- Email: [email protected]
Model-based evaluation of tetracycline hydrochloride removal and mineralization in an intimately coupled photocatalysis and biodegradation reactor
Chemical Engineering Journal 351 (2018) 967–975
Contents lists available at ScienceDirect
Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej
Model-based evaluation of tetracycline hydrochloride removal and mineralization in an intimately coupled photocatalysis and biodegradation reactor ⁎
Yue Maa, Houfeng Xiongb, Zhiquan Zhaoc, Yang Yuc,d, Dandan Zhouc,d, , Shuangshi Donga,
T
⁎
a
Key Laboratory of Groundwater Resources and Environment, Ministry of Education, Jilin University, Changchun 130021, China School of Chemistry and Environmental Engineering, Jiujiang University, Jiujiang 332005, China c State Key Laboratory of Urban Water Resource and Environment, Harbin Institute of Technology, Harbin 150090, China d Jilin Engineering Lab for Water Pollution Control and Resources Recovery, Northeast Normal University, Changchun 130117, China b
H I GH L IG H T S
G R A P H I C A L A B S T R A C T
models of intimately coupled • Kinetic photo-biological reaction (ICPB) were developed.
(TCH) removal kinetics • Tetracycline by ICPB were successfully validated. root square errors between ex• The perimental results and model simulations were at max 2.1%.
in ICPB played a cru• Biodegradation cial role in modeling TCH removal and mineralization.
A R T I C LE I N FO
A B S T R A C T
Keywords: Intimately coupled photocatalysis and biodegradation (ICPB) Modeling Refractory pollutants Kinetics Mechanisms
Intimately coupled photocatalysis and biodegradation (ICPB) shows great potential for treatment of refractory pollutants; however, no kinetics for modeling ICPB performance has been developed and the major challenge is to determine the relationship between photocatalysis and biodegradation. In this work, we developed a simplified kinetic model to predict removal and mineralization of a target pollutant (tetracycline hydrochloride; TCH) by hypothesizing that all of the biodegradable photocatalysis products are immediately bio-utilized. Combined with a second-order photocatalytic kinetic model and Monod-type biodegradation model, we observed the interactions between photocatalysis and biodegradation in ICPB. Parameters in the kinetic equations were estimated using the First Optimization software to fit the experimental data to the proposed model with nonlinear regression. Our experimental results showed that TCH and chemical oxygen demand (COD) removal were as high as 94% and 70% within 8 h, respectively. TCH was transformed to non-toxic intermediates in only 4 h. Significantly, the kinetic models could satisfactorily predict the TCH and COD removal, and agreed well with the experimental data with an R2 > 0.92. The models confirmed that biodegradation in ICPB played a major role in accelerating TCH and its intermediates removal and mineralization, as the kinetic coefficient k1 of ICPB was 10% greater than that of photocatalysis alone. The developed models accurately predicted the ICPB efficiencies, and revealed the mechanisms of ICPB operation.
⁎
Corresponding authors at: School of Environment, Northeast Normal University, Changchun 130117, China (D. Zhou). E-mail addresses: [email protected] (D. Zhou), [email protected] (S. Dong).
https://doi.org/10.1016/j.cej.2018.06.167 Received 11 February 2018; Received in revised form 16 May 2018; Accepted 25 June 2018 Available online 28 June 2018 1385-8947/ © 2018 Elsevier B.V. All rights reserved.
Chemical Engineering Journal 351 (2018) 967–975
Y. Ma et al.
Nomenclature
Rm Ks X RS [TiO2] Cmod Cexp P AD B PC ICPB
[TCH] TCH concentration (mg L−1) [CODT] total bulk chemical oxygen demand (mg L−1) [CODTCH] theoretical COD of the remaining TCH (mg L−1) [CODINT] amount of intermediates (=[CODT] – [CODTCH], mg L−1) k0 production rate constant of RS (h−1) k1 second-order rate constant for degradation of TCH (L h−1 mol−1) k2 second-order rate constant for degradation of Int. (L h−1 mol−1) k3 second-order rate constant for degradation of COD (L h−1 mol−1)
1. Introduction
maximum specific cell growth rate (h−1) TCH phenol half-saturation constant (mg L−1) biomass concentration at stable state (mg L−1) reactive species (mol L−1) TiO2 concentration (mg L−1) value of model calculated results (mg L−1) value of experimental results (mg L−1) protocol of photolysis alone protocol of adsorption alone protocol of biodegradation alone visible-light-induced photocatalysis alone intimately coupled photocatalysis and biodegradation
TCH was employed as model pollutant and the experimental data for TCH removal and mineralization were directly compared with the simulated results. This study is important for predicting removal and mineralization of refractory pollutants by ICPB technology, and achieving a direct scale-up and operation of ICPB reactors.
Intimately coupled photocatalysis and biodegradation (ICPB) technology shows great potential for enhancing the removal and mineralization of refractory pollutants [1–3]. In a typical ICPB system, millimeter-sized macro-porous carriers support a photocatalyst on their outer surface and a biofilm within their pores [3]. The concept is that bacteria in the carrier pores are protected from photocatalytic oxidation. This arrangement enables bacteria away from free radical damage but is close enough to rapid degradation of the biodegradable photocatalytic products. ICPB in treatment of bio-refractory pollutants overcomes the weaknesses of incomplete mineralization by photocatalysis alone or inhibition of toxic compounds by bioreactor alone [3–5]. Currently, ICPB has been successfully applied for the treatment of many refractory pollutants, such as phenols [1–3], dinitrotoluene [4] and dyes [5]. Recently, Xiong et al. [6] has used a visible-light-driven ICPB reactor to enhance TCH degradation and mineralization, with which the TCH and COD removal improved by 11% and 20% respectively by ICPB, compared to the photocatalysis alone. This was the first report of treatment of an antibiotic by ICPB. Notably, it was found that biofilm benefitted TCH degradation by reducing interaction of reactive species (RS) and TCH intermediates. Mathematical modeling serves as a both useful and reliable tool to probe into emerging technologies, while significantly reducing the workload related to process evaluation or optimization [7]. In particular, a kinetic model can be used to: (1) identify the key factors of degradation [8–10], (2) provide improved understanding of the complicated reaction mechanisms [11–13], (3) help to scale up relevant processes, and (4) predict the treatment duration and therefore running cost [14]. However, only the biodegrading kinetics of an ICPB reactor has previously been quantified, using an Aiba self-inhibition model [15]. This model lacks the structure of the real ICPB manifestation, i.e., mutually promotion between the photocatalysis and biodegradation [3–5]. The main challenge to determining an ICPB model is to integrate both photocatalytic and biological degradation, where the immediate consumption of photocatalytic intermediates by biodegradation has to be considered. To the best our knowledge, no mathematical model with the abovementioned real ICPB manifestation structure has been developed to date, consequently, the modeling evaluation of pollutants degradation in an ICPB remained unknown. In this work, we approached the mathematic model of ICPB for the first time, employing a strategy of hypothesizing that all the biodegradable photocatalysis products were immediately bio-utilized. This in fact constitutes an ideal ICPB reaction [6]. Modeling photocatalysis alone was achieved based on a secondorder kinetic model [6,16–18] and the biodegradation alone was simulated using the well-known Monod model [19]. Then, we developed the ICPB kinetic model by fully considering the transfer of the predicated photocatalytic products into the biodegradation models, under determination of the final removal and mineralization of pollutants.
2. Material and methods 2.1. Carriers, photocatalysts and biofilms We used a sol–gel method to prepare Ag/TiO2 nanoparticles, which are visible-light responsive photocatalysts. Briefly, Ag/TiO2 was prepared with titanium tetraisopropoxide [Ti(OCH(CH3)2)4] (SigmaAldrich Co. LLC., USA) and silver nitrate (Sigma-Aldrich Co. LLC., USA). Silver reduction and doping were conducted in sodium citrate tribasic dihydrate (Beijing Chemical Works, China) according to a previous report [20].This Ag/TiO2 sol was used for the following coating procedure. Polyurethane sponge carriers (2 mm × 2 mm × 2 mm) with a porosity of 87% and wet density of 0.89 g/mL (Hayi-diverse, Yixing, China) were coated with Ag/TiO2 by using the method of evaporation-induced self-assembly (EISA) [21]. Basically, Ag/TiO2 sol (50 mL) was ultrasonically dispersed in ethanol (150 mL) for 5 min and then the carriers (4 g) were added to the suspension. The carriers in the Ag/TiO2 sol mixture were further ultrasonically dispersed for another 5–10 min before drying in an oven at 80 °C by stirring with a glass rod every 30 min. After rinsing with deionized water in ultrasonic three times to remove the loosely attached Ag/TiO2 nanoparticles, porous carriers coated with photocatalysts were ready to use. The Ag/TiO2 coating amount was 0.39 g-catalyst/g-carrier, with a coating ratio of 28.3% (wt %). The coated carriers were immersed in an activated sludge (South WWTP, Changchun, China) for the initial attachment of microbes and the biofilm was developed at room temperature (∼20 °C) with a synthetic wastewater containing sodium acetate (330 mg/L), NH4Cl (29 mg/L), Na2HPO4·2H2O (8 mg/L), and NaH2PO4 (4 mg/L) in an internal loop airlift-driven reactor, which was described in our previous report [22]. When COD of the effluent was stabilized, the biofilm was considered to be mature. 2.2. Reactors and protocols Photolytic circulating-bed bioreactors of a working volume of 540 mL were used to simulate TCH degradation and mineralization in ICPB reactors with biofilm developed on the Ag/TiO2 coated sponge carriers, which was described in our previous publications [3], and schematically drawn in Fig. S1. Synthetic wastewater and biofilm carriers were circulated by bubbling air from the bottom of the reactor, supplied by a 35 W aeration pump (SOBO, Weifang, China) and controlled by a flow meter (superficial velocity in the draft tube was 968
Chemical Engineering Journal 351 (2018) 967–975
Y. Ma et al.
measurements, the samples were filtered through a 0.22 μm PVDF microporous membrane (GVHP04700, Millipore, USA) to remove suspended solids and biomass. TCH/COD was not adsorbed to the membrane due to its hydrophobic characteristic. Biomass concentrations were determined gravimetrically from the dry-weight loss of the biofilm attached to and detached from sponge carriers (oven temperature 80 °C). The sample morphology and surface elemental content were examined using a scanning electron microscope (SEM, Quanta 200 ESEM, FEI, USA) equipped with energy dispersive X-ray spectroscopy (EDX, Rigaku Corporation). The pretreatment of the samples for SEM was presented in the Supporting Informatiom (SI-1). Product toxicity was evaluated by the disc agar diffusion test (DAD test) based on the method of Sunil et al. [26] with modifications, which were detailed in SI-2. The Gram-positive strain Staphylococcus aureus (ATCC 6538) was purchased from Guangzhou industrial microbiology testing center and was used as model pathogen for the above testing, due to its sensitivity to antibiotics [26]. The bacterial inoculum was initially preserved in 25% v/v glycerin at −80 °C, and inoculated and cultivated in Luria broth (LB) media in an incubator for 16–18 h at 37 °C and 150 rpm. Prior to inoculation, glycerin for cell preservation was washed out via centrifugation (5000 rpm for 5 min) and suspension (in PBS) the inoculum thrice.
∼10 mm/s). An LED panel (42 W, Hueler, Guangdong, China), with light wavelengths ranging from 420 to 700 nm, was used to provide an incident light intensity of 5.38 × 10−5 einstein/(L·s) for visible illuminance; the illumination wavelength distribution was provided in Fig. S2. All of the experiments were performed at room temperature (∼20 °C). SEM imaging pretreat procedures, DAD tests procedure, the schematic representation of reactor, the illumination wavelength distribution of LED panel and SEM images of the Ag/TiO2 coated carrier, and EDX spectra and EDX-mappings for the elements in the photocatalyst film are all available in the Supporting Information. Supporting Information associated with this article can be found, in the online version, at https://doi.org/10.1016/j.cej.2018.06.167. When the reactor was illuminated, the exterior of the carriers was exposed to light and generated RS, which forced biofilms at the carrier surface to detach. Thereafter, the reactive species reacted with organic molecules in the bulk liquid more efficiently. When biodegradable products were generated, they immediately diffused into the interior of the microporous carriers, where biofilm bacteria were present and biodegraded all biodegradable products. Being inside the microporous carriers, the biofilm was protected from the biocidal effects of photocatalysis. This follows the processing of ICPB invented by the group of Prof. Rittmann [2,3,18,23,38]. Five sets of conditions were used to evaluate the removal and mineralization of TCH, including photolysis alone (P), adsorption alone (AD), visible-light-induced photocatalysis alone (PC), biodegradation alone (B), and intimately coupled photocatalysis and biodegradation (ICPB). All these conditions were tested in separate reactors, but with the same configuration as described above. Specifically, P was illuminated with LED only without photocatalysts, biofilm and carriers in the reactor; AD was conducted in the dark with catalyst-coated carriers but without biofilm; B was conducted in the dark with biofilm developed in the catalyst-coated carriers. PC and ICPB were conducted under LED illumination with the catalyst-coated carriers with and without biofilm, respectively. A synthetic wastewater was fed to the reactors, which was the same with the feed for biofilm development but replacing sodium acetate with various concentrations of TCH.
3. Modeling approach 3.1. Basic assumptions To develop kinetics for TCH removal and mineralization in ICPB, the following assumptions were proposed to simplify the models: (1) Biomass (in mg/L) attached to the carriers was constant at the ICPB steady stage with a balance of decaying and developing biofilm. (2) All photocatalysis intermediates (Int.) was mineralized to CO2 after a series of biodegradations. (3) The amount of Int., which accounted for both intermediates attacked by ROS and aerobic bio-oxidization generated products, was represented by CODINT. CODINT was calculated by subtracting CODTCH (the theoretical COD of the remained TCH) from the total bulk COD, CODT. (4) The strength of dominant RS represented the photocatalysis capability of ICPB.
2.3. Dominant RS identification A series of RS were generated during photocatalysis, including electrons (e−), holes (h+), %OH, H2O2, and %O2− [24]. Different scavengers were used to specifically quench each RS, and identify the dominant RS for TCH photocatalytic degradation. Sodium oxalate (0.5 mmol/L), isopropanol (0.5 mmol/L), Cr(VI) (0.05 mmol/L), Fe(II)EDTA (0.1 mmol/L), and TEMPOL (2 mmol/L) were used to scavenge h+, %OH, e−, H2O2, and %O2−, respectively [24].
3.2. Kinetic model of TCH removal TCH is usually non-biodegradable [27], thus it was considered that all TCH in ICPB were removed by photocatalysis. On the basis of assumptions (2) and (3), TCH-intermediates from photocatalysis were further oxidized and finally mineralized by microorganisms. Thus, TCH removal and mineralization were achieved through the following reactions where k1 was the second-order rate constant for degradation of TCH and k2 was the second-order rate constant for degradation of the photocatalysis intermediates
2.4. Photo-generated [h+] measurement The production of [h+] was evaluated in the above-mentioned reactors, which was operated according to the ICPB protocol. KI solution (50 mM; Sigma–Aldrich, USA) was used to stoichiometrically determine [h+], which should be twice the produced iodine concentration, according to the equation of 2I− + 2h+ → I2. The detailed procedure was described by Turolla et al. [25].
hv
(1)
TiO2→RS k1
TCH + RS → Int. 2.5. Analytical methods
k2
The TCH concentration was determined by an ultra-performance liquid chromatography system (UPLC; Waters, USA) equipped with a 2.1 × 50-mm, 1.7-μm BEH C18 column (Waters, USA), which was operated with a mobile phase of 10% acetonitrile and 90% ultra-water at a flow rate of 0.5 mL/min. The absorption wavelength for detection was 357 nm. COD was determined by potassium dichromate oxidation according to standard procedures (APHA, 2001). Prior to TCH and COD
(2)
Int. + RS → products
(3)
Int. + biodegradation → CO2 + H2 O
(4)
Then, the TCH removal kinetics were simulated with a second-order kinetic model [6,16–18], as shown in Eq. (5) PC
⎛− d[TCH] ⎞ dt ⎠ ⎝ 969
= k1 [RS][TCH]
(5)
Chemical Engineering Journal 351 (2018) 967–975
Y. Ma et al.
[RS]t = k 0 [TiO2 ] t+ [RS]0 e−k3 [CODT ]t
where [TCH] and [RS] were the TCH concentration and the dominant RS concentration, respectively. The superscript on the left of the equation represented the protocol that the model applied to. The same was labeled in the following equations. For the kinetic model of the Int. removal in the PC protocol, a similar protocol to Eq. (5) was used in the following; however, [CODINT] was the Int. amount accounting for COD [assumption (3)]
Therefore, COD removal by photocatalysis was expressed as Eq. (15) by adding Eq. (14) to Eq. (13), PC
⎛− d[CODT ] ⎞ dt ⎠ ⎝
= k1 [RS][TCH]−k2 [RS][CODINT]
(6)
−
In fact, Ints. were removed by photocatalysis and biodegradation simultaneously in the ICPB reactor. Hence, we obtained Eq. (7), where Int. removal by biodegradation was expressed by the Monod model [19], in which Rm was the maximum specific cell growth rate, Ks was the half-saturation constant, and X was biomass concentration at the steady state,
⎛ d[CODINT ] ⎞ dt ⎝ ⎠
ICPB
= k1 [RS][TCH]−k2 [RS][CODINT]−
Rm [CODINT ] X Ks + [CODINT ]
d[RS] = k 0 [TiO2]−(k1 [TCH] + k2 [CODINT ])[RS] dt
(9)
n
error(%) =
ln([TCH]t /[TCH]0 )
Took Eqs. (9) and (10) into Eqs. (5) and (7) to generate Eqs. (11) and (12) for the TCH removal kinetics in ICPB
(17)
2
⎛ Cmod−Cexp ⎞ ⎟ Cexp ⎝ ⎠
⎜
(18)
4. Results and discussion
ICPB
4.1. Feasibility of ICPB for TCH removal
= k1 (k 0 [TiO2] t+ [RS]0 e−(k1 [TCH] + A K1 [CODINT])t )[TCH]
All experiments were operated for at least six cycles and ICPB performances stabilized after three cycles, either in the current study or as previously reported [6]. Fig. 1 shows the course of successful ICPB fabrication. During the start-up term (the first three cycles), the biofilm decayed and detached from the outer surface due to photocatalytic damage (Fig. 1a and c). As a result, the Ag/TiO2 nanoparticles that uniformly coated on the carrier skeleton were well exposed. EDXmapping analysis on the photocatalysts-film confirmed the uniform distribution of Ag, Ti, O, and C elements, indicating that the Ag doped TiO2 photocatalysts were well coated on the sponge carrier (see Fig. S3. C represents the carrier). The prepared Ag/TiO2 nanoparticles had a size of ∼80 nm, and numerous Ag particles (3–6 nm in diameter) were uniformly distributed on their surface [18]. Such an intimate contact is advantageous for electron transfer and high stability of the photocatalyst [33,34]. Detailed characteristics of the TiO2 crystals have been provided in a previous report [18,35]. The EISA coating method favored the photocatalysts load on the carrier stability [6,36]. Thus, the prepared nanoparticles did not present an antibacterial effect [6,18]. In contrast, a thick biofilm matrix accumulated in the interior-pores of the carriers (Fig. 1d). During the steady stage, the suspended solids in the effluent did not exceed 20 mg/L, which indicated that the loss of the biofilms was negligible thereafter. Biofilm detachment from the outer surfaces is a prerequisite for successful ICPB fabrication and ensures
(11) and ICPB
= (k1 [TCH]−Ak1 [CODINT ])(k 0 [TiO2] t Rm [CODINT ] X Ks + [CODINT ] (12)
For the PC protocol, the second term on the right of Eq. (12) was absent. 3.3. The modified kinetic model of COD removal Similar to the TCH removal kinetics in the photocatalysis reaction, the COD removal were simulated with a second-order kinetic model [see Eqs. (5)] [30], where k3 was the second-order rate constant for degradation of COD, [CODT] was total bulk COD. PC
= k3 [RS][CODT ]
∑ i=1
(10)
⎛− d[CODT ] ⎞ dt ⎠ ⎝
Rm [CODINT ] X Ks + [CODINT ]
Eq. (11) and Eqs. (15)–(17) were solved by using the 1stOpt (First Optimization) software with Auto2Fit (Version 5.5) to achieve simulation, and a series of experimental results were used to validate the kinetic models of TCH removal and mineralization in the ICPB process. The root square errors between the experimental results and model simulations were analyzed according to a report by Wang et al. [32], where Cmod and Cexp represent the values of model calculated results and experimental results, respectively.
The relationship between k1 and k2 was expressed by the competition kinetic model [28], where A stood for the coefficient ln([CODINT ]t /[CODINT ]0 ) [29],
+ [RS]0 e−(k1 [TCH] + Ak1 [CODINT])t )−
= k3 (k 0 [TiO2 ] t+ [RS]0 e−k3 [CODT ]t )[CODT ]
3.4. Model implementation and estimation of parameters
where k0 was the production rate constant of RS from TiO2. The concentration of RS used for TCH removal at a specific time (t) after irradiation was given by
k2 = Ak1
(16)
ICPB
+
(8)
[RS]t = k 0 [TiO2 ] t+ [RS]0 e−(k1 [TCH] + k2 [CODINT ])t
d[CODT ] Rm [CODT ] = X dt Ks + [CODT ]
⎛− d[CODT ] ⎞ dt ⎝ ⎠
To determine the RS kinetics, the rates of RS formation from TiO2 and RS consumption on attacking TCH and its intermediates were considered together. Thus, the following Eq. (8) was developed to represent the kinetics of RS, in which the radical–radical termination reactions were neglected for simplification,
⎛ d[CODINT ] ⎞ dt ⎝ ⎠
(15)
where Rm was the maximum specific cell growth rate, Ks was the halfsaturation constant, and X was the biomass concentration at the steady state. In the ICPB process, photocatalysis and biodegradation simultaneously contributed to COD removal; hence, the COD removal kinetics were simulated by combination of Eqs. (15) and (16)
(7)
⎛ d[TCH] ⎞ ⎝ dt ⎠
= k3 (k 0 [TiO2 ] t+ [RS]0 e−k3 [CODT ]t )[CODT ]
COD removal by biodegradation was expressed by the Monod model [31] as
PC
⎛ d[CODINT ] ⎞ dt ⎝ ⎠
(14)
(13)
Hence, the concentration of RS used for COD degradation at a specific time (t) was, 970
Chemical Engineering Journal 351 (2018) 967–975
Y. Ma et al.
Fig. 1. SEM images of the sponge carriers before ICPB and at the steady stage of ICPB. (a) and (c) were the images at the outer skeletons, before ICPB and at the steady stage of ICPB, respectively; (b) and (d) were the images at the interior pores, before ICPB and at the steady stage of ICPB, respectively.
biofilm on the adsorption was negligible. In contrast, TCH removal was ∼94% for both PC and ICPB protocols, indicating the initial TCH degradation was mainly dependent on photocatalysis. Notably, the TCH removals by ICPB less than 4 h were greater than those by PC, presumably due to the fact that the photocatalysis products in ICPB were rapidly biodegraded [36]. As a result, these products did not compete as much with TCH for free radicals. Then, the free radicals from photocatalysis could more efficiently attack TCH itself, which contributed to the enhanced TCH removal in ICPB protocol [6]. More importantly, PC presented a weak capability for mineralization, however, mineralization was considerably enhanced by the effects
efficient photocatalysis at the carrier surface [1–3]; biofilms in the interior of the macro-porous carriers avoid direct damage from free radicals, but benefit from the non-toxic consumable TCH products of photocatalysis [4–6]. TCH removal and mineralization results for steady state operation (Cycles 4–6) for all protocols were shown in Fig. 2. TCH was almost not degraded by visible-light photolysis alone, as shown in Fig. 2a. According to previous reports, TCH could not be biodegraded [6,27]. Our results showed that B and AD followed the same trend regarding the TCH removal (∼30%). Therefore, TCH loss in B was basically due to adsorption of the sponge carriers, not biodegradation and the effect of
Fig. 2. TCH (a) and COD (b) removal at the steady state with initial TCH concentration of 20 mg/L. Conditions were: P, photolysis alone; AD, adsorption by the Ag/TiO2-loaded carriers alone; B, biodegradation alone with biofilms attached carrier and operated in darkness; PC, visible-light driven photocatalysis (no biofilm cultivation) alone; and ICPB, visible-light-driven photocatalysis coupled with biodegradation. Error bars represent the standard error for triplicate cycles at the steady stage.
971
Chemical Engineering Journal 351 (2018) 967–975
Y. Ma et al.
of biodegradation (Fig. 2b). The COD removal was only ∼45% for the PC protocol, which was only ∼15% greater than that of AD and B. This situation was similar to many previous works, where high TCH removal rates were accompanied by poor mineralization efficiencies for PC [36,37]. The COD removal increased to ∼70% when the photocatalysis was coupled with biodegradation for an initial TCH sample of 20 mg/L, which was approximately 25% greater than that achieved with the PC protocol. That is, the photocatalysis products are hard to be further photocatalytically mineralized [14], but could be consumed rapidly by the bacteria harbored inside the porous carrier [36,39,40]. Thus, we demonstrated the ICPB strategy for synergistic degradation of TCH, in which TCH was initially degraded by the Ag/TiO2 photocatalysis at the outer surface of the sponge carriers, and the photocatalytically generated TCH products were further biodegraded by biofilms harbored at the interior of the sponge carriers. This is a typical manifestation of ICPB [1,2,23]. To test the toxicity of TCH and its intermediates to microorganisms, the DAD test was carried out for S. aureus inhibition by using products from 20 mg/L TCH degradation at 0, 1, 4 and 8 h in B, PC, and ICPB. The corresponding inhibition distances were measured and are shown
in Fig. 3. At 0 h, the clear zones showed the same size of 2.77 mm for B, PC, and ICPB, indicating the intact TCH in these protocols had the same toxic effects. The clear zones of B decreased from 2.77 to 2.48 mm at 1 h, 2.05 mm at 4 h and 2.04 mm at 8 h, causing a ∼26% decrement which was attributed to TCH adsorption, as verified in Fig. 2a. The clear zones of PC increased from 2.77 mm to 4.36 mm (by 57%) at 1 h, before gradually decreasing to 3.64 mm at 4 h and 3.57 mm at 8 h, indicating an increase of toxicity when TCH was only treated by photocatalysis, this was consistent with a previous report [41]. In contrast, the clear zones of ICPB decreased from 2.77 mm to 1.76 mm at 1 h and to zero at 4 h and 8 h. For ICPB, biofilms harbored in the interior of carriers could adapt to the biodegradation of TCH-photocatalysis products. We previously reported that the biofilms were more enriched in Methylibium, Runella, Comamonas, and Pseudomonas when employing ICPB to treat TCH [6]. These are known to enable aromatic biodegradation and for being resistant to TCH. As a result, the toxicity of TCH and its intermediates in the ICPB effluent was significantly reduced and these residuals were almost non-toxic to S. aureus.
Fig. 3. Inhibition zones on S. aureus corresponding to the indicated time for 20 mg/L TCH with PC, B, and ICPB protocols at 0 h (A), 1 h (B), 4 h (C), and 8 h (D). B, biodegradation alone with biofilms attached carrier and operated in darkness; PC, visible-light driven photocatalysis (no biofilm cultivation) alone; and ICPB, visiblelight driven photocatalysis coupled with biodegradation. The figures represent the radius of inhibition zones subtracting the radius of the filter paper. 972
Chemical Engineering Journal 351 (2018) 967–975
Y. Ma et al.
scavenged, which reflects a ∼70% reduction of TCH removal compared with that of the non-scavenged case. The importance of h+ on TCH removal was confirmed by a small enhancement of 10% with e− scavenger when comparing with the protocol that no scavengers were used. Scavenging e− can avoid recombination of e− and h+, which enables more h+ to attack TCH [24]. Therefore, the h+ concentration in the ICPB-reactor was determined and used to represent the RS concentration [RS] in the model simulation process. 4.2.2. Other parameters [TiO2], the TiO2 concentration deposited on the sponge carrier, which was exposed to light for photocatalysis, was 175 mg/L. [TiO2] was obtained by multiplying the mass of TiO2 at per gram carrier (87.5 mg-TiO2/g-carrier) and the amount of the carriers in the ICPBreactor (2.0 g/L). [CODINT] at a specific time point (t) in modeling the TCH removal was calculated by solving Eqs. (11) and (12) and [CODINT] at a specific time point (t) in modeling COD removal was calculated by solving Eqs. (12) and (17).
Fig. 4. Photocatalytic removal efficiencies of TCH with different scavengers: 0.05 mmol/L Cr(VI) for e−, 0.5 mmol/L isopropanol for %OH; 0.5 mmol/L sodium oxalate for h+; 0.1 mmol/L Fe(II)-EDTA for H2O2; and 2 mmol/L TEMPOL for %O2−. TCH concentration = 20 mg/L. Error bars represent the standard error for triplicate cycles at the steady stage.
4.3. TCH removal kinetics To evaluate TCH removal, the kinetic parameter k1 in Eq. (11) was first optimized with the First Optimization software. As shown in Table 1, the kinetic coefficients of ICPB were basically ∼10% greater than those of PC, and the corresponding coefficients of determination (R2) were all greater than 0.95. This was attributed to the effects of biodegradation in ICPB, which accelerated the mineralization of the photocatalytic intermediates. This biological consumption lowered the competition between TCH and the intermediates for RS [6]. Here, we revealed an enhancement for the removal of targeted difficult-to-degrade pollutants by ICPB using our developed model for the first time. Furthermore, the applicability of the kinetic model [Eqs. (11) and (12)] to TCH removal was reasonably good based on the above obtained coefficient, as shown in Fig. 5. Considering the whole set of operating conditions used in this work, a root square error value (%) lower than 2.1% was obtained. Consistent with the above TCH removal kinetics, the simulated TCH removals (dashed lines) were obviously accelerated compared with those from PC, when the initial TCH concentrations were either 20 or 40 mg/L. Hence, the developed model was suitable for evaluating the removal of target pollutants in ICPB.
Table 1 Model parameters for TCH removal kinetics. 20 mg-TCH/L
k1(h R2
−1
)
40 mg-TCH/L
PC
ICPB
PC
ICPB
0.227 ± 0.002 0.96
0.254 ± 0.004 0.95
0.242 ± 0.004 0.97
0.265 ± 0.003 0.96
4.2. Modeling parameters 4.2.1. [RS] estimation To estimate the RS concentration, [RS], the dominant RS for TCH initial photocatalysis in the ICPB protocol were identified. We performed scavenger experiments to understand the function of each RS on TCH removal, as shown in Fig. 4. To elaborate the difference of the RS on TCH photocatalysis, we used Ag/TiO2 nanoparticles to conduct the scavenger experiments. Almost ∼90% of the TCH was removed within 20 min without any scavenging, indicating the good photocatalytic activity of Ag/TiO2 for TCH removal. However, there was a slight decrease (less than 10%) of TCH removal when %OH and H2O2 were scavenged, indicating that %OH and H2O2 were not the main RS for TCH removal. Rather, h+ was determined to be the most active species for TCH photocatalysis. The TCH removal was as low as ∼20% when h+ was
4.4. COD removal kinetics Model parameters for the COD removal kinetics are shown in Table 2. On the basis of the developed model [Eq. (17)], COD removal was well predicted and fitted with the experimental data (Fig. 6). The coefficients of determination (R2) generally remained above ∼0.92 and the root square error between experimental results and model Fig. 5. Comparison between experimental data (symbol) and model simulations (dashed lines) of TCH removal. TCH concentrations = 20 mg/L (a) and 40 mg/L (b). PC, visible-light driven photocatalysis (no biofilm cultivation) alone; and ICPB, visible-light driven photocatalysis coupled with biodegradation. Error bars represent the standard error for triplicate cycles at the steady stage.
973
Chemical Engineering Journal 351 (2018) 967–975
Y. Ma et al.
Table 2 Model parameters for COD removal kinetics. 20 mg-TCH/L
−1
k3(h ) Rm(h−1) Ks(mg·L−1) R2
40 mg-TCH/L
PC
ICPB
PC
0.132 ± 0.004 / / 0.97
0.178 ± 0.005 31 420 0.92
0.143 ± 0.003 / / 0.97
ICPB
0.204 ± 0.004 34 390 0.94
/ in the table indicates that corresponding parameter does not exist in the system. Fig. 6. COD removal comparison between experimental data (symbol) and simulations of the developed model (dashed lines, using Eq. (17)). TCH concentrations = 20 mg/L (a) and 40 mg/L (b). PC, visible-light driven photocatalysis (no biofilm cultivation) alone; and ICPB, visible-light driven photocatalysis coupled with biodegradation. Error bars represent the standard error for triplicate cycles at the steady stage.
predict the COD removal efficiencies of ICPB, at a root square error of 13.2%, which was 16 times greater than that of the developed model shown in Eq. (17). Moreover, the additive model over-estimated COD removal by ∼15%.
simulations was no larger than 0.9%. Both the model predicted and the experimental results showed that the COD removal by ICPB was enhanced by ∼25%, compared with that by PC. Accordingly, the biodegradation in ICPB improved the photocatalytic COD removal rate constant (k3) by 38% and 43%, when the initial TCH concentrations were 20 and 40 mg/L, respectively (Fig. 6). The maximum specific cell growth rate (Rm) also increased by around 10%. The above two facts verified that the intermediates generated from photocatalysis conversely promoted the biodegradation. As a result, the interaction between photocatalysis and biodegradation played a significant role in mineralization of TCH intermediates. Here, we clearly see the advantages of ICPB on the mineralization of refractory pollutants, which is consistent with previously reported experimental results [1–6,23,40]. The effect of an intimately coupled photocatalysis and biodegradation was not a simple addition of them. We predicted the total COD removal by simply adding the COD removals obtained via Eqs. (15) and (16), which resulted from the photocatalysis and biological consumption, respectively. As shown in Fig. 7, this method failed to
5. Conclusions TCH removal and mineralization using a novel ICPB system was investigated, and kinetic models were proposed for evaluating ICPB performances by combining a second-order photocatalytic kinetic model with a Monod-type biodegradation model. Experimental results indicated the successful operation of ICPB with enhanced TCH removal and mineralization as high as ∼94% and ∼70%, respectively, and the products were much less toxic. Such enhancements were attributed to the synergy effect of photocatalysis and biodegradation in ICPB. The developed models can satisfactorily predict the TCH and its mineralization, which agreed well with the experimental data (R2 > 0.9). The models further revealed that ICPB improved pollutants mineralization by the synergistic effects. Here, we successfully developed Fig. 7. COD removal comparison between experimental data (symbol) and simulations of the additive models (dashed lines, Eq. (15) + Eq. (16)). TCH concentrations = 20 mg/L (a) and 40 mg/L (b). Degradation times in Eqs. (15) and (16) were both 8 h. Error bars represent the standard error for triplicate cycles at the steady stage.
974
Chemical Engineering Journal 351 (2018) 967–975
Y. Ma et al.
models to accurately predict ICPB performances, and confirmed its working mechanisms. This is of significance for ICPB application in practical engineering.
[19] G. González, M.G. Herrera, M.T. García, M.M. Peña, Biodegradation of phenol in a continuous process: comparative study of stirred tank and fluidized-bed bioreactors, Bioresour. Technol. 76 (2001) 245–251, http://dx.doi.org/10.1016/S09608524(00)00092-4. [20] M.S. Lee, S.-S. Hong, M. Mohseni, Synthesis of photocatalytic nanosized TiO2–Ag particles with sol–gel method using reduction agent, J. Mol. Catal. A: Chem. 242 (2005) 135–140, http://dx.doi.org/10.1016/j.molcata.2005.07.038. [21] H. Zhou, C. Wang, Z. Feng, S. Li, B. Xu, Formation of grid-like mesoporous titania film via structural transformation and its surface superhydrophilicity conversion, Surf. Coat. Technol. 207 (2012) 34–41, http://dx.doi.org/10.1016/j.surfcoat.2012. 04.067. [22] D. Zhou, M. Liu, J. Wang, S. Dong, N. Cui, L. Gao, Granulation of activated sludge in a continuous flow airlift reactor by strong drag force, Biotechnol. Bioprocess Eng. 18 (2013) 289–299, http://dx.doi.org/10.1007/s12257-012-0513-4. [23] B.E. Rittmann, Biofilms, active substrata, and me, Water Res. 132 (2018) 135–145, http://dx.doi.org/10.1016/j.watres.2017.12.043. [24] D. Xia, T.W. Ng, T. An, G. Li, Y. Li, H.Y. Yip, H. Zhao, A. Lu, P.-K. Wong, A recyclable mineral catalyst for visible-light-driven photocatalytic inactivation of bacteria: natural magnetic sphalerite, Environ. Sci. Technol. 47 (2013) 11166–11173, http://dx.doi.org/10.1021/es402170b. [25] A. Turolla, A. Piazzoli, J. Farner Budarz, M.R. Wiesner, M. Antonelli, Experimental measurement and modelling of reactive species generation in TiO2 nanoparticle photocatalysis, Chem. Eng. J. 271 (2015) 260–268, http://dx.doi.org/10.1016/j. cej.2015.03.004. [26] S.K. Boda, J. Broda, F. Schiefer, J. Weber-Heynemann, M. Hoss, U. Simon, B. Basu, W. Jahnen-Dechent, Cytotoxicity of ultrasmall gold nanoparticles on planktonic and biofilm encapsulated gram-positive staphylococci, Small 11 (2015) 3183–3193, http://dx.doi.org/10.1002/smll.201403014. [27] S. Kim, P. Eichhorn, J.N. Jensen, A.S. Weber, D.S. Aga, Removal of antibiotics in wastewater: effect of hydraulic and solid retention times on the fate of tetracycline in the activated sludge process, Environ. Sci. Technol. 39 (2005) 5816–5823, http://dx.doi.org/10.1021/es050006u. [28] K. Hanna, S. Chiron, M.A. Oturan, Coupling enhanced water solubilization with cyclodextrin to indirect electrochemical treatment for pentachlorophenol contaminated soil remediation, Water Res. 39 (2005) 2763–2773, http://dx.doi.org/ 10.1016/j.watres.2005.04.057. [29] A.L. Boreen, B.L. Edhlund, J.B. Cotner, K. McNeill, Indirect photodegradation of dissolved free amino acids: the contribution of singlet oxygen and the differential reactivity of DOM from various sources, Environ. Sci. Technol. 42 (2008) 5492–5498, http://dx.doi.org/10.1021/es800185d. [30] A.A. Halim, H.A. Aziz, M.A.M. Johari, K.S. Ariffin, Comparison study of ammonia and COD adsorption on zeolite, activated carbon and composite materials in landfill leachate treatment, Desalination 262 (2010) 31–35, http://dx.doi.org/10.1016/j. desal.2010.05.036. [31] G. Nakhla, V. Liu, A. Bassi, Kinetic modeling of aerobic biodegradation of high oil and grease rendering wastewater, Bioresour. Technol. 97 (2006) 131–139, http:// dx.doi.org/10.1016/j.biortech.2005.02.003. [32] D. Wang, Y. Li, G. Li, C. Wang, W. Zhang, Q. Wang, Modeling of quantitative effects of water components on the photocatalytic degradation of 17α-ethynylestradiol in a modified flat plate serpentine reactor, 2013. doi: 10.1016/j.jhazmat.2013.03.049. [33] J. Di, X. Jiexiang, M. Ji, B. Wang, S. Yin, Y. Huang, Z. Chen, H. Li, New insight of Ag quantum dots with the improved molecular oxygen activation ability for photocatalytic applications, Appl. Catal. B 188 (2016) 376–387, http://dx.doi.org/10. 1016/j.apcatb.2016.01.062. [34] Y. Duan, M. Zhang, L. Wang, F. Wang, L. Yang, X. Li, C. Wang, Plasmonic AgTiO2−x nanocomposites for the photocatalytic removal of NO under visible light with high selectivity: the role of oxygen vacancies, Appl. Catal. B 204 (2017) 67–77, http://dx.doi.org/10.1016/j.apcatb.2016.11.023. [35] C. Zhang, L. Fu, Z. Xu, H. Xiong, D. Zhou, M. Huo, Contrasting roles of phenol and pyrocatechol on the degradation of 4-chlorophenol in a photocatalytic–biological reactor, Environ. Sci. Pollut. Res. 24 (2017) 24725–24731, http://dx.doi.org/10. 1007/s11356-017-0245-2. [36] S. Dong, S. Dong, X. Tian, Z. Xu, D. Ma, B. Cui, N. Ren, B.E. Rittmann, Role of selfassembly coated Er(3+): YAlO3/TiO2 in intimate coupling of visible-light-responsive photocatalysis and biodegradation reactions, J. Hazard. Mater. 302 (2016) 386–394, http://dx.doi.org/10.1016/j.jhazmat.2015.10.007. [37] Q. Chen, S. Wu, Y. Xin, Synthesis of Au–CuS–TiO2 nanobelts photocatalyst for efficient photocatalytic degradation of antibiotic oxytetracycline, Chem. Eng. J. 302 (2016) 377–387, http://dx.doi.org/10.1016/j.cej.2016.05.076. [38] E. Adamek, W. Baran, A. Sobczak, Photocatalytic degradation of veterinary antibiotics: biodegradability and antimicrobial activity of intermediates, Process Saf. Environ. Prot. 103 (2016) 1–9, http://dx.doi.org/10.1016/j.psep.2016.06.015. [39] M.D. Marsolek, M.J. Kirisits, K.A. Gray, B.E. Rittmann, Coupled photocatalyticbiodegradation of 2,4,5-trichlorophenol: effects of photolytic and photocatalytic effluent composition on bioreactor process performance, community diversity, and resistance and resilience to perturbation, Water Res. 50 (2014) 59–69, http://dx. doi.org/10.1016/j.watres.2013.11.043. [40] Y. Zhang, L. Chang, N. Yan, Y. Tang, R. Liu, B.E. Rittmann, UV photolysis for accelerating pyridine biodegradation, Environ. Sci. Technol. 48 (2014) 649–655, http://dx.doi.org/10.1021/es404399t. [41] S. Yahiat, F. Fourcade, S. Brosillon, A. Amrane, Removal of antibiotics by an integrated process coupling photocatalysis and biological treatment – Case of tetracycline and tylosin, Int. Biodeterior. Biodegrad. 65 (2011) 997–1003, http://dx.doi. org/10.1016/j.ibiod.2011.07.009.
Acknowledgement The authors thank the National Natural Science Foundation of China (51722803, 51678270) and Open Project of State Key Laboratory of Urban Water Resource and Environment, Harbin Institute of Technology (HC201712) for their financial support. References [1] M.D. Marsolek, C.I. Torres, M. Hausner, B.E. Rittmann, Intimate coupling of photocatalysis and biodegradation in a photocatalytic circulating-bed biofilm reactor, Biotechnol. Bioeng. 101 (2008) 83–92, http://dx.doi.org/10.1002/bit.21889. [2] G. Li, S. Park, D.-W. Kang, R. Krajmalnik-Brown, B.E. Rittmann, 2,4,5Trichlorophenol degradation using a novel TiO2-coated biofilm carrier: roles of adsorption, Photocatal. Biodegrad. Environ. Sci. Technol. 45 (2011) 8359–8367, http://dx.doi.org/10.1021/es2016523. [3] D. Zhou, Z. Xu, S. Dong, M. Huo, S. Dong, X. Tian, B. Cui, H. Xiong, T. Li, D. Ma, Intimate coupling of photocatalysis and biodegradation for degrading phenol using different light types: visible light vs UV light, Environ. Sci. Technol. 49 (2015) 7776–7783, http://dx.doi.org/10.1021/acs.est.5b00989. [4] D. Wen, G. Li, R. Xing, S. Park, B.E. Rittmann, 2,4-DNT removal in intimately coupled photobiocatalysis: the roles of adsorption, photolysis, photocatalysis, and biotransformation, Appl. Microbiol. Biotechnol. 95 (2012) 263–272, http://dx.doi. org/10.1007/s00253-011-3692-6. [5] G. Li, S. Park, B.E. Rittmann, Degradation of reactive dyes in a photocatalytic circulating-bed biofilm reactor, Biotechnol. Bioeng. 109 (2012) 884–893, http://dx. doi.org/10.1002/bit.24366. [6] H. Xiong, D. Zou, D. Zhou, S. Dong, J. Wang, B.E. Rittmann, Enhancing degradation and mineralization of tetracycline using intimately coupled photocatalysis and biodegradation (ICPB), Chem. Eng. J. 316 (2017) 7–14, http://dx.doi.org/10.1016/ j.cej.2017.01.083. [7] X. Chen, B.-J. Ni, Anaerobic conversion of hydrogen and carbon dioxide to fatty acids production in a membrane biofilm reactor: a modeling approach, Chem. Eng. J. 306 (2016) 1092–1098, http://dx.doi.org/10.1016/j.cej.2016.08.049. [8] S. Dominguez, P. Ribao, M.J. Rivero, I. Ortiz, Influence of radiation and TiO2 concentration on the hydroxyl radicals generation in a photocatalytic LED reactor. Application to dodecylbenzenesulfonate degradation, Appl. Catal. B: Environ. 178 (2015) 165–169, http://dx.doi.org/10.1016/j.apcatb.2014.09.072. [9] Y. Yang, J.J. Pignatello, J. Ma, W.A. Mitch, Comparison of halide impacts on the efficiency of contaminant degradation by sulfate and hydroxyl radical-based advanced oxidation processes (AOPs), Environ. Sci. Technol. 48 (2014) 2344–2351, http://dx.doi.org/10.1021/es404118q. [10] Z. Frontistis, V.M. Daskalaki, A. Katsaounis, I. Poulios, D. Mantzavinos, Electrochemical enhancement of solar photocatalysis: degradation of endocrine disruptor bisphenol-A on Ti/TiO2 films, Water Res. 45 (2011) 2996–3004, http:// dx.doi.org/10.1016/j.watres.2011.03.030. [11] I. Di Somma, L. Clarizia, S. Satyro, D. Spasiano, R. Marotta, R. Andreozzi, A kinetic study of the simultaneous removal of EDDS and cupric ions from acidic aqueous solutions by TiO2-based photocatalysis under artificial solar light irradiation and deaerated batch conditions, Chem. Eng. J. 270 (2015) 519–527, http://dx.doi.org/ 10.1016/j.cej.2015.01.056. [12] E. Fernandez-Fontaina, M. Carballa, F. Omil, J.M. Lema, Modelling cometabolic biotransformation of organic micropollutants in nitrifying reactors, Water Res. 65 (2014) 371–383, http://dx.doi.org/10.1016/j.watres.2014.07.048. [13] Y. Tang, A. Ontiveros-Valencia, L. Feng, C. Zhou, R. Krajmalnik-Brown, B.E. Rittmann, A biofilm model to understand the onset of sulfate reduction in denitrifying membrane biofilm reactors, Biotechnol. Bioeng. 110 (2013) 763–772, http://dx.doi.org/10.1002/bit.24755. [14] X. Zhou, Y. Zheng, J. Zhou, S. Zhou, Degradation kinetics of photoelectrocatalysis on landfill leachate using codoped TiO2 /Ti photoelectrodes, J. Nanomater. 2015 (2015) 7, http://dx.doi.org/10.1155/2015/810579. [15] N. Yan, L. Chang, L. Gan, Y. Zhang, R. Liu, B.E. Rittmann, UV photolysis for accelerated quinoline biodegradation and mineralization, Appl. Microbiol. Biotechnol. 97 (2013) 10555–10561, http://dx.doi.org/10.1007/s00253-0134804-2. [16] J. Brame, M. Long, Q. Li, P. Alvarez, Inhibitory effect of natural organic matter or other background constituents on photocatalytic advanced oxidation processes: mechanistic model development and validation, Water Res. 84 (2015) 362–371, http://dx.doi.org/10.1016/j.watres.2015.07.044. [17] M.J. Farré, X. Doménech, J. Peral, Combined photo-Fenton and biological treatment for Diuron and Linuron removal from water containing humic acid, J. Hazard. Mater. 147 (2007) 167–174, http://dx.doi.org/10.1016/j.jhazmat.2006.12.063. [18] H. Xiong, S. Dong, J. Zhang, D. Zhou, B.E. Rittmann, Roles of an easily biodegradable co-substrate in enhancing tetracycline treatment in an intimately coupled photocatalytic-biological reactor, Water Res. 136 (2018) 75–83, http://dx.doi.org/ 10.1016/j.watres.2018.02.061.
975
Contents lists available at ScienceDirect
Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej
Model-based evaluation of tetracycline hydrochloride removal and mineralization in an intimately coupled photocatalysis and biodegradation reactor ⁎
Yue Maa, Houfeng Xiongb, Zhiquan Zhaoc, Yang Yuc,d, Dandan Zhouc,d, , Shuangshi Donga,
T
⁎
a
Key Laboratory of Groundwater Resources and Environment, Ministry of Education, Jilin University, Changchun 130021, China School of Chemistry and Environmental Engineering, Jiujiang University, Jiujiang 332005, China c State Key Laboratory of Urban Water Resource and Environment, Harbin Institute of Technology, Harbin 150090, China d Jilin Engineering Lab for Water Pollution Control and Resources Recovery, Northeast Normal University, Changchun 130117, China b
H I GH L IG H T S
G R A P H I C A L A B S T R A C T
models of intimately coupled • Kinetic photo-biological reaction (ICPB) were developed.
(TCH) removal kinetics • Tetracycline by ICPB were successfully validated. root square errors between ex• The perimental results and model simulations were at max 2.1%.
in ICPB played a cru• Biodegradation cial role in modeling TCH removal and mineralization.
A R T I C LE I N FO
A B S T R A C T
Keywords: Intimately coupled photocatalysis and biodegradation (ICPB) Modeling Refractory pollutants Kinetics Mechanisms
Intimately coupled photocatalysis and biodegradation (ICPB) shows great potential for treatment of refractory pollutants; however, no kinetics for modeling ICPB performance has been developed and the major challenge is to determine the relationship between photocatalysis and biodegradation. In this work, we developed a simplified kinetic model to predict removal and mineralization of a target pollutant (tetracycline hydrochloride; TCH) by hypothesizing that all of the biodegradable photocatalysis products are immediately bio-utilized. Combined with a second-order photocatalytic kinetic model and Monod-type biodegradation model, we observed the interactions between photocatalysis and biodegradation in ICPB. Parameters in the kinetic equations were estimated using the First Optimization software to fit the experimental data to the proposed model with nonlinear regression. Our experimental results showed that TCH and chemical oxygen demand (COD) removal were as high as 94% and 70% within 8 h, respectively. TCH was transformed to non-toxic intermediates in only 4 h. Significantly, the kinetic models could satisfactorily predict the TCH and COD removal, and agreed well with the experimental data with an R2 > 0.92. The models confirmed that biodegradation in ICPB played a major role in accelerating TCH and its intermediates removal and mineralization, as the kinetic coefficient k1 of ICPB was 10% greater than that of photocatalysis alone. The developed models accurately predicted the ICPB efficiencies, and revealed the mechanisms of ICPB operation.
⁎
Corresponding authors at: School of Environment, Northeast Normal University, Changchun 130117, China (D. Zhou). E-mail addresses: [email protected] (D. Zhou), [email protected] (S. Dong).
https://doi.org/10.1016/j.cej.2018.06.167 Received 11 February 2018; Received in revised form 16 May 2018; Accepted 25 June 2018 Available online 28 June 2018 1385-8947/ © 2018 Elsevier B.V. All rights reserved.
Chemical Engineering Journal 351 (2018) 967–975
Y. Ma et al.
Nomenclature
Rm Ks X RS [TiO2] Cmod Cexp P AD B PC ICPB
[TCH] TCH concentration (mg L−1) [CODT] total bulk chemical oxygen demand (mg L−1) [CODTCH] theoretical COD of the remaining TCH (mg L−1) [CODINT] amount of intermediates (=[CODT] – [CODTCH], mg L−1) k0 production rate constant of RS (h−1) k1 second-order rate constant for degradation of TCH (L h−1 mol−1) k2 second-order rate constant for degradation of Int. (L h−1 mol−1) k3 second-order rate constant for degradation of COD (L h−1 mol−1)
1. Introduction
maximum specific cell growth rate (h−1) TCH phenol half-saturation constant (mg L−1) biomass concentration at stable state (mg L−1) reactive species (mol L−1) TiO2 concentration (mg L−1) value of model calculated results (mg L−1) value of experimental results (mg L−1) protocol of photolysis alone protocol of adsorption alone protocol of biodegradation alone visible-light-induced photocatalysis alone intimately coupled photocatalysis and biodegradation
TCH was employed as model pollutant and the experimental data for TCH removal and mineralization were directly compared with the simulated results. This study is important for predicting removal and mineralization of refractory pollutants by ICPB technology, and achieving a direct scale-up and operation of ICPB reactors.
Intimately coupled photocatalysis and biodegradation (ICPB) technology shows great potential for enhancing the removal and mineralization of refractory pollutants [1–3]. In a typical ICPB system, millimeter-sized macro-porous carriers support a photocatalyst on their outer surface and a biofilm within their pores [3]. The concept is that bacteria in the carrier pores are protected from photocatalytic oxidation. This arrangement enables bacteria away from free radical damage but is close enough to rapid degradation of the biodegradable photocatalytic products. ICPB in treatment of bio-refractory pollutants overcomes the weaknesses of incomplete mineralization by photocatalysis alone or inhibition of toxic compounds by bioreactor alone [3–5]. Currently, ICPB has been successfully applied for the treatment of many refractory pollutants, such as phenols [1–3], dinitrotoluene [4] and dyes [5]. Recently, Xiong et al. [6] has used a visible-light-driven ICPB reactor to enhance TCH degradation and mineralization, with which the TCH and COD removal improved by 11% and 20% respectively by ICPB, compared to the photocatalysis alone. This was the first report of treatment of an antibiotic by ICPB. Notably, it was found that biofilm benefitted TCH degradation by reducing interaction of reactive species (RS) and TCH intermediates. Mathematical modeling serves as a both useful and reliable tool to probe into emerging technologies, while significantly reducing the workload related to process evaluation or optimization [7]. In particular, a kinetic model can be used to: (1) identify the key factors of degradation [8–10], (2) provide improved understanding of the complicated reaction mechanisms [11–13], (3) help to scale up relevant processes, and (4) predict the treatment duration and therefore running cost [14]. However, only the biodegrading kinetics of an ICPB reactor has previously been quantified, using an Aiba self-inhibition model [15]. This model lacks the structure of the real ICPB manifestation, i.e., mutually promotion between the photocatalysis and biodegradation [3–5]. The main challenge to determining an ICPB model is to integrate both photocatalytic and biological degradation, where the immediate consumption of photocatalytic intermediates by biodegradation has to be considered. To the best our knowledge, no mathematical model with the abovementioned real ICPB manifestation structure has been developed to date, consequently, the modeling evaluation of pollutants degradation in an ICPB remained unknown. In this work, we approached the mathematic model of ICPB for the first time, employing a strategy of hypothesizing that all the biodegradable photocatalysis products were immediately bio-utilized. This in fact constitutes an ideal ICPB reaction [6]. Modeling photocatalysis alone was achieved based on a secondorder kinetic model [6,16–18] and the biodegradation alone was simulated using the well-known Monod model [19]. Then, we developed the ICPB kinetic model by fully considering the transfer of the predicated photocatalytic products into the biodegradation models, under determination of the final removal and mineralization of pollutants.
2. Material and methods 2.1. Carriers, photocatalysts and biofilms We used a sol–gel method to prepare Ag/TiO2 nanoparticles, which are visible-light responsive photocatalysts. Briefly, Ag/TiO2 was prepared with titanium tetraisopropoxide [Ti(OCH(CH3)2)4] (SigmaAldrich Co. LLC., USA) and silver nitrate (Sigma-Aldrich Co. LLC., USA). Silver reduction and doping were conducted in sodium citrate tribasic dihydrate (Beijing Chemical Works, China) according to a previous report [20].This Ag/TiO2 sol was used for the following coating procedure. Polyurethane sponge carriers (2 mm × 2 mm × 2 mm) with a porosity of 87% and wet density of 0.89 g/mL (Hayi-diverse, Yixing, China) were coated with Ag/TiO2 by using the method of evaporation-induced self-assembly (EISA) [21]. Basically, Ag/TiO2 sol (50 mL) was ultrasonically dispersed in ethanol (150 mL) for 5 min and then the carriers (4 g) were added to the suspension. The carriers in the Ag/TiO2 sol mixture were further ultrasonically dispersed for another 5–10 min before drying in an oven at 80 °C by stirring with a glass rod every 30 min. After rinsing with deionized water in ultrasonic three times to remove the loosely attached Ag/TiO2 nanoparticles, porous carriers coated with photocatalysts were ready to use. The Ag/TiO2 coating amount was 0.39 g-catalyst/g-carrier, with a coating ratio of 28.3% (wt %). The coated carriers were immersed in an activated sludge (South WWTP, Changchun, China) for the initial attachment of microbes and the biofilm was developed at room temperature (∼20 °C) with a synthetic wastewater containing sodium acetate (330 mg/L), NH4Cl (29 mg/L), Na2HPO4·2H2O (8 mg/L), and NaH2PO4 (4 mg/L) in an internal loop airlift-driven reactor, which was described in our previous report [22]. When COD of the effluent was stabilized, the biofilm was considered to be mature. 2.2. Reactors and protocols Photolytic circulating-bed bioreactors of a working volume of 540 mL were used to simulate TCH degradation and mineralization in ICPB reactors with biofilm developed on the Ag/TiO2 coated sponge carriers, which was described in our previous publications [3], and schematically drawn in Fig. S1. Synthetic wastewater and biofilm carriers were circulated by bubbling air from the bottom of the reactor, supplied by a 35 W aeration pump (SOBO, Weifang, China) and controlled by a flow meter (superficial velocity in the draft tube was 968
Chemical Engineering Journal 351 (2018) 967–975
Y. Ma et al.
measurements, the samples were filtered through a 0.22 μm PVDF microporous membrane (GVHP04700, Millipore, USA) to remove suspended solids and biomass. TCH/COD was not adsorbed to the membrane due to its hydrophobic characteristic. Biomass concentrations were determined gravimetrically from the dry-weight loss of the biofilm attached to and detached from sponge carriers (oven temperature 80 °C). The sample morphology and surface elemental content were examined using a scanning electron microscope (SEM, Quanta 200 ESEM, FEI, USA) equipped with energy dispersive X-ray spectroscopy (EDX, Rigaku Corporation). The pretreatment of the samples for SEM was presented in the Supporting Informatiom (SI-1). Product toxicity was evaluated by the disc agar diffusion test (DAD test) based on the method of Sunil et al. [26] with modifications, which were detailed in SI-2. The Gram-positive strain Staphylococcus aureus (ATCC 6538) was purchased from Guangzhou industrial microbiology testing center and was used as model pathogen for the above testing, due to its sensitivity to antibiotics [26]. The bacterial inoculum was initially preserved in 25% v/v glycerin at −80 °C, and inoculated and cultivated in Luria broth (LB) media in an incubator for 16–18 h at 37 °C and 150 rpm. Prior to inoculation, glycerin for cell preservation was washed out via centrifugation (5000 rpm for 5 min) and suspension (in PBS) the inoculum thrice.
∼10 mm/s). An LED panel (42 W, Hueler, Guangdong, China), with light wavelengths ranging from 420 to 700 nm, was used to provide an incident light intensity of 5.38 × 10−5 einstein/(L·s) for visible illuminance; the illumination wavelength distribution was provided in Fig. S2. All of the experiments were performed at room temperature (∼20 °C). SEM imaging pretreat procedures, DAD tests procedure, the schematic representation of reactor, the illumination wavelength distribution of LED panel and SEM images of the Ag/TiO2 coated carrier, and EDX spectra and EDX-mappings for the elements in the photocatalyst film are all available in the Supporting Information. Supporting Information associated with this article can be found, in the online version, at https://doi.org/10.1016/j.cej.2018.06.167. When the reactor was illuminated, the exterior of the carriers was exposed to light and generated RS, which forced biofilms at the carrier surface to detach. Thereafter, the reactive species reacted with organic molecules in the bulk liquid more efficiently. When biodegradable products were generated, they immediately diffused into the interior of the microporous carriers, where biofilm bacteria were present and biodegraded all biodegradable products. Being inside the microporous carriers, the biofilm was protected from the biocidal effects of photocatalysis. This follows the processing of ICPB invented by the group of Prof. Rittmann [2,3,18,23,38]. Five sets of conditions were used to evaluate the removal and mineralization of TCH, including photolysis alone (P), adsorption alone (AD), visible-light-induced photocatalysis alone (PC), biodegradation alone (B), and intimately coupled photocatalysis and biodegradation (ICPB). All these conditions were tested in separate reactors, but with the same configuration as described above. Specifically, P was illuminated with LED only without photocatalysts, biofilm and carriers in the reactor; AD was conducted in the dark with catalyst-coated carriers but without biofilm; B was conducted in the dark with biofilm developed in the catalyst-coated carriers. PC and ICPB were conducted under LED illumination with the catalyst-coated carriers with and without biofilm, respectively. A synthetic wastewater was fed to the reactors, which was the same with the feed for biofilm development but replacing sodium acetate with various concentrations of TCH.
3. Modeling approach 3.1. Basic assumptions To develop kinetics for TCH removal and mineralization in ICPB, the following assumptions were proposed to simplify the models: (1) Biomass (in mg/L) attached to the carriers was constant at the ICPB steady stage with a balance of decaying and developing biofilm. (2) All photocatalysis intermediates (Int.) was mineralized to CO2 after a series of biodegradations. (3) The amount of Int., which accounted for both intermediates attacked by ROS and aerobic bio-oxidization generated products, was represented by CODINT. CODINT was calculated by subtracting CODTCH (the theoretical COD of the remained TCH) from the total bulk COD, CODT. (4) The strength of dominant RS represented the photocatalysis capability of ICPB.
2.3. Dominant RS identification A series of RS were generated during photocatalysis, including electrons (e−), holes (h+), %OH, H2O2, and %O2− [24]. Different scavengers were used to specifically quench each RS, and identify the dominant RS for TCH photocatalytic degradation. Sodium oxalate (0.5 mmol/L), isopropanol (0.5 mmol/L), Cr(VI) (0.05 mmol/L), Fe(II)EDTA (0.1 mmol/L), and TEMPOL (2 mmol/L) were used to scavenge h+, %OH, e−, H2O2, and %O2−, respectively [24].
3.2. Kinetic model of TCH removal TCH is usually non-biodegradable [27], thus it was considered that all TCH in ICPB were removed by photocatalysis. On the basis of assumptions (2) and (3), TCH-intermediates from photocatalysis were further oxidized and finally mineralized by microorganisms. Thus, TCH removal and mineralization were achieved through the following reactions where k1 was the second-order rate constant for degradation of TCH and k2 was the second-order rate constant for degradation of the photocatalysis intermediates
2.4. Photo-generated [h+] measurement The production of [h+] was evaluated in the above-mentioned reactors, which was operated according to the ICPB protocol. KI solution (50 mM; Sigma–Aldrich, USA) was used to stoichiometrically determine [h+], which should be twice the produced iodine concentration, according to the equation of 2I− + 2h+ → I2. The detailed procedure was described by Turolla et al. [25].
hv
(1)
TiO2→RS k1
TCH + RS → Int. 2.5. Analytical methods
k2
The TCH concentration was determined by an ultra-performance liquid chromatography system (UPLC; Waters, USA) equipped with a 2.1 × 50-mm, 1.7-μm BEH C18 column (Waters, USA), which was operated with a mobile phase of 10% acetonitrile and 90% ultra-water at a flow rate of 0.5 mL/min. The absorption wavelength for detection was 357 nm. COD was determined by potassium dichromate oxidation according to standard procedures (APHA, 2001). Prior to TCH and COD
(2)
Int. + RS → products
(3)
Int. + biodegradation → CO2 + H2 O
(4)
Then, the TCH removal kinetics were simulated with a second-order kinetic model [6,16–18], as shown in Eq. (5) PC
⎛− d[TCH] ⎞ dt ⎠ ⎝ 969
= k1 [RS][TCH]
(5)
Chemical Engineering Journal 351 (2018) 967–975
Y. Ma et al.
[RS]t = k 0 [TiO2 ] t+ [RS]0 e−k3 [CODT ]t
where [TCH] and [RS] were the TCH concentration and the dominant RS concentration, respectively. The superscript on the left of the equation represented the protocol that the model applied to. The same was labeled in the following equations. For the kinetic model of the Int. removal in the PC protocol, a similar protocol to Eq. (5) was used in the following; however, [CODINT] was the Int. amount accounting for COD [assumption (3)]
Therefore, COD removal by photocatalysis was expressed as Eq. (15) by adding Eq. (14) to Eq. (13), PC
⎛− d[CODT ] ⎞ dt ⎠ ⎝
= k1 [RS][TCH]−k2 [RS][CODINT]
(6)
−
In fact, Ints. were removed by photocatalysis and biodegradation simultaneously in the ICPB reactor. Hence, we obtained Eq. (7), where Int. removal by biodegradation was expressed by the Monod model [19], in which Rm was the maximum specific cell growth rate, Ks was the half-saturation constant, and X was biomass concentration at the steady state,
⎛ d[CODINT ] ⎞ dt ⎝ ⎠
ICPB
= k1 [RS][TCH]−k2 [RS][CODINT]−
Rm [CODINT ] X Ks + [CODINT ]
d[RS] = k 0 [TiO2]−(k1 [TCH] + k2 [CODINT ])[RS] dt
(9)
n
error(%) =
ln([TCH]t /[TCH]0 )
Took Eqs. (9) and (10) into Eqs. (5) and (7) to generate Eqs. (11) and (12) for the TCH removal kinetics in ICPB
(17)
2
⎛ Cmod−Cexp ⎞ ⎟ Cexp ⎝ ⎠
⎜
(18)
4. Results and discussion
ICPB
4.1. Feasibility of ICPB for TCH removal
= k1 (k 0 [TiO2] t+ [RS]0 e−(k1 [TCH] + A K1 [CODINT])t )[TCH]
All experiments were operated for at least six cycles and ICPB performances stabilized after three cycles, either in the current study or as previously reported [6]. Fig. 1 shows the course of successful ICPB fabrication. During the start-up term (the first three cycles), the biofilm decayed and detached from the outer surface due to photocatalytic damage (Fig. 1a and c). As a result, the Ag/TiO2 nanoparticles that uniformly coated on the carrier skeleton were well exposed. EDXmapping analysis on the photocatalysts-film confirmed the uniform distribution of Ag, Ti, O, and C elements, indicating that the Ag doped TiO2 photocatalysts were well coated on the sponge carrier (see Fig. S3. C represents the carrier). The prepared Ag/TiO2 nanoparticles had a size of ∼80 nm, and numerous Ag particles (3–6 nm in diameter) were uniformly distributed on their surface [18]. Such an intimate contact is advantageous for electron transfer and high stability of the photocatalyst [33,34]. Detailed characteristics of the TiO2 crystals have been provided in a previous report [18,35]. The EISA coating method favored the photocatalysts load on the carrier stability [6,36]. Thus, the prepared nanoparticles did not present an antibacterial effect [6,18]. In contrast, a thick biofilm matrix accumulated in the interior-pores of the carriers (Fig. 1d). During the steady stage, the suspended solids in the effluent did not exceed 20 mg/L, which indicated that the loss of the biofilms was negligible thereafter. Biofilm detachment from the outer surfaces is a prerequisite for successful ICPB fabrication and ensures
(11) and ICPB
= (k1 [TCH]−Ak1 [CODINT ])(k 0 [TiO2] t Rm [CODINT ] X Ks + [CODINT ] (12)
For the PC protocol, the second term on the right of Eq. (12) was absent. 3.3. The modified kinetic model of COD removal Similar to the TCH removal kinetics in the photocatalysis reaction, the COD removal were simulated with a second-order kinetic model [see Eqs. (5)] [30], where k3 was the second-order rate constant for degradation of COD, [CODT] was total bulk COD. PC
= k3 [RS][CODT ]
∑ i=1
(10)
⎛− d[CODT ] ⎞ dt ⎠ ⎝
Rm [CODINT ] X Ks + [CODINT ]
Eq. (11) and Eqs. (15)–(17) were solved by using the 1stOpt (First Optimization) software with Auto2Fit (Version 5.5) to achieve simulation, and a series of experimental results were used to validate the kinetic models of TCH removal and mineralization in the ICPB process. The root square errors between the experimental results and model simulations were analyzed according to a report by Wang et al. [32], where Cmod and Cexp represent the values of model calculated results and experimental results, respectively.
The relationship between k1 and k2 was expressed by the competition kinetic model [28], where A stood for the coefficient ln([CODINT ]t /[CODINT ]0 ) [29],
+ [RS]0 e−(k1 [TCH] + Ak1 [CODINT])t )−
= k3 (k 0 [TiO2 ] t+ [RS]0 e−k3 [CODT ]t )[CODT ]
3.4. Model implementation and estimation of parameters
where k0 was the production rate constant of RS from TiO2. The concentration of RS used for TCH removal at a specific time (t) after irradiation was given by
k2 = Ak1
(16)
ICPB
+
(8)
[RS]t = k 0 [TiO2 ] t+ [RS]0 e−(k1 [TCH] + k2 [CODINT ])t
d[CODT ] Rm [CODT ] = X dt Ks + [CODT ]
⎛− d[CODT ] ⎞ dt ⎝ ⎠
To determine the RS kinetics, the rates of RS formation from TiO2 and RS consumption on attacking TCH and its intermediates were considered together. Thus, the following Eq. (8) was developed to represent the kinetics of RS, in which the radical–radical termination reactions were neglected for simplification,
⎛ d[CODINT ] ⎞ dt ⎝ ⎠
(15)
where Rm was the maximum specific cell growth rate, Ks was the halfsaturation constant, and X was the biomass concentration at the steady state. In the ICPB process, photocatalysis and biodegradation simultaneously contributed to COD removal; hence, the COD removal kinetics were simulated by combination of Eqs. (15) and (16)
(7)
⎛ d[TCH] ⎞ ⎝ dt ⎠
= k3 (k 0 [TiO2 ] t+ [RS]0 e−k3 [CODT ]t )[CODT ]
COD removal by biodegradation was expressed by the Monod model [31] as
PC
⎛ d[CODINT ] ⎞ dt ⎝ ⎠
(14)
(13)
Hence, the concentration of RS used for COD degradation at a specific time (t) was, 970
Chemical Engineering Journal 351 (2018) 967–975
Y. Ma et al.
Fig. 1. SEM images of the sponge carriers before ICPB and at the steady stage of ICPB. (a) and (c) were the images at the outer skeletons, before ICPB and at the steady stage of ICPB, respectively; (b) and (d) were the images at the interior pores, before ICPB and at the steady stage of ICPB, respectively.
biofilm on the adsorption was negligible. In contrast, TCH removal was ∼94% for both PC and ICPB protocols, indicating the initial TCH degradation was mainly dependent on photocatalysis. Notably, the TCH removals by ICPB less than 4 h were greater than those by PC, presumably due to the fact that the photocatalysis products in ICPB were rapidly biodegraded [36]. As a result, these products did not compete as much with TCH for free radicals. Then, the free radicals from photocatalysis could more efficiently attack TCH itself, which contributed to the enhanced TCH removal in ICPB protocol [6]. More importantly, PC presented a weak capability for mineralization, however, mineralization was considerably enhanced by the effects
efficient photocatalysis at the carrier surface [1–3]; biofilms in the interior of the macro-porous carriers avoid direct damage from free radicals, but benefit from the non-toxic consumable TCH products of photocatalysis [4–6]. TCH removal and mineralization results for steady state operation (Cycles 4–6) for all protocols were shown in Fig. 2. TCH was almost not degraded by visible-light photolysis alone, as shown in Fig. 2a. According to previous reports, TCH could not be biodegraded [6,27]. Our results showed that B and AD followed the same trend regarding the TCH removal (∼30%). Therefore, TCH loss in B was basically due to adsorption of the sponge carriers, not biodegradation and the effect of
Fig. 2. TCH (a) and COD (b) removal at the steady state with initial TCH concentration of 20 mg/L. Conditions were: P, photolysis alone; AD, adsorption by the Ag/TiO2-loaded carriers alone; B, biodegradation alone with biofilms attached carrier and operated in darkness; PC, visible-light driven photocatalysis (no biofilm cultivation) alone; and ICPB, visible-light-driven photocatalysis coupled with biodegradation. Error bars represent the standard error for triplicate cycles at the steady stage.
971
Chemical Engineering Journal 351 (2018) 967–975
Y. Ma et al.
of biodegradation (Fig. 2b). The COD removal was only ∼45% for the PC protocol, which was only ∼15% greater than that of AD and B. This situation was similar to many previous works, where high TCH removal rates were accompanied by poor mineralization efficiencies for PC [36,37]. The COD removal increased to ∼70% when the photocatalysis was coupled with biodegradation for an initial TCH sample of 20 mg/L, which was approximately 25% greater than that achieved with the PC protocol. That is, the photocatalysis products are hard to be further photocatalytically mineralized [14], but could be consumed rapidly by the bacteria harbored inside the porous carrier [36,39,40]. Thus, we demonstrated the ICPB strategy for synergistic degradation of TCH, in which TCH was initially degraded by the Ag/TiO2 photocatalysis at the outer surface of the sponge carriers, and the photocatalytically generated TCH products were further biodegraded by biofilms harbored at the interior of the sponge carriers. This is a typical manifestation of ICPB [1,2,23]. To test the toxicity of TCH and its intermediates to microorganisms, the DAD test was carried out for S. aureus inhibition by using products from 20 mg/L TCH degradation at 0, 1, 4 and 8 h in B, PC, and ICPB. The corresponding inhibition distances were measured and are shown
in Fig. 3. At 0 h, the clear zones showed the same size of 2.77 mm for B, PC, and ICPB, indicating the intact TCH in these protocols had the same toxic effects. The clear zones of B decreased from 2.77 to 2.48 mm at 1 h, 2.05 mm at 4 h and 2.04 mm at 8 h, causing a ∼26% decrement which was attributed to TCH adsorption, as verified in Fig. 2a. The clear zones of PC increased from 2.77 mm to 4.36 mm (by 57%) at 1 h, before gradually decreasing to 3.64 mm at 4 h and 3.57 mm at 8 h, indicating an increase of toxicity when TCH was only treated by photocatalysis, this was consistent with a previous report [41]. In contrast, the clear zones of ICPB decreased from 2.77 mm to 1.76 mm at 1 h and to zero at 4 h and 8 h. For ICPB, biofilms harbored in the interior of carriers could adapt to the biodegradation of TCH-photocatalysis products. We previously reported that the biofilms were more enriched in Methylibium, Runella, Comamonas, and Pseudomonas when employing ICPB to treat TCH [6]. These are known to enable aromatic biodegradation and for being resistant to TCH. As a result, the toxicity of TCH and its intermediates in the ICPB effluent was significantly reduced and these residuals were almost non-toxic to S. aureus.
Fig. 3. Inhibition zones on S. aureus corresponding to the indicated time for 20 mg/L TCH with PC, B, and ICPB protocols at 0 h (A), 1 h (B), 4 h (C), and 8 h (D). B, biodegradation alone with biofilms attached carrier and operated in darkness; PC, visible-light driven photocatalysis (no biofilm cultivation) alone; and ICPB, visiblelight driven photocatalysis coupled with biodegradation. The figures represent the radius of inhibition zones subtracting the radius of the filter paper. 972
Chemical Engineering Journal 351 (2018) 967–975
Y. Ma et al.
scavenged, which reflects a ∼70% reduction of TCH removal compared with that of the non-scavenged case. The importance of h+ on TCH removal was confirmed by a small enhancement of 10% with e− scavenger when comparing with the protocol that no scavengers were used. Scavenging e− can avoid recombination of e− and h+, which enables more h+ to attack TCH [24]. Therefore, the h+ concentration in the ICPB-reactor was determined and used to represent the RS concentration [RS] in the model simulation process. 4.2.2. Other parameters [TiO2], the TiO2 concentration deposited on the sponge carrier, which was exposed to light for photocatalysis, was 175 mg/L. [TiO2] was obtained by multiplying the mass of TiO2 at per gram carrier (87.5 mg-TiO2/g-carrier) and the amount of the carriers in the ICPBreactor (2.0 g/L). [CODINT] at a specific time point (t) in modeling the TCH removal was calculated by solving Eqs. (11) and (12) and [CODINT] at a specific time point (t) in modeling COD removal was calculated by solving Eqs. (12) and (17).
Fig. 4. Photocatalytic removal efficiencies of TCH with different scavengers: 0.05 mmol/L Cr(VI) for e−, 0.5 mmol/L isopropanol for %OH; 0.5 mmol/L sodium oxalate for h+; 0.1 mmol/L Fe(II)-EDTA for H2O2; and 2 mmol/L TEMPOL for %O2−. TCH concentration = 20 mg/L. Error bars represent the standard error for triplicate cycles at the steady stage.
4.3. TCH removal kinetics To evaluate TCH removal, the kinetic parameter k1 in Eq. (11) was first optimized with the First Optimization software. As shown in Table 1, the kinetic coefficients of ICPB were basically ∼10% greater than those of PC, and the corresponding coefficients of determination (R2) were all greater than 0.95. This was attributed to the effects of biodegradation in ICPB, which accelerated the mineralization of the photocatalytic intermediates. This biological consumption lowered the competition between TCH and the intermediates for RS [6]. Here, we revealed an enhancement for the removal of targeted difficult-to-degrade pollutants by ICPB using our developed model for the first time. Furthermore, the applicability of the kinetic model [Eqs. (11) and (12)] to TCH removal was reasonably good based on the above obtained coefficient, as shown in Fig. 5. Considering the whole set of operating conditions used in this work, a root square error value (%) lower than 2.1% was obtained. Consistent with the above TCH removal kinetics, the simulated TCH removals (dashed lines) were obviously accelerated compared with those from PC, when the initial TCH concentrations were either 20 or 40 mg/L. Hence, the developed model was suitable for evaluating the removal of target pollutants in ICPB.
Table 1 Model parameters for TCH removal kinetics. 20 mg-TCH/L
k1(h R2
−1
)
40 mg-TCH/L
PC
ICPB
PC
ICPB
0.227 ± 0.002 0.96
0.254 ± 0.004 0.95
0.242 ± 0.004 0.97
0.265 ± 0.003 0.96
4.2. Modeling parameters 4.2.1. [RS] estimation To estimate the RS concentration, [RS], the dominant RS for TCH initial photocatalysis in the ICPB protocol were identified. We performed scavenger experiments to understand the function of each RS on TCH removal, as shown in Fig. 4. To elaborate the difference of the RS on TCH photocatalysis, we used Ag/TiO2 nanoparticles to conduct the scavenger experiments. Almost ∼90% of the TCH was removed within 20 min without any scavenging, indicating the good photocatalytic activity of Ag/TiO2 for TCH removal. However, there was a slight decrease (less than 10%) of TCH removal when %OH and H2O2 were scavenged, indicating that %OH and H2O2 were not the main RS for TCH removal. Rather, h+ was determined to be the most active species for TCH photocatalysis. The TCH removal was as low as ∼20% when h+ was
4.4. COD removal kinetics Model parameters for the COD removal kinetics are shown in Table 2. On the basis of the developed model [Eq. (17)], COD removal was well predicted and fitted with the experimental data (Fig. 6). The coefficients of determination (R2) generally remained above ∼0.92 and the root square error between experimental results and model Fig. 5. Comparison between experimental data (symbol) and model simulations (dashed lines) of TCH removal. TCH concentrations = 20 mg/L (a) and 40 mg/L (b). PC, visible-light driven photocatalysis (no biofilm cultivation) alone; and ICPB, visible-light driven photocatalysis coupled with biodegradation. Error bars represent the standard error for triplicate cycles at the steady stage.
973
Chemical Engineering Journal 351 (2018) 967–975
Y. Ma et al.
Table 2 Model parameters for COD removal kinetics. 20 mg-TCH/L
−1
k3(h ) Rm(h−1) Ks(mg·L−1) R2
40 mg-TCH/L
PC
ICPB
PC
0.132 ± 0.004 / / 0.97
0.178 ± 0.005 31 420 0.92
0.143 ± 0.003 / / 0.97
ICPB
0.204 ± 0.004 34 390 0.94
/ in the table indicates that corresponding parameter does not exist in the system. Fig. 6. COD removal comparison between experimental data (symbol) and simulations of the developed model (dashed lines, using Eq. (17)). TCH concentrations = 20 mg/L (a) and 40 mg/L (b). PC, visible-light driven photocatalysis (no biofilm cultivation) alone; and ICPB, visible-light driven photocatalysis coupled with biodegradation. Error bars represent the standard error for triplicate cycles at the steady stage.
predict the COD removal efficiencies of ICPB, at a root square error of 13.2%, which was 16 times greater than that of the developed model shown in Eq. (17). Moreover, the additive model over-estimated COD removal by ∼15%.
simulations was no larger than 0.9%. Both the model predicted and the experimental results showed that the COD removal by ICPB was enhanced by ∼25%, compared with that by PC. Accordingly, the biodegradation in ICPB improved the photocatalytic COD removal rate constant (k3) by 38% and 43%, when the initial TCH concentrations were 20 and 40 mg/L, respectively (Fig. 6). The maximum specific cell growth rate (Rm) also increased by around 10%. The above two facts verified that the intermediates generated from photocatalysis conversely promoted the biodegradation. As a result, the interaction between photocatalysis and biodegradation played a significant role in mineralization of TCH intermediates. Here, we clearly see the advantages of ICPB on the mineralization of refractory pollutants, which is consistent with previously reported experimental results [1–6,23,40]. The effect of an intimately coupled photocatalysis and biodegradation was not a simple addition of them. We predicted the total COD removal by simply adding the COD removals obtained via Eqs. (15) and (16), which resulted from the photocatalysis and biological consumption, respectively. As shown in Fig. 7, this method failed to
5. Conclusions TCH removal and mineralization using a novel ICPB system was investigated, and kinetic models were proposed for evaluating ICPB performances by combining a second-order photocatalytic kinetic model with a Monod-type biodegradation model. Experimental results indicated the successful operation of ICPB with enhanced TCH removal and mineralization as high as ∼94% and ∼70%, respectively, and the products were much less toxic. Such enhancements were attributed to the synergy effect of photocatalysis and biodegradation in ICPB. The developed models can satisfactorily predict the TCH and its mineralization, which agreed well with the experimental data (R2 > 0.9). The models further revealed that ICPB improved pollutants mineralization by the synergistic effects. Here, we successfully developed Fig. 7. COD removal comparison between experimental data (symbol) and simulations of the additive models (dashed lines, Eq. (15) + Eq. (16)). TCH concentrations = 20 mg/L (a) and 40 mg/L (b). Degradation times in Eqs. (15) and (16) were both 8 h. Error bars represent the standard error for triplicate cycles at the steady stage.
974
Chemical Engineering Journal 351 (2018) 967–975
Y. Ma et al.
models to accurately predict ICPB performances, and confirmed its working mechanisms. This is of significance for ICPB application in practical engineering.
[19] G. González, M.G. Herrera, M.T. García, M.M. Peña, Biodegradation of phenol in a continuous process: comparative study of stirred tank and fluidized-bed bioreactors, Bioresour. Technol. 76 (2001) 245–251, http://dx.doi.org/10.1016/S09608524(00)00092-4. [20] M.S. Lee, S.-S. Hong, M. Mohseni, Synthesis of photocatalytic nanosized TiO2–Ag particles with sol–gel method using reduction agent, J. Mol. Catal. A: Chem. 242 (2005) 135–140, http://dx.doi.org/10.1016/j.molcata.2005.07.038. [21] H. Zhou, C. Wang, Z. Feng, S. Li, B. Xu, Formation of grid-like mesoporous titania film via structural transformation and its surface superhydrophilicity conversion, Surf. Coat. Technol. 207 (2012) 34–41, http://dx.doi.org/10.1016/j.surfcoat.2012. 04.067. [22] D. Zhou, M. Liu, J. Wang, S. Dong, N. Cui, L. Gao, Granulation of activated sludge in a continuous flow airlift reactor by strong drag force, Biotechnol. Bioprocess Eng. 18 (2013) 289–299, http://dx.doi.org/10.1007/s12257-012-0513-4. [23] B.E. Rittmann, Biofilms, active substrata, and me, Water Res. 132 (2018) 135–145, http://dx.doi.org/10.1016/j.watres.2017.12.043. [24] D. Xia, T.W. Ng, T. An, G. Li, Y. Li, H.Y. Yip, H. Zhao, A. Lu, P.-K. Wong, A recyclable mineral catalyst for visible-light-driven photocatalytic inactivation of bacteria: natural magnetic sphalerite, Environ. Sci. Technol. 47 (2013) 11166–11173, http://dx.doi.org/10.1021/es402170b. [25] A. Turolla, A. Piazzoli, J. Farner Budarz, M.R. Wiesner, M. Antonelli, Experimental measurement and modelling of reactive species generation in TiO2 nanoparticle photocatalysis, Chem. Eng. J. 271 (2015) 260–268, http://dx.doi.org/10.1016/j. cej.2015.03.004. [26] S.K. Boda, J. Broda, F. Schiefer, J. Weber-Heynemann, M. Hoss, U. Simon, B. Basu, W. Jahnen-Dechent, Cytotoxicity of ultrasmall gold nanoparticles on planktonic and biofilm encapsulated gram-positive staphylococci, Small 11 (2015) 3183–3193, http://dx.doi.org/10.1002/smll.201403014. [27] S. Kim, P. Eichhorn, J.N. Jensen, A.S. Weber, D.S. Aga, Removal of antibiotics in wastewater: effect of hydraulic and solid retention times on the fate of tetracycline in the activated sludge process, Environ. Sci. Technol. 39 (2005) 5816–5823, http://dx.doi.org/10.1021/es050006u. [28] K. Hanna, S. Chiron, M.A. Oturan, Coupling enhanced water solubilization with cyclodextrin to indirect electrochemical treatment for pentachlorophenol contaminated soil remediation, Water Res. 39 (2005) 2763–2773, http://dx.doi.org/ 10.1016/j.watres.2005.04.057. [29] A.L. Boreen, B.L. Edhlund, J.B. Cotner, K. McNeill, Indirect photodegradation of dissolved free amino acids: the contribution of singlet oxygen and the differential reactivity of DOM from various sources, Environ. Sci. Technol. 42 (2008) 5492–5498, http://dx.doi.org/10.1021/es800185d. [30] A.A. Halim, H.A. Aziz, M.A.M. Johari, K.S. Ariffin, Comparison study of ammonia and COD adsorption on zeolite, activated carbon and composite materials in landfill leachate treatment, Desalination 262 (2010) 31–35, http://dx.doi.org/10.1016/j. desal.2010.05.036. [31] G. Nakhla, V. Liu, A. Bassi, Kinetic modeling of aerobic biodegradation of high oil and grease rendering wastewater, Bioresour. Technol. 97 (2006) 131–139, http:// dx.doi.org/10.1016/j.biortech.2005.02.003. [32] D. Wang, Y. Li, G. Li, C. Wang, W. Zhang, Q. Wang, Modeling of quantitative effects of water components on the photocatalytic degradation of 17α-ethynylestradiol in a modified flat plate serpentine reactor, 2013. doi: 10.1016/j.jhazmat.2013.03.049. [33] J. Di, X. Jiexiang, M. Ji, B. Wang, S. Yin, Y. Huang, Z. Chen, H. Li, New insight of Ag quantum dots with the improved molecular oxygen activation ability for photocatalytic applications, Appl. Catal. B 188 (2016) 376–387, http://dx.doi.org/10. 1016/j.apcatb.2016.01.062. [34] Y. Duan, M. Zhang, L. Wang, F. Wang, L. Yang, X. Li, C. Wang, Plasmonic AgTiO2−x nanocomposites for the photocatalytic removal of NO under visible light with high selectivity: the role of oxygen vacancies, Appl. Catal. B 204 (2017) 67–77, http://dx.doi.org/10.1016/j.apcatb.2016.11.023. [35] C. Zhang, L. Fu, Z. Xu, H. Xiong, D. Zhou, M. Huo, Contrasting roles of phenol and pyrocatechol on the degradation of 4-chlorophenol in a photocatalytic–biological reactor, Environ. Sci. Pollut. Res. 24 (2017) 24725–24731, http://dx.doi.org/10. 1007/s11356-017-0245-2. [36] S. Dong, S. Dong, X. Tian, Z. Xu, D. Ma, B. Cui, N. Ren, B.E. Rittmann, Role of selfassembly coated Er(3+): YAlO3/TiO2 in intimate coupling of visible-light-responsive photocatalysis and biodegradation reactions, J. Hazard. Mater. 302 (2016) 386–394, http://dx.doi.org/10.1016/j.jhazmat.2015.10.007. [37] Q. Chen, S. Wu, Y. Xin, Synthesis of Au–CuS–TiO2 nanobelts photocatalyst for efficient photocatalytic degradation of antibiotic oxytetracycline, Chem. Eng. J. 302 (2016) 377–387, http://dx.doi.org/10.1016/j.cej.2016.05.076. [38] E. Adamek, W. Baran, A. Sobczak, Photocatalytic degradation of veterinary antibiotics: biodegradability and antimicrobial activity of intermediates, Process Saf. Environ. Prot. 103 (2016) 1–9, http://dx.doi.org/10.1016/j.psep.2016.06.015. [39] M.D. Marsolek, M.J. Kirisits, K.A. Gray, B.E. Rittmann, Coupled photocatalyticbiodegradation of 2,4,5-trichlorophenol: effects of photolytic and photocatalytic effluent composition on bioreactor process performance, community diversity, and resistance and resilience to perturbation, Water Res. 50 (2014) 59–69, http://dx. doi.org/10.1016/j.watres.2013.11.043. [40] Y. Zhang, L. Chang, N. Yan, Y. Tang, R. Liu, B.E. Rittmann, UV photolysis for accelerating pyridine biodegradation, Environ. Sci. Technol. 48 (2014) 649–655, http://dx.doi.org/10.1021/es404399t. [41] S. Yahiat, F. Fourcade, S. Brosillon, A. Amrane, Removal of antibiotics by an integrated process coupling photocatalysis and biological treatment – Case of tetracycline and tylosin, Int. Biodeterior. Biodegrad. 65 (2011) 997–1003, http://dx.doi. org/10.1016/j.ibiod.2011.07.009.
Acknowledgement The authors thank the National Natural Science Foundation of China (51722803, 51678270) and Open Project of State Key Laboratory of Urban Water Resource and Environment, Harbin Institute of Technology (HC201712) for their financial support. References [1] M.D. Marsolek, C.I. Torres, M. Hausner, B.E. Rittmann, Intimate coupling of photocatalysis and biodegradation in a photocatalytic circulating-bed biofilm reactor, Biotechnol. Bioeng. 101 (2008) 83–92, http://dx.doi.org/10.1002/bit.21889. [2] G. Li, S. Park, D.-W. Kang, R. Krajmalnik-Brown, B.E. Rittmann, 2,4,5Trichlorophenol degradation using a novel TiO2-coated biofilm carrier: roles of adsorption, Photocatal. Biodegrad. Environ. Sci. Technol. 45 (2011) 8359–8367, http://dx.doi.org/10.1021/es2016523. [3] D. Zhou, Z. Xu, S. Dong, M. Huo, S. Dong, X. Tian, B. Cui, H. Xiong, T. Li, D. Ma, Intimate coupling of photocatalysis and biodegradation for degrading phenol using different light types: visible light vs UV light, Environ. Sci. Technol. 49 (2015) 7776–7783, http://dx.doi.org/10.1021/acs.est.5b00989. [4] D. Wen, G. Li, R. Xing, S. Park, B.E. Rittmann, 2,4-DNT removal in intimately coupled photobiocatalysis: the roles of adsorption, photolysis, photocatalysis, and biotransformation, Appl. Microbiol. Biotechnol. 95 (2012) 263–272, http://dx.doi. org/10.1007/s00253-011-3692-6. [5] G. Li, S. Park, B.E. Rittmann, Degradation of reactive dyes in a photocatalytic circulating-bed biofilm reactor, Biotechnol. Bioeng. 109 (2012) 884–893, http://dx. doi.org/10.1002/bit.24366. [6] H. Xiong, D. Zou, D. Zhou, S. Dong, J. Wang, B.E. Rittmann, Enhancing degradation and mineralization of tetracycline using intimately coupled photocatalysis and biodegradation (ICPB), Chem. Eng. J. 316 (2017) 7–14, http://dx.doi.org/10.1016/ j.cej.2017.01.083. [7] X. Chen, B.-J. Ni, Anaerobic conversion of hydrogen and carbon dioxide to fatty acids production in a membrane biofilm reactor: a modeling approach, Chem. Eng. J. 306 (2016) 1092–1098, http://dx.doi.org/10.1016/j.cej.2016.08.049. [8] S. Dominguez, P. Ribao, M.J. Rivero, I. Ortiz, Influence of radiation and TiO2 concentration on the hydroxyl radicals generation in a photocatalytic LED reactor. Application to dodecylbenzenesulfonate degradation, Appl. Catal. B: Environ. 178 (2015) 165–169, http://dx.doi.org/10.1016/j.apcatb.2014.09.072. [9] Y. Yang, J.J. Pignatello, J. Ma, W.A. Mitch, Comparison of halide impacts on the efficiency of contaminant degradation by sulfate and hydroxyl radical-based advanced oxidation processes (AOPs), Environ. Sci. Technol. 48 (2014) 2344–2351, http://dx.doi.org/10.1021/es404118q. [10] Z. Frontistis, V.M. Daskalaki, A. Katsaounis, I. Poulios, D. Mantzavinos, Electrochemical enhancement of solar photocatalysis: degradation of endocrine disruptor bisphenol-A on Ti/TiO2 films, Water Res. 45 (2011) 2996–3004, http:// dx.doi.org/10.1016/j.watres.2011.03.030. [11] I. Di Somma, L. Clarizia, S. Satyro, D. Spasiano, R. Marotta, R. Andreozzi, A kinetic study of the simultaneous removal of EDDS and cupric ions from acidic aqueous solutions by TiO2-based photocatalysis under artificial solar light irradiation and deaerated batch conditions, Chem. Eng. J. 270 (2015) 519–527, http://dx.doi.org/ 10.1016/j.cej.2015.01.056. [12] E. Fernandez-Fontaina, M. Carballa, F. Omil, J.M. Lema, Modelling cometabolic biotransformation of organic micropollutants in nitrifying reactors, Water Res. 65 (2014) 371–383, http://dx.doi.org/10.1016/j.watres.2014.07.048. [13] Y. Tang, A. Ontiveros-Valencia, L. Feng, C. Zhou, R. Krajmalnik-Brown, B.E. Rittmann, A biofilm model to understand the onset of sulfate reduction in denitrifying membrane biofilm reactors, Biotechnol. Bioeng. 110 (2013) 763–772, http://dx.doi.org/10.1002/bit.24755. [14] X. Zhou, Y. Zheng, J. Zhou, S. Zhou, Degradation kinetics of photoelectrocatalysis on landfill leachate using codoped TiO2 /Ti photoelectrodes, J. Nanomater. 2015 (2015) 7, http://dx.doi.org/10.1155/2015/810579. [15] N. Yan, L. Chang, L. Gan, Y. Zhang, R. Liu, B.E. Rittmann, UV photolysis for accelerated quinoline biodegradation and mineralization, Appl. Microbiol. Biotechnol. 97 (2013) 10555–10561, http://dx.doi.org/10.1007/s00253-0134804-2. [16] J. Brame, M. Long, Q. Li, P. Alvarez, Inhibitory effect of natural organic matter or other background constituents on photocatalytic advanced oxidation processes: mechanistic model development and validation, Water Res. 84 (2015) 362–371, http://dx.doi.org/10.1016/j.watres.2015.07.044. [17] M.J. Farré, X. Doménech, J. Peral, Combined photo-Fenton and biological treatment for Diuron and Linuron removal from water containing humic acid, J. Hazard. Mater. 147 (2007) 167–174, http://dx.doi.org/10.1016/j.jhazmat.2006.12.063. [18] H. Xiong, S. Dong, J. Zhang, D. Zhou, B.E. Rittmann, Roles of an easily biodegradable co-substrate in enhancing tetracycline treatment in an intimately coupled photocatalytic-biological reactor, Water Res. 136 (2018) 75–83, http://dx.doi.org/ 10.1016/j.watres.2018.02.061.
975