JES-01390; No of Pages 9

J O U RN A L OF E N V I RO N ME N TA L S CI EN CE S X X (2 0 1 8 ) XX X–XXX

Available online at www.sciencedirect.com

ScienceDirect
www.elsevier.com/locate/jes

1Q2 Structural characteristic and ammonium and

2 manganese catalytic activity of two types of filter

F
3 media in groundwater treatment

O
Q44Q3 Ya Cheng, Tinglin Huang⁎, Lijie Cheng, Yuankui Sun, Laisheng Zhu, Ye Li

O
5 Key Laboratory of Northwest Resource, Environment and Ecology, MOE, Xi'an University of Architecture and Technology, Xi'an 710055, China
6 Shaanxi Key Laboratory of Environmental Engineering, Xi'an University of Architecture and Technology, Xi'an 710055, China

R
7

P
10 AR TIC LE I NFO ABSTR ACT

11 Article history: Two types of filter media in groundwater treatment were conducted for a comparative 16
D
12 Received 3 November 2017 study of surface structure and catalytic performance. Natural filter media was adopted from 17
13 Revised 22 December 2017 a conventional aeration–filtration groundwater treatment plant, and active filter media as 18
E
14 Accepted 22 December 2017 a novel and promising filter media was also adopted. The physicochemical properties of 19
15 Available online xxxx these two kinds of filter media were characterized using numerous analytical techniques, 20
such as X-Ray diffraction (XRD), scanning electron microscope (SEM), energy dispersive 21
T

38 Keywords: X-ray (EDX), X-ray photoelectron spectroscopy (XPS) and Zeta potential. The catalytic 22
39 Natural filter media
C

activities of these filter media were evaluated for ammonium and manganese oxidation. 23
40 Active filter media XRD data showed that both active filter media and natural filter media belonged to 24
41 Manganese and ammonium removal birnessite family. A new manganese dioxide (MnO2) phase (PDF#72-1982) was found in the 25
E

42 Manganese oxides structure of natural filter media. The SEM micrograph of natural filter media showed 26
43 Characterization honeycomb structures and the active filter media presented plate structures and consisted 27
R

44 Activity of stacked particle. These natural filter media presented lower level of some trace elements 28
45 such as calcium and magnesium, lower degree of crystallinity, lower Mn(III) content and 29
R

lattice oxygen content than that of active filter media, which were associated with its 30
poor ammonium and manganese catalytic activities. In addition, some γ-Fe2O3 and MnCO3 31
O

were found in the coating which may hinder the ammonium and manganese catalytic 32
oxidation. This study provides a thorough and comprehensive understanding about the 33
most commonly used filter media in water treatment, which can provide a theoretical guide 34
C

to practical applications. 35
© 2018 The Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences. 36
N

Published by Elsevier B.V. 37

U

51 Introduction operational problems including discoloration of water, un- 56

pleasant metallic taste and odor, increased turbidity and 57
52 Manganese (Mn2+) and ammonium (NH+4) are common inor- biofouling of pipelines as well as staining of laundry and 58
53 ganic pollutants with salient features present in groundwater plumbing fixtures (Cai et al., 2015; Hasan et al., 2013). The 59
54 (Du et al., 2016; Tekerlekopoulou et al., 2013). Presence of presence of ammonium in water systems also leads to oxygen 60
55 manganese in finished water may cause organoleptic and depletion, eutrophication of surface water and toxicity for fish 61

⁎ Corresponding author. E-mail: huangtinglin@xauat.edu.cn (Tinglin Huang).

https://doi.org/10.1016/j.jes.2017.12.014
1001-0742/© 2018 The Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences. Published by Elsevier B.V.

Please cite this article as: Cheng, Y., et al., Structural characteristic and ammonium and manganese catalytic activity of two types
of filter media in groundwater treatment, J. Environ. Sci. (2018), https://doi.org/10.1016/j.jes.2017.12.014
 2 J O U RN A L OF E N V I RO N ME N TA L S CIE N CE S X X (2 0 1 8 ) XXX –XXX

62 (Cai et al., 2015; Cheng et al., 2017a). The increasing and very 2017b, 2017c). This filter media have special ammonium and 122
63 often uncontrollable use of fertilizers has led to increased manganese catalytic oxidation ability (Cheng et al., 2017b). 123
64 amounts of ammonium and manganese in potable water, which A review of water research publications was conducted in 124
65 often exceed the upper permitted limit (Tekerlekopoulou and this research effort to seek out any previously performed 125
66 Vayenas, 2012). Most countries have their own maximum con- studies related to the comparative investigation of structure 126
67 centration limits as a guideline to produce good quality drinking and activity of these two types of filter media. However, up to 127
68 water. The World Health Organization (WHO, 2004) has regulated now, rarely, work has been done on the comparative study of 128
Q5
69 that standard limits for NH+4–N and Mn2+ in drinking water the above filter media. 129
70 should be below 1.5 and 0.1 mg/L, respectively (SEPA, 2002). In In this study, we make a comparative study of the active 130
71 China, the maximum concentration limits for NH+4–N and Mn2+ filter media and natural filter media which have been running 131
72 in treated drinking water are 0.1 and 0.5 mg/L, respectively. for more than one year. Structural (X-ray diffraction (XRD)), 132
73 Manganese and ammonium removal becomes a problem when morphological (scanning electron microscope (SEM)), compo- 133

F
74 using groundwater as a drinking water, especially in country sitional (energy dispersive X-ray (EDX) spectroscopy) and 134
75 sites, outside big cities (Taffarel and Rubio, 2010). Therefore, elements state (X-ray photoelectron spectroscopy (XPS)) 135

O
76 controlling the NH+4–N and Mn2+ content in drinking water has characterization of the filter film have been used to under- 136
77 become a great public health issue. stand differences among these filter media and their effects 137

O
78 Processes for the removal of ammonium ions from potable on catalytic activity. Catalytic activity tests were carried out 138
79 water include chlorination, ion exchange, membrane filtra- with contaminant such as ammonium and manganese. 139
80 tion and biological treatment processes (Oehmen et al., 2006;

R
81 Vaaramaa and Lehto, 2003; Fu et al., 2011). Chlorination is
82 usually limited due to the high volumes of sludge and the 1. Materials and methods 141
140

P
83 cost of the oxidants and ion exchange, and membrane
84 filtration are limited due to their higher capital and opera- 1.1. Source of filter media 142
85 tional costs (Tekerlekopoulou et al., 2013; Cai et al., 2015). D
86 Biological treatment processes as an additional and effective The active filter media were adopted from pilot-scale water 143
87 advanced biological system for the simultaneous removal of treatment plant. The initiating process of this type filter was 144
E
88 NH+4-N and Mn2+ in the treatment of drinking water attracts carried out by oxidation–reduction method. The potassium 145
89 many attentions of researchers. However, actually numerous permanganate (KMnO4) solution, the manganese chloride 146
90
T

groundwater treatment facilities have opted to reduce their (MnCl2) solution and ferrous chloride (FeCl2) solution were 147
91 costs by applying origin quartz sand filter media without any separately pumped into the tube. After mixed by static mixer, 148
pretreatment, which is called “conventional aeration-rapid
C

92 the oxidation reaction of manganese and iron with potassium 149

93 sand filtration”. It is reported that more than 95% of the permanganate are listed as follows: 150
94 water treatment plants (WTPs) in China were still using the
E

95 conventional water treatment process as of 2010 (Feng et al.,

3Fe2þ þ MnO−4 þ 7H2 O → 3FeðOHÞ3 þ MnO2 þ 5Hþ ð1Þ
96 2012). Also, in European countries, the removal of manganese
R

97 from groundwater is commonly achieved by aeration-rapids 152

151
98 and filtration (Bruins et al., 2015a). The effect, over time, is 3Mn2þ þ 2MnO−4 þ 2H2 O → 5MnO2 þ 4Hþ ð2Þ
R

99 the generation of a manganese oxide film on the media

100 surface of the quartz sand (hereafter called the “natural filter During the process, manganese and iron were dosed at 154
153
155
O

101 media”). However, as general, in-rapid sand filters used for the concentration of 2 and 1 mg/L, respectively. The concen- 156
102 groundwater treatment, conditions for chemical formation of tration of potassium permanganate was determined by the 157
103 auto-catalytically active manganese oxides on virgin filter stoichiometry, and then, iron-manganese oxides were gener- 158
C

104 media, which can adsorb and subsequently oxidize dissolved ated. The iron-manganese oxides along with the raw water 159
105 manganese, are poor since most groundwater has a low redox was carried to the filter and deposited on the surface of the 160
N

106 potential and pH (Bruins et al., 2015a). Limited removal quartz sands. With more and more iron-manganese oxides 161
107 of manganese and ammonium could be achieved. Bruins deposited during the initiating process, an active film finally 162
108 163
U

et al. (2015a) have paid much excellent effort on how the filter was formed. The active filter has running continuously for
109 media ripening process and confirmed the origin of the 1 year. The active filter media samples were taken from the 164
110 manganese oxide film. Some other researchers proposed the top of the filter bed. 165
111 property of the manganese oxide depends on the valence of The natural filter media were adopted from large water 166
112 manganese in the oxide (Anschutz et al., 2005). Despite this, treatment plant. The characteristics of feed water are 167
113 the relationship between the structure and ammonium and as follows: Fe 0.05–0.10 mg/L, Mn 0.05–0.10 mg/L, NH+4-N 168
114 manganese catalytic activity of the mature filter media has 0.01–0.10 mg/L, pH 7.8–8.0, Redox potential 200–280 mv and 169
115 not been established. hardness190–200 mg/L. The feed water is faced with sea- 170
116 Besides, another novel and promising active manganese sonal iron and manganese pollution and ammonium in 171
117 oxide filter media (hereafter called the “active filter media”) the feed water is always below the limited concentration 172
118 has been developed for simultaneous removal of ammonium (< 0.5 mg/L). This plant has been running for more than 173
119 and manganese in recent years and these filter media exhibit 1 year. A thick film was formed on the surface of the quartz 174
120 some advantages, such as short start-up time and good sand. All filter media samples were taken from the top of the 175
Q6
121 low temperature resistance (Guo et al., 2017; Cheng et al., filter bed. 176

Please cite this article as: Cheng, Y., et al., Structural characteristic and ammonium and manganese catalytic activity of two types
of filter media in groundwater treatment, J. Environ. Sci. (2018), https://doi.org/10.1016/j.jes.2017.12.014
 J O U RN A L OF E N V I RO N ME N TA L S CI EN CE S X X (2 0 1 8 ) XX X–XXX 3

177 1.2. The operation of the filters Table 1 – Atomic ratios of the active filter media and natural t1:1
filter media determined by energy dispersive X-ray (EDX). t1:2
178 t1:3
Two identical filters were adopted for the ammonium and Element Si O Mg Ca Mn Fe t1:4
179 manganese catalytic oxidation tests. Each filter consisted of
Active filter media 15.58% 62.67% 1.05% 2.39% 11.71% 6.45% t1:5
180 a 200-cm-high Plexiglas column with an inside diameter of
Natural filter media 15.34% 76.19% 0.68% 0.95% 4.96% 1.87% t1:6
181 10 cm. Natural filter media and active filter media were used
182 as the filtration media, respectively. The filter bed was 120 cm
183 high. Eight sampling points (located at 0, 20, 45, 60, 75, 90, 105,
184 120 cm from bottom to top) were attached along the height of 2.1.2. Surface morphology and Zeta potential behavior analysis 229
185 the column. Groundwater pumped from a water source well Significant difference in appearance can be seen with the 230
186 was used as the feeding water. The concentration of dissolved naked eye, as can be seen from photographs shown in Fig. 1a 231
187 oxygen (DO) in the influent was between 6.0 and 7.0 mg/L. The and b. The active filter media appears black, in comparison 232

F
188
Q7 influent ammonium and manganese concentrations were with the red brown natural filter media which may attribute 233
189 kept at 1.5 and 2.0 mg/L, respectively. to the presence of a certain amount of ferric oxide (Fe2O3) on 234

O
the surface of the natural filter media. It could be further 235
190 1.3. Analytic methods verified by Fe 2p XPS analysis (in Section 2.1.4). Figure 1c and d 236

O
shows the SEM of the active filter media and the natural filter 237
191 During all the experiments, samples were collected in 500-mL media taken at 50000× magnification, respectively. The SEM 238
192 polypropylene bottles and immediately analyzed after sam- micrograph of active filter media presents plate structures 239

R
193 pling. NH+4, nitrite (NO−2), nitrate (NO−3) and Mn2+ were detected and consists of stacked particle. The SEM micrograph of 240
194 by spectrophotometry. Temperature (T), dissolved oxygen (DO) natural filter media shows honeycomb structures. According 241

P
195 and pH were determined by a portable instrument (HACH, to Bruins et al. (2015a), biologically produced birnessite has 242
196 HQ30d, USA). Alkalinity was determined by acid–base titration. a fluffy plate structure in contrast to physico-chemically formed 243
birnessite that has a more coral or sponge type structure. The
D 244
197 1.4. Characterization morphological feature of filter media collected in our study was 245
differing from them. The morphologies of manganese oxide 246
E
198 The structures of the samples were examined by XRD using were various in a natural environment and were closely related 247
199 Rigaku Ultiman IV with Cu Kα radiation (1.54 Å) at room with its forming environment. In our study, we can't determine 248
T

200 temperature, which provided the structural investigation of the origin of the iron-manganese oxide film according to the 249
201 the samples with data collected in the range of 5 to 75° at a morphological feature. 250
C

202 scan rate of 5°/min. These experiments were operating at Figure 2 shows the Zeta potential behavior of natural 251
203 40 kV and 40 mA and a step width of 0.02°. The morphology filter media and active filter media as a function of pH. The 252
204 of the samples was observed using a field emission SEM (FEI surface charge of the filter media becomes more negative 253
E

205 Quanta 600F). EDX spectroscopy was performed to determine when the medium pH is increased as a result of increased 254
206 the chemical composition in the filter media samples using the ratio OH− to the H+ bounds. When the pH ranged from 7.0 to 255
R

207 same instrument with SEM with an Oxford INCA/ENERGY-350. 9.0, the Zeta potential absolute value of the natural filter 256
208 XPS (K-Alpha XPS, Thermo Electron Corporation) was carried media was slightly less than that of the active filter media, 257
R

209 out to investigate the chemical oxidation states of elements implying the surface charge of the active filter media was a 258
210 Charging effects were corrected by adjusting the binding energy little more negative during the water treatment process. 259
O

211 of C 1s to 284.6 eV. Zeta potential measurements were carried

212 out with a Zetasizer Nano ZS90 (Malvern Instruments). 2.1.3. Structure analysis 260
XRD was used to identify phases and crystallinity of the 261
C

oxide film samples (Fig. 3) (Kumar et al., 2015; Kuo et al., 2015). 262
214
213 2. Results and discussion The XRD pattern obtained for active filter media samples 263
N

compares well with previous studies, which have a mixed 264

215 2.1. Characterization of the filter media phase structure (birnessite, buserite PDF#86-1630). X-ray 265
U

patterns of natural filter media samples also confirm to have 266

216 2.1.1. Component analysis a mixed phase structure (birnessite, buserite PDF#86-1630, 267
217 EDX spectroscopy was used to investigate the surface element MnO2 PDF#72-1982). It has to be mentioned that the phase 268
218 composition. As can be seen from Table 1, the major elements of silicon dioxide (SiO2) (PDF#32-1128) are the experimental 269
219 constituting the deposits on the filter media were manganese, artifacts related to quartz sand, and thus, they should 270
220 iron, calcium, silicon, magnesium and oxygen. Manganese not be considered in the analysis of the filter film patterns. 271
221 and iron contents accounted for a larger proportion of the The big difference between active filter media samples 272
222 deposits of active filter media (11.71% and 6.45%) than that and natural filter media samples was that a new MnO2 273
223 of natural filter media (4.96% and 1.87%), which was called phase (PDF#72-1982) was found in the structure of natural 274
224 iron-manganese oxide film. Besides, it is important to note filter media samples. In addition, in comparison with active 275
225 that a major difference between active filter media and filter media, XRD patterns of natural filter media samples 276
226 natural filter media is that more calcium and magnesium showed low peak intensities and peak broadening. This 277
227 was detected on the surface of active filter media than that on indicated low crystallinity of the natural filter media (Iyer 278
228 the surface of natural filter media. et al., 2015). A comparison of XRD patterns in Fig. 3 showed 279

Please cite this article as: Cheng, Y., et al., Structural characteristic and ammonium and manganese catalytic activity of two types
of filter media in groundwater treatment, J. Environ. Sci. (2018), https://doi.org/10.1016/j.jes.2017.12.014
 4 J O U RN A L OF E N V I RO N ME N TA L S CIE N CE S X X (2 0 1 8 ) XXX –XXX

F
O
O
R
P
D
E
T

Fig. 1 – (a) Photographs of the active filter media, (b) photographs of the natural filter media, (c) scanning electron microscopy
(SEM) of the active filter media (50000×), (d) surface of natural filter media (50000×).
C
E

280 that both of the filter media were belong to birnessite (Cheng the high resolution XPS spectrum of Mn2p can be deconvoluted 287
R

281 et al., 2017b). into three different components to reveal a characteristic of 288
a mixed-valence manganese system. According to Batis et al. 289
282 2.1.4. XPS analysis (2005), the peak positions of Mn2p3/2 level of MnO, Mn2O3 290
R

283 XPS was conducted to identify surface manganese elemental and MnO2 were 640.6, 641.9 and 642.2 eV, respectively. The 291
284 binding energies (Zhang et al., 2012; Park, 2002). Both active filter results indicate the presence of three manganese species, 292
O

285 media samples and natural filter media samples collected at i.e., Mn(II), Mn(III) and Mn(IV), differently concentrated in the 293
286 the 3rd and 7th day were analyzed by XPS. As shown in Fig. 4, two types of filter media samples. According to XPS data, the 294
C
N
U

Fig. 2 – Zeta potential of the active filter media and the Fig. 3 – X-ray diffraction patterns of the active filter film and
natural filter media as a function of pH. natural filter film.

Please cite this article as: Cheng, Y., et al., Structural characteristic and ammonium and manganese catalytic activity of two types
of filter media in groundwater treatment, J. Environ. Sci. (2018), https://doi.org/10.1016/j.jes.2017.12.014
 J O U RN A L OF E N V I RO N ME N TA L S CI EN CE S X X (2 0 1 8 ) XX X–XXX 5

F
O
O
R
P
D
E
T
C

Fig. 4 – X-ray photoelectron spectroscopy (XPS) spectra of (a) Mn 2p3/2 of the active filter media collected at the 3rd day, (b) Mn
2p3/2 of the natural filter media collected at the 3rd day, (c) Mn 2p3/2 of the active filter media collected at the 7th day, (d) Mn
E

2p3/2 of the natural filter media collected at the 7th day.

R

295 Mn(II), Mn(III) and Mn(IV) atomic concentrations (%) in active were involved in natural filter media samples, and also it is 319
R

296 filter media samples collected at the 3rd day (Fig. 4a) were noteworthy that more Mn(II) exist in these samples according 320
297 12.3%, 64.5% and 23.1%, whereas in the natural filter media to XPS Mn2p analysis. Therefore, we speculate the presence of 321
O

298 samples collected at the 3rd day (Fig. 4b), they were 31.1%, larger amounts of manganese carbonate (MnCO3) particles 322
299 27.2%, and 41.7%, respectively, which indicate more Mn(III) deposited on the filter media in the filter column. In addition, 323
300 were involved in active filter media samples. In addition, more the XPS of the Ca2p level was also analyzed as shown in Fig. 6a. 324
C

301 Mn(IV) was involved in natural filter media samples which According to the literature, two doublets of the spectrum 325
302 was in accordance with the XRD analysis. A new MnO2 phase indicated the presence of calcium in different chemical 326
N

303 (PDF#72-1982) was found in the structure of natural filter surroundings and it might be due to the presence of cal- 327
304 media samples. In addition, it was found that the content of cium carbonate CaCO3 (Batis et al., 2005; Zhang-Steenwinkel 328
U

305 Mn(III) in both active filter media samples and natural filter et al., 2002). The result was in accordance with the O1s 329
306 media samples collected at the 7th day (Fig. 4c and d) varied spectrum where an additional peak was observed at BE of 330
307 slightly with that of the filter media samples collected at 530.9–531.4 eV. According to the redox-pH diagram of aqueous 331
308 the 3rd day (Fig. 4a and b), which indicated that the content manganese (Fig. 6b) (Bruins et al., 2015a, 2015b; Stumm and 332
309 of Mn(III) in both active filter media samples and natural Morgan, 2015), the pH and redox potential of feed water to 333
310 filter media samples remained unchanged during the water the pilot (black dot) is in the transition zone of conditions 334
311 treatment. for dissolved manganese and conditions theoretically re- 335
312 Deconvolutions of the O1s XPS spectrum were also per- quired to form MnCO3. This assumption was confirmed by the 336
313 formed and are shown in Fig. 5. The O1s peak is attributed to O above analysis, where some MnCO3 were found in the coating, 337
314 species derived from different oxygen sources. The spectrum especially in natural filter film. In addition, a higher lattice 338
315 can be deconvoluted into three components, which are related oxygen density was found in active filter media (Fig. 5a and c) 339
316 to lattice oxygen (528.9–529.2 eV), carbonates (530.9–531.4 eV), than those in natural filter media (Fig. 5b and d). The content 340
317 and hydroxyl groups or adsorbed O− species (532.6–532.8 eV) of lattice oxygen species in both active filter media samples 341
318 (Batis et al., 2005). The results indicated that more carbonates and natural filter media samples collected at the 7th day 342

Please cite this article as: Cheng, Y., et al., Structural characteristic and ammonium and manganese catalytic activity of two types
of filter media in groundwater treatment, J. Environ. Sci. (2018), https://doi.org/10.1016/j.jes.2017.12.014
 6 J O U RN A L OF E N V I RO N ME N TA L S CIE N CE S X X (2 0 1 8 ) XXX –XXX

F
O
O
R
P
D
E
T

Fig. 5 – X-ray photoelectron spectroscopy (XPS) spectra of (a) O1s of the active filter media collected at the 3rd day, (b) O1s of the
C

natural filter media collected at the 3rd day, (c) O1s of the active filter media collected at the 7th day, (d) O1s of the natural filter
media collected at the 7th day.
E
R

343 (Fig. 5c and d) varied slightly with that of the filter media Figure 7 presents the XPS spectra of Fe 2p core level. XPS 348
344 samples collected at the 3rd day (Fig. 5a and b), which has been known to be sensitive to Fe(II) and Fe(III). Above all 349
R

345 indicated that the content of lattice oxygen species in both things, the key characteristic which distinguishes γ-Fe2O3 350
346 active filter media samples and natural filter media samples from Fe3O4 is the presence of the satellite peak at a binding 351
O

347 also remained unchanged during the water treatment. energy of 719.0 eV (Sang et al., 2014; Gota et al., 1999). In our 352
C
N
U

Fig. 6 – (a) XPS spectra of Ca 2p core level attained from active filter media and natural filter media collected at the 3rd day.
(b) Electron activity (pε) or redox potential (Eh in e V) -pH diagram for aqueous manganese (adopted from Stumm and Morgan,
Q1 1996), compared to feed water conditions of the groundwater of the GWTPs (black dot).

Please cite this article as: Cheng, Y., et al., Structural characteristic and ammonium and manganese catalytic activity of two types
of filter media in groundwater treatment, J. Environ. Sci. (2018), https://doi.org/10.1016/j.jes.2017.12.014
 J O U RN A L OF E N V I RO N ME N TA L S CI EN CE S X X (2 0 1 8 ) XX X–XXX 7

with the active filter media demonstrated better ammonium 371

and manganese removal ability (Fig. 8a and b). The effluent 372
NH+4-N and Mn2+ could hardly be detected. Ammonium 373
evolution in the filter system was confirmed by analyses of 374
NO−3 and NO−2 in the solution. As illustrated in Fig. 8a, the 375
decrease in NH+4-N concentration was accompanied with an 376
elevation of NO−3, NO−2 formation and DO consumption in 377
active filter which was consistent with our previous study 378
(Cheng et al., 2017b, 2017c). In contrast, the concentration of 379
NO−3, NO−2 and DO along the height of the natural filter column 380
remained almost constant (data not shown). 381
In addition, the evolution of NH+4-N and Mn2+ along the 382

F
height of these two filters was also determined. As can be 383
seen from Fig. 9, the relative constant NH+4-N and Mn2+ 384

O
concentration along the height of the pilot column filled 385
with the natural filter media further confirmed that under 386
Fig. 7 – XPS spectra of Fe 2p core level attained from active

O
the adopted reaction conditions there was no occurrence of 387
filter media and natural filter media collected at the 3rd day.
NH+4-N and Mn2+ catalytic reaction. In contrast, the NH+4-N and 388
Mn2+ concentration decreased along the height of the active 389

R
filter, and at a depth of 60 cm (total 120 cm), NH+4-N could go 390
Q8
353 case, the satellite peak was observable at a binding energy of below the maximum concentration limits (< 0.5 mg/L) (Fig. 9a), 391

P
354 719.2 eV as indicated with an arrow in Fig. 7. Therefore, it is and almost all Mn2+ could have been removed (Fig. 9b). 392
355 certain that non-negligible amount of γ-Fe2O3 were found in Obviously, natural filter media was less active than active 393
356 the natural filter film, which was closely related to the red filter media for NH+4-N and Mn2+ catalytic oxidation.
D 394
357 brown layer coated on the surface of filter media (Fig. 1b).
2.3. The relationship between structure and activity 395
E
358 2.2. Activity of the filter media
In the aqueous system, the metal oxides have surface 396
T

359 These two filters have operated for 7 days. Performance in hydroxyl groups that have acidic and basic characteristics 397
360 terms of ammonium and manganese removal was systemat- (Taffarel and Rubio, 2010). According to the analysis of Zeta 398
C

361 ically studied, and comparisons between this active filter potential, the surface charge of the active filter media was 399
362 media and natural filter media were carried out. just a little more negative than that of the natural filter 400
363 Figure 8 shows the performance of the natural filter media during the water treatment process. Even so, the little 401
E

364 and active filter with respect to simultaneous removal of difference in the surface charge couldn't determine the 402
365 NH+4-N and Mn2+ during the operation period. There was little ammonium and manganese removal performance. Therefore, 403
R

366 difference (0.1–0.2 mg/L) between the influent and effluent we speculate the difference between the two types of the filter 404
367 Mn2+ concentration in the natural filter (as shown in Fig. 8b), media in ammonium and manganese catalytic activity did not 405
R

368 indicating the limited manganese removal ability of the arise from the difference of the surface hydroxyl groups. 406
Q9
369 natural filter media. Similarly, the NH+4-N removal perfor- It is well known that the catalytic activity of iron-manganese 407
O

370 mance was poor (as shown in Fig. 8a). Unlike it, the filter filled oxide is associated with its composition and structure. As 408
C
N
U

Fig. 8 – The running performance of the natural filter and active filter with respect to simultaneous removal of ammonium
(NH+4-N) (a) and manganese (Mn2 +) (b). The inset shows the profiles about the concentrations of nitrite (NO−2-N), nitrate (NO−3-N)
and dissolved oxygen (DO), respectively, along the height of the active filter.

Please cite this article as: Cheng, Y., et al., Structural characteristic and ammonium and manganese catalytic activity of two types
of filter media in groundwater treatment, J. Environ. Sci. (2018), https://doi.org/10.1016/j.jes.2017.12.014
 8 J O U RN A L OF E N V I RO N ME N TA L S CIE N CE S X X (2 0 1 8 ) XXX –XXX

F
O
Fig. 9 – Concentration profiles of ammonium (NH+4-N) (a) and manganese (Mn2+) (b) along the height of the pilot column filled

O
with the natural filter media and the active filter media.

R
409 revealed by the XPS investigation, more Mn(III) were involved in (2015, 2017) reported that the content of water and interlayer 449

P
410 active filter media. Our previous research has shown that the cations in birnessite had dramatic effect on its activity due 450
411 highly catalytic activity of active birnessite involved in ammo- to their influence on surface hydroxyl, which acted as active 451
412 nium catalytic oxidation was proposed to be intermediated species in formaldehyde oxidation. Some γ-Fe2O3 and MnCO3
D 452
413 by Mn(III) (Cheng et al., 2017b). In addition, Peluso et al. (2008) were found in the coating, especially in natural filter media. 453
414 studied the manganese dioxide catalyzed oxidation of ethanol, According to Bruins et al. (2015a), MnCO3 has no auto-catalytical 454
E
415 and suggested that a high concentration of Mn(III) could adsorption and oxidation properties, which consequently hin- 455
Q10
416 result in weak Mn\O bonds, which might ensure good catalytic dered the ammonium and manganese catalytic oxidation. 456
417 457
T

activity. XRD data showed that both active filter media and Therefore, we concluded that the poor catalytic perfor-
418 natural filter media belonged to the birnessite family. The mance of natural filter media was mainly related to the low 458
C

419 difference of phase constitution and crystallinity helped to Mn(III) content, the lattice oxygen and heteroatom doping. In 459
420 explain the ammonium and manganese catalytic activity addition, the presence of MnCO3 may hinder the ammonium 460
421 differences between them. Compared with the active filter and manganese catalytic oxidation. 461
E

422 media, a new MnO2 phase (PDF#72-1982) was found in the

423 structure of natural filter media. Mn(IV) phase was less active
R

424 than Mn(III) phase. What's more, a higher lattice oxygen density 3. Conclusion 463
462
425 was found in active filter media. According to the literature,
R

426 structural defect is beneficial for the activation of oxygen The following conclusions can be drawn based on the results 464
427 molecules to active oxygen adspecies (Wang et al., 2012). emerging from this study: (1) Analysis of XRD data showed 465
O

428 Specifically, lattice oxygen species with high mobility enhanced that both active filter media and natural filter media belong to 466
429 the formation of surface active oxygen species through the the birnessite family. A new MnO2 phase (PDF#72-1982) was 467
430 complex migration between surface lattice oxygen and oxygen found in the structure of natural filter media samples. (2) The 468
C

431 vacancy with molecular oxygen (Zhu et al., 2017). Therefore, morphological feature of active filter media was differing 469
432 the Mn(III) content and the oxygen adspecies concentration from natural filter media. The SEM micrograph of active filter 470
N

433 relevant to the surface oxygen vacancy density were correlative media presented plate structures and consisted of stacked 471
434 with the catalytic activity of these filter media samples. In particle, and the natural filter media showed honeycomb 472
435 473
U

addition, the content of Mn(III) and concentration of lattice structures. (3) Natural filter media was less active than active
436 oxygen species in both active filter media samples and natural filter media for ammonium and manganese catalytic oxida- 474
437 filter media samples remained almost unchanged during the tion. (4) More Mn(III) and a higher lattice oxygen density 475
438 water treatment, which suggested that there was no loss of were found in active filter media, and some γ-Fe2O3 and 476
439 the active components. It was in accordance with the stable MnCO3 were found in the coating of natural filter media, 477
440 removal efficiency of ammonium and manganese during the which was correlative with the catalytic activity of these filter 478
441 operation period (Fig. 8). media samples. 479
442 Another major difference between active filter media and
443 natural filter media was that more calcium and magnesium
444 was detected on the surface of active filter media. A survey of Acknowledgments 481
480
445 the literature available showed that heteroatom doping in
446 the manganese oxides can also improve catalytic activities to This work was supported by the National Key Research and 482
447 some extent, which is ascribed to the generated defects or Development Program of China (No. 2016YFC00400706), the 483
448 synergistic effects (Wang et al., 2017). In addition, Wang et al. National Natural Science Foundation of China (No. 51778521) 484

Please cite this article as: Cheng, Y., et al., Structural characteristic and ammonium and manganese catalytic activity of two types
of filter media in groundwater treatment, J. Environ. Sci. (2018), https://doi.org/10.1016/j.jes.2017.12.014
 J O U RN A L OF E N V I RO N ME N TA L S CI EN CE S X X (2 0 1 8 ) XX X–XXX 9

485 and the Natural Science Basic Research Plan in Shaanxi Kumar, R.V., Ghoshal, A.K., Pugazhenthi, G., 2015. Fabrication 549
486 Province of China (No. 2017JQ2014). of zirconia composite membrane by in-situ hydrothermal 550
technique and its application in separation of methyl orange. 551
Ecotoxicol. Environ. Saf. 121 (12), 73–79. 552
Kuo, C.H., Mosa, I.M., Poyraz, A.S., Biswas, S., El-Sawy, A.M., Song, 553
487 REFERENCES W., Luo, Z., Chen, S.Y., Rusling, J.F., He, J., 2015. Robust 554
mesoporous manganese oxide catalysts for water oxidation. 555
488 ACS Catal. 5 (3), 1693–1699. 556
489 Anschutz, P., Dedieu, K., Desmazes, F., Chaillou, G., 2005. Oehmen, A., Viegas, R., Velizarov, S., Reis, M.A.M., Crespo, J.G., 557
490 Speciation, oxidation state, and reactivity of particulate 2006. Removal of heavy metals from drinking water supplies 558
491 manganese in marine sediments. Chem. Geol. 218 (3), 265–279. through the ion exchange membrane bioreactor. Desalination 559
492 Batis, N.H., Delichere, P., Batis, H., 2005. Physicochemical 199 (1–3), 405–407. 560
493 and catalytic properties in methane combustion of Park, H.J., 2002. Interfacial properties of thin film hetero-structure: 561
494 La1−xCaxMnO3 ± y (0 ≤ x ≤ 1; − 0.04 ≤ y ≤ 0.24) perovskite-type copper-oxides of hafnium-silicon. Diss. Abstr. Int. 64 (03), 1445. 562

F
495 oxide. Appl. Catal. A Gen. 282 (1), 173–180. Peluso, M.A., Gambaro, L.A., Pronsato, E., Gazzoli, D., Thomas, H.J., 563
496 Bruins, J.H., Petrusevski, B., Slokar, Y.M., Huysman, K., Joris, K., Sambeth, J.E., 2008. Synthesis and catalytic activity of 564

O
497 Kruithof, J.C., Kennedy, M.D., 2015a. Biological and manganese dioxide (type OMS-2) for the abatement of 565
498 physico-chemical formation of birnessite during the ripening oxygenated VOCs. Catal. Today 133–135 (1), 487–492. 566
499 of manganese removal filters. Water Res. 69 (7053), 154. Sang, Y.L., Kim, D.H., Choi, S.C., Lee, D.J., Ji, Y.C., Kim, H.D., 2014. 567

O
500 Bruins, J.H., Petrusevski, B., Slokar, Y.M., Kruithof, J.C., Kennedy, M.D., Porous multi-walled carbon nanotubes by using catalytic 568
501 2015b. Manganese removal from groundwater: characterization oxidation via transition metal oxide. Microporous Mesoporous 569
502 of filter media coating. Desalin. Water Treat. 55 (7), 1–13. Mater. 194 (4), 46–51. 570

R
503 Cai, Y., Li, D., Liang, Y., Luo, Y., Zeng, H., Zhang, J., 2015. Effective SEPA (State Environmental Protection Administration), 2002. 571
504 start-up biofiltration method for Fe, Mn, and ammonia Monitoring and Analytical Methods for Water and Wastewater. 572
505

P
removal and bacterial community analysis. Bioresour. fourth ed. China Environmental Science Press, Beijing 573
506 Technol. 176 (16), 149–155. (in Chinese). 574
507 Cheng, Y., Huang, T., Shi, X., Wen, G., Sun, Y., 2017a. Removal of Stumm, W., Morgan, J.J., 2015. Aquatic Chemistry: Chemical 575
508 ammonium ion from water by Na-rich birnessite: performance DEquilibria and Rates in Naturalwaters. Cram101 Textbook 576
509 and mechanisms. J. Environ. Sci. 57 (7), 402–410. Outlines to Accompany 179 (11) p. A277. 577
510 Cheng, Y., Huang, T., Sun, Y., Shi, X., 2017b. Catalytic oxidation Taffarel, S.R., Rubio, J., 2010. Removal of Mn2+ from aqueous 578
511 removal of ammonium from groundwater by manganese
E
solution by manganese oxide coated zeolite. Miner. Eng. 23 579
512 oxides filter: performance and mechanisms. Chem. Eng. J. 322 (14), 1131–1138. 580
513 (15), 82–89. Tekerlekopoulou, A.G., Vayenas, D.V., 2012. Ammonia, iron and 581
T

514 Cheng, Y., Li, Y., Huang, T., Sun, Y., Shi, X., Shao, Y., 2017c. A manganese removal from potable water using trickling filters. 582
515 comparison study of the start-up of a MnOx filter for catalytic Desalination 210 (1), 225–235. 583
C

516 oxidative removal of ammonium from groundwater and Tekerlekopoulou, A.G., Pavlou, S., Vayenas, D.V., 2013. Removal of 584
517 surface water. J. Environ. Sci. https://doi.org/10.1016/ ammonium, iron and manganese from potable water in 585
518 j.jes.2017.07.008. biofiltration units: a review. J. Chem. Technol. Biotechnol. 88 586
E

519 Du, X., Liu, G., Qu, F., Li, K., Shao, S., Li, G., Liang, H., 2016. Removal (5), 751–773. 587
520 of iron, manganese and ammonia from groundwater using a Vaaramaa, K., Lehto, J., 2003. Removal of metals and anions from 588
R

521 PAC-MBR system: the anti-pollution ability, microbial drinking water by ion exchange. Desalination 155 (2), 157–170. 589
522 population and membrane fouling. Desalination 403, 97–106. Wang, F., Dai, H., Deng, J., Bai, G., Ji, K., Liu, Y., 2012. Manganese 590
523 Feng, S., Xie, S., Zhang, X., Yang, Z., Ding, W., Liao, X., Liu, Y., Chen, oxides with rod-, wire-, tube-, and flower-like morphologies: 591
R

524 C., 2012. Ammonium removal pathways and microbial highly effective catalysts for the removal of toluene. Environ. 592
525 community in GAC-sand dual media filter in drinking water Sci. Technol. 46 (7), 4034–4041. 593
526 treatment. J. Environ. Sci. 24 (9), 1587.
O

Wang, J., Zhang, P., Li, J., Jiang, C., Yunus, R., Kim, J., 2015. 594
527 Fu, J.X., Shang, J., Zhao, Y.H., 2011. Simultaneous removal of iron, Room-temperature oxidation of formaldehyde by layered 595
528 manganese and ammonia from groundwater in single biofilter manganese oxide: effect of water. Environ. Sci. Technol. 49 596
529
C

layer using baf. Adv. Mater. Res. 183-185, 442–446. (20), 12372. 597
530 Gota, S., Guiot, E., Henriot, M., Gautier-Soyer, M., 1999. Wang, J., Li, J., Jiang, C., Zhou, P., Zhang, P., Yu, J., 2017. The 598
531 Atomic-oxygen-assisted MBE growth of α −Fe2O3 on α −Al2O3 effect of manganese vacancy in birnessite-type MnO2 on 599
N

532 (0001): metastable FeO(111)-like phase at subnanometer room-temperature oxidation of formaldehyde in air. Appl. 600
533 thicknesses. Phys. Rev. B 60 (20), 14387–14395. Catal. B Environ. 204, 147–155. 601
534 Guo, Y., Huang, T., Wen, G., Cao, X., 2017. The simultaneous
U

WHO, 2004. Guideline for Drinking Water Quality. third ed. vol. 1. 602
535 removal of ammonium and manganese from groundwater by World Health Organization, Geneva. 603
536 iron-manganese co-oxide filter film: the role of chemical Zhang, T., Yan, X., Sun, D.D., 2012. Hierarchically multifunctional 604
537 catalytic oxidation for ammonium removal. Chem. Eng. J. 308, K-OMS-2/TiO2/Fe3O4 heterojunctions for the photocatalytic 605
538 322–329. oxidation of humic acid under solar light irradiation. J. Hazard. 606
539 Hasan, H.A., Abdullah, S.R.S., Kamarudin, S.K., Kofli, N.T., 2013. Mater. 243 (4), 302–310. 607
540 On–off control of aeration time in the simultaneous removal of Zhang-Steenwinkel, Y., Beckers, J., Bliek, A., 2002. Surface 608
541 ammonia and manganese using a biological aerated filter properties and catalytic performance in CO oxidation of 609
542 system. Process. Saf. Environ. Prot. 91 (5), 415–422. cerium substituted lanthanum–manganese oxides. Appl. 610
543 Iyer, A., Delpilar, J., King'Ondu, C.K., Kissel, E., Garces, H.F., Huang, Catal. A Gen. 235 (1–2), 79–92. 611
544 H., Elsawy, A.M., Dutta, P.K., Suib, S.L., 2015. Water oxidation Zhu, L., Wang, J., Rong, S., Wang, H., Zhang, P., 2017. Cerium 612
545 catalysis using amorphous manganese oxides, octahedral modified birnessite-type MnO2 for gaseous formaldehyde 613
546 molecular sieves (OMS-2), and octahedral layered (OL-1) oxidation at low temperature. Appl. Catal. B Environ. 211 (15), 614
547 manganese oxide structures. J. Phys. Chem. C 116 (10), 212–221. 615
548 6474–6483.
616
617

Please cite this article as: Cheng, Y., et al., Structural characteristic and ammonium and manganese catalytic activity of two types
of filter media in groundwater treatment, J. Environ. Sci. (2018), https://doi.org/10.1016/j.jes.2017.12.014