- Email: [email protected]
Ammonia removal from ammonia-rich wastewater by air stripping using a rotating packed bed
Accepted Manuscript Title: Ammonia removal from ammonia-rich wastewater by air stripping using a rotating packed bed Author: Min-Hao Yuan Yi-Hung Chen Jhih-Ying Tsai Ching-Yuan Chang PII: DOI: Reference:
S0957-5820(16)30107-0 http://dx.doi.org/doi:10.1016/j.psep.2016.06.021 PSEP 809
To appear in:
Process Safety and Environment Protection
Received date: Revised date: Accepted date:
27-2-2016 22-4-2016 11-6-2016
Please cite this article as: Yuan, M.-H., Chen, Y.-H., Tsai, J.-Y., Chang, C.Y.,Ammonia removal from ammonia-rich wastewater by air stripping using a rotating packed bed, Process Safety and Environment Protection (2016), http://dx.doi.org/10.1016/j.psep.2016.06.021 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Graphical Abstract
Clear Air
Continuous inflow of Ammonia-rich Wastewater
Neutralizer of Stripping Ammonia T
ip t
1
Continuous Supply of Fresh Air
D
us
2
Clear Effluent
cr
Rotating Packed Bed
3
an
4 5
Ac ce p
te
d
M
6
1
Page 1 of 30
6
Highlights
7
A continuous-flow rotating packed bed (RPB) was evaluated for ammonia stripping.
8
The RPB holds the advantages of small size and short retention time for ammonia stripping.
9
The roles of air and liquid flow rate in mass transfer and stripping efficiencies of RPB were
12
ip t
11
examined.
The continuous-flow RPB is a promising alternative stripping process for ammonia rich wastewater.
cr
10
13
us
14
Ac ce p
te
d
M
an
15
2
Page 2 of 30
Ammonia removal from ammonia-rich wastewater
16
by air stripping using a rotating packed bed
cr
ip t
15
us
17
Min-Hao Yuan,a Yi-Hung Chen,a,* Jhih-Ying Tsaia and Ching-Yuan Changb,c
19
a
Department of Chemical Engineering and Biotechnology, National Taipei University of
b
24 25 26
Taiwan c
Department of Chemical Engineering, National Taiwan University, Taipei 106, Taiwan
Ac ce p
23
Graduate Institute of Environmental Engineering, National Taiwan University, Taipei 106,
te
22
M
Technology, Taipei 106, Taiwan
d
20 21
an
18
27
*Corresponding Author. Tel.: +886 2 27712171 ext. 2539; fax: +886 2 87724328.
28
E-mail: [email protected] (Y.H. Chen)
29
3
Page 3 of 30
KEYWORDS
30
Air stripping; ammonia removal; rotating packed bed; continuous flow; mass transfer coefficient
31
ABSTRACT
32
Air stripping of ammonia from an ammonia-rich stream (1000 mg/L) was performed in a
33
continuous-flow rotating packed bed (RPB) at temperatures from 25 to 40 °C. The effects of the
34
major operating variables (rotational speed (ω), liquid flow rate (QL), gas flow rate (QG), and
35
stripping temperature (T)) on the volumetric liquid mass-transfer coefficient (KLa) and stripping
36
efficiency (η) were elucidated. The results indicate that the RPB exhibits higher mass-transfer
37
performance (12.3–18.4 1/h) compared with those of stripping tanks, packed towers, and other
38
advanced gas–liquid contactors (0.42–1.2 1/h). At QL = 0.05 L/min, QG/QL = 1600, and ω = 1200
39
rpm, η values for the RPB at 30 and 40 °C respectively reached 69% and 81% within 13.3 s. In
40
contrast, conventional ammonia stripping processes with liquid recirculation in larger towers
41
usually take hours to achieve the same values. The proposed dimensionless models describe the
42
relationship between KLa and the major parameters for ammonia stripping in the RPB. KLa
43
showed the greatest increase with increasing QG followed by the increase in QL, ω, and T.
44
However, the operating conditions that would make the technology economically viable and the
45
optimal conditions for efficient ammonia removal must be further studied.
Ac ce p
te
d
M
an
us
cr
ip t
29
46 47
1. Introduction
48
The removal of ammonia from ammonia-rich streams has gained increasing attention in recent
49
years because of the more stringent discharge limits for ammonium nitrogen (NH3-N) that have
50
been steadily imposed on wastewater treatment plants (WWTPs) worldwide. A substantial 4
Page 4 of 30
amount of NH3-N released into bodies of water may cause several problems such as toxicity to
52
sensitive aquatic biota, oxygen depletion, and eutrophication (Camargo et al., 2005; Ding et al.,
53
2014; Rabalais, 2002; Xiao et al., 2015). For these reasons, the US Environmental Protection
54
Agency updated national recommended criteria for water quality for NH3-N discharge (USEPA,
55
2013a). In addition, the Chinese government set out a goal of 10% reduction in total NH3-N in
56
the 12th Five-Year Plan (China MEP, 2012). In 2014, the Taiwan Environmental Protection
57
Administration implemented new restrictions on NH3-N discharge, specifying different
58
maximum values and grace periods for petrochemical, semiconductor, and optoelectronic
59
industries, as well as for science parks that discharge ammonia wastewater (Taiwan EPA, 2014).
60
There is thus an urgent need for improvement of treatment of wastewater containing ammonia by
61
removal or conversion of ammonia into a more stable and fixed form.
M
an
us
cr
ip t
51
The treatment of high-strength ammonia wastewater in typical WWTPs or biological
63
processing is often challenging. NH3-N levels in excess of allowable limits in raw water result in
64
an increase in oxygen demand and interfere with the chlorination and manganese filtration
65
processes, thus impairing the performance of typical WWTPs (Hasan et al., 2011 and 2013).
66
These problems have caused shutdowns of WWTPs, leading to a shortage of domestic water in a
67
certain region of Malaysia (Hasan et al., 2011). Biological processing that combines nitrification
68
and denitrification is the conventional method for ammonia wastewater treatment (Ioannou et al.,
69
2015; Sun et al., 2015; Zhao et al., 1999). However, it requires a relatively long retention time
70
and large spatial requirements compared with those of other methods, making its implementation
71
difficult in existing enterprises that have space limitations (Zhao et al., 1999). Moreover, this
72
technology is sensitive to shock, toxic loads, and cold weather conditions, and it does not allow
73
for the recovery of ammonia.
Ac ce p
te
d
62
5
Page 5 of 30
Ammonia recovery from ammonia-rich wastewater is preferable, as ammonia can be used to
75
produce fertilizer for agricultural use and is thus an additional revenue source for WWTP
76
operators (Zhao et al., 1999). In practice, ammonia can be recovered by ion exchange (Jorgensen
77
and Weatherley, 2003), struvite precipitation (Liu et al., 2013), and air stripping (Basakcilardan-
78
Kabakci et al., 2007; Chang et al., 2013; Cheung et al., 1997; Deĝermenci et al., 2012; Jiang et
79
al., 2014; Kutzer et al., 1995; Laureni et al., 2013; Le et al., 2006; Liu et al., 2015; Zeng et al.,
80
2006; Zhang and Jahng, 2010). Among these techniques, ion exchange requires an extremely low
81
concentration of solids in the effluent to prevent fouling. Struvite precipitation occurs at
82
equimolecular concentrations of Mg2+, NH4+, and PO43− under slightly alkaline conditions.
83
Hence, the addition of a source of Mg2+ or PO43− salts is essential to optimize the struvite
84
crystallization process when wastewater contains less magnesium and phosphate as compared
85
with nitrogen. The final technique, air stripping, is a physical process for ammonia recovery,
86
tolerating some degree of solids and requiring mainly temperature and pH controls. As long as
87
the air temperature and pH remain stable, the operation of ammonia stripping is relatively simple
88
and is unaffected by wastewater fluctuation and toxic loads. However, the lime used for raising
89
the pH of the effluent to 10.8–11.5 often results in unwanted fouling in the packed beds due to
90
calcium carbonate deposition (Kutzer et al., 1995; Liu et al., 2015).
Ac ce p
te
d
M
an
us
cr
ip t
74
91
The air stripping process has been successfully implemented to remove ammonia from
92
different ammonia-rich streams, e.g., pig slurry (Laureni et al., 2013; Zhang and Jahng, 2010),
93
cattle and fermented chicken manure (Jiang et al., 2014; Zeng et al., 2006), human urine
94
(Basakcilardan-Kabakci et al., 2007; Le et al., 2006), landfill leachate (Cheung et al., 1997), and
95
sour water from oil refineries (Chang et al., 2013). The process involves two packed towers for
96
transferring volatile ammonia from wastewater into a gas phase and then converting the
6
Page 6 of 30
ammonia gas with an acid solution to give stable ammonium salts for use as mineral fertilizer.
98
However, the common packed depth of each tower is about 6.1–7.6 m (USEPA, 2013b), which
99
poses a challenge to existing treatment plants that expanding capacity is not feasible due to space
100
limitations. Therefore, efforts to improve the design and miniaturization of ammonia stripping,
101
e.g., the use of jet loops (Deĝermenci et al., 2012) and aerocyclone reactors (Quan et al., 2009),
102
are underway.
cr
ip t
97
Previous studies on the absorption of CO2 (Jassim et al., 2007; Pan et al., 2013 and 2015) and
104
volatile organic compounds (VOCs) (Chen and Liu, 2002; Lin et al., 2003 and 2006), distillation
105
(Kellehe and Fair, 1996; Lin et al., 2002), VOC stripping (Gudena et al., 2012; Lin et al., 2004;
106
Liu et al., 1996; Singh et al., 1992), ozonation (Chen et al., 2004), and esterification (Chen et al.,
107
2010) in a rotating packed bed (RPB) have remarkable potential to reduce the size of the packed
108
tower. This potential is due to the intensified mass transfer at the gas–liquid interface generated
109
in a RPB via high centrifugal force. The high centrifugal force (300–10000 m/s2) is usually 1–3
110
orders of magnitude greater than gravitational acceleration (Rao et al., 2004). An RPB using
111
basic oxygen furnace (BOF) slag can remove approximately 96–99% of the CO2 in flue gas
112
stream with 30% CO2 within a short reaction time (1 min) at 25 °C and 1 atm (Pan et al., 2013).
113
The conventional stripper height necessary to achieve a performance similar to that of a RPB for
114
CO2 absorption is greater by a factor of 8.4, while the diameter is greater by a factor of 11.3
115
(Jassim et al., 2007). Hence, the results suggest a significant reduction in equipment size and
116
space with use of the RPB. Trichloroethylene (TCE) removal from groundwater at low gas/liquid
117
ratios, a preferred outcome that is difficult to obtain in conventional packed beds, has been
118
clearly enhanced by air stripping by a RPB (Gudena et al., 2012). Moreover, the total capital cost
119
of the RPB can be significantly reduced according to the estimate for TCE stripping. In
Ac ce p
te
d
M
an
us
103
7
Page 7 of 30
120
summary, the unique features of the RPB, such as high volumetric gas–liquid mass-transfer
121
coefficients (KLa), reduced tendency for fouling, and reduced equipment size and cost, can
122
potentially alleviate the limitations associated with conventional ammonia stripping. Although ammonia stripping is technically feasible for the removal of ammonia from various
124
ammonia-rich streams, there is still an urgent need for a reduction in equipment size and space
125
for its implementation in existing enterprises due to the more stringent discharge limits for NH3-
126
N in wastewater. Yuan et al. (2016) had showed that the remarkable removal efficiencies and
127
small height of transfer units for the ammonia stripping were obtained from both laboratory and
128
pilot scale RPB systems at ambient temperature. This study follows up with the previous work
129
and shows the effects of temperature for ammonia stripping in a high-voidage RPB system. The
130
evaluation of various operating conditions should identify the critical factors that affect the KLa
131
values and stripping efficiency (η). High-voidage packing of stainless steel wire mesh was used
132
to provide a low pressure drop and to reduce the tendency for fouling (Lin et al., 2004). Effects
133
of the major operating variables (rotational speed (ω), liquid flow rate (QL), the ratio of flow rate
134
of air (QG) to QL, and temperature (T)) on KLa and η values were investigated and elucidated.
135
Dimensionless equations are proposed to predict KLa and η values for ammonia stripping. These
136
results provide useful information for the practical design of ammonia stripping using RPBs at
137
various operating conditions.
139
cr
us
an
M
d
te
Ac ce p
138
ip t
123
2. Materials and methods
140
2.1. Apparatus and operating conditions
141
A schematic diagram of ammonia stripping using the RPB is shown in Fig. 1. The system
142
consists of the RPB, gas and liquid feed controls, analyzer of effluents, and neutralizer of
8
Page 8 of 30
effluent gas. The inner radius (ri), outer radius (r0), and axial height (zB) of the packed-bed
144
rotator were 0.0355, 0.0850, and 0.0215 m, respectively. The thickness of the RPB shell and the
145
distance from the side of RPB to the shell were 0.008 and 0.011 m, respectively. The reactor
146
volume (VB) of the RPB was 410−4 m3. It was used as an estimate of the volume of the packed-
147
bed rotator, π(r02 − ri2)zB. The bed rotation was controlled by a motor, and ω was varied from 300
148
to 1200 rpm. This resulted in a centrifugal force equivalent to 6.25 to 100 gravitational
149
acceleration. The packing material was made of grade 304 stainless-steel wires in the shape of
150
annular rings, which comprise interconnected wires with a mean diameter of 0.22 mm. Stainless
151
steel has critical surface tension higher than that of glass, ceramic, and acrylic, thus facilitating
152
gas-liquid mass transfer in the RPB (Chen et al., 2006). High-voidage packing (0.954 m3/m3) of
153
the wire mesh was used to lower the pressure drop and to reduce the tendency for fouling (Singh
154
et al., 1992).
d
M
an
us
cr
ip t
143
Aqueous ammonia was pumped into the RPB at a controlled feed rate after the solution
156
concentration was adjusted to 1000 mg/L with ammonium chloride (99%, Jin Yih Chemical Co.,
157
Ltd., Taipei, Taiwan) and distilled water, and the pH value was adjusted to 11 with sodium
158
hydroxide solution (45%, Jin Yih Chemical Co., Ltd., Taipei, Taiwan). The ammonia in the RPB
159
was then evaporated rapidly through acceleration by centrifugal force and by mixing with air
160
flow to achieve a high η value.
Ac ce p
te
155
161
For the experiments, a diaphragm liquid pump (KNF 1.300 TT 18S, KNF, Sursee,
162
Switzerland) was used to constantly introduce ammonia solution into the housing at a QL of
163
0.025–0.1 L/min. An air compressor (DT-175-2C, Tong Cheng Iron Works Co., Ltd., Taipei,
164
Taiwan) and an air flow meter (Dwyer Instruments Inc., Michigan City, IN, USA) were used to
165
control QG to within 30–60 L/min. The QG/QL ratio was set between 750 and 1800. The
9
Page 9 of 30
temperatures of the solution and the RPB with the ammonia solution were set to 25, 30, and 40
167
°C, by using a hot plate and magnetic stirrer (C-MAG HS7 S1, IKA Lab Equipment, Staufen,
168
Germany). Similarly, Teflon heating tape (Kwo-Yi Co., Taipei, Taiwan) was used for the RPB
169
housing and type-K thermocouples (TES-1306, TES Electrical Electronic Corp., Taipei,
170
Taiwan).
ip t
166
cr
171 2.2. Analytical methods
173
The ammonia concentrations of the inlet and outlet solutions were determined by using an ion-
174
selective electrode (HI4101, Hanna Instruments, Woonsocket, RI, USA). A pH meter (HI1131B,
175
Hanna Instruments, Woonsocket, RI, USA) was used to measure the pH value of the ammonia
176
solution. To neutralize the effluent gas, the vaporized or purged ammonia gas was introduced
177
into a sulfuric acid solution and converted to ammonium sulfate (Sigma-Aldrich, St. Louis, MO,
178
USA) before it was expelled as exhaust. For each reaction condition, more than 15 min of
179
operation was allowed for the system to reach a steady state prior to assessment.
an
M
d
te
Ac ce p
180
us
172
181
2.3. Theoretical considerations
182
In this study, η and overall KLa values were used to assess the performance of the RPB for
183
ammonia stripping. The η value was computed from the ammonia concentration in the feed
184
(CL,in) and effluent (CL,out) liquid streams.
185
C 1 L ,out C L ,in
100%
(1)
186
In the centrifugal vapor–liquid contactor, KLa was calculated from mass-balance concepts and
187
from the two-film theory by Singh (1989), Singh et al. (1992), and Chen et al. (2005a and 2005b).
10
Page 10 of 30
188
The mass-balance equation for a thin annulus section within the RPB at steady state is as
189
follows:
192
(2)
The concentration at gas-liquid equilibrium (CL*) is assumed to follow Henry’s law: C*L
ip t
191
Q L dC L K L a C*L C L 2rz Bdr
CG HC
(3)
cr
190
where HC is the dimensionless Henry constant for ammonia. HC values are given in Table 1.
194
According to the two-film theory, the variable gas-phase concentration is expressed as
198
an
(4)
1 C 1 ln 1 L ,in Q S C L ,out S K La L 1 VB 1 S
M
Solving Eqs. 2–4 simultaneously gives the following equation for computing KLa:
d
197
QL C L CL,in QG
(5)
te
196
CG
where S is the stripping factor and VB is the reactor volume of the RPB:
Ac ce p
195
us
193
QG QL
199
S HC
200
VB r02 ri2 z B
(6)
(7)
201
VB is approximately 410−4 m3. Notably, the overall liquid-to-gas mass transfer consists of
202
resistances associated with the stagnant liquid and gas interfacial films. Therefore, the resistances
203
of both the liquid and the gas films have been taken into consideration in the calculation of KLa.
204
On the basis of KLa, prediction of η using Eq. 1 may be modified further as follows:
11
Page 11 of 30
(8)
ip t
205
1 1 S 100% 1 VB 1 1 1 exp K L a Q L S S
206 3. Results and discussion
cr
207
3.1. The effects of ω and T
209
Fig. 2 shows the dependence of KLa and η on ω and T under operation at QL = 0.05 L/min and
210
QG/QL = 1200. As seen in Fig. 2a, KLa increased as ω increased from 300 to 1200 rpm. The KLa
211
value was 0.00257 1/s at ω = 1200 rpm and T = 25 °C, which is 1.5 times higher than that at ω =
212
300 rpm (0.00167 1/s). According to the two-film model, the resistance to transfer in each phase
213
is localized in the thin films at the interface. Packing in the RPB via centrifugal force under high
214
ω facilitates division of the ammonia solution into smaller droplets and thinner films. The
215
consequent lower resistance at the stagnant liquid film results in a higher KLa value. KLa values
216
were 0.00257, 0.00326, and 0.00404 1/s at ω = 1200 rpm at T values of 25, 30, and 40 °C,
217
respectively. The observed dependence of KLa on T is predicted by the increase in Henry’s law
218
constants for ammonia with T (25–40 °C in this case), which caused more ammonia to evaporate
219
easily into the gas phase.
Ac ce p
te
d
M
an
us
208
220
As shown in Fig. 2b, the increase in ω and T in terms of η also enhanced ammonia removal.
221
According to the theory, a higher η is obtained by increasing KLa at the same stripping time. The
222
stripping time is the liquid hydraulic retention time (tL = VL/QL or LVB/QL) in the RPB, where
223
VL and L are the liquid hold-up and relative liquid hold-up, respectively (Chen et al., 2004). In
224
the experimental conditions in the present study, t L ranged from 11 to 20 s, as determined by the
225
calculation of L (0.03–0.06) based on the empirical equation proposed by Chen et al. (2004). At 12
Page 12 of 30
226
QL = 0.05 L/min, QG/QL = 1600, and ω =1200 rpm, only 13.3 s was required to achieve 69% and
227
81% of η in the RPB at 30 and 40 °C, respectively. Table 2 lists conventional and advanced processes for stripping ammonia from wastewater
229
according to their reaction conditions and performance (Basakcilardan-Kabakci et al., 2007;
230
Deĝermenci et al., 2012; Le et al., 2006; Quan et al., 2009). In contrast to the other processes, the
231
continuous-flow RPB does not require liquid recirculation to extend the stripping time in order to
232
achieve the desired η value. As seen in Table 2, mass transfer (0.00342–0.00510 1/s or 12.3–18.4
233
1/h) was high and had a very short stripping time as compared with other stripping processes
234
such as those in the stirrer tank, packed tower, jet loop, and aerocyclone reactors. These
235
processes often take hours and liquid recirculation to achieve a similar level of η value.
an
us
cr
ip t
228
It is noted that the longer stripping time was required since the single-stage RPB did not
237
achieve the discharge standards (e.g. 20 mg/L-NH3-N) for wastewater with high ammonia
238
concentration (as this study of 1000 mg/L). Hence, a large size of RPB or a series configuration
239
of continuous-flow RPB was suggested for the on-site application. On the basis of current
240
experimental results, the effluent concentration of NH3 can be reduced from 1000 mg/L to 16.8
241
mg/L at 25 °C by a series of four-stage RPB and to 6.8 mg/L at 40 °C by a series of three-stage
242
RPB. Consequently, the stripping time for the satisfied effluent concentration was three to four
243
times of the single one, which was smaller than the required time with other stripping processes
244
by two orders of magnitude.
Ac ce p
te
d
M
236
245 246
3.2. The effect of QG
247
Fig. 3 shows the effect of the QG/QL ratio (or QG) and ω on KLa and η values at QL = 0.05
248
L/min and T = 25 and 40 °C. The results indicate that a higher QG/QL ratio or a higher QG value
13
Page 13 of 30
at a given QL value increases KLa and η for ammonia stripping by multiplying the stripping
250
factor (S). It is noted that the S is the ratio of the slope of equilibrium line and the slope of the
251
operating line (Liu et al., 2015). The increasing supply of continuous flow of fresh air in the RPB
252
facilitated the transfer of ammonia molecules from the water to the air, thus resulting in larger
253
values of S and transfer units. At T = 25 °C and ω = 1200 rpm, the KLa value at a QG/QL ratio of
254
1800 (0.00342 1/s) was 1.8 times higher than that at a QG/QL ratio of 750 (0.00186 1/s), while
255
the η value at a QG/QL ratio of 1800 (64.3%) was 1.7 times higher than that at a QG/QL ratio of
256
400 (36.8%). A similar trend was observed at high T (40 °C) and at ω = 1200 rpm. The KLa
257
values were 0.00227 and 0.00510 1/s at QG/QL ratios of 400 and 1600, respectively; likewise, the
258
η values were 39.5% and 81.3% at QG/QL ratios of 400 and 1600, respectively. Although a
259
higher QG/QL ratio or QG also resulted in a shorter time of contact between the liquid and gas
260
currents, introduction of abundant air apparently enhanced the mass transfer of ammonia by
261
increasing the refresh rate of the gas–liquid interface. Consequently, a shorter exposure time was
262
required to attain the desired high efficiency of the RPB in ammonia stripping.
cr
us
an
M
d
te
Ac ce p
263
ip t
249
264
3.3. The effect of QL
265
The effect of QL (0.02–0.1 L/min) on KLa and η values at QG values of 40–75 L/min (T = 25–
266
40 °C and ω = 1200 rpm) is shown in Fig. 4. The results differ from the effects of QG, ω, and T
267
on KLa and η. The η value decreased with QL, in contrast to the tendency of KLa with QL (Fig.
268
4a). Changes in KLa due to these operational factors are normally directly reflected in the
269
separation efficiency. At QG = 40 L/min, ω = 1200 rpm, and T = 40 °C, the experimental KLa
270
increased by 2.2 times, from 0.00256 to 0.00583 1/s, while the QL value increased four times,
271
from 0.025 to 0.1 L/min. In contrast, the η value decreased from 81.4% to 41.8% when the QL
14
Page 14 of 30
value increased from 0.025 to 0.1 L/min. The results indicate that the increased QL leads to
273
considerable compensatory effects due to the decrease in tL and S, resulting in a decrease in η for
274
ammonia stripping. Note that the tL decreased from 17.2 to 10.4 s when QL increased from 0.025
275
to 0.1 L/min at the same level of QG = 40 L/min, ω = 1200 rpm, and T = 40 °C.
ip t
272
276 3.4. Prediction models for KLa and η
278
In order to explore the effect of the operating variables QL, QG, ω, and T on KLa within the
279
range of experimental conditions, this study used the empirical dimensionless model that has
280
been proposed for the prediction of KLa in ethanol stripping (Kelleher and Fair, 1996), VOC
281
absorption (Chen and Liu, 2002), and ozonation (Chen et al., 2004) processes in the RPB system.
282
The dimensionless parameters of the model include the Reynolds numbers of the liquid (ReL)
283
and gas (ReG), the Grashof number of the liquid based on the average bed radius (GrL,avg), and
284
the dimensionless Henry constant for ammonia (HC). All of these parameters may be computed
285
from operating variables and other physical properties. The correlation model for KLa is obtained
286
by using the observational data in Figs. 2–4 (R2 = 0.8075):
us
an
M
d
te
Ac ce p
287
cr
277
K L ad p D La p
0.024 Re L
0.352
Re G
0.613
GrL,avg
0.198
HC
0.159
(9)
288
where dP is the stainless-steel-wire diameter of the packing (0.22 m), ap is the specific area of
289
packing per unit volume of the packed bed (840 m2/m3), and DL is the molecular liquid diffusion
290
coefficient of ammonia. ReL, ReG, and GrL,avg are defined as follows:
291
r L Q L ln o ri Re L 2Z B ro ri a P L
(10)
15
Page 15 of 30
292
r G Q G ln o ri Re G 2Z B ro ri a P G
(11)
ravg2 ravg ri
3
GrL , avg
L
(12)
2
ip t
293
The values of the physical properties of the dimensionless groups are shown in Table 1. The
295
range of the dimensionless groups in the empirical model is 0.12 < KLadP/DLaP < 0.85, 0.10 <
296
ReL < 0.40, 3.11 < ReG < 15.23, 1.35 × 109 < GrL,avg < 4.27 × 1010, and 6.96 × 10−4 < HC < 1.28 ×
297
10−4. As shown in Eq. 9, KLa is more sensitive to changes in ReG than to changes in ReL, GrL,avg,
298
and HC, in accordance with its power. Hence, the strong enhancement in mass transfer through
299
the supply of fresh air is attributable to the fast refresh rate, which directly increases the QG value
300
within a short tL.
M
an
us
cr
294
Figs. 5a and 5b illustrate the parity of the experimental and predicted values of KLa and η,
302
respectively. Note that the determination of η is based on the substitution of the predicted KLa
303
values into Eq. 8. As can be seen, most of the experimental values of KLa and η lie within 20%
304
of their predicted values. For comparison, the predicted curves are also plotted in Figs. 2–4.
305
Except for the underestimated KLa values at QL (0.1 L/min), ω (1200 rpm), and T (30 and 40 °C)
306
values in Fig. 4a, the experimental and predicted values of KLa and η indicate a good degree of
307
fit for the dimensionless groups ReL, ReG, GrL,avg, and HC.
309
te
Ac ce p
308
d
301
4. Conclusions
310
In summary, the continuous-flow RPB used in this study appears to be very efficient at air
311
stripping of ammonia from an ammonia-rich stream (1000 mg/L). The overall KLa values are
312
12.3–18.4 1/h, which are significantly higher than the values (0.42–1.2 1/h) obtained by using
16
Page 16 of 30
stripping tanks, packed towers, and other advanced gas–liquid contactors. For industrial
314
application of wastewater with high ammonia concentration, a large size of RPB or a series
315
configuration of continuous-flow RPB can be used in order to satisfy the more stringent
316
discharge standards. The results of the empirical dimensionless model show that QG and QL are
317
the most sensitive ammonia stripping parameters for liquid–gas mass transfer in the RPB, while
318
ω and T are less-sensitive parameters.
cr
ip t
313
us
319 Acknowledgment
321
The authors gratefully acknowledge the financial support provided by the Ministry of Science
322
and Technology of Taiwan (102-2622-E-027-024-CC3).
M
an
320
323 Abbreviations
325
aP
Specific area of packing per unit volume of a packed-bed, m2/m3
326
CL,in
Liquid concentration of ammonia of inlet liquid, mg/L
327
CL,out
328
DL
329
dP
330
g
331
HC
332
KLa
Overall liquid volumetric mass transfer coefficient, 1/s
333
ravg
Average radius of a packed bed, (ri + ro)/2, m
334
ri
Inner radius of a packed bed, m
Ac ce p
te
d
324
Liquid concentration of ammonia of outlet liquid, mg/L Molecular liquid diffusion coefficient of ammonia Stainless wire diameter of bed, 0.22 m Gravitational acceleration, m/s2 Dimensionless Henry’s law constant of ammonia
17
Page 17 of 30
Outer radius of a packed bed, m
336
RPB
Rotating packed bed
337
S
Stripping factor, HCQG/QL
338
tL
liquid hydraulic retention time, VL/QL or LVB/QL, s, min or h
339
T
Stripping temperature, °C
340
TCE
Trichloroethylene
341
VOCs
Volatile organic compounds
342
QG
Gas flow rate, L/min
343
QG/QL
Ratio of gas and liquid flow rate, -
344
QL
Liquid flow rate, L/min
345
VB
Volume of a packed bed, π(ro2 – ri2)ZB, m3
346
VL
Liquid hold-up, m3
347
WWTPs
Wastewater treatment plants
348
ZB
Axial height of a packed bed, m
us an
M
d
te
Ac ce p
349
ip t
ro
cr
335
350
Greek Symbols
351
L
352
μG
353
μL
Liquid viscosity
354
L
Dynamic liquid viscosity
355
π
Circular ratio
Relative liquid hold-up Gas viscosity
18
Page 18 of 30
356
ρG
Density of gas
357
ρL
Density of liquid
358
ω
Rotational speed, rpm or rad/s (for Gravg calculation)
ip t
359 Dimensionless Groupings
361
Gravg
Grashof number of the liquid based on the average bed radius, ravgω2(ravg – ri)3/νL2
362
ReG
Reynolds number of the gas, ρGQGln(ro/ri)/[2πZB(ro – ri)apμG]
363
ReL
Reynolds number of the liquid, ρLQLln(ro/ri)/[2πZB(ro – ri)apμL]
an
us
cr
360
364 References
366
Basakcilardan-Kabakci, S., Ipekoglu, A.N., Talini, I., 2007, Recovery of ammonia from human
369
d
Camargo, J.A. Alonso, A., Salamanca, A., 2005, Nitrate toxicity to aquatic animals: A review
te
368
urine by stripping and absorption. Environ. Eng. Sci. 24, 615–624.
with new data for freshwater invertebrates. Chemosphere 58, 1255–1267.
Ac ce p
367
M
365
370
Chang, S.Y., Tu, M.Y., Hsiao, A.F., Tsao, L.K., Huang, T.L., 2013, A real case of ammonia and
371
nitrate removal improvement in refinery wastewater treatment. J. Petrol. 49, 91–110. (In
372
Chinese)
373 374
Chen, Y.S., Liu, H.S., 2002, Absorption of VOCs in a rotating packed bed. Ind. Eng. Chem. Res. 41, 1583–1588.
375
Chen, Y.H., Chang, C.Y., Su, W.L., Chen, C.C., Chiu, C.Y., Yu, Y.H., Chiang, P.C., Chiang,
376
S.I.M., 2004, Modeling ozone contacting process in a rotating packed bed. Ind. Eng. Chem.
377
Res. 43, 228–236.
378 379
Chen, Y.S., Lin, C.C., Liu, H.S., 2005a, Mass transfer in a rotating packed bed with viscous Newtonian and non-Newtonian fluids. Ind. Eng. Chem. Res. 44, 1043–1051.
19
Page 19 of 30
383 384 385 386 387 388
Chen, Y.S., Lin, F.Y., Lin, C.C., Tai, C.Y.D., Liu, H.S., 2006, Packing characteristics for mass transfer in a rotating packed bed. Ind. Eng. Chem. Res. 45, 6846–6853.
ip t
382
of the bed. Ind. Eng. Chem. Res. 44, 7868–7875.
Chen, Y.H., Wang, L.C., Tsai, C.H., Shang, N.C., 2010, Continuous-flow esterification of free fatty acids in a rotating packed bed. Ind. Eng. Chem. Res. 49, 4117–4122.
cr
381
Chen, Y.S., Lin, C.C., Liu, H.S., 2005b, Mass transfer in a rotating packed bed with various radii
Cheung, K.C., Chu, L.M., Wong, M.H., 1997, Ammonia stripping as a pretreatment for landfill leachate. Water Air Soil. Pollut. 94, 209–221.
us
380
China Ministry of Environmental Protection (MEP), 2012, MEP explains the National 12th FiveYear
390
http://english.mep.gov.cn/News_service/infocus/201202/t20120207_223194.htm
392
for
environmental
protection,
China
MEP,
Beijing,
China.
Deĝermenci, N., Ata, O.N., Yildiz, E., 2012, Ammonia removal by air stripping in a semi-batch
M
391
Plan
an
389
jet loop reactor. J. Ind. Eng. Chem. 18, 399–404.
Ding, L.J., An, X.L., Li, S., Zhang, G.L., Zhu, Y.G., 2014 Nitrogen loss through anaerobic
394
ammonium oxidation coupled to iron reduction from paddy soils in a chronosequence.
395
Environ. Sci. Technol. 48, 10641−10647.
397
te
Ac ce p
396
d
393
Gudena, K., Rangaiah, G.P., Lakshminarayanan, S., 2012, Optimal design of a rotating packed bed for VOC stripping from contaminated groundwater. Ind. Eng. Chem. Res. 51, 835–847.
398
Hasan, H.A., Abdullah, S.R.S., Kamarudin, S.K., Kofli, N.T., 2011, Problems of ammonia and
399
manganese in Malaysian drinking water treatments. World Appl. Sci. J. 12, 1890–1896.
400
Hasan, H.A., Abdullah, S.R.S., Kamarudin, S.K., Kofli, N.T., 2013, On-off control of aeration
401
time in the simultaneous removal of ammonia and manganese using a biological aerated
402
filter system. Process Saf. Environ. 91, 415–422.
403
Ioannou, L.A., Li Puma, G., Fatta-Kassinos, D., 2015, Treatment of winery wastewater by
404
physicochemical, biological and advanced processes: A review. J. Hazard. Mater. 286, 343–
405
368.
20
Page 20 of 30
406 407
Jorgensen, T.C., Weatherley, L.R., 2003, Ammonia removal from wastewater by ion exchange in the presence of organic contaminants. Water Res. 37, 1723–1728. Liu, Y.H., Kumar, S., Kwag, J.H., Ra, C.S., 2013, Magnesium ammonium phosphate formation,
409
recovery and its application as valuable resources: A review. J. Chem. Technol. Biot. 88,
410
181–189.
ip t
408
Liu, B., Giannis, A., Zhang, J., Chang, V.W.C., Wang, J.Y., 2015, Air stripping process for
412
ammonia recovery from source-separated urine: Modeling and optimization. J. Chem.
413
Technol. Biot. 90, 2208–2217.
us
cr
411
Laureni, M., Palatsi, J., Llovera, M., Bonmati, A., 2013, Influence of pig slurry characteristics on
415
ammonia stripping efficiencies and quality of the recovered ammonium-sulfate solution. J.
416
Chem. Technol. Biot. 88, 1654–1662.
an
414
Jassim, M.S., Rochelle, G., Eimer, D., Ramshaw, C., 2007, Carbon dioxide absorption and
418
desorption in aqueous monoethanolamine solutions in a rotating packed bed. Ind. Eng.
419
Chem. Res. 46, 2823–2833.
d
M
417
Jiang, A., Zhang, T., Zhao, Q.B., Li, X., Chen, S., Frear, C.S., 2014, Evaluation of an integrated
421
ammonia stripping, recovery, and biogas scrubbing system for use with anaerobically
422
digested dairy manure. Biosystems Eng. 119, 117–126.
424 425 426 427 428 429 430
Ac ce p
423
te
420
Kelleher, Y., Fair, J.R., 1996, Distillation studies in a high-gravity contactor. Ind. Eng. Chem. Res. 12, 4646–4655.
Kutzer, S., Wintrich, H., Mersmann, A., 1995, Air stripping – A method for treatment of wastewater. Chem. Eng. Technol. 18, 149–155. Le, L., Wang, H.W., Lu, H.H., 2006, Nitrogen removal using air stripping tower in urban wastewater treatment plant. China Water Wastewater 22, 92–95. Lin, C.C., Ho, T.J., Liu, W.T., Liu, H.S., 2002, Distillation in a rotating packed bed. J. Chem. Eng. Jpn. 35, 1298–1304.
21
Page 21 of 30
432 433 434
Lin, C.C., Liu, W.T., Tan, C.S., 2003, Removal of carbon dioxide by absorption in a rotating packed bed. Ind. Eng. Chem. Res. 42, 2381–2386. Lin, C.C., Wei, T.Y., Liu, W.T., Shen, K.P., 2004, Removal of VOCs from gaseous streams in a high-voidage rotating packed bed. J. Chem. Eng. Jpn. 37, 1471–1477.
ip t
431
Lin, C.C., Wei, T.Y., Hsu, S.K., Liu, W.T., 2006, Performance of a pilot-scale cross-flow
436
rotating packed bed in removing VOCs from waste gas streams. Sep. Purif. Technol. 52,
437
274–279.
439
Liu, H.S., Lin, C.C., Wu, S.C., Hsu, H.W., 1996, Characteristics of a rotating packed bed. Ind.
us
438
cr
435
Eng. Chem. Res. 35, 3590–3596.
Pan, S.Y., Chiang, P.C., Chen, Y.H., Tan, C.S., Chang, E.E., 2013, Ex situ CO2 capture by
441
carbonation of steelmaking slag coupled with metalworking wastewater in a rotating packed
442
bed. Environ. Sci. Technol. 47, 3308–3315.
M
an
440
Pan, S.Y., Chen, Y.H., Chen, C.D., Shan, A.L., Lin, M., Chiang, P.C., 2015, High-gravity
444
carbonation process for enhancing CO2 fixation and utilization exemplified by the
445
steelmaking industry. Environ. Sci. Technol. 49, 12380–12387.
te
447
Quan, X., Wang, F., Zhao, Q., Zhao, T., Xiang, J., 2009, Air stripping of ammonia in a watersparged aerocyclone reactor. J. Hazard. Mater. 170, 983–988.
Ac ce p
446
d
443
448
Rabalais, N.N., 2002, Nitrogen in aquatic ecosystems. Ambio 31, 102–112.
449
Rao, D.P., Bhowal, A., Goswami, P.S., 2004, Process intensification in rotating packed beds
450
(HIGEE): An appraisal. Ind. Eng. Chem. Res. 43, 1150–1162.
451
Singh, S.P., 1989, Air stripping of volatile organic compounds from groundwater: An evaluation
452
of a centrifugal vapor-liquid contactor. Ph.D. Dissertation, University of Tennessee,
453
Knoxville, TN.
454
Singh, S.P., Wilson, J.H., Counce, R.M., Villiers-Fisher, J.F., Jennings, H.L., Lucero, A.J., Reed,
455
G.D., Ashworth, R.A., Elliott, M.G., 1992, Removal of volatile organic compounds from
456
groundwater using a rotary air stripper. Ind. Eng. Chem. Res. 31, 574–580.
22
Page 22 of 30
457
Sun, F., Sun, B., Hu, J., He, Y., Wu, W., 2015, Organics and nitrogen removal from textile
458
auxiliaries wastewater with A2O-MBR in a pilot-scale. J. Hazard. Mater. 286, 416–424.
459
Taiwan Environmental Protection Administration (EPA), 2014, Effluent standards for chemical industry announced. Environ. Policy Mon. 17, 6–7.
ip t
460
United States Environmental Protection Agency (EPA), 2013, Aquatic life ambient water quality
462
criteria for ammonia – Freshwater; EPA 822-R-13-001, USEPA, Washington, DC.
463
http://water.epa.gov/scitech/swguidance/standards/criteria/aqlife/ammonia/upload/aquatic-
464
life-ambient-water-quality-criteria-for-ammonia-freshwater-2013.pdf
us
465
cr
461
USEPA, 2013, Wastewater technology fact sheet: Ammonia stripping, USEPA, Washington, DC. http://water.epa.gov/scitech/wastetech/upload/2002_06_28_mtb_ammonia_stripping.
467
pdf
469
Xiao, Y., Araujo, C.D., Sze, C.C., Stuckey, D.C., 2015, Toxicity measurement in biological
M
468
an
466
wastewater treatment processes: A review. J. Hazard. Mater. 286, 15−29. Yuan, M.H., Chen, Y.H., Tsai, J.Y., Chang, C.Y., 2016, Removal of ammonia from wastewater by
471
air stripping process in laboratory and pilot scales using a rotating packed bed at ambient
472
temperature. J. Taiwan Inst. Chem. Eng. 60, 488–495.
474 475 476
te
Zeng, L., Mangan, C., Li, X., 2006, Ammonia recovery from anaerobically digested cattle
Ac ce p
473
d
470
manure by steam stripping. Water Sci. Technol. 54, 137–145. Zhang, L., Jahng, D., 2010, Enhanced anaerobic digestion of piggery wastewater by ammonia stripping: Effects of alkali types. J. Hazard. Mater. 182, 536–543.
477
Zhao, H.W., Mavinic, D.S., Oldham, W.K., Koch, F.A., 1999, Controlling factors for
478
simultaneous nitrification and denitrification in a two-stage intermittent aeration process
479
treating domestic sewage. Water Res. 33, 961–970.
480
23
Page 23 of 30
480
Table 1. Physical properties of dimensionless groups
Unit
Dimensionless Henry constant for ammonia, HC
Temperature 30 °C
–
0.000560
0.000858
0.00128
Density of liquid, ρL
kg/m3
997.1
995.7
992.3
Liquid viscosity, μL
10-3 kg/(m-s)
0.890
0.798
0.653
Density of gas, ρG
kg/m3
1.1855
Gas viscosity, μG
10-5 kg/(m-s)
1.85
Liquid diffusion coefficient, DL
10-9 m2/s
1.71
of
Standards
us
482
http://webbook.nist.gov/cgi/cbook.cgi?Name=ammonia&Units=SI
and
1.166
1.127
1.87
1.91
1.90
2.41 (NIST),
Ammonia
te
d
Technology
Ac ce p
484
Institute
an
Ref:
M
481
483
National
40 °C
cr
25 °C
ip t
Item
24
Page 24 of 30
ip t
Experimental conditions
Type of liquid flow
pH
VB
QL
QG
(L)
(L/min)
(L/min)
Stripping Air consumption,
us
Equipments
cr
Table 2. A comparison of stripping time, air consumption and KLa in different equipments for ammonia stripping
T
time (h)
(°C)
M an
485
QG/QLb (–) or
KLa (1/h)
QG/VBc (1/min)
η (%)
Continuous
11
0.4
0.05
90
25
0.0037a
1800b
12.3
64
bed
flow
11
0.4
0.05
80
30
0.0037a
1600b
13.9
69
11
0.4
0.05
80
40
0.0037a
1600b
18.4
81
12
0.05
–d
4.5
20
9
96b
0.48
98
Aerocyclone
ce pt
Jet loop reactor
Recirculation 10.5–11 Recirculation Recirculation
11
11–12
Ac
Packed tower
Batch
ed
Rotating packed
Stirrer tank
1000 9
10
e
8.3
Reference
This work
(BasakcilardanKabakci et al., 2007)
25000
15
3.5
b
c
3000 or 25
0.42
75
(Le et al., 2006)
50e
50
20
5.8
5.6c
0.63
97
(Deĝermenci et al., 2012)
–f
114
25
3.5
11.4c
1.2
486
a
487
ratio in terms of QG/VB. dNot applicable. eLiquid circulation rate for reactor type of liquid recirculation. fNot available.
98
(Quan et al., 2009)
As liquid hydraulic retention time (tL) estimated by Chen et al. (2004). bAir consumption ratio in terms of QG/QL. cAir consumption
488
25
Page 25 of 30
To hood
3
11
Liquid in
1
Gas out
T
T
8 9
5
Gas in
D
6
7
us
489
cr
Liquid out
10
4
ip t
2
490
Fig. 1. Schematic diagram of the rotating packed bed (RPB) system for ammonia stripping.
492
Components: 1. hot-plate magnetic-stirrer device; 2. ammonia-storage tank; 3. Pump; 4.
493
Thermocouple; 5. RPB (stainless steel wire mesh); 6. In situ ammonia monitor; 7. Motor; 8. RPB
494
shell; 9. Air flow meter; 10. Air compressor; 11. Neutralization tank.
M
an
491
te Ac ce p
496
d
495
26
Page 26 of 30
0.005
(a) KLa for ω and T at Q G/Q L of 1200
ip t
0.003
cr
KLa (1/s)
0.004
us
0.002
0.001
90
0
500
1000
1500
an
(b) for ω and T at Q G/QL of 1200 80
M
(%)
70 60
d
50
te
40 30
496
Ac ce p
0
500
1000
1500
(rpm)
497
Fig. 2. Effects of ω and T on (a) KLa and (b) η for ammonia stripping at QL of 0.05 L/min and
498
QG/QL = 1200. Symbols: T of 25 (), 30 () and 40 () °C, respectively. Solid lines:
499
Predictions for KLa and η were based on Eqs. 9 and 8, respectively.
500 501
27
Page 27 of 30
0.006
0.006
0.004
0.004
0.003
0.003
0.002
0.002 0.001
0.001 90 400
800
1200
1600
2000 90 400
(b) for and QG/QL at 25 °C
80
70
1200
1600
2000
d) for and Q G/QL at 40 °C
70
60
60
M
(%)
800
an
80
50
50 40
40
20 400
800
te
d
30
1200
QG/QL (-)
1600
30 20 2000 400
800
1200
1600
2000
QG/QL (-)
Ac ce p
501
cr
0.005
us
KLa (s-1)
0.005
ip t
(c) KLa for and QG/Q L at 40 °C
(a) KLa for and Q G/QL at 25 °C
502
Fig. 3. Effects of ω and QG/QL on KLa (a and c) and η (b and d) for ammonia stripping at QL of
503
0.05 L/min and T of 25 (a and b) and 40 (c and d) °C. Symbols: ω = 300 (), 600 () and 1200
504
() rpm. Solid lines: Predictions for KLa and η were based on Eqs. 9 and 8, respectively.
505
28
Page 28 of 30
0.007
(a) K La for Q L and T at
0.006
ip t
0.004
cr
KLa (1/s)
0.005
0.003
us
0.002
0.001 0.04
0.08
0.12
0.04 0.08 Q L (L min-1)
0.12
(b) for Q L and T at
M
90 80
d
70 60
te
(%)
an
100 0
50
Ac ce p
40 30 20
0
505 506
Fig. 4. Effects of QL and T on (a) KLa and (b) η for ammonia stripping at ω of 600 (empty
507
symbol and dot line) and 1200 (solid symbol and line) rpm. Symbols: QG of 75 L/min and T of
508
25 °C ( or ); QG of 60 L/min and T of 30 °C ( or ); QG of 40 L/min and T of 40 °C ( or
509
). Dot and solid lines: Predictions for KLa and η were based on Eqs. 9 and 8, respectively.
510
29
Page 29 of 30
0.008
ip t
0.006
0.004
cr
0.002
us
Predicted KLa (1/s)
(a)
0 0
0.002
0.004
0.008
an
100
(b)
M
80 60 40
d
Predicted (%)
0.006
Experimental KLa (1/s)
te
20
Ac ce p
0
0
510
20
40
60
80
Experimental (%)
100
511
Fig. 5. Diagonal graphs of experimental and predicted (a) KLa and (b) η. Dotted lines:
512
Prediction values within 20%.
513
30
Page 30 of 30
S0957-5820(16)30107-0 http://dx.doi.org/doi:10.1016/j.psep.2016.06.021 PSEP 809
To appear in:
Process Safety and Environment Protection
Received date: Revised date: Accepted date:
27-2-2016 22-4-2016 11-6-2016
Please cite this article as: Yuan, M.-H., Chen, Y.-H., Tsai, J.-Y., Chang, C.Y.,Ammonia removal from ammonia-rich wastewater by air stripping using a rotating packed bed, Process Safety and Environment Protection (2016), http://dx.doi.org/10.1016/j.psep.2016.06.021 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Graphical Abstract
Clear Air
Continuous inflow of Ammonia-rich Wastewater
Neutralizer of Stripping Ammonia T
ip t
1
Continuous Supply of Fresh Air
D
us
2
Clear Effluent
cr
Rotating Packed Bed
3
an
4 5
Ac ce p
te
d
M
6
1
Page 1 of 30
6
Highlights
7
A continuous-flow rotating packed bed (RPB) was evaluated for ammonia stripping.
8
The RPB holds the advantages of small size and short retention time for ammonia stripping.
9
The roles of air and liquid flow rate in mass transfer and stripping efficiencies of RPB were
12
ip t
11
examined.
The continuous-flow RPB is a promising alternative stripping process for ammonia rich wastewater.
cr
10
13
us
14
Ac ce p
te
d
M
an
15
2
Page 2 of 30
Ammonia removal from ammonia-rich wastewater
16
by air stripping using a rotating packed bed
cr
ip t
15
us
17
Min-Hao Yuan,a Yi-Hung Chen,a,* Jhih-Ying Tsaia and Ching-Yuan Changb,c
19
a
Department of Chemical Engineering and Biotechnology, National Taipei University of
b
24 25 26
Taiwan c
Department of Chemical Engineering, National Taiwan University, Taipei 106, Taiwan
Ac ce p
23
Graduate Institute of Environmental Engineering, National Taiwan University, Taipei 106,
te
22
M
Technology, Taipei 106, Taiwan
d
20 21
an
18
27
*Corresponding Author. Tel.: +886 2 27712171 ext. 2539; fax: +886 2 87724328.
28
E-mail: [email protected] (Y.H. Chen)
29
3
Page 3 of 30
KEYWORDS
30
Air stripping; ammonia removal; rotating packed bed; continuous flow; mass transfer coefficient
31
ABSTRACT
32
Air stripping of ammonia from an ammonia-rich stream (1000 mg/L) was performed in a
33
continuous-flow rotating packed bed (RPB) at temperatures from 25 to 40 °C. The effects of the
34
major operating variables (rotational speed (ω), liquid flow rate (QL), gas flow rate (QG), and
35
stripping temperature (T)) on the volumetric liquid mass-transfer coefficient (KLa) and stripping
36
efficiency (η) were elucidated. The results indicate that the RPB exhibits higher mass-transfer
37
performance (12.3–18.4 1/h) compared with those of stripping tanks, packed towers, and other
38
advanced gas–liquid contactors (0.42–1.2 1/h). At QL = 0.05 L/min, QG/QL = 1600, and ω = 1200
39
rpm, η values for the RPB at 30 and 40 °C respectively reached 69% and 81% within 13.3 s. In
40
contrast, conventional ammonia stripping processes with liquid recirculation in larger towers
41
usually take hours to achieve the same values. The proposed dimensionless models describe the
42
relationship between KLa and the major parameters for ammonia stripping in the RPB. KLa
43
showed the greatest increase with increasing QG followed by the increase in QL, ω, and T.
44
However, the operating conditions that would make the technology economically viable and the
45
optimal conditions for efficient ammonia removal must be further studied.
Ac ce p
te
d
M
an
us
cr
ip t
29
46 47
1. Introduction
48
The removal of ammonia from ammonia-rich streams has gained increasing attention in recent
49
years because of the more stringent discharge limits for ammonium nitrogen (NH3-N) that have
50
been steadily imposed on wastewater treatment plants (WWTPs) worldwide. A substantial 4
Page 4 of 30
amount of NH3-N released into bodies of water may cause several problems such as toxicity to
52
sensitive aquatic biota, oxygen depletion, and eutrophication (Camargo et al., 2005; Ding et al.,
53
2014; Rabalais, 2002; Xiao et al., 2015). For these reasons, the US Environmental Protection
54
Agency updated national recommended criteria for water quality for NH3-N discharge (USEPA,
55
2013a). In addition, the Chinese government set out a goal of 10% reduction in total NH3-N in
56
the 12th Five-Year Plan (China MEP, 2012). In 2014, the Taiwan Environmental Protection
57
Administration implemented new restrictions on NH3-N discharge, specifying different
58
maximum values and grace periods for petrochemical, semiconductor, and optoelectronic
59
industries, as well as for science parks that discharge ammonia wastewater (Taiwan EPA, 2014).
60
There is thus an urgent need for improvement of treatment of wastewater containing ammonia by
61
removal or conversion of ammonia into a more stable and fixed form.
M
an
us
cr
ip t
51
The treatment of high-strength ammonia wastewater in typical WWTPs or biological
63
processing is often challenging. NH3-N levels in excess of allowable limits in raw water result in
64
an increase in oxygen demand and interfere with the chlorination and manganese filtration
65
processes, thus impairing the performance of typical WWTPs (Hasan et al., 2011 and 2013).
66
These problems have caused shutdowns of WWTPs, leading to a shortage of domestic water in a
67
certain region of Malaysia (Hasan et al., 2011). Biological processing that combines nitrification
68
and denitrification is the conventional method for ammonia wastewater treatment (Ioannou et al.,
69
2015; Sun et al., 2015; Zhao et al., 1999). However, it requires a relatively long retention time
70
and large spatial requirements compared with those of other methods, making its implementation
71
difficult in existing enterprises that have space limitations (Zhao et al., 1999). Moreover, this
72
technology is sensitive to shock, toxic loads, and cold weather conditions, and it does not allow
73
for the recovery of ammonia.
Ac ce p
te
d
62
5
Page 5 of 30
Ammonia recovery from ammonia-rich wastewater is preferable, as ammonia can be used to
75
produce fertilizer for agricultural use and is thus an additional revenue source for WWTP
76
operators (Zhao et al., 1999). In practice, ammonia can be recovered by ion exchange (Jorgensen
77
and Weatherley, 2003), struvite precipitation (Liu et al., 2013), and air stripping (Basakcilardan-
78
Kabakci et al., 2007; Chang et al., 2013; Cheung et al., 1997; Deĝermenci et al., 2012; Jiang et
79
al., 2014; Kutzer et al., 1995; Laureni et al., 2013; Le et al., 2006; Liu et al., 2015; Zeng et al.,
80
2006; Zhang and Jahng, 2010). Among these techniques, ion exchange requires an extremely low
81
concentration of solids in the effluent to prevent fouling. Struvite precipitation occurs at
82
equimolecular concentrations of Mg2+, NH4+, and PO43− under slightly alkaline conditions.
83
Hence, the addition of a source of Mg2+ or PO43− salts is essential to optimize the struvite
84
crystallization process when wastewater contains less magnesium and phosphate as compared
85
with nitrogen. The final technique, air stripping, is a physical process for ammonia recovery,
86
tolerating some degree of solids and requiring mainly temperature and pH controls. As long as
87
the air temperature and pH remain stable, the operation of ammonia stripping is relatively simple
88
and is unaffected by wastewater fluctuation and toxic loads. However, the lime used for raising
89
the pH of the effluent to 10.8–11.5 often results in unwanted fouling in the packed beds due to
90
calcium carbonate deposition (Kutzer et al., 1995; Liu et al., 2015).
Ac ce p
te
d
M
an
us
cr
ip t
74
91
The air stripping process has been successfully implemented to remove ammonia from
92
different ammonia-rich streams, e.g., pig slurry (Laureni et al., 2013; Zhang and Jahng, 2010),
93
cattle and fermented chicken manure (Jiang et al., 2014; Zeng et al., 2006), human urine
94
(Basakcilardan-Kabakci et al., 2007; Le et al., 2006), landfill leachate (Cheung et al., 1997), and
95
sour water from oil refineries (Chang et al., 2013). The process involves two packed towers for
96
transferring volatile ammonia from wastewater into a gas phase and then converting the
6
Page 6 of 30
ammonia gas with an acid solution to give stable ammonium salts for use as mineral fertilizer.
98
However, the common packed depth of each tower is about 6.1–7.6 m (USEPA, 2013b), which
99
poses a challenge to existing treatment plants that expanding capacity is not feasible due to space
100
limitations. Therefore, efforts to improve the design and miniaturization of ammonia stripping,
101
e.g., the use of jet loops (Deĝermenci et al., 2012) and aerocyclone reactors (Quan et al., 2009),
102
are underway.
cr
ip t
97
Previous studies on the absorption of CO2 (Jassim et al., 2007; Pan et al., 2013 and 2015) and
104
volatile organic compounds (VOCs) (Chen and Liu, 2002; Lin et al., 2003 and 2006), distillation
105
(Kellehe and Fair, 1996; Lin et al., 2002), VOC stripping (Gudena et al., 2012; Lin et al., 2004;
106
Liu et al., 1996; Singh et al., 1992), ozonation (Chen et al., 2004), and esterification (Chen et al.,
107
2010) in a rotating packed bed (RPB) have remarkable potential to reduce the size of the packed
108
tower. This potential is due to the intensified mass transfer at the gas–liquid interface generated
109
in a RPB via high centrifugal force. The high centrifugal force (300–10000 m/s2) is usually 1–3
110
orders of magnitude greater than gravitational acceleration (Rao et al., 2004). An RPB using
111
basic oxygen furnace (BOF) slag can remove approximately 96–99% of the CO2 in flue gas
112
stream with 30% CO2 within a short reaction time (1 min) at 25 °C and 1 atm (Pan et al., 2013).
113
The conventional stripper height necessary to achieve a performance similar to that of a RPB for
114
CO2 absorption is greater by a factor of 8.4, while the diameter is greater by a factor of 11.3
115
(Jassim et al., 2007). Hence, the results suggest a significant reduction in equipment size and
116
space with use of the RPB. Trichloroethylene (TCE) removal from groundwater at low gas/liquid
117
ratios, a preferred outcome that is difficult to obtain in conventional packed beds, has been
118
clearly enhanced by air stripping by a RPB (Gudena et al., 2012). Moreover, the total capital cost
119
of the RPB can be significantly reduced according to the estimate for TCE stripping. In
Ac ce p
te
d
M
an
us
103
7
Page 7 of 30
120
summary, the unique features of the RPB, such as high volumetric gas–liquid mass-transfer
121
coefficients (KLa), reduced tendency for fouling, and reduced equipment size and cost, can
122
potentially alleviate the limitations associated with conventional ammonia stripping. Although ammonia stripping is technically feasible for the removal of ammonia from various
124
ammonia-rich streams, there is still an urgent need for a reduction in equipment size and space
125
for its implementation in existing enterprises due to the more stringent discharge limits for NH3-
126
N in wastewater. Yuan et al. (2016) had showed that the remarkable removal efficiencies and
127
small height of transfer units for the ammonia stripping were obtained from both laboratory and
128
pilot scale RPB systems at ambient temperature. This study follows up with the previous work
129
and shows the effects of temperature for ammonia stripping in a high-voidage RPB system. The
130
evaluation of various operating conditions should identify the critical factors that affect the KLa
131
values and stripping efficiency (η). High-voidage packing of stainless steel wire mesh was used
132
to provide a low pressure drop and to reduce the tendency for fouling (Lin et al., 2004). Effects
133
of the major operating variables (rotational speed (ω), liquid flow rate (QL), the ratio of flow rate
134
of air (QG) to QL, and temperature (T)) on KLa and η values were investigated and elucidated.
135
Dimensionless equations are proposed to predict KLa and η values for ammonia stripping. These
136
results provide useful information for the practical design of ammonia stripping using RPBs at
137
various operating conditions.
139
cr
us
an
M
d
te
Ac ce p
138
ip t
123
2. Materials and methods
140
2.1. Apparatus and operating conditions
141
A schematic diagram of ammonia stripping using the RPB is shown in Fig. 1. The system
142
consists of the RPB, gas and liquid feed controls, analyzer of effluents, and neutralizer of
8
Page 8 of 30
effluent gas. The inner radius (ri), outer radius (r0), and axial height (zB) of the packed-bed
144
rotator were 0.0355, 0.0850, and 0.0215 m, respectively. The thickness of the RPB shell and the
145
distance from the side of RPB to the shell were 0.008 and 0.011 m, respectively. The reactor
146
volume (VB) of the RPB was 410−4 m3. It was used as an estimate of the volume of the packed-
147
bed rotator, π(r02 − ri2)zB. The bed rotation was controlled by a motor, and ω was varied from 300
148
to 1200 rpm. This resulted in a centrifugal force equivalent to 6.25 to 100 gravitational
149
acceleration. The packing material was made of grade 304 stainless-steel wires in the shape of
150
annular rings, which comprise interconnected wires with a mean diameter of 0.22 mm. Stainless
151
steel has critical surface tension higher than that of glass, ceramic, and acrylic, thus facilitating
152
gas-liquid mass transfer in the RPB (Chen et al., 2006). High-voidage packing (0.954 m3/m3) of
153
the wire mesh was used to lower the pressure drop and to reduce the tendency for fouling (Singh
154
et al., 1992).
d
M
an
us
cr
ip t
143
Aqueous ammonia was pumped into the RPB at a controlled feed rate after the solution
156
concentration was adjusted to 1000 mg/L with ammonium chloride (99%, Jin Yih Chemical Co.,
157
Ltd., Taipei, Taiwan) and distilled water, and the pH value was adjusted to 11 with sodium
158
hydroxide solution (45%, Jin Yih Chemical Co., Ltd., Taipei, Taiwan). The ammonia in the RPB
159
was then evaporated rapidly through acceleration by centrifugal force and by mixing with air
160
flow to achieve a high η value.
Ac ce p
te
155
161
For the experiments, a diaphragm liquid pump (KNF 1.300 TT 18S, KNF, Sursee,
162
Switzerland) was used to constantly introduce ammonia solution into the housing at a QL of
163
0.025–0.1 L/min. An air compressor (DT-175-2C, Tong Cheng Iron Works Co., Ltd., Taipei,
164
Taiwan) and an air flow meter (Dwyer Instruments Inc., Michigan City, IN, USA) were used to
165
control QG to within 30–60 L/min. The QG/QL ratio was set between 750 and 1800. The
9
Page 9 of 30
temperatures of the solution and the RPB with the ammonia solution were set to 25, 30, and 40
167
°C, by using a hot plate and magnetic stirrer (C-MAG HS7 S1, IKA Lab Equipment, Staufen,
168
Germany). Similarly, Teflon heating tape (Kwo-Yi Co., Taipei, Taiwan) was used for the RPB
169
housing and type-K thermocouples (TES-1306, TES Electrical Electronic Corp., Taipei,
170
Taiwan).
ip t
166
cr
171 2.2. Analytical methods
173
The ammonia concentrations of the inlet and outlet solutions were determined by using an ion-
174
selective electrode (HI4101, Hanna Instruments, Woonsocket, RI, USA). A pH meter (HI1131B,
175
Hanna Instruments, Woonsocket, RI, USA) was used to measure the pH value of the ammonia
176
solution. To neutralize the effluent gas, the vaporized or purged ammonia gas was introduced
177
into a sulfuric acid solution and converted to ammonium sulfate (Sigma-Aldrich, St. Louis, MO,
178
USA) before it was expelled as exhaust. For each reaction condition, more than 15 min of
179
operation was allowed for the system to reach a steady state prior to assessment.
an
M
d
te
Ac ce p
180
us
172
181
2.3. Theoretical considerations
182
In this study, η and overall KLa values were used to assess the performance of the RPB for
183
ammonia stripping. The η value was computed from the ammonia concentration in the feed
184
(CL,in) and effluent (CL,out) liquid streams.
185
C 1 L ,out C L ,in
100%
(1)
186
In the centrifugal vapor–liquid contactor, KLa was calculated from mass-balance concepts and
187
from the two-film theory by Singh (1989), Singh et al. (1992), and Chen et al. (2005a and 2005b).
10
Page 10 of 30
188
The mass-balance equation for a thin annulus section within the RPB at steady state is as
189
follows:
192
(2)
The concentration at gas-liquid equilibrium (CL*) is assumed to follow Henry’s law: C*L
ip t
191
Q L dC L K L a C*L C L 2rz Bdr
CG HC
(3)
cr
190
where HC is the dimensionless Henry constant for ammonia. HC values are given in Table 1.
194
According to the two-film theory, the variable gas-phase concentration is expressed as
198
an
(4)
1 C 1 ln 1 L ,in Q S C L ,out S K La L 1 VB 1 S
M
Solving Eqs. 2–4 simultaneously gives the following equation for computing KLa:
d
197
QL C L CL,in QG
(5)
te
196
CG
where S is the stripping factor and VB is the reactor volume of the RPB:
Ac ce p
195
us
193
QG QL
199
S HC
200
VB r02 ri2 z B
(6)
(7)
201
VB is approximately 410−4 m3. Notably, the overall liquid-to-gas mass transfer consists of
202
resistances associated with the stagnant liquid and gas interfacial films. Therefore, the resistances
203
of both the liquid and the gas films have been taken into consideration in the calculation of KLa.
204
On the basis of KLa, prediction of η using Eq. 1 may be modified further as follows:
11
Page 11 of 30
(8)
ip t
205
1 1 S 100% 1 VB 1 1 1 exp K L a Q L S S
206 3. Results and discussion
cr
207
3.1. The effects of ω and T
209
Fig. 2 shows the dependence of KLa and η on ω and T under operation at QL = 0.05 L/min and
210
QG/QL = 1200. As seen in Fig. 2a, KLa increased as ω increased from 300 to 1200 rpm. The KLa
211
value was 0.00257 1/s at ω = 1200 rpm and T = 25 °C, which is 1.5 times higher than that at ω =
212
300 rpm (0.00167 1/s). According to the two-film model, the resistance to transfer in each phase
213
is localized in the thin films at the interface. Packing in the RPB via centrifugal force under high
214
ω facilitates division of the ammonia solution into smaller droplets and thinner films. The
215
consequent lower resistance at the stagnant liquid film results in a higher KLa value. KLa values
216
were 0.00257, 0.00326, and 0.00404 1/s at ω = 1200 rpm at T values of 25, 30, and 40 °C,
217
respectively. The observed dependence of KLa on T is predicted by the increase in Henry’s law
218
constants for ammonia with T (25–40 °C in this case), which caused more ammonia to evaporate
219
easily into the gas phase.
Ac ce p
te
d
M
an
us
208
220
As shown in Fig. 2b, the increase in ω and T in terms of η also enhanced ammonia removal.
221
According to the theory, a higher η is obtained by increasing KLa at the same stripping time. The
222
stripping time is the liquid hydraulic retention time (tL = VL/QL or LVB/QL) in the RPB, where
223
VL and L are the liquid hold-up and relative liquid hold-up, respectively (Chen et al., 2004). In
224
the experimental conditions in the present study, t L ranged from 11 to 20 s, as determined by the
225
calculation of L (0.03–0.06) based on the empirical equation proposed by Chen et al. (2004). At 12
Page 12 of 30
226
QL = 0.05 L/min, QG/QL = 1600, and ω =1200 rpm, only 13.3 s was required to achieve 69% and
227
81% of η in the RPB at 30 and 40 °C, respectively. Table 2 lists conventional and advanced processes for stripping ammonia from wastewater
229
according to their reaction conditions and performance (Basakcilardan-Kabakci et al., 2007;
230
Deĝermenci et al., 2012; Le et al., 2006; Quan et al., 2009). In contrast to the other processes, the
231
continuous-flow RPB does not require liquid recirculation to extend the stripping time in order to
232
achieve the desired η value. As seen in Table 2, mass transfer (0.00342–0.00510 1/s or 12.3–18.4
233
1/h) was high and had a very short stripping time as compared with other stripping processes
234
such as those in the stirrer tank, packed tower, jet loop, and aerocyclone reactors. These
235
processes often take hours and liquid recirculation to achieve a similar level of η value.
an
us
cr
ip t
228
It is noted that the longer stripping time was required since the single-stage RPB did not
237
achieve the discharge standards (e.g. 20 mg/L-NH3-N) for wastewater with high ammonia
238
concentration (as this study of 1000 mg/L). Hence, a large size of RPB or a series configuration
239
of continuous-flow RPB was suggested for the on-site application. On the basis of current
240
experimental results, the effluent concentration of NH3 can be reduced from 1000 mg/L to 16.8
241
mg/L at 25 °C by a series of four-stage RPB and to 6.8 mg/L at 40 °C by a series of three-stage
242
RPB. Consequently, the stripping time for the satisfied effluent concentration was three to four
243
times of the single one, which was smaller than the required time with other stripping processes
244
by two orders of magnitude.
Ac ce p
te
d
M
236
245 246
3.2. The effect of QG
247
Fig. 3 shows the effect of the QG/QL ratio (or QG) and ω on KLa and η values at QL = 0.05
248
L/min and T = 25 and 40 °C. The results indicate that a higher QG/QL ratio or a higher QG value
13
Page 13 of 30
at a given QL value increases KLa and η for ammonia stripping by multiplying the stripping
250
factor (S). It is noted that the S is the ratio of the slope of equilibrium line and the slope of the
251
operating line (Liu et al., 2015). The increasing supply of continuous flow of fresh air in the RPB
252
facilitated the transfer of ammonia molecules from the water to the air, thus resulting in larger
253
values of S and transfer units. At T = 25 °C and ω = 1200 rpm, the KLa value at a QG/QL ratio of
254
1800 (0.00342 1/s) was 1.8 times higher than that at a QG/QL ratio of 750 (0.00186 1/s), while
255
the η value at a QG/QL ratio of 1800 (64.3%) was 1.7 times higher than that at a QG/QL ratio of
256
400 (36.8%). A similar trend was observed at high T (40 °C) and at ω = 1200 rpm. The KLa
257
values were 0.00227 and 0.00510 1/s at QG/QL ratios of 400 and 1600, respectively; likewise, the
258
η values were 39.5% and 81.3% at QG/QL ratios of 400 and 1600, respectively. Although a
259
higher QG/QL ratio or QG also resulted in a shorter time of contact between the liquid and gas
260
currents, introduction of abundant air apparently enhanced the mass transfer of ammonia by
261
increasing the refresh rate of the gas–liquid interface. Consequently, a shorter exposure time was
262
required to attain the desired high efficiency of the RPB in ammonia stripping.
cr
us
an
M
d
te
Ac ce p
263
ip t
249
264
3.3. The effect of QL
265
The effect of QL (0.02–0.1 L/min) on KLa and η values at QG values of 40–75 L/min (T = 25–
266
40 °C and ω = 1200 rpm) is shown in Fig. 4. The results differ from the effects of QG, ω, and T
267
on KLa and η. The η value decreased with QL, in contrast to the tendency of KLa with QL (Fig.
268
4a). Changes in KLa due to these operational factors are normally directly reflected in the
269
separation efficiency. At QG = 40 L/min, ω = 1200 rpm, and T = 40 °C, the experimental KLa
270
increased by 2.2 times, from 0.00256 to 0.00583 1/s, while the QL value increased four times,
271
from 0.025 to 0.1 L/min. In contrast, the η value decreased from 81.4% to 41.8% when the QL
14
Page 14 of 30
value increased from 0.025 to 0.1 L/min. The results indicate that the increased QL leads to
273
considerable compensatory effects due to the decrease in tL and S, resulting in a decrease in η for
274
ammonia stripping. Note that the tL decreased from 17.2 to 10.4 s when QL increased from 0.025
275
to 0.1 L/min at the same level of QG = 40 L/min, ω = 1200 rpm, and T = 40 °C.
ip t
272
276 3.4. Prediction models for KLa and η
278
In order to explore the effect of the operating variables QL, QG, ω, and T on KLa within the
279
range of experimental conditions, this study used the empirical dimensionless model that has
280
been proposed for the prediction of KLa in ethanol stripping (Kelleher and Fair, 1996), VOC
281
absorption (Chen and Liu, 2002), and ozonation (Chen et al., 2004) processes in the RPB system.
282
The dimensionless parameters of the model include the Reynolds numbers of the liquid (ReL)
283
and gas (ReG), the Grashof number of the liquid based on the average bed radius (GrL,avg), and
284
the dimensionless Henry constant for ammonia (HC). All of these parameters may be computed
285
from operating variables and other physical properties. The correlation model for KLa is obtained
286
by using the observational data in Figs. 2–4 (R2 = 0.8075):
us
an
M
d
te
Ac ce p
287
cr
277
K L ad p D La p
0.024 Re L
0.352
Re G
0.613
GrL,avg
0.198
HC
0.159
(9)
288
where dP is the stainless-steel-wire diameter of the packing (0.22 m), ap is the specific area of
289
packing per unit volume of the packed bed (840 m2/m3), and DL is the molecular liquid diffusion
290
coefficient of ammonia. ReL, ReG, and GrL,avg are defined as follows:
291
r L Q L ln o ri Re L 2Z B ro ri a P L
(10)
15
Page 15 of 30
292
r G Q G ln o ri Re G 2Z B ro ri a P G
(11)
ravg2 ravg ri
3
GrL , avg
L
(12)
2
ip t
293
The values of the physical properties of the dimensionless groups are shown in Table 1. The
295
range of the dimensionless groups in the empirical model is 0.12 < KLadP/DLaP < 0.85, 0.10 <
296
ReL < 0.40, 3.11 < ReG < 15.23, 1.35 × 109 < GrL,avg < 4.27 × 1010, and 6.96 × 10−4 < HC < 1.28 ×
297
10−4. As shown in Eq. 9, KLa is more sensitive to changes in ReG than to changes in ReL, GrL,avg,
298
and HC, in accordance with its power. Hence, the strong enhancement in mass transfer through
299
the supply of fresh air is attributable to the fast refresh rate, which directly increases the QG value
300
within a short tL.
M
an
us
cr
294
Figs. 5a and 5b illustrate the parity of the experimental and predicted values of KLa and η,
302
respectively. Note that the determination of η is based on the substitution of the predicted KLa
303
values into Eq. 8. As can be seen, most of the experimental values of KLa and η lie within 20%
304
of their predicted values. For comparison, the predicted curves are also plotted in Figs. 2–4.
305
Except for the underestimated KLa values at QL (0.1 L/min), ω (1200 rpm), and T (30 and 40 °C)
306
values in Fig. 4a, the experimental and predicted values of KLa and η indicate a good degree of
307
fit for the dimensionless groups ReL, ReG, GrL,avg, and HC.
309
te
Ac ce p
308
d
301
4. Conclusions
310
In summary, the continuous-flow RPB used in this study appears to be very efficient at air
311
stripping of ammonia from an ammonia-rich stream (1000 mg/L). The overall KLa values are
312
12.3–18.4 1/h, which are significantly higher than the values (0.42–1.2 1/h) obtained by using
16
Page 16 of 30
stripping tanks, packed towers, and other advanced gas–liquid contactors. For industrial
314
application of wastewater with high ammonia concentration, a large size of RPB or a series
315
configuration of continuous-flow RPB can be used in order to satisfy the more stringent
316
discharge standards. The results of the empirical dimensionless model show that QG and QL are
317
the most sensitive ammonia stripping parameters for liquid–gas mass transfer in the RPB, while
318
ω and T are less-sensitive parameters.
cr
ip t
313
us
319 Acknowledgment
321
The authors gratefully acknowledge the financial support provided by the Ministry of Science
322
and Technology of Taiwan (102-2622-E-027-024-CC3).
M
an
320
323 Abbreviations
325
aP
Specific area of packing per unit volume of a packed-bed, m2/m3
326
CL,in
Liquid concentration of ammonia of inlet liquid, mg/L
327
CL,out
328
DL
329
dP
330
g
331
HC
332
KLa
Overall liquid volumetric mass transfer coefficient, 1/s
333
ravg
Average radius of a packed bed, (ri + ro)/2, m
334
ri
Inner radius of a packed bed, m
Ac ce p
te
d
324
Liquid concentration of ammonia of outlet liquid, mg/L Molecular liquid diffusion coefficient of ammonia Stainless wire diameter of bed, 0.22 m Gravitational acceleration, m/s2 Dimensionless Henry’s law constant of ammonia
17
Page 17 of 30
Outer radius of a packed bed, m
336
RPB
Rotating packed bed
337
S
Stripping factor, HCQG/QL
338
tL
liquid hydraulic retention time, VL/QL or LVB/QL, s, min or h
339
T
Stripping temperature, °C
340
TCE
Trichloroethylene
341
VOCs
Volatile organic compounds
342
QG
Gas flow rate, L/min
343
QG/QL
Ratio of gas and liquid flow rate, -
344
QL
Liquid flow rate, L/min
345
VB
Volume of a packed bed, π(ro2 – ri2)ZB, m3
346
VL
Liquid hold-up, m3
347
WWTPs
Wastewater treatment plants
348
ZB
Axial height of a packed bed, m
us an
M
d
te
Ac ce p
349
ip t
ro
cr
335
350
Greek Symbols
351
L
352
μG
353
μL
Liquid viscosity
354
L
Dynamic liquid viscosity
355
π
Circular ratio
Relative liquid hold-up Gas viscosity
18
Page 18 of 30
356
ρG
Density of gas
357
ρL
Density of liquid
358
ω
Rotational speed, rpm or rad/s (for Gravg calculation)
ip t
359 Dimensionless Groupings
361
Gravg
Grashof number of the liquid based on the average bed radius, ravgω2(ravg – ri)3/νL2
362
ReG
Reynolds number of the gas, ρGQGln(ro/ri)/[2πZB(ro – ri)apμG]
363
ReL
Reynolds number of the liquid, ρLQLln(ro/ri)/[2πZB(ro – ri)apμL]
an
us
cr
360
364 References
366
Basakcilardan-Kabakci, S., Ipekoglu, A.N., Talini, I., 2007, Recovery of ammonia from human
369
d
Camargo, J.A. Alonso, A., Salamanca, A., 2005, Nitrate toxicity to aquatic animals: A review
te
368
urine by stripping and absorption. Environ. Eng. Sci. 24, 615–624.
with new data for freshwater invertebrates. Chemosphere 58, 1255–1267.
Ac ce p
367
M
365
370
Chang, S.Y., Tu, M.Y., Hsiao, A.F., Tsao, L.K., Huang, T.L., 2013, A real case of ammonia and
371
nitrate removal improvement in refinery wastewater treatment. J. Petrol. 49, 91–110. (In
372
Chinese)
373 374
Chen, Y.S., Liu, H.S., 2002, Absorption of VOCs in a rotating packed bed. Ind. Eng. Chem. Res. 41, 1583–1588.
375
Chen, Y.H., Chang, C.Y., Su, W.L., Chen, C.C., Chiu, C.Y., Yu, Y.H., Chiang, P.C., Chiang,
376
S.I.M., 2004, Modeling ozone contacting process in a rotating packed bed. Ind. Eng. Chem.
377
Res. 43, 228–236.
378 379
Chen, Y.S., Lin, C.C., Liu, H.S., 2005a, Mass transfer in a rotating packed bed with viscous Newtonian and non-Newtonian fluids. Ind. Eng. Chem. Res. 44, 1043–1051.
19
Page 19 of 30
383 384 385 386 387 388
Chen, Y.S., Lin, F.Y., Lin, C.C., Tai, C.Y.D., Liu, H.S., 2006, Packing characteristics for mass transfer in a rotating packed bed. Ind. Eng. Chem. Res. 45, 6846–6853.
ip t
382
of the bed. Ind. Eng. Chem. Res. 44, 7868–7875.
Chen, Y.H., Wang, L.C., Tsai, C.H., Shang, N.C., 2010, Continuous-flow esterification of free fatty acids in a rotating packed bed. Ind. Eng. Chem. Res. 49, 4117–4122.
cr
381
Chen, Y.S., Lin, C.C., Liu, H.S., 2005b, Mass transfer in a rotating packed bed with various radii
Cheung, K.C., Chu, L.M., Wong, M.H., 1997, Ammonia stripping as a pretreatment for landfill leachate. Water Air Soil. Pollut. 94, 209–221.
us
380
China Ministry of Environmental Protection (MEP), 2012, MEP explains the National 12th FiveYear
390
http://english.mep.gov.cn/News_service/infocus/201202/t20120207_223194.htm
392
for
environmental
protection,
China
MEP,
Beijing,
China.
Deĝermenci, N., Ata, O.N., Yildiz, E., 2012, Ammonia removal by air stripping in a semi-batch
M
391
Plan
an
389
jet loop reactor. J. Ind. Eng. Chem. 18, 399–404.
Ding, L.J., An, X.L., Li, S., Zhang, G.L., Zhu, Y.G., 2014 Nitrogen loss through anaerobic
394
ammonium oxidation coupled to iron reduction from paddy soils in a chronosequence.
395
Environ. Sci. Technol. 48, 10641−10647.
397
te
Ac ce p
396
d
393
Gudena, K., Rangaiah, G.P., Lakshminarayanan, S., 2012, Optimal design of a rotating packed bed for VOC stripping from contaminated groundwater. Ind. Eng. Chem. Res. 51, 835–847.
398
Hasan, H.A., Abdullah, S.R.S., Kamarudin, S.K., Kofli, N.T., 2011, Problems of ammonia and
399
manganese in Malaysian drinking water treatments. World Appl. Sci. J. 12, 1890–1896.
400
Hasan, H.A., Abdullah, S.R.S., Kamarudin, S.K., Kofli, N.T., 2013, On-off control of aeration
401
time in the simultaneous removal of ammonia and manganese using a biological aerated
402
filter system. Process Saf. Environ. 91, 415–422.
403
Ioannou, L.A., Li Puma, G., Fatta-Kassinos, D., 2015, Treatment of winery wastewater by
404
physicochemical, biological and advanced processes: A review. J. Hazard. Mater. 286, 343–
405
368.
20
Page 20 of 30
406 407
Jorgensen, T.C., Weatherley, L.R., 2003, Ammonia removal from wastewater by ion exchange in the presence of organic contaminants. Water Res. 37, 1723–1728. Liu, Y.H., Kumar, S., Kwag, J.H., Ra, C.S., 2013, Magnesium ammonium phosphate formation,
409
recovery and its application as valuable resources: A review. J. Chem. Technol. Biot. 88,
410
181–189.
ip t
408
Liu, B., Giannis, A., Zhang, J., Chang, V.W.C., Wang, J.Y., 2015, Air stripping process for
412
ammonia recovery from source-separated urine: Modeling and optimization. J. Chem.
413
Technol. Biot. 90, 2208–2217.
us
cr
411
Laureni, M., Palatsi, J., Llovera, M., Bonmati, A., 2013, Influence of pig slurry characteristics on
415
ammonia stripping efficiencies and quality of the recovered ammonium-sulfate solution. J.
416
Chem. Technol. Biot. 88, 1654–1662.
an
414
Jassim, M.S., Rochelle, G., Eimer, D., Ramshaw, C., 2007, Carbon dioxide absorption and
418
desorption in aqueous monoethanolamine solutions in a rotating packed bed. Ind. Eng.
419
Chem. Res. 46, 2823–2833.
d
M
417
Jiang, A., Zhang, T., Zhao, Q.B., Li, X., Chen, S., Frear, C.S., 2014, Evaluation of an integrated
421
ammonia stripping, recovery, and biogas scrubbing system for use with anaerobically
422
digested dairy manure. Biosystems Eng. 119, 117–126.
424 425 426 427 428 429 430
Ac ce p
423
te
420
Kelleher, Y., Fair, J.R., 1996, Distillation studies in a high-gravity contactor. Ind. Eng. Chem. Res. 12, 4646–4655.
Kutzer, S., Wintrich, H., Mersmann, A., 1995, Air stripping – A method for treatment of wastewater. Chem. Eng. Technol. 18, 149–155. Le, L., Wang, H.W., Lu, H.H., 2006, Nitrogen removal using air stripping tower in urban wastewater treatment plant. China Water Wastewater 22, 92–95. Lin, C.C., Ho, T.J., Liu, W.T., Liu, H.S., 2002, Distillation in a rotating packed bed. J. Chem. Eng. Jpn. 35, 1298–1304.
21
Page 21 of 30
432 433 434
Lin, C.C., Liu, W.T., Tan, C.S., 2003, Removal of carbon dioxide by absorption in a rotating packed bed. Ind. Eng. Chem. Res. 42, 2381–2386. Lin, C.C., Wei, T.Y., Liu, W.T., Shen, K.P., 2004, Removal of VOCs from gaseous streams in a high-voidage rotating packed bed. J. Chem. Eng. Jpn. 37, 1471–1477.
ip t
431
Lin, C.C., Wei, T.Y., Hsu, S.K., Liu, W.T., 2006, Performance of a pilot-scale cross-flow
436
rotating packed bed in removing VOCs from waste gas streams. Sep. Purif. Technol. 52,
437
274–279.
439
Liu, H.S., Lin, C.C., Wu, S.C., Hsu, H.W., 1996, Characteristics of a rotating packed bed. Ind.
us
438
cr
435
Eng. Chem. Res. 35, 3590–3596.
Pan, S.Y., Chiang, P.C., Chen, Y.H., Tan, C.S., Chang, E.E., 2013, Ex situ CO2 capture by
441
carbonation of steelmaking slag coupled with metalworking wastewater in a rotating packed
442
bed. Environ. Sci. Technol. 47, 3308–3315.
M
an
440
Pan, S.Y., Chen, Y.H., Chen, C.D., Shan, A.L., Lin, M., Chiang, P.C., 2015, High-gravity
444
carbonation process for enhancing CO2 fixation and utilization exemplified by the
445
steelmaking industry. Environ. Sci. Technol. 49, 12380–12387.
te
447
Quan, X., Wang, F., Zhao, Q., Zhao, T., Xiang, J., 2009, Air stripping of ammonia in a watersparged aerocyclone reactor. J. Hazard. Mater. 170, 983–988.
Ac ce p
446
d
443
448
Rabalais, N.N., 2002, Nitrogen in aquatic ecosystems. Ambio 31, 102–112.
449
Rao, D.P., Bhowal, A., Goswami, P.S., 2004, Process intensification in rotating packed beds
450
(HIGEE): An appraisal. Ind. Eng. Chem. Res. 43, 1150–1162.
451
Singh, S.P., 1989, Air stripping of volatile organic compounds from groundwater: An evaluation
452
of a centrifugal vapor-liquid contactor. Ph.D. Dissertation, University of Tennessee,
453
Knoxville, TN.
454
Singh, S.P., Wilson, J.H., Counce, R.M., Villiers-Fisher, J.F., Jennings, H.L., Lucero, A.J., Reed,
455
G.D., Ashworth, R.A., Elliott, M.G., 1992, Removal of volatile organic compounds from
456
groundwater using a rotary air stripper. Ind. Eng. Chem. Res. 31, 574–580.
22
Page 22 of 30
457
Sun, F., Sun, B., Hu, J., He, Y., Wu, W., 2015, Organics and nitrogen removal from textile
458
auxiliaries wastewater with A2O-MBR in a pilot-scale. J. Hazard. Mater. 286, 416–424.
459
Taiwan Environmental Protection Administration (EPA), 2014, Effluent standards for chemical industry announced. Environ. Policy Mon. 17, 6–7.
ip t
460
United States Environmental Protection Agency (EPA), 2013, Aquatic life ambient water quality
462
criteria for ammonia – Freshwater; EPA 822-R-13-001, USEPA, Washington, DC.
463
http://water.epa.gov/scitech/swguidance/standards/criteria/aqlife/ammonia/upload/aquatic-
464
life-ambient-water-quality-criteria-for-ammonia-freshwater-2013.pdf
us
465
cr
461
USEPA, 2013, Wastewater technology fact sheet: Ammonia stripping, USEPA, Washington, DC. http://water.epa.gov/scitech/wastetech/upload/2002_06_28_mtb_ammonia_stripping.
467
469
Xiao, Y., Araujo, C.D., Sze, C.C., Stuckey, D.C., 2015, Toxicity measurement in biological
M
468
an
466
wastewater treatment processes: A review. J. Hazard. Mater. 286, 15−29. Yuan, M.H., Chen, Y.H., Tsai, J.Y., Chang, C.Y., 2016, Removal of ammonia from wastewater by
471
air stripping process in laboratory and pilot scales using a rotating packed bed at ambient
472
temperature. J. Taiwan Inst. Chem. Eng. 60, 488–495.
474 475 476
te
Zeng, L., Mangan, C., Li, X., 2006, Ammonia recovery from anaerobically digested cattle
Ac ce p
473
d
470
manure by steam stripping. Water Sci. Technol. 54, 137–145. Zhang, L., Jahng, D., 2010, Enhanced anaerobic digestion of piggery wastewater by ammonia stripping: Effects of alkali types. J. Hazard. Mater. 182, 536–543.
477
Zhao, H.W., Mavinic, D.S., Oldham, W.K., Koch, F.A., 1999, Controlling factors for
478
simultaneous nitrification and denitrification in a two-stage intermittent aeration process
479
treating domestic sewage. Water Res. 33, 961–970.
480
23
Page 23 of 30
480
Table 1. Physical properties of dimensionless groups
Unit
Dimensionless Henry constant for ammonia, HC
Temperature 30 °C
–
0.000560
0.000858
0.00128
Density of liquid, ρL
kg/m3
997.1
995.7
992.3
Liquid viscosity, μL
10-3 kg/(m-s)
0.890
0.798
0.653
Density of gas, ρG
kg/m3
1.1855
Gas viscosity, μG
10-5 kg/(m-s)
1.85
Liquid diffusion coefficient, DL
10-9 m2/s
1.71
of
Standards
us
482
http://webbook.nist.gov/cgi/cbook.cgi?Name=ammonia&Units=SI
and
1.166
1.127
1.87
1.91
1.90
2.41 (NIST),
Ammonia
te
d
Technology
Ac ce p
484
Institute
an
Ref:
M
481
483
National
40 °C
cr
25 °C
ip t
Item
24
Page 24 of 30
ip t
Experimental conditions
Type of liquid flow
pH
VB
QL
QG
(L)
(L/min)
(L/min)
Stripping Air consumption,
us
Equipments
cr
Table 2. A comparison of stripping time, air consumption and KLa in different equipments for ammonia stripping
T
time (h)
(°C)
M an
485
QG/QLb (–) or
KLa (1/h)
QG/VBc (1/min)
η (%)
Continuous
11
0.4
0.05
90
25
0.0037a
1800b
12.3
64
bed
flow
11
0.4
0.05
80
30
0.0037a
1600b
13.9
69
11
0.4
0.05
80
40
0.0037a
1600b
18.4
81
12
0.05
–d
4.5
20
9
96b
0.48
98
Aerocyclone
ce pt
Jet loop reactor
Recirculation 10.5–11 Recirculation Recirculation
11
11–12
Ac
Packed tower
Batch
ed
Rotating packed
Stirrer tank
1000 9
10
e
8.3
Reference
This work
(BasakcilardanKabakci et al., 2007)
25000
15
3.5
b
c
3000 or 25
0.42
75
(Le et al., 2006)
50e
50
20
5.8
5.6c
0.63
97
(Deĝermenci et al., 2012)
–f
114
25
3.5
11.4c
1.2
486
a
487
ratio in terms of QG/VB. dNot applicable. eLiquid circulation rate for reactor type of liquid recirculation. fNot available.
98
(Quan et al., 2009)
As liquid hydraulic retention time (tL) estimated by Chen et al. (2004). bAir consumption ratio in terms of QG/QL. cAir consumption
488
25
Page 25 of 30
To hood
3
11
Liquid in
1
Gas out
T
T
8 9
5
Gas in
D
6
7
us
489
cr
Liquid out
10
4
ip t
2
490
Fig. 1. Schematic diagram of the rotating packed bed (RPB) system for ammonia stripping.
492
Components: 1. hot-plate magnetic-stirrer device; 2. ammonia-storage tank; 3. Pump; 4.
493
Thermocouple; 5. RPB (stainless steel wire mesh); 6. In situ ammonia monitor; 7. Motor; 8. RPB
494
shell; 9. Air flow meter; 10. Air compressor; 11. Neutralization tank.
M
an
491
te Ac ce p
496
d
495
26
Page 26 of 30
0.005
(a) KLa for ω and T at Q G/Q L of 1200
ip t
0.003
cr
KLa (1/s)
0.004
us
0.002
0.001
90
0
500
1000
1500
an
(b) for ω and T at Q G/QL of 1200 80
M
(%)
70 60
d
50
te
40 30
496
Ac ce p
0
500
1000
1500
(rpm)
497
Fig. 2. Effects of ω and T on (a) KLa and (b) η for ammonia stripping at QL of 0.05 L/min and
498
QG/QL = 1200. Symbols: T of 25 (), 30 () and 40 () °C, respectively. Solid lines:
499
Predictions for KLa and η were based on Eqs. 9 and 8, respectively.
500 501
27
Page 27 of 30
0.006
0.006
0.004
0.004
0.003
0.003
0.002
0.002 0.001
0.001 90 400
800
1200
1600
2000 90 400
(b) for and QG/QL at 25 °C
80
70
1200
1600
2000
d) for and Q G/QL at 40 °C
70
60
60
M
(%)
800
an
80
50
50 40
40
20 400
800
te
d
30
1200
QG/QL (-)
1600
30 20 2000 400
800
1200
1600
2000
QG/QL (-)
Ac ce p
501
cr
0.005
us
KLa (s-1)
0.005
ip t
(c) KLa for and QG/Q L at 40 °C
(a) KLa for and Q G/QL at 25 °C
502
Fig. 3. Effects of ω and QG/QL on KLa (a and c) and η (b and d) for ammonia stripping at QL of
503
0.05 L/min and T of 25 (a and b) and 40 (c and d) °C. Symbols: ω = 300 (), 600 () and 1200
504
() rpm. Solid lines: Predictions for KLa and η were based on Eqs. 9 and 8, respectively.
505
28
Page 28 of 30
0.007
(a) K La for Q L and T at
0.006
ip t
0.004
cr
KLa (1/s)
0.005
0.003
us
0.002
0.001 0.04
0.08
0.12
0.04 0.08 Q L (L min-1)
0.12
(b) for Q L and T at
M
90 80
d
70 60
te
(%)
an
100 0
50
Ac ce p
40 30 20
0
505 506
Fig. 4. Effects of QL and T on (a) KLa and (b) η for ammonia stripping at ω of 600 (empty
507
symbol and dot line) and 1200 (solid symbol and line) rpm. Symbols: QG of 75 L/min and T of
508
25 °C ( or ); QG of 60 L/min and T of 30 °C ( or ); QG of 40 L/min and T of 40 °C ( or
509
). Dot and solid lines: Predictions for KLa and η were based on Eqs. 9 and 8, respectively.
510
29
Page 29 of 30
0.008
ip t
0.006
0.004
cr
0.002
us
Predicted KLa (1/s)
(a)
0 0
0.002
0.004
0.008
an
100
(b)
M
80 60 40
d
Predicted (%)
0.006
Experimental KLa (1/s)
te
20
Ac ce p
0
0
510
20
40
60
80
Experimental (%)
100
511
Fig. 5. Diagonal graphs of experimental and predicted (a) KLa and (b) η. Dotted lines:
512
Prediction values within 20%.
513
30
Page 30 of 30