transfer units, (dimensionless) stripping factor, G/(L 3 H), (dimensionless) absolute temperature, (K) mg/kg in liquid phase mole fraction in liquid phase equilibrium mole fraction height of the stripping tower, (m)

Greek Symbols g = overall tray efciency (%) k = mGM/LM qL = molal density of liquid phase, (mol/m3) h = column efciency (%)
ACKNOWLEDGMENTS

The authors gratefully acknowledge the contribution of Ms. Jill Rounds from CSIRO Minerals library for obtaining the many references. Thanks are also extended to the AMC staff for supplying the plant solution and for analyzing the solutions for methanol, glycol, and 2-CE.
LITERATURE CITED

G = air rate, (mol/h) H = Henrys law constant, (kPa) Hcc(MeOH) = Henrys law constant for methanol, [(moleiG/m3 ) (moleiL/m3 )21] G L = pollutant removal efciency (%) Kspecies = Henrys law constant for NH3, KH(NH3) (mol kg21 atm21) = overall volumetric liquid phase mass KLa transfer coefcient (h21) L = liquid rate, (mol/h) = liquid ow rate per bubbling area, LM (mol/(h m22) m = slope of equilibrium curve, 5 dy/dx = actual number of stages or trays, NA (dimensionless)
Environmental Progress (Vol.26, No.4) DOI 10.1002/ep

1. Sheehan, G.J., Wong, F.-S., Hourn, M.M., Kodama, M., & Jenkins, D.H. (1995). Anhydrous magnesium chloride. PCT International; WO 95/11859. 2. Menser, N. (2001). Building a magnesium industry. Engineers Australia, 73, 3639. 3. Slechta, A.F. & Culp, G.L. (1967). Water reclamation studies at the South Tahoe public utility district. Journal of Water Pollution Control Federation, 39, 787814. 4. Ball, W.P., Jones, M.D., & Kavanaugh, M.C. (1984). Mass transfer of volatile organic compounds in packed tower aeration. Journal of Water Pollution Control Federation, 56, 127135. 5. Rothon, R.N., McCallum, D.A., & Elsner, D. (1997). The development of a new route to ame retardant ller quality magnesium hydroxide. In International conference on lled polymers and llers: Eurollers 97, (pp. 1923), Manchester, UK: University of Manchester, Materials Science Centre. 6. Blackwell, B.R., Murray, F.E., & Oldam, W.K. (1979). Air stripping kraft foul condensates to remove methanol. Pulp and Paper Canada, 80, T266T269. 7. McCabe, W.L. & Thiele, E.W. (1925). Graphical design of fractionating columns. Industrial and Engineering Chemistry, 17, 605611. 8. Clegg, S.L. & Brimblecombe, P. (1989). Solubility of ammonia in pure aqueous and multicomponent solutions. Journal of Physical Chemistry, 93, 72377248. 9. Staudinger, J. & Roberts, P.V. (2000). A Critical compilation of Henrys law constant temperature
December 2007 373

10.

11.

12. 13.

14.

15. 16.

dependence relations for organic compounds in dilute aqueous solutions. Chemosphere, 44, 561 576. Sander, R. (2006). Compilation of Henrys law constants for inorganic and organic species of potential importance in environmental chemistry.Available at http://www.mpch-mainz.mpg. ` de/sander/res/henry.html Yaffe, D., Cohen, Y., Espinosa, G., Arenas, A., & Giralt, F. (2003). A fuzzy ARTMAP-based quantitative structure2property relationship (QSPR) for the Henrys law constant of organic compounds. Journal of Chemical Information and Computer Sciences, 3, 85112. Sander, R. (2006). Available at http://www. mpch-mainz.mpg.de/$sander/res/henry-conv. html (accessed January 2006). Kutzer, S., Wintrich, H., & Mersmann, A. (1995). Air strippingA method for treatment of wastewater. Chemical Engineering and Technology, 18, 149155. Narbaibz, R.M., Mayorga, W.J., Torres, P., Greeneld, J.H., Amy, G.L., & Minear, R.A. (2002). Evaluating aeration-stripping media on the pilot scale. Journal of the American Water Works Association, 94, 97111. Won, K.W. (1983). Sour-water stripper efciency. Plant/Operations Progress, 2, 108113. Blackwell, B.R., Murray, F.E., & Oldam, W.K. (1984). Analysis of simultaneous heat and mass transfer in the stripping of methanol from aque-

17.

18. 19. 20.

21. 22. 23.

24. 25.

ous solutions. Canadian Journal of Chemical Engineering, 62, 845856. EPA (2000). Wastewater technology fact sheet Ammonia stripping. United States Environmental Protection Agency, Ofce of Water, Washington, DC. EPA 832-F-00-019. Victorian Climate Services Centre (2001). Victorian Regional Ofce, Bureau of Meteorology, Docklands Victoria. Station 86071. Djebbar, Y. & Narbaitz, R.M. (1995). Mass transfer correlations for air stripping towers. Environmental Progress, 14, 137145. American Institute of Chemical Engineers (1958). Bubble-tray design manual, Prediction of fractionation efciency. The Science Press, Inc., New York. Diab, S.Y. & Maddox, R.N. (1982). Absorption. Chemical Engineering, 89, 3856. Benneworth, N.E. & Morris, N.G. (1972). Removal of ammonia by air stripping. Water Pollution Control, 485492. Velaga, A., Thibodeaux, L.J., Valsaraj, K.T., Eldridge, R.B., Moncada, D.M., & Cho, J.S. (1988). Packed crisscross ow cascade tower efciencies for methanol-water separations. Experimental versus Calculated Values Based on Counter Current Flow Correlations. Ind Eng Chem Res, 27, 14811487. Bravo, J.L. (1998). Column internals. Chemical Engineering, 105, 7683. Bravo, J.L. (1993). Effectively ght fouling of packing. Chemical Engineering Progress, 89, 7276.

374 December 2007

Environmental Progress (Vol.26, No.4) DOI 10.1002/ep