Environ Sci Pollut Res DOI 10.1007/s11356-014-3166-3
RESEARCH ARTICLE
Performance evaluation of a continuous flow photocatalytic reactor for wastewater treatment Mohammad Rezaei & Fariborz rashidi & Sayed Javid Royaee & Morteza Jafarikojour
Received: 22 March 2014 / Accepted: 5 June 2014 # Springer-Verlag Berlin Heidelberg 2014
Abstract A novel photocatalytic reactor for wastewater treatment was designed and constructed. The main part of the reactor was an aluminum tube in which 12 stainless steel circular baffles and four quartz tube were placed inside of the reactor like shell and tube heat exchangers. Four UV–C lamps were housed within the space of the quartz tubes. Surface of the baffles was coated with TiO2. A simple method was employed for TiO2 immobilization, while the characterization of the supported photocatalyst was based on the results obtained through performing some common analytical methods such as X-ray diffraction (XRD), scanning electron microscope (SEM), and BET. Phenol was selected as a model pollutant. A solution of a known initial concentration (20, 60, and 100 ppmv) was introduced to the reactor. The reactor also has a recycle flow to make turbulent flow inside of the reactor. The selected recycle flow rate was 7×10−5 m3.s−1, while the flow rate of feed was 2.53×10−7, 7.56×10−7, and 1.26× 10−6 m3.s−1, respectively. To evaluate performance of the reactor, response surface methodology was employed. A four-factor three-level Box–Behnken design was developed to evaluate the reactor performance for degradation of phenol. Effects of phenol inlet concentration (20–100 ppmv), pH (3– 9), liquid flow rate (2.53×10−7−1.26×10−6 m3.s−1), and TiO2 loading (8.8–17.6 g.m−2) were analyzed with this method. The adjusted R2 value (0.9936) was in close agreement with that of corresponding R2 value (0.9961). The maximum predicted degradation of phenol was 75.50 % at the optimum processing Responsible editor: Bingcai Pan M. Rezaei : F. rashidi (*) : M. Jafarikojour Chemical Engineering Department, Amirkabir University of Technology, Tehran, Iran e-mail: [email protected] S. J. Royaee Petroleum Refining Technology Development Division, Research Institute of Petroleum Industry, Tehran, Iran
conditions (initial phenol concentration of 20 ppmv, pH∼ 6.41, and flow rate of 2.53×10−7 m3.s−1 and catalyst loading of 17.6 g.m−2). Experimental degradation of phenol determined at the optimum conditions was 73.7 %. XRD patterns and SEM images at the optimum conditions revealed that crystal size is approximately 25 nm and TiO2 nanoparticles with visible agglomerates distribute densely and uniformly over the surface of stainless steel substrate. BET specific surface area of immobilized TiO2 was 47.2 and 45.8 m2 g−1 before and after the experiments, respectively. Reduction in TOC content, after steady state condition, showed that maximum phenol decomposition occurred at neutral condition (pH∼6). Keywords Photocatalysis . Novel-designed photoreactor . TiO2 . Photocatalytic degradation . Response surface methodology . Phenol
Introduction The phenolic compounds are quite stable and remain in the environment for longer period. Due to their toxicity and carcinogenic character, they are dangerous to the ecosystem in water bodies and human health (Royaee and Sohrabi 2010). In recent years, advanced oxidation processes (AOPs), in general, and photocatalytic oxidation techniques, in particular, with the help of semiconductors such as TiO2 have attracted considerable attention among treating methods such as conventional physical and chemical treatments, biological methods, etc.(Royaee et al. 2012, Royaee et al. 2011). Although photocatalytic oxidation (PCO) is an attractive technique, certain difficulties related to their efficiency, cost, and scale-up have not yet been solved completely. Several laboratory-scale reactors have been designed and built to test and develop the UV/TiO2 technology for the treatment of
Environ Sci Pollut Res
organic pollutants in water. Photoreactors for water treatment are typically based on slurry and immobilized systems. Requirements of catalyst separation and low-light utilization efficiencies due to the scattering light by the catalyst particles may be regarded as disadvantages of slurry systems (Du et al. 2008, Royaee et al. 2012). Hence, much interest has been devoted to utilize immobilized photoreactors for degradation of organic pollutants in aqueous solutions. The main advantageous of immobilized systems are: (a) easier separation of the supported TiO2 from the test system and (b) reusing immobilized photocatalyst (Souzanchi et al. 2013). Several materials have been tested as supports for TiO2 immobilization, e.g., glass beads (Karches et al. 2002), glass tubes (Lee et al. 2002), glass fiber (Horikoshi et al. 2002), quartz (Martyanov and Klabunde 2004), borosilicate glass (Pelaez et al. 2012), silica (Vohra and Tanaka 2003), activated carbon (Matos et al. 2007), aluminum (Chen et al. 2006), pumice stone (Venkata Subba Rao et al. 2003), perlite granules (Hosseini et al. 2007), and stainless steel (Chen and Dionysiou 2006b). Also a variety of techniques have been tested for TiO2 immobilization, such as spray coating (Yi et al. 2008), dip coating (Choi et al. 2006), chemical vapor deposition (CVD; Jung et al. 2005), sol–gel (Wang et al. 2011), and electrophoretic deposition (Zhang et al. 2010). Annular, corrugated plate photocatalytic reactor, monolith honeycomb reactors, horizontal circulating bed photoreactor, recirculating flow CPC reactors, rotating tube photocatalytic reactor, spinning disc reactor, multiple tube reactor, and tube light reactor are among reactors equipped with immobilized photocatalyst. Computer simulation of these reactors was performed to determine effects of flow rates, diffusion coefficients, reaction rate constants, and inlet species concentrations (Alrousan et al. 2012; Behnajady et al. 2007; Boiarkina et al. 2013; Boyjoo et al. 2014; Damodar and Swaminathan 2008; Du et al. 2008; Kumar and Bansal 2013; Pareek et al. 2003; Passalía et al. 2011; Ray 2009; Wang et al. 2009; Wang et al. 2012; Zhang et al. 2004). The configuration of UV lamps with respect to reaction area individualizes these photoreactors which have different degradation capabilities (Mo et al. 2009). In general, an efficient photoreactor should have a high specific surface area, small pass-through channels, and direct light irradiation to the reaction area (Mo et al. 2009). The efficiency of photocatalytic reactors depends on a large number of parameters such as initial concentration of pollutants (Konstantinou and Albanis 2004), photocatalyst loading (Mohammadi et al. 2014, Mozia et al. 2007), flow rate, pH (Konstantinou and Albanis 2004), light intensity (Zacarías et al. 2010), temperature (Henri J.M. 1999), dissolved oxygen (Kabra et al. 2004), etc.
Therefore, comprehensive experimental design and sensitivity analysis are required to investigate not only the effect of various parameters but also to study their possible interaction with each other (Chen et al. 2012). The response surface methodology (RSM) is one of the feasible and efficient experimental design techniques that can estimate linear, interaction and quadratic effects of the factors. It can also estimate a prediction model for the responses (Jiang et al. 2013; Khataee et al. 2010). In this study, a novel continuous flow photoreactor with immobilized photocatalyst named annular baffle type photoreactor (ABPR) was designed and constructed. The reactor provides turbulent flow by means of 12 baffles which were placed inside the reactor. The reactor also has quite high specific surface area for photocatalytic reactions. Moreover, this unique design provides high illumination of photocatalyst. The novel reactor configuration shows certain features which are vital for the viability of the process in large scale application; these include: treatment large amount of wastewater with only trace amounts of photocatalyst. Using immobilized photocatalyst removes the filtration problems and cost. Narrow path of water provide high specific area contact of photocatalyst and pollutant. Subsequently, phenol was selected as a model pollutant to examine the performance of this novel photoreactor. Box–Behnken design (BBD) was applied for modeling, optimization, and also to study the influence of different parameters such as phenol inlet concentration, pH, flow rate, and TiO2 loading in the photocatalytic oxidation of phenol. The TiO2 films were characterized by X-ray diffraction (XRD), scanning electron microscope (SEM), and BET. Reduction of TOC content of discharging reactor was measured to investigate the effect of pH on destruction of phenol and intermediates more accurately.
Materials and methods Materials and preparation of the immobilized TiO2 Titanium dioxide nanoparticles Evonik P-25 powder (BET specific surface area 50±15m2g−1; a mixture of anatase and rutile; average primary particle size of 21 nm) was chosen as TiO 2 source. Ethanol (C 2 H 5 OH − Merck) as solvent, acetylacetone (CH3(CO)CH2(CO)CH3,Merck) for uniform suspension, and sodium dodecyl sulfate (SDS−CH3(CH2)11OSO3,Merck) for better dispersion were used in the preparation of the Titania slurry. The baffles were prepared for stainless steel sheets by cutting stainless steel (316 L) with a water jet cutter
Environ Sci Pollut Res
(DWJKing Company), and then blades were attached to them by a laser spot welding (power unique DN 63). An aluminum tube was used as the reactor which was cut horizontally by a wire cut instrument because of special shape of the reactor. In the first step, 3 g of SDS was added very slowly (in 1 h) to 100 mL of ethanol under vigorous stirring to make a uniform solution. Thirty grams of TiO2 powder was added slowly into the solution to minimize formation of large agglomerates. Slurry was stirred for 24 h at room temperature. Then 5 mL of acetylacetone was added, and the system was sonicated for 0.5 h in order to obtain more uniform slurry (Hilscher 150 W). At second step, the uncoated stainless steel and aluminum substrates were cleaned using hot diluted sulfuric acid solution, boiling distillated water, acetone, and deionized water in sequential order, and subsequently, were dried under atmospheric conditions (Souzanchi et al. 2013). The slurry was sprayed onto the stainless steel and aluminum substrates with a spray gun (Mini Gravity feed spray gun HS-F75), while the air pressure, slurry flow rate, and distance of the substrate and gun were fixed at 4 bar, 0.24 L.h−1, and 0.1 m, respectively. This technique, which is known as spray painted method, with minor differences, was used before and is applied as a standard method of coating. Schematic diagram of coating process is shown in Fig. 1. The spraying process was carried out three times to obtain three levels of catalyst loading. Subsequently, all substrates were dried at room temperature for 12 h and then placed into a furnace. The furnace temperature was incremented at a ramp rate of 0.25 °C/h until 350 °C and then calcined for 0.5 h. It has been reported that BET surface area decreases when calcination temperature increases (Chen and Dionysiou 2006a). The total mass of TiO2 deposited on the supports was determined through weighing the baffles and aluminum tube before and after immobilization to estimate thickness of TiO2 films with different catalyst loadings (Han and Bai 2011; Souzanchi et al. 2013).
Fig. 1 Schematic diagram of coating process: 1 air cylinder, 2 needle valve, 3 feedstock suspension, 4 peristaltic pump, 5 spray gun, and 6 stainless steel substrate
was cut (Fig. 2). These baffles were then arranged in a way to make zigzag arranged in a way to make zigzag pattern for liquid flow along the reactor length. Figure 2 shows configuration of reactor. This reactor has a large specific surface area coated with TiO2 photocatalyst and illuminated with four UV lamps. Moreover, fluid flow in the reactor is turbulent because of high amount of recycle flow and zigzag pattern provided with baffles, and as a result, it provides high mass transfer coefficient inside the reactor which is permanent parameter in reactor with immobilized catalyst (Ray 2009). To clarify this, Reynolds number in the channels is calculated according to the equation (1).
Photoreactor and experimental set-up The reactor was made of an aluminum tube (157 mm inside diameter and 300 mm in length) with four quartz tubes (23 mm outside diameter and 400 mm in length). Four UV lamps (16 W—Philips with λmax centered around 254 nm, 15 mm in diameter, and 303 mm in length) were located at the axis of the quartz tubes. Twelve stainless steel circular baffles (1 mm thickness) with specific geometry that were coated with TiO 2 nanoparticles were placed inside the reactor. The distance between sheets was 20 mm. In order to make flow route for liquid to pass, the top part of the baffles
ρud ð1Þ μ Where ρ is density of waste, u is velocity (u ¼ QA ), and A is surface area. In the present study, surface area was calculated to be 15cm2 and flow rate is approximately equal to recycle flow rate (7 × 10− 5m3. s− 1). Therefore, velocity of waste water in the channels is obtained as 0.06 m.s− 1. d is hydraulic diameter of the channel d ¼ 4 rH ; rH ¼ Sp , S is surface where flow passes Re ¼
through it and p is circumference that gets wet (d = 2.75cm). Therefore, Reynolds becomes 2,526. However,
Environ Sci Pollut Res Fig. 2 Dimension and arrangement of the photoreactor (not to scale)
considering that the reactor has baffles, the real Reynolds will be somewhat bigger than 2,526. Which suggests transient flow regime. Subsequently, Sherwood number Sh ¼ K:d D , which represents the ratio of convective to diffusive mass transfer may be evaluated. D=1.013×10− 9m2.s− 1 which is molecular diffusion coefficient was calculated elsewhere (Plugatyr and Svishchev 2011) Alternatively, Sherwood number may be calculated according to the equation (2) in pipes if flow regime is not laminar. (Treybal and Treybal Robert 1968): 1
Sh ¼ 0:023:Re0:83 :Sc 3
ð2Þ
Schmidt number for much diluted phenol-water solution is 1,900 (Kothandaraman 2004); therefore, Sherwood becomes 189.9. Finally, K is obtained as 7×10− 6 m.s− 1 which shows a moderate to good mass transfer coefficient. The experimental set-up as shown in Fig. 3 consists of liquid feed delivery system that is saturated with oxygen with an air compressor, the photoreactor and the output port. A centrifugal pump was used to pump the wastewater from a 50L reservoir at prescribed flow rate. Another centrifugal pump was located below the reactor and provided an adjustable
recycling system, feeding from the end of the reactor, and being discharged at the inlet. Flow of recycle was fixed at 7× 10−5 m3.s−1 during the experiments. Analyses The crystalline structure of calcined TiO2 films coated on stainless steel baffles was studied by XRD using Philips P1 140 XRD analyzer with a Cu Kα radiation. The surface structure and particle size of the prepared Ti O 2 f i l m s a m p l e s w e r e o b s e r v e d b y a V E G A \crTESCAN SEM followed by AU-coated through sputtering method. The specific surface area of the TiO2 films was measured by BET (Brunauer–Emmett– Teller) method, using Quantachrome Autosorb-1 BET analyzer. The concentration of phenol was measured by a UV–vis spectrophotometer (Dr 2,800 Hach Co.). It was used to determine the absorbance (A) of the samples (λmax =495 nm with pH=3) using specific reagents and following the instructions provided by the supplier (phenol Test Kit—Merck Spectroquant). Reduction in TOC contents after photocatalytic degradation was also determined through the TOC measurements spectrophotometrically at 605 nm using the TOC cell tests and following the instructions provided by Merck.
Environ Sci Pollut Res Fig. 3 Experimental set-up: 1 air compressor, 2 feed tank, 3 air distributor, 4 inlet valve, 5 centrifuge water pump, 6 centrifuge pump for recycling, 7 valve for controlling recycle flow rate, 8 recycle flow meter, 9 inlet flow meter, 10 photocatalytic reactor, 11 outlet valve, 12 flow meter to controlling outlet flow rate, and 13 outlet container
Experimental design method A series of preliminary conventional experiments were carried out to investigate the effect of each parameter on the degradation efficiency of phenol. Phenol inlet concentration, pH, flow rate, and TiO2 loading were considered as main parameters. Optimization was carried out according to the preliminary experimental results using RSM. The Box–Behnken design (BBD) method was employed to analyze the simultaneous effect of phenol inlet concentration (20–100 ppmv), pH (3–9), flow rate (2.53×10−7−1.26×10−6 m3.s−1) and TiO2 loading (8.8–17.6 g.m−2) on the phenol conversion and also to evaluate the interactions between the studied parameters. The current design, which is a modified central composite experimental design, is more efficient than the three-level full factorial designs (Kayan and Gözmen 2012; Zhang et al. 2011). Optimum conditions and interactive effects of the four parameters were determined using regression and graphical analysis by the Design Expert® Software (Version 7.6.1). Response (R) during the experimental design is phenol conversion.
Results and discussion Characterization of the immobilized TiO2 film XRD is an analytical technique that is used to examine the crystalline structure of materials by determining bond length and angles (i.e., the diffraction pattern of X-ray as electromagnetic radiation depends on the right
order of spacing which exists in crystals where distances between atoms are of the order of few angstroms). Figure 4 shows XRD patterns obtained for the pure TiO2 and immobilized TiO2 films on stainless steel and aluminum after being used in photo catalytic experiments. As it is seen, the main peaks are at two-theta (2θ) angle of 24.9° and 53.3°, which are corresponded to anatase and rutile crystalline phases, respectively. The content of the anatase (A) and rutile (R) phases in TiO2 sample can be determined based on the intensities of the peaks in the XRD pattern as follows (Qiu and Zheng 2007; Souzanchi et al. 2013: Spurr and Myers 1957): IR 0:79 IA XR ¼ IR 0:79 1þ IA
ð3Þ
Where XR is the weight fraction of the rutile phase in the sample, IA and IR are the intensities of the reflection peaks of anatase and rutile phases in the diffractograms, respectively. It is clear that the intensity of the radiation from a given crystalline structure of the element (for example, “A” or “R” phase) is depended on the amount of the structural element present in the crystal sample. In this study, percentage of anatase phase of the pure TiO2 and TiO2 after immobilization on stainless steel and aluminum and being used in the photocatalytic experiments was 80.0 %, 78.3 %, and 76.8 %, respectively. In fact, crystalline composition of anatase-based TiO2 particles was remained almost constant after immobilization on the stainless steel and aluminum
Environ Sci Pollut Res
he main grain size (L) of the TiO2 sample can also be estimated using XRD data and applying Scherrer’s formula as follows (Hosseini et al. 2007, Patterson 1939): L¼
Fig. 4 X-ray diffraction spectra of TiO2 coatings prepared in optimum conditions (loading of 17.6 g.m−2): a stainless steel substrate, b aluminum substrate, and cpure TiO2
support and use in photocatalytic experiments. A higher content of rutile phase of TiO2 catalyst is related to a lower photocatalytic activity of TiO2 (Fernández et al. 1995; Mohammadi et al. 2014; Qiu and Zheng 2007). According to equation 3 and XRD analysis of TiO2 thin films on the stainless steel and aluminum substrate photocatalyst thin film contain from roughly 80 % anatase phase of TiO2 and 20 % rutile phase of TiO2 and because the amount of SDS is little we do not see any sharp peak of SDS. Density of the anatase phase of TiO2 is 3.89, and density of rutile phase is 4.23. Total density of TiO2 film can be obtained according to the following equation. densityt ¼ 0:2 densityrutile þ 0:8 densityanatase
ð4Þ
And it was 3.89. Although, the film density (after coating process) and pure TiO2 are not necessarily the same, this is a usual assumption used in other studies (Han and Bai 2010, 2011). After calculation of density, TiO2 film thickness was roughly estimated by dividing TiO2 loading rate with the density. The film thicknesses, which were obtained by this method, are 2.2, 3.3, and 4.4 μm, respectively. In previous study photocatalytic activity of TiO2 for degradation of malic acid had better performance on quartz support and this phenomenon might be because e_/h + recombination at the TiO2 surface due to the presence of metal ions impurities at high calcination temperature (Fernández et al. 1995). Due to this fact in this study, calcination temperature was selected 350 °C, and no considerable rutilization of the TiO 2 was detected (Fig. 4)
0:89λ βcosθ
ð5Þ
Where λ the X-ray wavelength and β is the full width at the half of the maximum peak. The crystal size for the TiO2 in the present work for pure TiO2 and TiO2 after the immobilization on stainless steel and aluminum and use in the experiments was 23.9, 26.9, and 27.1 nm respectively. These were in agreement with the size reported by others who used stainless steel support and applying sol–gel method for preparing the TiO2 film (Chen and Dionysiou 2006b). In general, coating process has not brought any significant size changes to the P-25 material used in the coating process. SEM analysis (Fig. 5) was employed to examine the surface morphology of the TiO2 films that was prepared in the present work. The structure and morphology of TiO2 coating baffles in optimal condition and before being used in photocatalytic experiments is shown in Fig. 5. From the Fig.5a, it is noted that surface of stainless steel support is completely covered with TiO2 films. It can be seen that TiO2 particles have been distributed densely and uniformly over the surface of stainless steel substrate forming a relatively integrated layer with many micro pores among the TiO2 particles. Figure 5b shows that surfaces of films are full of visible agglomerates which are caused by incorporation of P-25 powders in the film. Moreover, it can be clearly seen that large agglomerates are composed of many grains with average size of approximately 25 nm which is nearly equal to crystal sizes obtained from XRD pattern. The surface of the TiO2 films after photocatalytic degradation of phenol in the ABPR have been changed considerably (Fig. 5c). It is clear from the image (Fig. 5c) that TiO2 nanoparticles aggregated and at the same time some micro pores were formed in the TiO2 films. Developing inter-aggregate creates new areas at the photocatalyst surface and this may affect process positively with respect to absorption of UV light as well as better access of phenol and oxygen molecules to active sites within the TiO2 particles (Souzanchi et al. 2013). Specific surface area is also one of the important factors to analyze morphology of the immobilized TiO2 films as used in the photocatalytic degradation experiments (Chen and Dionysiou 2006a). Study of Chen and Dionysiou ( 2006a) on TiO2 films coated on stainless steel support confirmed the positive correlation between high BET specific surface area
Environ Sci Pollut Res
Fig. 5 SEM images of TiO2 coatings prepared in optimum conditions a before performing photocatalytic experiments SEM MAG=20,000, b before performing photocatalytic experiments SEM MAG=50,000, and
c after performing photocatalytic experiments SEM MAG=50,000 (loading of 17.6 g.m−2)
and photocatalytic activity. In the present study, the BET surface area was 47.2 and 45.8 m2 g−1 for the supported TiO2 films before and after the photocatalytic degradation process inside reactor, respectively. Work of other researchers showed that as the calcination temperature increased from 400 °C to 600 °C, the BET surface area for the TiO2 film on the stainless steel support decreased from 49.6 to less than 2 m2 g−1 (Chen and Dionysiou 2006a, Souzanchi et al. 2013). In this study, BET surface area was high and remained constant after experiments. Results of weighting method used for measuring loading of the photocatalyst inside the reactor showed no considerable change of the photocatalyst amount before and after the experiments. After last experiment for each coating, the first experiment for that coating was repeated to check the consistency of results and results showed no change again. Due to all of the results that were shown above, it is expected that photocatalytic activity and stability of the coated TiO2 in ABPR would be reasonably good. Also photocatalytic activity loss was measured in optimal condition. Phenol concentration in the outlet was measured every 0.5 h and is reported in Table 1. No sharp reduction in photocatalytic degradation was observed after 9 h. Note that system reaches steady state after about 6 h.
design were processed using Design Expert® 7.6.1 software and were fitted to a quadratic model. Finally, the best model equation in terms of coded factors (95 % confidence level (p4.0) indicates adequate precision. The coefficient of determination is defined as the ratio of the explained variation to the total variation and
is used as a measure of degree of fit of the model. The predicted R2 value of 0.9961 for phenol removal is in good agreement with the corresponding adjusted R2 value of 0.9936. Therefore, the quadratic polynomial equation could be used to predict the degradation of phenol in the experimental range (Chen et al. 2012, Cheng et al. 2012, Kayan and Gözmen 2012, Khataee et al. 2010). The predicted values
Table 3 Experimental design matrix and responses based on experimental runs proposed by BBD design Run
C Inlet concentration (ppmv)
P pH
F Flow rate (m3/s)
L Catalyst loading (g/m2)
Degradation % (experimental)
Degradation % (predicted)
1 2 3
20 100 60
6 6 6
7.56×10−7 7.56×10−7 7.56×10−7
17.6 17.6 13.2
45.07 12.43 22.50
44.84 12.88 21.14
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
60 60 60 60 60 100 60 60 60 60 60 60 60 60 100 20 20 60 60
9 6 3 3 9 6 6 3 6 6 3 6 6 6 6 6 3 6 9
7.56×10−7 2.53×10−7 2.53×10−7 1.26×10−6 7.56×10−7 7.56×10−7 7.56×10−7 7.56×10−7 7.56×10−7 7.56×10−7 7.56×10−7 1.26×10−6 7.56×10−7 1.26×10−6 2.53×10−7 1.26×10−6 7.56×10−7 7.56×10−7 2.53×10−7
17.6 17.6 13.2 13.2 8.8 8.8 8.8 17.6 13.2 13.2 8.8 17.6 13.2 8.8 13.2 13.2 13.2 13.2 13.2
15.30 46.85 27.03 3.60 7.92 7.10 28.72 14.00 21.69 20.92 5.40 11.34 19.80 10.26 20.63 20.63 30.88 20.78 31.23
16.49 45.75 29.24 2.36 7.38 8.17 28.12 14.68 21.14 21.14 5.57 10.35 21.14 9.77 20.66 21.35 30.39 21.14 31.05
23 24 25 26 27 28 29
60 100 20 20 20 100 100
9 6 6 6 9 9 3
1.26×10−6 1.26×10−6 2.53×10−7 7.56×10−7 7.56×10−7 7.56×10−7 7.56×10−7
13.2 13.2 13.2 8.8 13.2 13.2 13.2
3.30 10.61 66.56 30.96 32.21 5.97 4.17
4.17 11.74 66.18 31.34 32.2 4.64 2.83
Environ Sci Pollut Res Table 4 Analysis of variance regression model for phenol degradation Source
Sum of squares
Model C pH F L C.F. C.L. F.L. C2 pH2 F2 L2 Residual Lack of fit Pure error
df
Mean square
F value
p value Prob>F
12 1 1 1 1 1 1 1 1 1 1 1 16 12 4
486.59 2,011.39 10.94 2,167.87 252.36 322.2 17.56 72.68 320.42 342.04 129.3 9.27 0.39 0.48 0.14
1,235.21 5,105.93 27.78 5,503.15 640.61 817.91 44.57 184.49 813.4 868.26 328.22 23.53
RESEARCH ARTICLE
Performance evaluation of a continuous flow photocatalytic reactor for wastewater treatment Mohammad Rezaei & Fariborz rashidi & Sayed Javid Royaee & Morteza Jafarikojour
Received: 22 March 2014 / Accepted: 5 June 2014 # Springer-Verlag Berlin Heidelberg 2014
Abstract A novel photocatalytic reactor for wastewater treatment was designed and constructed. The main part of the reactor was an aluminum tube in which 12 stainless steel circular baffles and four quartz tube were placed inside of the reactor like shell and tube heat exchangers. Four UV–C lamps were housed within the space of the quartz tubes. Surface of the baffles was coated with TiO2. A simple method was employed for TiO2 immobilization, while the characterization of the supported photocatalyst was based on the results obtained through performing some common analytical methods such as X-ray diffraction (XRD), scanning electron microscope (SEM), and BET. Phenol was selected as a model pollutant. A solution of a known initial concentration (20, 60, and 100 ppmv) was introduced to the reactor. The reactor also has a recycle flow to make turbulent flow inside of the reactor. The selected recycle flow rate was 7×10−5 m3.s−1, while the flow rate of feed was 2.53×10−7, 7.56×10−7, and 1.26× 10−6 m3.s−1, respectively. To evaluate performance of the reactor, response surface methodology was employed. A four-factor three-level Box–Behnken design was developed to evaluate the reactor performance for degradation of phenol. Effects of phenol inlet concentration (20–100 ppmv), pH (3– 9), liquid flow rate (2.53×10−7−1.26×10−6 m3.s−1), and TiO2 loading (8.8–17.6 g.m−2) were analyzed with this method. The adjusted R2 value (0.9936) was in close agreement with that of corresponding R2 value (0.9961). The maximum predicted degradation of phenol was 75.50 % at the optimum processing Responsible editor: Bingcai Pan M. Rezaei : F. rashidi (*) : M. Jafarikojour Chemical Engineering Department, Amirkabir University of Technology, Tehran, Iran e-mail: [email protected] S. J. Royaee Petroleum Refining Technology Development Division, Research Institute of Petroleum Industry, Tehran, Iran
conditions (initial phenol concentration of 20 ppmv, pH∼ 6.41, and flow rate of 2.53×10−7 m3.s−1 and catalyst loading of 17.6 g.m−2). Experimental degradation of phenol determined at the optimum conditions was 73.7 %. XRD patterns and SEM images at the optimum conditions revealed that crystal size is approximately 25 nm and TiO2 nanoparticles with visible agglomerates distribute densely and uniformly over the surface of stainless steel substrate. BET specific surface area of immobilized TiO2 was 47.2 and 45.8 m2 g−1 before and after the experiments, respectively. Reduction in TOC content, after steady state condition, showed that maximum phenol decomposition occurred at neutral condition (pH∼6). Keywords Photocatalysis . Novel-designed photoreactor . TiO2 . Photocatalytic degradation . Response surface methodology . Phenol
Introduction The phenolic compounds are quite stable and remain in the environment for longer period. Due to their toxicity and carcinogenic character, they are dangerous to the ecosystem in water bodies and human health (Royaee and Sohrabi 2010). In recent years, advanced oxidation processes (AOPs), in general, and photocatalytic oxidation techniques, in particular, with the help of semiconductors such as TiO2 have attracted considerable attention among treating methods such as conventional physical and chemical treatments, biological methods, etc.(Royaee et al. 2012, Royaee et al. 2011). Although photocatalytic oxidation (PCO) is an attractive technique, certain difficulties related to their efficiency, cost, and scale-up have not yet been solved completely. Several laboratory-scale reactors have been designed and built to test and develop the UV/TiO2 technology for the treatment of
Environ Sci Pollut Res
organic pollutants in water. Photoreactors for water treatment are typically based on slurry and immobilized systems. Requirements of catalyst separation and low-light utilization efficiencies due to the scattering light by the catalyst particles may be regarded as disadvantages of slurry systems (Du et al. 2008, Royaee et al. 2012). Hence, much interest has been devoted to utilize immobilized photoreactors for degradation of organic pollutants in aqueous solutions. The main advantageous of immobilized systems are: (a) easier separation of the supported TiO2 from the test system and (b) reusing immobilized photocatalyst (Souzanchi et al. 2013). Several materials have been tested as supports for TiO2 immobilization, e.g., glass beads (Karches et al. 2002), glass tubes (Lee et al. 2002), glass fiber (Horikoshi et al. 2002), quartz (Martyanov and Klabunde 2004), borosilicate glass (Pelaez et al. 2012), silica (Vohra and Tanaka 2003), activated carbon (Matos et al. 2007), aluminum (Chen et al. 2006), pumice stone (Venkata Subba Rao et al. 2003), perlite granules (Hosseini et al. 2007), and stainless steel (Chen and Dionysiou 2006b). Also a variety of techniques have been tested for TiO2 immobilization, such as spray coating (Yi et al. 2008), dip coating (Choi et al. 2006), chemical vapor deposition (CVD; Jung et al. 2005), sol–gel (Wang et al. 2011), and electrophoretic deposition (Zhang et al. 2010). Annular, corrugated plate photocatalytic reactor, monolith honeycomb reactors, horizontal circulating bed photoreactor, recirculating flow CPC reactors, rotating tube photocatalytic reactor, spinning disc reactor, multiple tube reactor, and tube light reactor are among reactors equipped with immobilized photocatalyst. Computer simulation of these reactors was performed to determine effects of flow rates, diffusion coefficients, reaction rate constants, and inlet species concentrations (Alrousan et al. 2012; Behnajady et al. 2007; Boiarkina et al. 2013; Boyjoo et al. 2014; Damodar and Swaminathan 2008; Du et al. 2008; Kumar and Bansal 2013; Pareek et al. 2003; Passalía et al. 2011; Ray 2009; Wang et al. 2009; Wang et al. 2012; Zhang et al. 2004). The configuration of UV lamps with respect to reaction area individualizes these photoreactors which have different degradation capabilities (Mo et al. 2009). In general, an efficient photoreactor should have a high specific surface area, small pass-through channels, and direct light irradiation to the reaction area (Mo et al. 2009). The efficiency of photocatalytic reactors depends on a large number of parameters such as initial concentration of pollutants (Konstantinou and Albanis 2004), photocatalyst loading (Mohammadi et al. 2014, Mozia et al. 2007), flow rate, pH (Konstantinou and Albanis 2004), light intensity (Zacarías et al. 2010), temperature (Henri J.M. 1999), dissolved oxygen (Kabra et al. 2004), etc.
Therefore, comprehensive experimental design and sensitivity analysis are required to investigate not only the effect of various parameters but also to study their possible interaction with each other (Chen et al. 2012). The response surface methodology (RSM) is one of the feasible and efficient experimental design techniques that can estimate linear, interaction and quadratic effects of the factors. It can also estimate a prediction model for the responses (Jiang et al. 2013; Khataee et al. 2010). In this study, a novel continuous flow photoreactor with immobilized photocatalyst named annular baffle type photoreactor (ABPR) was designed and constructed. The reactor provides turbulent flow by means of 12 baffles which were placed inside the reactor. The reactor also has quite high specific surface area for photocatalytic reactions. Moreover, this unique design provides high illumination of photocatalyst. The novel reactor configuration shows certain features which are vital for the viability of the process in large scale application; these include: treatment large amount of wastewater with only trace amounts of photocatalyst. Using immobilized photocatalyst removes the filtration problems and cost. Narrow path of water provide high specific area contact of photocatalyst and pollutant. Subsequently, phenol was selected as a model pollutant to examine the performance of this novel photoreactor. Box–Behnken design (BBD) was applied for modeling, optimization, and also to study the influence of different parameters such as phenol inlet concentration, pH, flow rate, and TiO2 loading in the photocatalytic oxidation of phenol. The TiO2 films were characterized by X-ray diffraction (XRD), scanning electron microscope (SEM), and BET. Reduction of TOC content of discharging reactor was measured to investigate the effect of pH on destruction of phenol and intermediates more accurately.
Materials and methods Materials and preparation of the immobilized TiO2 Titanium dioxide nanoparticles Evonik P-25 powder (BET specific surface area 50±15m2g−1; a mixture of anatase and rutile; average primary particle size of 21 nm) was chosen as TiO 2 source. Ethanol (C 2 H 5 OH − Merck) as solvent, acetylacetone (CH3(CO)CH2(CO)CH3,Merck) for uniform suspension, and sodium dodecyl sulfate (SDS−CH3(CH2)11OSO3,Merck) for better dispersion were used in the preparation of the Titania slurry. The baffles were prepared for stainless steel sheets by cutting stainless steel (316 L) with a water jet cutter
Environ Sci Pollut Res
(DWJKing Company), and then blades were attached to them by a laser spot welding (power unique DN 63). An aluminum tube was used as the reactor which was cut horizontally by a wire cut instrument because of special shape of the reactor. In the first step, 3 g of SDS was added very slowly (in 1 h) to 100 mL of ethanol under vigorous stirring to make a uniform solution. Thirty grams of TiO2 powder was added slowly into the solution to minimize formation of large agglomerates. Slurry was stirred for 24 h at room temperature. Then 5 mL of acetylacetone was added, and the system was sonicated for 0.5 h in order to obtain more uniform slurry (Hilscher 150 W). At second step, the uncoated stainless steel and aluminum substrates were cleaned using hot diluted sulfuric acid solution, boiling distillated water, acetone, and deionized water in sequential order, and subsequently, were dried under atmospheric conditions (Souzanchi et al. 2013). The slurry was sprayed onto the stainless steel and aluminum substrates with a spray gun (Mini Gravity feed spray gun HS-F75), while the air pressure, slurry flow rate, and distance of the substrate and gun were fixed at 4 bar, 0.24 L.h−1, and 0.1 m, respectively. This technique, which is known as spray painted method, with minor differences, was used before and is applied as a standard method of coating. Schematic diagram of coating process is shown in Fig. 1. The spraying process was carried out three times to obtain three levels of catalyst loading. Subsequently, all substrates were dried at room temperature for 12 h and then placed into a furnace. The furnace temperature was incremented at a ramp rate of 0.25 °C/h until 350 °C and then calcined for 0.5 h. It has been reported that BET surface area decreases when calcination temperature increases (Chen and Dionysiou 2006a). The total mass of TiO2 deposited on the supports was determined through weighing the baffles and aluminum tube before and after immobilization to estimate thickness of TiO2 films with different catalyst loadings (Han and Bai 2011; Souzanchi et al. 2013).
Fig. 1 Schematic diagram of coating process: 1 air cylinder, 2 needle valve, 3 feedstock suspension, 4 peristaltic pump, 5 spray gun, and 6 stainless steel substrate
was cut (Fig. 2). These baffles were then arranged in a way to make zigzag arranged in a way to make zigzag pattern for liquid flow along the reactor length. Figure 2 shows configuration of reactor. This reactor has a large specific surface area coated with TiO2 photocatalyst and illuminated with four UV lamps. Moreover, fluid flow in the reactor is turbulent because of high amount of recycle flow and zigzag pattern provided with baffles, and as a result, it provides high mass transfer coefficient inside the reactor which is permanent parameter in reactor with immobilized catalyst (Ray 2009). To clarify this, Reynolds number in the channels is calculated according to the equation (1).
Photoreactor and experimental set-up The reactor was made of an aluminum tube (157 mm inside diameter and 300 mm in length) with four quartz tubes (23 mm outside diameter and 400 mm in length). Four UV lamps (16 W—Philips with λmax centered around 254 nm, 15 mm in diameter, and 303 mm in length) were located at the axis of the quartz tubes. Twelve stainless steel circular baffles (1 mm thickness) with specific geometry that were coated with TiO 2 nanoparticles were placed inside the reactor. The distance between sheets was 20 mm. In order to make flow route for liquid to pass, the top part of the baffles
ρud ð1Þ μ Where ρ is density of waste, u is velocity (u ¼ QA ), and A is surface area. In the present study, surface area was calculated to be 15cm2 and flow rate is approximately equal to recycle flow rate (7 × 10− 5m3. s− 1). Therefore, velocity of waste water in the channels is obtained as 0.06 m.s− 1. d is hydraulic diameter of the channel d ¼ 4 rH ; rH ¼ Sp , S is surface where flow passes Re ¼
through it and p is circumference that gets wet (d = 2.75cm). Therefore, Reynolds becomes 2,526. However,
Environ Sci Pollut Res Fig. 2 Dimension and arrangement of the photoreactor (not to scale)
considering that the reactor has baffles, the real Reynolds will be somewhat bigger than 2,526. Which suggests transient flow regime. Subsequently, Sherwood number Sh ¼ K:d D , which represents the ratio of convective to diffusive mass transfer may be evaluated. D=1.013×10− 9m2.s− 1 which is molecular diffusion coefficient was calculated elsewhere (Plugatyr and Svishchev 2011) Alternatively, Sherwood number may be calculated according to the equation (2) in pipes if flow regime is not laminar. (Treybal and Treybal Robert 1968): 1
Sh ¼ 0:023:Re0:83 :Sc 3
ð2Þ
Schmidt number for much diluted phenol-water solution is 1,900 (Kothandaraman 2004); therefore, Sherwood becomes 189.9. Finally, K is obtained as 7×10− 6 m.s− 1 which shows a moderate to good mass transfer coefficient. The experimental set-up as shown in Fig. 3 consists of liquid feed delivery system that is saturated with oxygen with an air compressor, the photoreactor and the output port. A centrifugal pump was used to pump the wastewater from a 50L reservoir at prescribed flow rate. Another centrifugal pump was located below the reactor and provided an adjustable
recycling system, feeding from the end of the reactor, and being discharged at the inlet. Flow of recycle was fixed at 7× 10−5 m3.s−1 during the experiments. Analyses The crystalline structure of calcined TiO2 films coated on stainless steel baffles was studied by XRD using Philips P1 140 XRD analyzer with a Cu Kα radiation. The surface structure and particle size of the prepared Ti O 2 f i l m s a m p l e s w e r e o b s e r v e d b y a V E G A \crTESCAN SEM followed by AU-coated through sputtering method. The specific surface area of the TiO2 films was measured by BET (Brunauer–Emmett– Teller) method, using Quantachrome Autosorb-1 BET analyzer. The concentration of phenol was measured by a UV–vis spectrophotometer (Dr 2,800 Hach Co.). It was used to determine the absorbance (A) of the samples (λmax =495 nm with pH=3) using specific reagents and following the instructions provided by the supplier (phenol Test Kit—Merck Spectroquant). Reduction in TOC contents after photocatalytic degradation was also determined through the TOC measurements spectrophotometrically at 605 nm using the TOC cell tests and following the instructions provided by Merck.
Environ Sci Pollut Res Fig. 3 Experimental set-up: 1 air compressor, 2 feed tank, 3 air distributor, 4 inlet valve, 5 centrifuge water pump, 6 centrifuge pump for recycling, 7 valve for controlling recycle flow rate, 8 recycle flow meter, 9 inlet flow meter, 10 photocatalytic reactor, 11 outlet valve, 12 flow meter to controlling outlet flow rate, and 13 outlet container
Experimental design method A series of preliminary conventional experiments were carried out to investigate the effect of each parameter on the degradation efficiency of phenol. Phenol inlet concentration, pH, flow rate, and TiO2 loading were considered as main parameters. Optimization was carried out according to the preliminary experimental results using RSM. The Box–Behnken design (BBD) method was employed to analyze the simultaneous effect of phenol inlet concentration (20–100 ppmv), pH (3–9), flow rate (2.53×10−7−1.26×10−6 m3.s−1) and TiO2 loading (8.8–17.6 g.m−2) on the phenol conversion and also to evaluate the interactions between the studied parameters. The current design, which is a modified central composite experimental design, is more efficient than the three-level full factorial designs (Kayan and Gözmen 2012; Zhang et al. 2011). Optimum conditions and interactive effects of the four parameters were determined using regression and graphical analysis by the Design Expert® Software (Version 7.6.1). Response (R) during the experimental design is phenol conversion.
Results and discussion Characterization of the immobilized TiO2 film XRD is an analytical technique that is used to examine the crystalline structure of materials by determining bond length and angles (i.e., the diffraction pattern of X-ray as electromagnetic radiation depends on the right
order of spacing which exists in crystals where distances between atoms are of the order of few angstroms). Figure 4 shows XRD patterns obtained for the pure TiO2 and immobilized TiO2 films on stainless steel and aluminum after being used in photo catalytic experiments. As it is seen, the main peaks are at two-theta (2θ) angle of 24.9° and 53.3°, which are corresponded to anatase and rutile crystalline phases, respectively. The content of the anatase (A) and rutile (R) phases in TiO2 sample can be determined based on the intensities of the peaks in the XRD pattern as follows (Qiu and Zheng 2007; Souzanchi et al. 2013: Spurr and Myers 1957): IR 0:79 IA XR ¼ IR 0:79 1þ IA
ð3Þ
Where XR is the weight fraction of the rutile phase in the sample, IA and IR are the intensities of the reflection peaks of anatase and rutile phases in the diffractograms, respectively. It is clear that the intensity of the radiation from a given crystalline structure of the element (for example, “A” or “R” phase) is depended on the amount of the structural element present in the crystal sample. In this study, percentage of anatase phase of the pure TiO2 and TiO2 after immobilization on stainless steel and aluminum and being used in the photocatalytic experiments was 80.0 %, 78.3 %, and 76.8 %, respectively. In fact, crystalline composition of anatase-based TiO2 particles was remained almost constant after immobilization on the stainless steel and aluminum
Environ Sci Pollut Res
he main grain size (L) of the TiO2 sample can also be estimated using XRD data and applying Scherrer’s formula as follows (Hosseini et al. 2007, Patterson 1939): L¼
Fig. 4 X-ray diffraction spectra of TiO2 coatings prepared in optimum conditions (loading of 17.6 g.m−2): a stainless steel substrate, b aluminum substrate, and cpure TiO2
support and use in photocatalytic experiments. A higher content of rutile phase of TiO2 catalyst is related to a lower photocatalytic activity of TiO2 (Fernández et al. 1995; Mohammadi et al. 2014; Qiu and Zheng 2007). According to equation 3 and XRD analysis of TiO2 thin films on the stainless steel and aluminum substrate photocatalyst thin film contain from roughly 80 % anatase phase of TiO2 and 20 % rutile phase of TiO2 and because the amount of SDS is little we do not see any sharp peak of SDS. Density of the anatase phase of TiO2 is 3.89, and density of rutile phase is 4.23. Total density of TiO2 film can be obtained according to the following equation. densityt ¼ 0:2 densityrutile þ 0:8 densityanatase
ð4Þ
And it was 3.89. Although, the film density (after coating process) and pure TiO2 are not necessarily the same, this is a usual assumption used in other studies (Han and Bai 2010, 2011). After calculation of density, TiO2 film thickness was roughly estimated by dividing TiO2 loading rate with the density. The film thicknesses, which were obtained by this method, are 2.2, 3.3, and 4.4 μm, respectively. In previous study photocatalytic activity of TiO2 for degradation of malic acid had better performance on quartz support and this phenomenon might be because e_/h + recombination at the TiO2 surface due to the presence of metal ions impurities at high calcination temperature (Fernández et al. 1995). Due to this fact in this study, calcination temperature was selected 350 °C, and no considerable rutilization of the TiO 2 was detected (Fig. 4)
0:89λ βcosθ
ð5Þ
Where λ the X-ray wavelength and β is the full width at the half of the maximum peak. The crystal size for the TiO2 in the present work for pure TiO2 and TiO2 after the immobilization on stainless steel and aluminum and use in the experiments was 23.9, 26.9, and 27.1 nm respectively. These were in agreement with the size reported by others who used stainless steel support and applying sol–gel method for preparing the TiO2 film (Chen and Dionysiou 2006b). In general, coating process has not brought any significant size changes to the P-25 material used in the coating process. SEM analysis (Fig. 5) was employed to examine the surface morphology of the TiO2 films that was prepared in the present work. The structure and morphology of TiO2 coating baffles in optimal condition and before being used in photocatalytic experiments is shown in Fig. 5. From the Fig.5a, it is noted that surface of stainless steel support is completely covered with TiO2 films. It can be seen that TiO2 particles have been distributed densely and uniformly over the surface of stainless steel substrate forming a relatively integrated layer with many micro pores among the TiO2 particles. Figure 5b shows that surfaces of films are full of visible agglomerates which are caused by incorporation of P-25 powders in the film. Moreover, it can be clearly seen that large agglomerates are composed of many grains with average size of approximately 25 nm which is nearly equal to crystal sizes obtained from XRD pattern. The surface of the TiO2 films after photocatalytic degradation of phenol in the ABPR have been changed considerably (Fig. 5c). It is clear from the image (Fig. 5c) that TiO2 nanoparticles aggregated and at the same time some micro pores were formed in the TiO2 films. Developing inter-aggregate creates new areas at the photocatalyst surface and this may affect process positively with respect to absorption of UV light as well as better access of phenol and oxygen molecules to active sites within the TiO2 particles (Souzanchi et al. 2013). Specific surface area is also one of the important factors to analyze morphology of the immobilized TiO2 films as used in the photocatalytic degradation experiments (Chen and Dionysiou 2006a). Study of Chen and Dionysiou ( 2006a) on TiO2 films coated on stainless steel support confirmed the positive correlation between high BET specific surface area
Environ Sci Pollut Res
Fig. 5 SEM images of TiO2 coatings prepared in optimum conditions a before performing photocatalytic experiments SEM MAG=20,000, b before performing photocatalytic experiments SEM MAG=50,000, and
c after performing photocatalytic experiments SEM MAG=50,000 (loading of 17.6 g.m−2)
and photocatalytic activity. In the present study, the BET surface area was 47.2 and 45.8 m2 g−1 for the supported TiO2 films before and after the photocatalytic degradation process inside reactor, respectively. Work of other researchers showed that as the calcination temperature increased from 400 °C to 600 °C, the BET surface area for the TiO2 film on the stainless steel support decreased from 49.6 to less than 2 m2 g−1 (Chen and Dionysiou 2006a, Souzanchi et al. 2013). In this study, BET surface area was high and remained constant after experiments. Results of weighting method used for measuring loading of the photocatalyst inside the reactor showed no considerable change of the photocatalyst amount before and after the experiments. After last experiment for each coating, the first experiment for that coating was repeated to check the consistency of results and results showed no change again. Due to all of the results that were shown above, it is expected that photocatalytic activity and stability of the coated TiO2 in ABPR would be reasonably good. Also photocatalytic activity loss was measured in optimal condition. Phenol concentration in the outlet was measured every 0.5 h and is reported in Table 1. No sharp reduction in photocatalytic degradation was observed after 9 h. Note that system reaches steady state after about 6 h.
design were processed using Design Expert® 7.6.1 software and were fitted to a quadratic model. Finally, the best model equation in terms of coded factors (95 % confidence level (p4.0) indicates adequate precision. The coefficient of determination is defined as the ratio of the explained variation to the total variation and
is used as a measure of degree of fit of the model. The predicted R2 value of 0.9961 for phenol removal is in good agreement with the corresponding adjusted R2 value of 0.9936. Therefore, the quadratic polynomial equation could be used to predict the degradation of phenol in the experimental range (Chen et al. 2012, Cheng et al. 2012, Kayan and Gözmen 2012, Khataee et al. 2010). The predicted values
Table 3 Experimental design matrix and responses based on experimental runs proposed by BBD design Run
C Inlet concentration (ppmv)
P pH
F Flow rate (m3/s)
L Catalyst loading (g/m2)
Degradation % (experimental)
Degradation % (predicted)
1 2 3
20 100 60
6 6 6
7.56×10−7 7.56×10−7 7.56×10−7
17.6 17.6 13.2
45.07 12.43 22.50
44.84 12.88 21.14
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
60 60 60 60 60 100 60 60 60 60 60 60 60 60 100 20 20 60 60
9 6 3 3 9 6 6 3 6 6 3 6 6 6 6 6 3 6 9
7.56×10−7 2.53×10−7 2.53×10−7 1.26×10−6 7.56×10−7 7.56×10−7 7.56×10−7 7.56×10−7 7.56×10−7 7.56×10−7 7.56×10−7 1.26×10−6 7.56×10−7 1.26×10−6 2.53×10−7 1.26×10−6 7.56×10−7 7.56×10−7 2.53×10−7
17.6 17.6 13.2 13.2 8.8 8.8 8.8 17.6 13.2 13.2 8.8 17.6 13.2 8.8 13.2 13.2 13.2 13.2 13.2
15.30 46.85 27.03 3.60 7.92 7.10 28.72 14.00 21.69 20.92 5.40 11.34 19.80 10.26 20.63 20.63 30.88 20.78 31.23
16.49 45.75 29.24 2.36 7.38 8.17 28.12 14.68 21.14 21.14 5.57 10.35 21.14 9.77 20.66 21.35 30.39 21.14 31.05
23 24 25 26 27 28 29
60 100 20 20 20 100 100
9 6 6 6 9 9 3
1.26×10−6 1.26×10−6 2.53×10−7 7.56×10−7 7.56×10−7 7.56×10−7 7.56×10−7
13.2 13.2 13.2 8.8 13.2 13.2 13.2
3.30 10.61 66.56 30.96 32.21 5.97 4.17
4.17 11.74 66.18 31.34 32.2 4.64 2.83
Environ Sci Pollut Res Table 4 Analysis of variance regression model for phenol degradation Source
Sum of squares
Model C pH F L C.F. C.L. F.L. C2 pH2 F2 L2 Residual Lack of fit Pure error
df
Mean square
F value
p value Prob>F
12 1 1 1 1 1 1 1 1 1 1 1 16 12 4
486.59 2,011.39 10.94 2,167.87 252.36 322.2 17.56 72.68 320.42 342.04 129.3 9.27 0.39 0.48 0.14
1,235.21 5,105.93 27.78 5,503.15 640.61 817.91 44.57 184.49 813.4 868.26 328.22 23.53
