Accepted Manuscript

Sequestration of phosphorus from wastewater by cement-based or alternative

cementitious materials

Xinjun Wang , Jiding Chen , Yaping Kong , Xianming Shi

PII: S0043-1354(14)00374-1
DOI: 10.1016/[Link].2014.05.021
Reference: WR 10675

To appear in: Water Research

Received Date: 14 March 2014

Revised Date: 12 May 2014
Accepted Date: 13 May 2014

Please cite this article as: Wang, X., Chen, J., Kong, Y., Shi, X., Sequestration of phosphorus from
wastewater by cement-based or alternative cementitious materials, Water Research (2014), doi:
10.1016/[Link].2014.05.021.

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.
 ACCEPTED MANUSCRIPT
For Table of Contents Only

PT
RI
U SC
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

1 Sequestration of phosphorus from wastewater by cement-based or alternative

2 cementitious materials

3 Xinjun Wanga,b, Jiding Chen b, Yaping Kong b, Xianming Shia,*.

a
4 Corrosion and Sustainable Infrastructure Laboratory, Western Transportation

PT
5 Institute and Civil Engineering Department, PO Box 174250, College of Engineering,

RI
6 Montana State University, Bozeman, MT 59717-4250, USA

b
7 Laboratory of Road Ecology and Pollution Emergency Response, Research Center

SC
8 for Environmental Protection and Transportation Safety, China Academy of

U
9 Transportation Sciences, Beijing 100029, China
AN
10 *Corresponding author. Tel.: +1 406 994 6486. E-mail address:

11 xianming_s@[Link]
M

12 Abstract
D

13 Cement-based and alternative cementitious materials were tested in the laboratory for
TE

14 their capability of removing phosphate from wastewater. The results demonstrated

15 that both Langmuir and Freundlich adsorption isotherms were suitable for describing
EP

16 the adsorption characteristics of these materials. Among the four types of filter media
C

17 tested, the cement-based mortar A has the highest value of maximum adsorption
AC

18 (30.96 mg g-1). The P-bonding energy (KL) and adsorption capacity (K) exhibited a

19 positive correlation with the total content of Al2O3 and Fe2O3 in each mortar. The

20 maximum amount of P adsorbed (Qm) and adsorption intensity (1/n) exhibited a

21 positive correlation with the CaO content in each mortar. For three of them, the

22 P-sequestration rates were in excess of 94 percent for phosphorus concentrations

 ACCEPTED MANUSCRIPT

23 ranging from 20 to 1000 mg L-1. The underlying mechanisms were examined using

24 field emission scanning microscopy (FESEM), coupled with energy-dispersive X-ray

25 spectroscopy (EDX) and X-ray powder diffraction (XRD). The results reveal that the

26 removal of phosphate predominantly followed a precipitation mechanism in addition

PT
27 to weak physical interactions between the surface of adsorbent filter media and the

RI
28 metallic salts of phosphate. The use of cement-based or alternative cementitious

29 materials in the form of ground powder shows great promise for developing a

SC
30 cost-effective and environmentally sustainable technology for P-sequestration and for

U
31 wastewater treatment.
AN
32 Keywords: wastewater treatment; mortars; fly ash; constructed wetland; adsorption

33 isotherms; bio-retention tank; FESEM; EDX; XRD

M

34
D

35 1. Introduction
TE

36 Phosphorus (P) is a key element for causing eutrophication and algae bloom, which

37 lead to short-term or long-term environmental risks in streams, lakes and other water
EP

38 bodies. A host of wastewater treatment techniques have been employed for phosphate
C

39 removal, in order to meet effluent quality standards and avoid P-pollution. Septic tank,
AC

40 mechanical aeration system and constructed wetland system (CWS) are the primary

41 methods used for wastewater treatment in rural areas or in urban areas that are not

42 connected to a public sewer system (Brix, 1993; De-Bashan and Bashan, 2004;

43 Vymazal, 2013; Wang et al., 2012). One appropriate approach to P-sequestration is

44 filtration through materials with high affinity for P-binding. To this end, a wide array
 ACCEPTED MANUSCRIPT

45 of low-cost and readily available materials (natural materials, by-products and

46 man-made products) have been investigated in batch experiments or in field practices

47 (Ahmaruzzaman, 2010; Vohla et al., 2011; Westholm 2006; Wu and Sansalone, 2013).

48 Cement-based materials (e.g., mortars and concretes) are commonly used to

PT
49 construct a significant portion of current wastewater systems, including septic tank

RI
50 and CWS. If such materials are endowed with the additional function of

51 P-sequestration, it would greatly improve the wastewater treatment performance.

SC
52 However, little research has been conducted to test the feasibility of this approach,

U
53 despite some recent reports of pervious concretes with promising stormwater
AN
54 management performance (Horst et al., 2010; Luck et al., 2009).

55 Furthermore, environmentally friendly concretes that utilize industrial wastes as

M

56 supplementary cementitious materials (SCMs) can be employed to further enhance the

D

57 environmental benefits in the aforementioned approach to P-sequestration. Currently,

TE

58 about 12 million t of the coal fly ashes in the U.S. are utilized in concretes and

59 mortars each year (American Coal Ash Association, 2004), as SCM or replacement of
EP

60 fine aggregate. The beneficial use of this byproduct of coal-fired electricity generating
C

61 plants can translate to cost and environmental benefits without sacrificing the
AC

62 engineering properties of the concrete (Atis, 2003; Mehta, 2004; Shi et al., 2011,

63 2012). Coal fly ash has also been used as a low-cost adsorbent for the removal of

64 organic and inorganic elements in wastewater (Ahmaruzzaman, 2010; Wang and Wu,

65 2006).

66 In this context, this study investigates the potential of using cement-based and fly
 ACCEPTED MANUSCRIPT

67 ash-based alternative cementitious materials for the sequestration of phosphorus from

68 various synthetic wastewater solutions. The objectives of this study are to: (1)

69 evaluate the adsorption capacity of and P-sequestration efficiency by these materials

70 as filter media; (2) elucidate the P-sequestration mechanism(s); and (3) discuss the

PT
71 potential for practical applications. The laboratory data will shed light on the potential

RI
72 of using low-cost cementitious materials for P-sequestration and wastewater

73 treatment.

SC
74

U
75 2. Materials and methods
AN
76 2.1 Materials

77 The cement-based and alternative cementitious materials used in this study were
M

78 fabricated in the Corrosion and Sustainable Infrastructure laboratory at the Montana

D

79 State University. Mortar A is made with an ASTM specification C150-07 Type I/II
TE

80 low-alkali Portland cement obtained from ASH Grove Cement Company (Clancy, MT,

81 USA) as the sole binder, with water/binder ratio of 0.45, and a surfactant (sodium
EP

82 dodecyl sulfate, or SDS) was admixed at 0.005 % by weight of cement. Fine

C

83 aggregates with a maximum size of 1.18 mm in diameter were used, at a sand/binder

AC

84 ratio of 2.0. Mortar B is a modified mix design of mortar A by admixing a

85 polysiloxane-modified montmorillonite (1.44PSTM, bulk density: 0.251 g/cm3,

86 Nanocor Inc., Chicago, IL, USA) and a coated cellulose microfiber (Buckeye

87 UltraFiber 500™, Buckeye Technologies Inc., Memphis, TN, USA), at 0.5% and 0.25%

88 by weight of cement, respectively. Mortar C is made with a Class C fly ash obtained
 ACCEPTED MANUSCRIPT

89 from the Corette electric power plant (Billings, MT, USA) as the sole binder, with

90 water/binder ratio of 0.28, and a set retarder (borax, Na2B4O7), a water reducer

91 (Glenium®, BASF Corporation) and a nanoclay (1.44PS) were admixed at 0.2 %,

92 0.43 % and 0.25 % by weight of fly ash, respectively. Fine aggregates with a

PT
93 maximum size of 1.18 mm in diameter were used, along with light weight aggregates

RI
94 (LWA, crushed fines of expanded shale, 10 mesh, Utelite Corporation, Coalville,

95 Utah), at a LWA/sand/binder ratio of 1:3:5. Mortar D featured the same

SC
96 LWA/sand/binder ratio and borax/Glenium/fly ash ratio as Mortar C, but with a

U
97 modified water/binder ratio of 0.33, a nanoclay/fly ash ratio of 0.53 %, and
AN
98 dehydrated gypsum (CaSO4.1/2H2O) at 5 % by weight in place of fly ash as binder

99 and a spent petrochemical catalyst (nano-sized zeolite) at 3 % by weight of fly ash as

M

100 aggregate.
D

101 For all the mortar samples, the sample production and curing procedures are
TE

102 described elsewhere.14 For each mix design, twelve replicate 2″ by 4″ (diameter 51

103 mm × length 102 mm) cylinders were prepared. Cement (or fly ash) is mixed with
EP

104 water (containing borax, surfactant, water reducer, and/or well-dispersed nanoclay, if
C

105 applicable) at a low-speed hand mixer for 5 min. Subsequently, fine aggregates
AC

106 (premixed with zeolite, LWA, microfiber, and/or dehydrated gypsum, if applicable)

107 were added, after which the slurries were stirred for 3 min. The fine aggregates were

108 prepared to surface-saturated-dry (SSD) condition in advance. All the slurries were

109 cast into hollow poly(vinyl chloride) piping cylinders and then carefully compacted to

110 minimize the amount of entrapped air. The cylindrical samples were demolded after
 ACCEPTED MANUSCRIPT

111 curing for 24 h in a chamber with over 90% relative humidity. After demolding, the

112 samples were cured in the moist cure chamber (with over 90% relative humidity) for

113 another 89 days before testing. All of the mortar samples exhibited an average

114 compressive strength of at least 27.5 MPa at 28 days.

PT
115 2.2 P-adsorption experiments

RI
116 Mortars A, B, C and D were ground and sieved to a maximum grain size of 0.45 mm

117 before being sampled for P-adsorption experiments. A standard phosphate solution

SC
118 (1000 mg L-1 P) was prepared from anhydrous KH2PO4 with 0.01 mol L-1 KCl. A

U
119 small dosage of sulfuric acid (5 g L-1) was added to ensure its stability over time.
AN
120 Other standard phosphate solutions were prepared by diluting this specific solution.

121 The adsorption studies were carried out with phosphate solutions at concentrations of
M

122 0-1000 mg L-1 (0, 10, 20, 40, 80, 160, 320, 600, and 1000 mg L-1 P). A given amount
D

123 of each sample (1.5 g) was placed in a 50-mL volumetric tube and mixed with 30 mL
TE

124 of phosphate solution. The mixture was continuously shaken at 175 rpm on a gyratory

125 shaker for 24 h under controlled temperature conditions (25 ºC). All experiments were
EP

126 conducted in triplicates. In order to extract supernatant for phosphate concentration

C

127 test, the mixtures were centrifuged at 2000 rpm. The phosphate concentration of each
AC

128 solution was determined according to the molybdenum-blue ascorbic acid method

129 (APHA et al., 1995), using a UV Spectrophotometer (UV1206, Shimadzu, Japan).

130

131 P-adsorption capacities (q, mg g-1) were calculated form equation (1) as

 
132 follows:
= 
× (1)
 ACCEPTED MANUSCRIPT

133 Where C0 is the initial P concentration (mg L-1), Ce is the P concentration in the

134 supernatant at equilibrium (mg L-1), M is the mass of sample (g), and V is the volume

135 of the solution (L).

136 Removal rates of phosphate (PR) by the absorbent were evaluated as percentage as

PT
137 follows:
 

RI
138 R = × 100 (2)


139 2.3 P-adsorption isotherms

SC
140 Langmuir and Freundlich adsorption isotherm equations were used for describing the

U
141 adsorption characteristics of the materials in this study. The Langmuir isotherm
AN
142 equation is the best known of all equations describing adsorption, and it has been

143 successfully applied to many adsorption processes (Drizo et al., 1999). It is expressed
M

144 as follows:
D

  
145 =  
(3)
TE

146 and its linear equation is presented as follows:

  
147 = + (4)
EP

   

148 Where C is the equilibrium P concentration (mg L-1), and Qm (mg g-1) and KL (l
C

149 mg-1) are constants related to the maximum adsorption capacity and energy of
AC

150 adsorption. The value of Qm and KL can be determined by the linear equation (4).

151 The Freundlich isotherm equation is most frequently used to describe the

152 adsorption of inorganic and organic components in solution and it is expressed as

153 follows:

154 =   (5)
 ACCEPTED MANUSCRIPT

155 And its linear equation is presented as follows:


156 log =   + log  (6)


157 Where K is an indicator of adsorption capacity and 1/n is the adsorption intensity.

158 The value of K and n can be calculated by the linear equation (6).

PT
159 2.4 P-desorption experiments

RI
160 After decanting the supernatant from the adsorption experiment, 30 mL of 0.01 mol

161 L-1 KCl was placed in a 50-mL volumetric tube and continuously shaken at 175 rpm

SC
162 on a gyratory shaker for 24 h under controlled temperature conditions (25 ºC). The

U
163 mixtures were centrifuged at 2000 rpm to extract supernatant for the P concentration
AN
164 test. Phosphate was determined by the same method described above. P desorption

165 capacities (qd, mg g-1) were calculated form equation (7) as follows:
M

"
166
! = (7)

D

167 Where C is the equilibrium P concentration in the supernatant (mg L-1), M is the
TE

168 mass of sample (g), and V is the volume of the solution (L).

169 An exponential growth model with two parameters was employed to describe the
EP

170 amount of P desorbed from the mortar filter media. The model is as follows:
C

171
! = #$ %& (8)

AC

172 Where q and qd are the same as the parameters in equation (1) and (7), respectively,

173 and a (mg kg-1) and b (g mg-1) are constants related to desorption potential and

174 desorption rate, respectively.

175 2.5 FESEM/EDX measurements

176 Each mortar powder sample before and after the P-adsorption experiments was
 ACCEPTED MANUSCRIPT

177 subjected to field emission scanning electron microscopy (FESEM) and

178 energy-dispersive X-ray spectroscopy (EDX), in order to examine its localized

179 morphology and elemental distributions at the microscopic scale. A Zeiss Supra

180 PGT/HKL system coupled with the energy-dispersive X-ray analyzer was used in

PT
181 these investigations. The FESEM was used to investigate the samples’ morphology,

RI
182 by collecting data from at least three randomly selected sites. Care was taken in the

183 examination of each sample and the random selection of the sites avoided aggregates

SC
184 or overlapping areas. SEM micrographs were taken using an SE2 detector, at

U
185 magnifications beginning at 100 times and up to 20, 000 times. The EDX data were
AN
186 obtained using a micro-analytical unit that featured the ability to detect small variation

187 in trace element contents. For the EDX analysis, an accelerating voltage of 20 kV was
M

188 used with a scan time of 45-120 s per sampling area. Areas used for EDX analysis
D

189 corresponded directly to the FESEM morphological examination, and at least three
TE

190 areas were analyzed for each sample.

191 2.6 XRD analyses

EP

192 X-ray diffraction (XRD) was used to characterize the precipitates that had formed in
C

193 the P-adsorption experiments. The ground mortar samples were removed from the
AC

194 P-adsorption experiments and dried, then the dry samples were sieved to eliminate the

195 sand in them. The obtained solid was smeared on a slide and analyzed with a

196 D/max-RB diffractometer (Rigaku, Japan) using Cu-Kα radiation (λ= 1.5418 Å) in the

197 2θ range of 5° to 60°.

198
 ACCEPTED MANUSCRIPT

199 3. Results and discussion

200 3.1 Chemical composition of the mortars

201 The EDX analyses confirmed that the main chemical components of these four

202 hardened mortars include SiO2, CaO, Al2O3, and Fe2O3 (Table 1). The percent CaO

PT
203 and Al/Ca and Si/Ca ratios fall in reasonable ranges observed before in the mortar

RI
204 phase of Portland cement concrete (Shi et al., 2011). Mortars A and B exhibited

205 significantly higher CaO content than mortar C (mean value at 15.9 % and 14.1 % vs.

SC
206 9.9 %), which is attributable to the higher CaO content in their binder (63.3 % in

U
207 cement vs. 30.6 % in fly ash). Mortar D also exhibited significantly higher CaO
AN
208 content than mortar C (mean value at 13.9 % vs. 9.9 %), which is attributable to the

209 inclusion of gypsum in its mix design. The SiO2 contents and the Si/Ca ratios may be
M

210 complicated by the presence of fine aggregates and thus excluded from discussion.
D

211 Mortar C featured the highest mean Al/Ca and Fe/Ca ratios (0.62 and 0.35), followed
TE

212 by mortar D (0.43 and 0.25), mortar B (0.18 and 0.12), and mortar A (0.11 and 0.07).

213 The high Al2O3 and Fe2O3 contents in mortars C and D are attributable to the higher
EP

214 Al2O3 and Fe2O3 contents in their binder (17.3 % and 6.5 % in fly ash vs. 3.7 % and
C

215 3.4 % in cement, respectively).

AC

216 3.2 P-adsorption isotherms

217 The phosphate adsorption isotherms of the mortar powders used in the study were

218 determined by plotting the equilibrium concentration of phosphate (C) against the

219 amount of phosphate adsorbed (Q). The adsorption isotherms reveal that the four

220 hardened mortar samples exhibited different adsorption characteristics (Fig. 1). It is
 ACCEPTED MANUSCRIPT

221 evident from all the curves that the rate of P-adsorption increased with the P

222 concentration yet levelled off at higher P concentration as the ground mortar reached

223 its capacity to adsorb P.

224 Langmuir and Freundlich isotherm equations were both found to be suitable for

PT
225 describing P-adsorption characteristics of the mortar samples. However, Langmuir

RI
226 isotherm equations are better suited to mortars A, B and C, while Freundlich isotherm

227 equations are better suited to mortar D. This is illustrated by the relatively high

SC
228 R-square values shown in Table 2. Table 2 also presents the results of model

U
229 estimation, which show that P-bonding energy (KL) and adsorption capacity (K)
AN
230 follows the order of C>D>B>A for the four mortar mixes investigated. This coincides

231 with the order of Al/Ca and Fe/Ca ratios in the hardened mortars, implying that the KL
M

232 and K values have a positive correlation with the total content of Al2O3 and Fe2O3 in
D

233 each mortar.

TE

234 Comparing the maximum amount of P adsorbed in the samples (Qm), it is evident

235 that each mortar mix features different capacities to adsorb P (Table 2). The
EP

236 experimental results indicate that mortar A has the highest value of maximum
C

237 adsorption at 30.96 mg g-1, followed by mortar B, mortar D and mortar C with
AC

238 maximum P adsorption values of 23.75, 20.20 and 12.7 mg g-1 respectively. This

239 confirms the previous report that the Qm value has a positive correlation with the CaO

240 content in each mortar (Barca, 2012). In addition, the 1/n value (adsorption intensity)

241 follows the same trend as Qm, likely sharing a positive correlation with the CaO

242 content in each mortar.

 ACCEPTED MANUSCRIPT

243 The experimental results from this study are comparable with those reported in

244 the literature for other materials based on batch experiments. Many studies

245 demonstrated that slag and fly ash are good filters for use in CWS (Barca 2012;

246 Bowden et al., 2009; Xu et al., 2006). The maximum P-sorption of various fly ashes

PT
247 ranged from 5.51 to 42.55 mg g-1 (Chen et al., 2007). The highest P-adsorption

capacity of furnace slag was 8.89 mg g -1, while that of sand was only 0.13 mg g-1 (Xu

RI
248

249 et al., 2006).

SC
250 Before extending the P-adsorption experimental results to field practices,

U
251 however, a longer-term investigation should be conducted as suggested by recent
AN
252 studies (Drizo et al., 2002, 2006; Vohla et al., 2011; Westholm, 2006). The Langmuir

253 equation has been widely utilized in P-adsorption studies for the calculation of
M

254 theoretical P-adsorption maxima (Drizo et al., 1999). Yet, the P saturation potential
D

255 derived from the Langmuir isotherm may lead to a biased and unrealistic estimate of
TE

256 the adsorption parameter (Drizo et al., 2002). Despite the limitations, the current study

257 reveals valuable information about adsorption characteristics of ground mortar

EP

258 samples and their potential use for removal of phosphorous from wastewater.
C

259 3.3 P-desorption from synthetic solutions

AC

260 Fig. 2 illustrates that there exhibited a good fit of P-desorption data with the

261 two-parameter exponential growth model used to describe desorption kinetics (see

262 Eqn. (8)). Note that chloride anions were added to replace the phosphate anions

263 adsorbed on the ground mortar. When the amount of P adsorbed (Q) was small, the

264 amount of P desorbed from the mortar increased slowly with Q. Once the amount of P
 ACCEPTED MANUSCRIPT

265 adsorbed approaches the maximum capacity (Qm), the amount of P desorbed from the

266 mortar increased sharply with Q. Although the parameters a and b in Eqn. (8) are

267 related to desorption potential and desorption rate respectively, one cannot judge the

268 desorption potential directly by the value of two parameters. Instead, a linear

PT
269 relationship was obtained by plotting the logarithm of P-adsorption capacities against

RI
270 the logarithm of P desorption capacities, the slope of which was applied to determine

271 the desorption potential of each mortar mix (see Fig. S1 and Table S1 in the

SC
272 Supporting Information). The results reveal that the desorption potential of the four

U
273 ground mortar mixes follows the order of B>A>C>D. This suggests that the
AN
274 phosphates adsorbed onto the ground mortar D remained more stable than those onto

275 other mortar samples.

M

276 The mechanisms underlying the observed differences in the P-desorption results
D

277 of these mortar mixes merit further investigation. With the limited data from this study,
TE

278 one can hypothesize the P-immobilization benefits of using fly ash instead of cement

279 as the sole binder and a preference to inclusion of dehydrated gypsum (with additional
EP

280 CaO) and possibly a higher water/binder ratio and/or inclusion of nano-sized zeolite.
C

281 Unhydrated fly ash has been widely demonstrated as an outstanding adsorbent for
AC

282 removing P from wastewater (Westholm, 2006). It was reported that fly ash

283 immobilized phosphate via adsorption or binding much more preferentially than

284 ordinary Portland cement (Agyei et al., 2002).

285 The phosphate anions desorbed by Cl- are mainly in an exchangeable form

286 adsorbed on the surface of mineral or metal oxide particles, and this fraction of P is
 ACCEPTED MANUSCRIPT

287 considered to be reactive and thus a risk of secondary pollution. However, the results

288 shown in Table 3 demonstrate that only a small fraction of phosphates can be

289 desorbed. This fraction is only 0.3 % for mortar mix D, and does not exceed 4.3 % for

290 the other mixes.

PT
291 3.4 P-sequestration from synthetic solutions and underlying mechanisms

RI
292 The results of P-sequestration from synthetic solutions are shown in Fig. 3. All

293 the ground powder of the four hardened mortar mixes exhibited outstanding

SC
294 P-sequestration capabilities across the wide range of initial phosphate concentrations.

U
295 All the samples (except mortar C) achieved more than 95% removal over the range of
AN
296 20 to 600 mg L-1 P and more than 94 % removal at the 1000 mg L-1 of P concentration.

297 The amount of P adsorbed (Q) increased as the initial P concentration increased from
M

298 20 to 1000 mg L-1, suggesting that a limit in P-adsorption was not reached. On the
D

299 other hand, the removal rate decreased once the initial P concentration exceeded 320
TE

300 mg L-1, suggesting that a limit in P-sequestration was reached.

301 Several relative ratios of critical elements indicative of the mortar samples and
EP

302 their possible reaction products with synthetic solutions (600 mg L-1 P) were analyzed
C

303 based on EDX results (Fig. 4). The Si/Ca and Al/Ca ratios of the mortar powders all
AC

304 increased after exposure to synthetic solutions without P added, suggesting the

305 preferential leaching of Ca2+ and Al3+ from the mortar surface. The subsequent

306 addition of KH2PO4 into the synthetic solutions led to significant decreases in the

307 Si/Ca and Al/Ca ratios of the mortar powders, suggesting the presence of reaction

308 products of Ca2+ and Al3+ with PO43- etc.

 ACCEPTED MANUSCRIPT

309 The mortar surface before and after the P-sequestration experiments was

310 examined by SEM and EDX analyses. Figures S2 to S5 illustrate the surface

311 morphology and chemical composition of mortars A, B, C, and D, respectively, both

312 before and after the P-adsorption experiments. For all the mortar mixes, the SEM

PT
313 results reveal that the exposure of mortar powder to the synthetic P-containing

RI
314 wastewater led to the formation of crystalline phases on the mortar surface. The EDX

315 data show that the surface of original mortars A and B mainly consisted of Si, Ca and

SC
316 O, whereas their surface after P-adsorption was dominated by a layer rich in Ca, Si, O

U
317 and P. Moreover, the XRD data (see Fig. S6 in the Supporting Information) identified
AN
318 SiO2 and CaH2P2O7 as the predominant crystalline phases, indicating that

319 P-sequestration was mainly based on precipitation of Ca-PO4-P complexes. This

M

320 finding is consistent with a previous study (Bowden et al., 2009). For mortar C, the
D

321 EDX data show that its original surface mainly consisted of Si, Ca, O and Al, whereas
TE

322 its surface after P-adsorption was dominated by a layer rich in Ca, Si, Al, O, and P.

323 The EDX data of mortar D were similar to those of mortar C, except the significant
EP

324 presence of S as a result of gypsum in the mix design. The XRD data identified SiO2,
C

325 CaHPO4·2H2O and AlPO4 were the predominant crystalline phases. This confirms the
AC

326 precipitation of AlPO4 as another mechanism for P-sequestration (Oǧuz et al., 2003).

327 These results support the mechanism that phosphate anions react with calcium cations

328 and at high phosphate levels lead to the precipitation of CaHPO4·2H2O (or CaH2P2O7)

329 onto the Ca-rich hydration products. This mechanism is consistent with a previous

330 study (Demirbas, 2003). However, these findings do not support the other proposed
 ACCEPTED MANUSCRIPT

331 mechanism where tricalcium phosphate was the main precipitation in ordinary

332 Portland cement exposed to phosphate solution for a relatively short time (Agyei et al.,

333 2003). This may be attributed to the low pH of the phosphate solutions (1.69 for P

334 concentration of 1000 mg L-1). Tricalcium phosphate and dicalcium phosphate have

PT
335 been documented as precursors of hydroxyapatite (HAP) in a number of experimental

RI
336 studies, in the neutral to acidic pH range (Valsami-Jones, 2001). Thus, HAP has been

337 recognized, typically after 3-6 months, by a number of authors (Barca 2012; Bowden

SC
338 et al., 2009; Forbes et al., 2004).

U
339 In summary, the main mechanism of P-sequestration by mortar powders consists
AN
340 of two reactive steps. First, the interaction between mortar powder and water releases

341 Ca2+, Al3+ and OH- ions. Then, there is a precipitation mechanism in addition to weak
M

342 physical interactions between the surface of adsorbent filter media and the metallic
D

343 salts of phosphate.

TE

344 3.5 Practical implications

345 When designing a CWS for P-sequestration, the selection of material for a
EP

346 wetland substrate plays a crucial role (Brix et al., 2001). CWSs have been extensively
C

347 investigated and regarded as a promising way for treating various types of wastewater
AC

348 (Vohla et al., 2011; Westholm, 2006). This study has demonstrated that hydrated

349 mortar can be used as a type of promising substrate for CWS. The estimated cost of

350 hydrated mortar is approximately $23 per ton, which is not substantially higher than

351 that of the commonly used materials ($15 per ton for peal gravel and $10 to $15 per

352 ton for crushed rock). It is speculated that the powdered mortar could be packaged by
 ACCEPTED MANUSCRIPT

353 geotextile or other materials to avoid being washed away. It is also desirable to

354 remove fly ashes from the waste stream and use them for valued-added applications

355 such as wastewater best management practices (BMPs). Besides retention and

356 precipitation by substrates, the removal mechanisms for phosphorus in CWSs also

PT
357 include: plant uptake, accretions of wetland soils, and microbial immobilization (Diaz

RI
358 et al, 1994; Greenway and Woolley, 1999; Kadlec 1997; Newbold et al., 1983; Reddy

359 et al., 1999; Tanner et al., 1998). Since P is retained within the wetland, its ultimate

SC
360 removal from the CWS is achieved by periodically harvesting plants and removing

U
361 the saturated filter media. The saturated filter media (phosphate saturated mortar
AN
362 powder) may be used as partial replacement of fine aggregate in concrete, thus

363 minimizing the risk of secondary pollution. Future research will explore the
M

364 engineering and environmental performance of concrete mixes containing such

D

365 recycled fine aggregate. Another line of research will explore methods to reduce the
TE

366 energy consumption in producing such mortar powders, without sacrificing their

367 effectiveness in P-sequestration.

EP

368 When planning to use the filter in a wetland system to obtain efficient and
C

369 cost-effective P-sequestration, the longevity of the system is one of the major
AC

370 questions that the designer must address. The lifetime of CWS may be calculated

371 using the maximum P sorption capacity of substrate and the area of CWS (Xu et al.,

372 2006). Given that a total of 7 g PO4 (or 2.3 g of P) per person per day is discharged

373 and assuming that the land area required to treat wastewater from one person is 5 m2

374 and the depth of the bed is 0.6 m, 3 m3 of substrate would be needed to treat the
 ACCEPTED MANUSCRIPT

375 effluent from one person (Drizo et al., 1999; EC/EWPCA, 1990; Laak, 1986). If

376 saturation is reached at 30.96 g kg-1 P (P-adsorption maximum of mortar A), the

377 amount of mortar A (3 m3 or 6.3 tonnes with a density of 2.1 g cm-3) would adsorb

378 195.05 kg P, which translates to a theoretical lifetime of 232 yr. However, if one uses

PT
379 maximal P adsorption capacity calculated from the Langmuir equation, it is easy to

RI
380 make an overestimation (Vohla et al., 2011). In addition, many studies have stated that

381 materials that show very high phosphorus removal in the laboratory do not give

SC
382 similarly good results in field systems (Àdàm et al., 2006; Arias et al., 2001). One

U
383 possible reason is the presence of competing ions in the field systems which degrade
AN
384 the effectiveness of P-adsorption by the filter material. Therefore, further

385 investigation of these materials is needed before their appropriate use in the field
M

386 environment can be recommended.

D

387 Finally, cementitious materials investigated in the present study can also be used
TE

388 as construction materials. If such materials were applied in the construction of septic

389 tanks or bio-retention tanks for wastewater or runoff treatment, it would be able to
EP

390 enhance the effect of P-sequestration. In that case, it is desirable to use the
C

391 investigated binders for pervious concrete or highly porous concrete so as to increase
AC

392 the hydraulic conductivity of the concrete and improve the performance of wastewater

393 treatment. Such concrete may be used for the inner structure of septic tanks and

394 bio-retention tanks to allow water to penetrate deeper into the materials.

395 4. Conclusions

396 The effects of cement-based and alternative cementitious materials on

 ACCEPTED MANUSCRIPT

397 phosphorus-sequestration from wastewater were investigated. For all four hydrated

398 and powdered mortars investigated, the P-sequestration rates were in excess of 80

399 percent for phosphorus concentrations ranging from 20 to 1000 mg L-1, while the

400 ratios of desorption were less than 5 percent. The P-bonding energy (KL) and

PT
401 adsorption capacity (K) exhibited a positive correlation with the total content of Al2O3

RI
402 and Fe2O3 in each mortar. The maximum amount of P adsorbed (Qm) and adsorption

403 intensity (1/n) exhibited a positive correlation with the CaO content in each mortar.

SC
404 The removal of phosphate predominantly followed a precipitation mechanism in

U
405 addition to weak physical interactions between the surface of adsorbent filter media
AN
406 and the metallic salts of phosphate. The use of these cementitious materials in the

407 form of ground powder shows great promise as a filter material for constructed
M

408 wetland system.

D

409
TE

410 Acknowledgements

411 The work was financially supported by Construction & Technology Projects in
EP

412 Ministry of Transportation (20113186701290 and 2013318490010)，Key Projects in

C

413 the National Science & Technology Pillar Program (2014BAG05B06), Construction
AC

414 & Technology Projects in Department of Transportation of Jiangxi Province

415 (2011C00005). The authors acknowledge postdoctoral visiting scholar grant for Dr.

416 Xinjun Wang provided by the Chinese Scholarship Council. In addition, the authors

417 would like to thank Dr. Yudong Dang from Tongji University and Dr. Ning Xie form

418 Harbin Institute of Technology for their professional assistance.

 ACCEPTED MANUSCRIPT

419

420 References

421 Àdàm, K., Søvik, A.K., Krogstad, T., 2006. Sorption of phosphorous to Filtralite-P

TM
422 —the effect of different scales. Water Research 40(6), 1143–1154.

PT
423 Agyei, N.M., Strydom, C.A., Potgeiter, J.H., 2002. The removal of phosphate ions

RI
424 from aqueous solution by fly ash, slag, ordinary portland cement and related

425 blends. Cement and Concrete Research 32(12), 1889–1897.

SC
426 Agyei, N.M., Strydom C.A., Potgieter, J.H., 2003. Reply to the discussion by A.

U
427 Demirbas of the paper "The removal of phosphate ions from aqueous solution by
AN
428 fly ash, slag, ordinary Portland cement and related blends”. Cement and Concrete

429 Research 33(6), 937-937.

M

430 Ahmaruzzaman, M., 2010. A review on the utilization of fly ash. Progress in Energy
D

431 and Combustion Science 36 (3), 327-363.

TE

432 American Coal Ash Association., 2004. 2003 Coal combustion product (CCP)

433 production and use survey, [Link] Portals/9/

EP

434 Files/PDFs/2003CCP Survey (10-1-04).pdf. Accessed October 1, 2004.

C

435 APHA., AWWA., WCF., 1995. Standard methods for the examination of water and
AC

436 wastewater, 19th ed. American Public Health Association, Washington DC, pp

437 4-112.

438 Arias, C.A., Del Bubba, M., Brix, H., 2001. Phosphorus removal by sands for use as

439 media in subsurface flow constructed reed beds. Water Research 35 (5), 1159–

440 1168.
 ACCEPTED MANUSCRIPT

441 Atis, C.D., 2003. High-volume fly ash concrete with high strength and low drying

442 shrinkage. Journal of materials in civil engineering 15(2), 153-156.

443 Barca, C., 2012. Phosphate removal from synthetic and real wastewater using steel

444 slags produced in Europe. Water Research 46(7), 2376-2384.

PT
445 Bowden, L.I., Jarvis, A.P., Younger, P.L., Johnson, K.L., 2009. Phosphorus removal

RI
446 from wastewaters using basic oxygen steel slag. Environmental Science and

447 Technology 43 (7), 2476-2481.

SC
448 Brix, H., 1993. Wastewater treatment in constructed wetlands: system design, removal

U
449 processes, and treatment performance. In: Constructed wetlands for water quality
AN
450 improvement. Edited by Moshiri G. A., CRC Press. Boca Raton, FL., pp 9-22.

451 Brix, H., Arias, C.A., Del Bubba, M., 2001. Media selection for sustainable
M

452 phosphorus removal in subsurface flow constructed wetlands. Water Science and
D

453 Technology 44 (11-12), 47-54.

TE

454 Chen, J., Kong, H., Wu, D., Chen, X., Zhang, D., Sun, Z., 2007. Phosphate

455 immobilization from aqueous solution by fly ashes in relation to their

EP

456 composition. Journal of Hazardous Materials 139(2), 293-300.

C

457 De-Bashan, L.E., Bashan, Y., 2004. Recent advances in removing phosphorus from
AC

458 wastewater and its future use as fertilizer (1997–2003). Water Research 38(19),

459 4222-4246.

460 Demirbas, A.A., 2003. Discussion of the paper "The removal of phosphate ions from

461 aqueous solutions by fly ash, slag, ordinary Portland cement and related blends"

462 By N.M. Agyei, C. A. Strydom and J.H. Potgieter.". Cement and Concrete
 ACCEPTED MANUSCRIPT

463 Research 33(6), 935-935.

464 Diaz, O.A., Reddy, K.R., Moore, P.A., 1994. Solubility of inorganic phosphorus in

465 stream water as influenced by pH and calcium concentration. Water Research

466 28(8), 1755–1763.

PT
467 Drizo, A., Comeau, Y., Forget, C., Chapuis, R.P., 2002. Phosphorus saturation

RI
468 potential: a parameter for estimating the longevity of constructed wetland

469 systems. Environmental Science and Technology 36(21), 4642-4648.

SC
470 Drizo, A., Forget, C., Chapuis, R.P., Comeau, Y., 2006. Phosphorus removal by

U
471 electric arc furnace steel slag and serpentinite. Water Research, 40(8),
AN
472 1547-1554.

473 Drizo, A., Frost, C.A., Grace, J., Smith, K.A., 1999. Physico-chemical screening of
M

474 phosphate-removing substrates for use in constructed wetland systems. Water

D

475 Research 33(17), 3595-3602.

TE

476 EC/EWPCA., 1990. European design and operations guidelines for reed bed treatment

477 systems. Presented to the Conference ''Constructed Wetlands in Water Pollution

EP

478 Control'', Cambridge, UK.

C

479 Forbes, M.G., Dickinson, K.R., Golden, T.D., Hudak, P., Doyle, R.D., 2004.
AC

480 Dissolved Phosphorus Retention of Light-Weight Expanded Shale and Masonry

481 Sand in Subsurface Flow Treatment Wetlands. Environmental Science and

482 Technology 38(3), 892–898.

483 Greenway, M., Woolley, A., 1999. Constructed wetlands in Queensland: performance

484 efficiency and nutrient bioaccumulation. Ecological Engineering 12(1-2), 39–55.

 ACCEPTED MANUSCRIPT

485 Horst, M., Welker, A.L., Traver, R.G., 2010. Multiyear performance of a pervious

486 concrete infiltration basin BMP. Journal of Irrigation and Drainage Engineering

487 137(6), 352-358.

488 Kadlec, R.H., 1997. An autobiotic wetland phosphorus model. Ecological Engineering

PT
489 8(2), 145–172.

RI
490 Laak, R., 1986. Wastewater Engineering Design for Unsewered Areas, Technomic

491 Publishing, The Hague, Netherlands.

SC
492 Luck, J.D., Workman, S.R., Coyne, M.S., Higgins, S.F., 2009. Consequences of

U
493 manure filtration through pervious concrete during simulated rainfall events.
AN
494 Biosystems Engineering 102(4), 417-423.

495 Mehta, P.K., 2004. High-performance, high-volume fly ash concrete for sustainable
M

496 development. In Proceedings of the international workshop on sustainable

D

497 development and concrete technology, pp. 3-14.

TE

498 Newbold, D.J., Elwood, J.W., O’Neil, R.V., Sheldon, A.L., 1983. Phosphorus

499 dynamics in a woodland stream ecosystem: a study of nutrient spiraling. Ecology

EP

500 64(5), 1249–1263.

C

501 Oǧuz, E., Gurses, A., Canpolat, N., 2003. Removal of phosphate from wastewaters.
AC

502 Cement and Concrete Research 33(8), 1109-1112.

503 Reddy, K.R., Kadlec, R.H., Flaig, E., Gale, P.M., 1999. Phosphorus retention in

504 streams and wetlands: a review. Critical Review in Environmental Science and

505 Technology 29(1), 83–146.

506 Shi, X., Fay, L., Peterson, M.M.; Berry, M., Mooney, M.A., 2011. FESEM/EDX
 ACCEPTED MANUSCRIPT

507 investigation into how continuous deicer exposure affects the chemistry of

508 portland cement concrete. Construction and Building Materials 25(2), 957-966.

509 Shi, X., Xie, N., Fortune, K., Gong, J., 2012. Durability of steel reinforced concrete in

510 chloride environments: An overview. Construction and Building Materials 30,

PT
511 125-138.

RI
512 Shi, X., Yang, Z., Liu, Y., Cross, D., 2011. Strength and corrosion properties of

513 portland cement mortar and concrete with mineral admixtures. Construction and

SC
514 Building Materials 25(8), 3245-3256.

U
515 Tanner, C.C., Sukias, J.P.S., Upsdell, M.P., 1998. Relationships between loadingrates
AN
516 and pollutant removal during maturation of gravel-bed constructed wetlands.

517 Journal of Environmental Quality 27(2), 448–458.

M

518 Valsami-Jones, E., 2001. Mineralogical controls on phosphorus recovery from

D

519 wastewaters. Mineralogical Magazine 65 (5), 611-620.

TE

520 Vohla, C., Kõiv, M., Bavor, H.J., Chazarenc, F., Mander, U., 2011. Filter materials for

521 phosphorus removal from wastewater in treatment wetlands - a review.

EP

522 Ecological Engineering 37 (1), 70-89.

C

523 Vymazal, J., 2013. The use of hybrid constructed wetlands for wastewater treatment
AC

524 with special attention to nitrogen removal: A review of a recent development.

525 Water Research 47(14), 4795-4811.

526 Wang, S., Wu, H., 2006. Environmental-benign utilization of fly ash as low-cost

527 adsorbents. Journal of Hazardous Materials 136(3), 482-501.

528 Wang, X.J., Kong, Y.P., Chen, J.D., Liu, X.X., 2012. Present status about wastewater
 ACCEPTED MANUSCRIPT

529 treatment at highway affiliated facilities of China. Advanced Material Research

530 356-360, 2673-2677.

531 Westholm L.J., 2006. Substrates for phosphorus removal - potential benefits for

532 on-site wastewater treatment. Water research 40(1), 23-36.

PT
533 Wu, T., Sansalone, J., 2013. Phosphorus equilibrium. II: Comparing filter media,

RI
534 models, and leaching. Journal of Environmental Engineering 139(11), 1325-1335.

535 Xu, D.F., Xu, J.M., Wu J.J., Muhammad, A., 2006. Studies on the phosphorus

SC
536 sorption capacity of substrates used in constructed wetland systems.

U
537 Chemosphere 63(2), 344-352.
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

Table 1. Chemical composition of the hardened mortars, based on the EDX data

PT
A B C D
wt.%

RI
Min Max Mean Min Max Mean Min Max Mean Min Max Mean
SiO2 14.3 25.9 18.4 13.1 38.2 21.8 7.3 19.5 14.6 14.0 19.2 17.2

SC
CaO 11.1 20.1 15.9 11.1 18.7 14.1 7.3 12.2 9.9 13 15.2 13.9
Al2O3 1.6 3.0 2.5 3.0 3.9 3.3 7.7 9.5 8.3 6.8 9.4 8.1

U
Fe2O3 0.8 1.2 1.1 0.8 3.7 1.8 1.8 4.8 3.5 2.9 4.3 3.5

AN
MgO 0.2 0.7 0.3 -- -- -- -- -- -- 3.1 3.8 3.4
Si/Ca 0.84 0.84 0.76 0.77 1.34 1.01 0.65 1.05 0.96 0.70 0.83 0.81
Al/Ca 0.11 0.11 0.11 0.20 0.15 0.18 0.78 0.58 0.62 0.39 0.46 0.43

M
Fe/Ca 0.07 0.06 0.07 0.07 0.19 0.12 0.24 0.38 0.35 0.22 0.28 0.25
Mg/Ca 0.02 0.03 0.02 -- -- -- -- -- -- 0.20 0.21 0.21

D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

Table 2. Langmuir and Freundlich parameters for P adsorption on mortar

Langmuir model Freundlich model

-1
Mortar Qm (mg g ) KL (l mg -1) R 2
K 1/n R2
A 30.96 0.034 0.9676 0.98 0.8987 0.8781
B 23.75 0.065 0.9313 1.24 0.8147 0.9281
C 12.7 0.088 0.9546 1.83 0.3488 0.7858

PT
D 20.2 0.079 0.8727 1.43 0.6871 0.9547

RI
U SC
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT
Table 3. The ratio of desorption and maximum adsorption of phosphorus in saturated
mortar

Mortar A B C D
Qm/mg g-1 30.96 23.75 12.70 20.20
maximum desorption 1010.1 1020.5 505.3 64.0
of phosphorus/mg Kg-1
The ratio of 3.26 4.30 3.98 0.32

PT
desorption/%

RI
U SC
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

A B

PT
RI
C D

U SC
AN
M

Figure 1. Phosphorus adsorption isotherms of mortar A, B, C and D

(……. Langmuir model; — Freundlich model)
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

1200
A B C
Desorbed P (mg/kg) 1000 D

800

600

PT
400

RI
200

SC
0
0 5 10 15 20 25
Q (mg/g)

U
Figure 2. Desorption of adsorbed P from mortar A, B, C and D
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

20 100
18

b-Percentage P removal (%)

16 A 96
14 B
12 C 92
a-Q (mg/g)

10 D

PT
8 A 88
6 B
4 C 84

RI
D
2
0 80

SC
20 40 80 160 320 600 1000
Initial P concentration (mg/L)
Figure 3. Effect of range of P concentration on P adsorption capacity (a) and P

U
removal (b) of mortar A, B, C and D.
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

A B

PT
RI
SC
C D

U
AN
M
D
TE

Figure 4. The chemical changes of mortar A, B, C and D induced by synthetic

solutions. 1. mortar samples before P removal experiment, 2. mortar samples
subjected to synthetic solutions without P, 3. mortar samples subjected to P solution.
C EP
AC
 ACCEPTED MANUSCRIPT

Highlights

1. We tested the P-sequestration capability of hydrated cementitious materials.

2. Both Langmuir and Freundlich adsorption isotherms were suitable.

3. Four mortars showed maximum P-adsorption of 30.96., 23.75, 20.20, and 12.7

PT
mg g-1 respectively.

RI
4. P-sequestration rates of three mortars exceeded 94% for P solutions at 20-1000

mg/L.

SC
5. The underlying mechanisms were examined using FESEM/EDX/XRD.

U
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

Supporting Information

Sequestration of phosphorus from wastewater by cement-based or alternative

cementitious materials

Xinjun Wanga,b, Jiding Chen b, Yaping Kong b, Xianming Shia,*.

PT
a
Corrosion and Sustainable Infrastructure Laboratory, Western Transportation

RI
Institute and Civil Engineering Department, PO Box 174250, College of Engineering,

Montana State University, Bozeman, MT 59717-4250, USA

SC
b
Laboratory of Road Ecology and Pollution Emergency Response, Research Center

U
for Environmental Protection and Transportation Safety, China Academy of
AN
Transportation Sciences, Beijing 100029, China

*Corresponding author. Tel.: +1 406 994 6486. E-mail address:

M

xianming_s@[Link]
D
TE

Table S1. P-desorption equation parameters

a (mg kg-1) b (g mg -1) R2

EP

A 1.7534 0.3077 0.911

B 3.4251 0.297 0.926
C 7.574 0.1835 0.828
C

D 1.0779 0.1925 0.661

AC
 ACCEPTED MANUSCRIPT

A B

PT
RI
SC
C D

U
AN
M
D

Figure S1. Desorption potential of Mortar A, B,C and D

A linear equation was obtained by plotting logarithmic of P absorption capacities (Q)
TE

against logarithmic of P desorption capacities(q). The slope was applied to determine

the desorption potential of each mortar mix.
C EP
AC
 ACCEPTED MANUSCRIPT
a b

PT
c d

RI
U SC
Figure S2. SEM and EDX analyses of mortar A. (a) surface of mortar A before P
AN
removal; (b) chemical analyses of mortar A before P removal; (c) surface of mortar A
before P removal; (d) chemical analyses of mortar A after P removal.
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

a b

PT
c d

RI
U SC
Figure S3. SEM and EDX analyses of mortar B. (a) surface of mortar B before P
AN
removal; (b) chemical analyses of mortar B before P removal; (c) surface of mortar B
before P removal; (d) chemical analyses of mortar B after P removal.
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

a b

PT
c d

RI
U SC
AN
Figure S4. SEM and EDX analyses of mortar C. (a) surface of mortar C before P
removal; (b) chemical analyses of mortar C before P removal; (c) surface of mortar
Cbefore P removal; (d) chemical analyses of mortar C after P removal.
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT
1
2
3
4
a b
5
6
7
8

PT
9
10
11

RI
12
c d
13

SC
14
15
16
17

U
18
AN
19
20
21
Figure S5. SEM and EDX analyses of mortar D. (a) surface of mortar D before P
M

22
23 removal; (b) chemical analyses of mortar D before P removal; (c) surface of mortar D
24 before P removal; (d) chemical analyses of mortar D after P removal.
D

25
26
TE

27
28
29
EP

30
31
32
C

33
34
AC

35
36
37
38
39
40
41
42
43
44
 ACCEPTED MANUSCRIPT
1
2
3 A B
4
5
6
7
8

PT
9
10
11

RI
12
C D
13

SC
14
15
16
17

U
18
AN
19
20
21
Figure S6. XRD analyses of mortar A, B, C and D. A and B: Crystals are mainly SiO2
M

22
23 and CaH2P2O7; C and D: Crystals are mainly SiO2, CaHPO4·2H2O and AlPO4.
D
TE
C EP
AC