Modelling of succinic acid heterogeneous catalytic ozonation on metallic foam

Chemical Engineering Journal 279 (2015) 1004–1009

Contents lists available at ScienceDirect

Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej

Modelling of succinic acid heterogeneous catalytic ozonation on metallic foam Aude Audirac a,b, Florence Pontlevoy b, Nathalie Karpel Vel Leitner a,⇑ a University of Poitiers, Institute of Chemistry of Materials and Natural Resources, UMR CNRS 7285, Department of Water, Geochemistry and Health, ENSIP, 1 rue Marcel Doré, 86073 Poitiers Cedex 9, France b Technavox, ENSIP, Plate-forme Eaux, Bâtiment B16, 7 rue Marcel Doré, 86073 Poitiers Cedex, France

h i g h l i g h t s  A metallic foam was used in a catalytic ozonation device.  The concentration of the model molecule decreased linearly with ozonation time.  Modelization of the concentration pattern was consistent with the experimental results.  Above a given concentration of compound, the rate reached a constant value.  Results suggest catalytic ozonation proceeds through two successive reaction steps.

a r t i c l e

i n f o

Article history: Received 19 February 2015 Received in revised form 4 May 2015 Accepted 8 May 2015 Available online 15 May 2015 Keywords: Ozone Solid foam catalyst Modelling

a b s t r a c t This work aims at envisaging the use of a solid catalyst in the catalytic ozonation process to avoid the separation step needed after the treatment for catalyst recovery when powdered material is involved. This work devoted to the removal of succinic acid (SA), a low molecular weight refractory organic acid whose heterogeneous catalytic ozonation was previously studied using powder form catalysts. Heterogeneous catalytic ozonation of succinic acid was carried out in a reactor fitted with a recirculation loop. This study investigated if the location of the solid catalyst, i.e. inside the reactor in the zone of ozone transfer or in a cartridge placed on the recirculation loop, had an influence on the amount of SA removed or on the elimination rate expression. No influence of the catalyst location was detected. The rates of SA removal by catalytic ozonation were found to depend on both the catalyst concentration and the initial SA concentration for weak SA initial concentrations, but only on the catalyst concentration for greater initial SA concentrations. The pattern of SA concentration from modelization closely matched the experimental results. This implementation of the catalytic ozonation process with solid catalyst is of interest for the application at industrial scale. The 1.8 ratio between the rates of SA and DOC removal showed that mineralization to CO2 did not occur simultaneously with SA removal. Ó 2015 Elsevier B.V. All rights reserved.

1. Introduction Heterogeneous catalytic ozonation is viewed as a promising way to treat waters containing refractory compounds such as dyes, some aromatic structures, low molecular weight substances, etc. [1–4]. There is growing interest in the choice and preparation of appropriate catalysts, as indicated by the numerous studies devoted to this subject [5–8]. However, only a few articles have assessed catalytic ozonation at semi-industrial or industrial scales. Most laboratory studies are performed on powder form catalysts poured in stirred vessels [9–11], and few of them optimize the ⇑ Corresponding author. Tel.: +33 (0)5 49 45 39 16. E-mail address: [email protected] (N. Karpel Vel Leitner). http://dx.doi.org/10.1016/j.cej.2015.05.030 1385-8947/Ó 2015 Elsevier B.V. All rights reserved.

operating conditions from a process standpoint [12]. Several patented processes based on heterogeneous catalytic ozonation, with a powder catalyst operating in slurry-type reactor [13–15] or with pellets operating in a trickle-bed-type reactor, have also been reported [16–19]. However, no published data are available regarding the design and optimization of the operating conditions. Using a powder catalyst implies a separation step to recover the catalyst. This step requires time, penalizing the whole process and being prohibitive to its use in industrial field. Furthermore, industrial effluents often contain suspended solids which could disturb the separation process whatever the technology is used. Using a solid catalyst, fixed in the system makes the separation step unnecessary and would allow the treatment of large flowrate.

1005

A. Audirac et al. / Chemical Engineering Journal 279 (2015) 1004–1009

This study was devoted to the optimization of operating conditions in a heterogeneous catalytic ozonation system through a study of the catalyst location in the process. The model molecule was succinic acid (SA), a low molecular weight refractory organic acid whose heterogeneous catalytic ozonation was previously studied by Ernst et al. [20], Karpel Vel Leitner et al. [21] and Delanoë et al. [22] using powder form catalysts. It was found that the presence of 0.2 g L1 of a Ru/CeCO2 catalyst enabled 100% removal of succinic acid (2 mM) within 90 min ozonation [22]. Total mineralization of a 1.25 mmol L1 solution was achieved in about 1 h with 50 g L1 of Al2O3 [20]. In the present study, the catalyst location and kinetic parameters were studied, while focusing particularly on the influence of succinic acid and catalyst concentrations. The catalyst used in this study was supported on a metallic porous structure.

2.2. Methods All experiments were performed in demineralised water (DOC = 0.1 ± 0.05 mg C L1, pH = 5.7, conductivity = 1.6 lS cm). Succinic acid of over 99% purity was supplied by Sigma–AldrichÒ. Its concentration was measured by HPLC for different ozonation times, using an ion exchange resin-packed column composed of sulfonated di-vinylbenzene polystyrene (SUPELCOGEL C-610) and a UV-detector set at 210 nm. The mobile phase was ultrapure water acidified with 1‰ (v/v) of phosphoric acid. To assess the degree of mineralization of the succinic acid, DOC was measured using a TOC-Vcsh analyzer (Shimadzu Corporation). Dissolved ozone was measured by the Indigo Trisulfonate method [23]. 3. Results and discussion

2. Materials and methods 3.1. Preliminary validation experiments 2.1. Material The reactor containing 17 L of aqueous solution was continuously fed with gaseous ozone through a fine bubble diffuser. The device included a recirculation loop for the solution, as shown in Fig. 1. Ozone was produced from oxygen by a CFS-1A ozone generator (Ozonia). The volumetric flow rate was 0.16 Nm3 h1, with an ozone concentration of 130 g Nm3. Various recirculation flow rates between 80 and 160 L h1 were tested. When the system was operated in continuous mode, the solution flow rate was 10 L h1. The ozone gas concentration was monitored online using a BMT-694-BT ozone analyser (BMT Messtechnik). The catalyst was located either directly inside the reactor or in the recirculation loop and the system was operated in continuous or semi-continuous mode. When the catalyst was located directly inside the reactor, it was placed in a stainless steel basket submerged in the liquid phase. The catalyst was placed in a cartridge when it was located in the recirculation loop. The catalyst was stuck on a disk shaped metallic foam (40 mm of diameter). The foam pore diameter was 2.3 mm, and the disk thickness was 5 mm. The mass of active material on the solid was around 400 mg per disk. One or three disks were used for the experiments, corresponding to 20.3 and 64.7 mg of catalyst per litre of solution in the reactor, respectively.

Destructor

(b)

(a)

3.1.1. Mixing efficiency Several experiments were performed to examine if: (i) the concentration was similar at a given time in different points of the reactor and in the recirculation loop, and (ii) the time needed to achieve a perfect mixing was sufficiently short to be considered negligible. The mixing efficiency was assessed under semi-continuous conditions using Acid Red 94 (Rose Bengal) dye (AldrichÒ) as tracer. For these experiments, gaseous ozone was replaced by oxygen. At t0, the initial time, 10 mL of Acid Red solution (5.2 g L1) were poured in the reactor containing 17 L of pure water. The recirculation flow rate was set at 120 L h1. Three different sampling points were chosen (Fig. 1): within the reactor (1), before (2) and after the catalyst-loaded cartridge (3) containing the catalyst. The absorbance of the samples withdrawn at different times was measured at 549 nm. The results are shown in Fig. 2. These results were clearly repeatable: tests 1 and 2 exhibited the same time-course pattern. The dye concentration variations were the same in samples from the reactor and from the recirculation loop (sampling points: 1 and 2, respectively, in Fig. 1). It appeared that after 40 s approximately 90% of the final dye concentration was reached inside the reactor and in the recirculation loop and perfect mixing was obtained within only 1 min. Consequently, injecting fine bubbles of ozone gas (0.16 Nm3 h1) with a recirculation flow rate of 120 L h1 could provide sufficient

Ozone generator

1.2

1

(3)

(2)

[dye]/[dye]f

0.8

Feeding tank

0.6

0.4

Test 1 Reactor Test 1 Loop

(1) Treated water tank

0.2

Test 2 Reactor Test 2 Loop

0 0

100

200

300

400

Time (s) Fig. 1. Experimental reactor system with the catalyst located directly in the reactor, or in the recirculation loop. Samples can be drawn from three sampling points (1–3) and in the treated water tank.

Fig. 2. Time-course variations in the dye concentration in the reactor and in the recirculation loop (gas flow rate: 0.16 Nm3 h1, recirculation flow rate: 120 L h1).

1006

A. Audirac et al. / Chemical Engineering Journal 279 (2015) 1004–1009

mixing to get a similar concentration throughout the reaction system. The system hydraulics were also validated during ozonation of pure water without catalyst (not presented). During this experiment, it was checked that the concentration of dissolved ozone was similar inside the reactor (point 1, Fig. 1) and in the recirculation loop (points 2 and 3, Fig. 1) (8 mg L1). Moreover, it was confirmed that without catalyst SA was refractory to ozone since no variations in the SA concentration were observed. 3.1.2. Ozone concentration Preliminary catalytic ozonation experiments were performed on aqueous solutions of SA ([SA]0 = 1 mmol L1) in order to be sure that, when the catalyst was placed in the recirculation loop, the concentration of dissolved ozone was not limiting for the reaction. The concentrations of dissolved ozone measured upstream (point 2, Fig. 1) and downstream (point 3, Fig. 1) from the catalyst during ozonation showed a rapid increase followed by only weak variations. The values obtained at the two sampling locations were very close. The concentration of dissolved ozone at the cartridge outlet represented 85–92% of the inlet, i.e., the reaction was found to consume only 8–15% of dissolved ozone through the cartridge. Therefore, there was enough ozone for the reaction. Moreover, the ozone gas concentration measured successively at the reactor inlet (point (a)) and in the vents (point (b)) showed that the ozone gas concentration in the vent also quickly levelled off during the experiment (Fig. 3). The global transfer efficiency was 11% due to the fact that ozone gas was introduced in large excess. As the conditions were similar in terms of mixing and ozone concentration wherever the catalyst was located, the reaction rates at the two catalyst locations were compared.

where r is the rate of SA removal in mmol L1 min1; k is the rate constant of the a

reaction

in

c L1þaþbþc mmolO3 mmolSA min mg cata ; [O3] is the dissolved ozone concentration in mmol L1; [SA] is the succinic acid concentration in mmol L1; [catalyst] is the catalyst concentration in mg L1; a, b, c are the partial kinetic orders regarding dissolved ozone, SA and catalyst, respectively. 1b

1

With ozone being continuously introduced in the reactor, and in large excess, it was previously found (see Section 3.1.2) that the dissolved ozone concentration was constant during the reaction time. The dissolved ozone concentration was found to be 22 ± 2 mg L1 for all the experiments performed, so this constant could be included in the rate constant. In the course of the ozonation experiments, the rate expression (3.1) became:

r ¼ k1 ½SAb ½catalyst

c

ð3:2Þ

c where k1 ¼ k½O3 a in L1þbþc min mg cata mmolSA . The results obtained under the semi-continuous conditions presented in Fig. 4 showed that the time-course variations in the succinic acid concentration were linear. Consequently, the model expression could be further simplified by considering b = 0 regarding SA: 1

r ¼ k2 ½catalyst

1b

c

ð3:3Þ

where k2 is the apparent rate constant of the modellized reaction in c mmolSA L1þc min mg cata and k1 is equal to k2. These results are similar to the data obtained by Beltràn et al. [12] with TiO2 catalyst and by Gonçalves et al. [24] with multi-walled nanotubes. These authors also obtained zero-order kinetics regarding a low molecular weight organic acid. 1

3.2. Influence of the catalyst location Experiments carried out for 2 h without ozone and with 64.7 mg L1 of catalyst showed that the SA concentration did not vary (results not shown), indicating negligible adsorption phenomena on the catalyst. Therefore, in the following experiments, the removal of SA could be considered to be only due to a heterogeneous catalytic reaction. The two catalyst locations tested were: (i) directly in the reactor in a totally immerged stainless steel basket, and (ii) in a cartridge placed in the recirculation loop. In semi-continuous mode, initial SA concentrations between 0.9 and 7.7 mmol L1 were tested and the catalyst concentration was 64.7 mg L1. When the catalyst was placed in the recirculation loop, the recirculation flow rate (Qloop) was varied from 80 to 160 L h1 to assess its influence on the catalytic activity. The results obtained for an initial SA concentration of 4 mmol L1 are presented in the insert of Fig. 4. The removed concentration was identical for the three recirculation flow rates, indicating that the recirculation flow rate did not influence the catalyst activity between 80 and 160 L h1. From an industrial standpoint, pump operational costs could be reduced by lowering the recirculation flow rate. For the different SA concentrations, the results did not depend on the catalyst location (Fig. 4). For example, for initial SA concentrations of 0.9 and 4 mmol L1, the removed concentrations were 0.4 and 1.6 mmol L1, respectively, for 90 min reaction time. It was shown that the converted SA concentration was similar for [SA]0 = 5.6 mM and 7.7 mM and was equal to 2.2 mmol L1. 3.3. Catalytic ozonation modelization The general rate expression for catalytic ozonation of SA was considered as follows:

r ¼ k½O3 a ½SAb ½catalyst

c

ð3:1Þ

3.3.1. Catalyst weight influence To determine the expression of the rate for various catalyst weights, semi-continuous experiments were carried out with different volumes of SA solution ([SA]0 = 1 mmol L1), ranging from 3 to 17 L in the reaction system, corresponding to 360– 64.7 mg L1 of catalyst, respectively. An additional experiment was performed with 20.3 mg L1 of catalyst. Fig. 5 shows the linear SA concentration pattern for the different amounts of catalyst that enabled the determination of the corresponding rates. The insert of Fig. 5 presents the rates versus the catalyst concentration for an initial SA concentration of 1 mmol L1, indicating the proportional relation (with c = 1) as follows:

r ¼ 0:07 ½catalyst for ½SA0 ¼ 1 mM

ð3:4Þ 1

where [catalyst] is the catalyst concentration in mg L

.

3.3.2. Influence of the initial succinic acid concentration To complete the modelling, from the semi-continuous experiments performed with different initial SA concentrations (0.9– 7.7 mmol L1) and a catalyst concentration of 64.7 mg L1 (presented in Fig. 4), the rates were calculated. It appeared that in the 0.9–5.6 mmol L1 range, the rate increased linearly with the initial SA concentration (Fig. 6). The slope of the straight line led to the expression (3.5):

r ¼ 4:5  103 ½SA0 where

ð3:5Þ

1007

A. Audirac et al. / Chemical Engineering Journal 279 (2015) 1004–1009 140 Gaseous Ozone Concentration (g/Nm3)

Dissolved ozone concentration (mg/L)

30

25

20

15

10

dissolved ozone after cartridge (3)

5

120 100

80 60

40 20

CgIN

Dissolved ozone before cartridge (2)

CgOUT

0

0 0

25

50

75

100

125

150

0

25

50

Time (min)

75

100

125

150

Time (min)

Fig. 3. Concentrations of dissolved ozone (left), inlet and outlet ozone gas (right) during catalytic ozonation of SA. [SA]0 = 1 mmol L1, [O3]g = 130 g Nm3, Qg = 160 L h1, Qloop = 100 L h1.

2.5 A

B

D

1.5 1.0 0.5

1.5 0

0.0 100

50

1.0

0.5

0.0 0

20

40 60 time (min)

80

100

Fig. 4. Catalytic ozonation of SA for different initial concentrations with [O3]g = 130 g Nm3, Qg = 160 L h1. Full symbols: catalyst in the reactor, empty symbols: catalyst in the cartridge. } [SA]0 = 0.9 mM; x [SA]0 = 1.3 mM; D [SA]0 = 4 mM; h [SA]0 = 5.6 mM; s [SA]0 = 7.7 mM. Insert: [SA]0 = 4 mM; A: Qloop = 120 L h1, B: Qloop = 120 L h1, C: Qloop = 60 L h1, D: Qloop = 160 L h1.

Above [SA]0 = 5.6 mmol L1, as shown in Fig. 6, the rate would reach a maximum value of 25  103 mmol L1 min1 (for [catalyst] = 64.7 mg L1) and be independent of the initial SA concentration. The amount of SA removed in the system for [catal yst] = 64.7 mg L1 reached a maximum value of 23.5 mmol h1. Complementary semi-continuous experiments were also performed with a 20.3 mg L1 catalyst concentration (Fig. 6). The same phenomenon as with a catalyst concentration of 64.7 mg L1 was observed. The rate increased linearly with the initial SA concentration, for initial SA concentrations below 5 mmol L1, and then remained constant. The slope was found to be 1.4 min1 in the 1–5 mmol L1 concentration range. The value of the rate for the initial SA concentration above 5 mmol L1 was 3 1 1 7.5  10 mmol L min . With a 20.3 mg L1 catalyst concentration, the maximal amount of SA which could be removed was 7.5 mmol h1. Table 1 summarizes the rate expressions in the 0.9–11 mmol L1 initial SA concentration range for 20.3 and 64.7 mg L1 catalyst.

1.0 A

B

C

D

E

F

30

25

Rate (micromol/L/min)

0.8

0.6

12 10 8 6 4 2 0

v (mmol/L/min)

SA (mmol/L)

2.5

0.4

0.2

y = 0.07x R2 = 0.9927 0

0.0 0

20

40 60 time (min)

20 1.5 continuous/reactor continuous/cartridge batch/reactor batch/cartridge

15

80

100

Fig. 5. Time-course variations of succinic acid concentration with [SA]0 = 1 mmol L1. Insert: Rate against the catalyst concentration with [SA]0 = 1 mmol L1. (A = 275 mg L1; B = 157 mg L1; C = 103 mg L1; D = 92 mg L1; E = 65 mg L1 and F = 20 mg L1.)

1.0

10 0.5

20.3 mg L-1

5

100 200 [catalyst] (mg/L)

2.0

64.7 mg L-1

Removed concentration (mmol/L)

2.0 SA0-SA (mmol/L)

r is the rate of SA removal in mmol L1 min1; [SA]0 is the initial succinic acid concentration in mmol L1 (for [SA]0 < 5.6 mM).

2.0 C

0.0

0 0

2 4 6 8 10 Succinic acid initial concentration (mmol/L)

12

Fig. 6. Rate of SA removal by catalytic ozonation for different initial SA concentrations ([catalyst] = 64.7 and 20.3 mg L1); Right scale: SA removed for 90 min reaction time.

1008

A. Audirac et al. / Chemical Engineering Journal 279 (2015) 1004–1009

Table 1 Rate expressions for [SA]0 between 0.9 and 11 mM and [catalyst] of 20.3 and 64.7 mg L1.

[SA]0 (mmol L ) Rate (mmol L1 min1)

64.7 <5.6 4.5  103 [SA]0 0.07  103 [SA]0 [catalyst]

20.3 >5.6 25  103 0.37  103 [catalyst]

3.3.3. Rate modelling By pooling the results from Sections 3.3.1 and 3.3.2, the global modelling for catalytic ozonation of SA by metal foam was determined. Based on the results of Section 3.3.1, whereby the rate of SA removal was proportional to the catalyst concentration, by dividing the expressions of Table 1 by the catalyst concentration used, i.e. 64.7 or 20.3 mg L1, rate constants of 3 1 1 0.07  10 L mg min were obtained for SA concentrations below [SA]lim. Above [SA]lim, the rate of SA removal was a constant function of the catalyst concentration only (r = 0.37  103 [catalyst]). For the two catalyst weights tested, [SA]lim can be easily calculated and is 5.3 mmol L1 (i.e. 0.37/0.07). Considering the data of Section 3.3.1 above obtained at various catalyst concentrations, dividing relations (3.4) by the initial concentration of SA (1 mM) yielded rate constants of 0.07  103 L mg1 min1. This value is in agreement with the data in Table 1. Then, the general rate modelling (3.6) could be proposed:

r ¼ 0:07  103 ½SA0 ½catalyst

ð3:6Þ

This equation is valid up to an initial SA concentration ([SA]0 lim) above which the rates reached a constant value (Fig. 6) of r = 0.07  103 [SA]0 lim [cata]. The [SA]0 lim value was found to be 5.3 mmol L1 and independent of the catalyst concentration. This original behaviour could be explained by the presence of a limiting step in the reaction system that would not involve the catalyst, such as the probable contribution of the formation of an intermediate with ozone. Limitation by ozone seemed not consistent with our results, as ozone transfer from gas to liquid reached a constant value within the first minutes of experiments and was found constant during our experiments, whatever the amount of succinic acid was removed. Moreover concerning the transfer to the catalyst surface, it was observed that the concentration of dissolved ozone was independent of the operating conditions (recirculation flowrate). Fig. 7 shows a comparison between the experimental and modelled rates. The modelled results closely matched the data obtained. Regardless of the initial succinic acid or catalyst concentration, the modelled results were ±15% equal to the experimental rate. The kinetic parameters obtained in this study were consistent with previously published values for zero order kinetics. Indeed, Beltràn et al. [12], found rates between 7.2 and 44 lmol L1 min1 for oxalic acid with TiO2 and Gonçalves et al. [24], found values between 11.5 and 18 lmol L1 min1 for oxalic acid with 0.14 g L1 of carbon nanotubes. In a previous work from our laboratory [21], rates between 2.5 and 5.5 lmol L1 min1 were obtained for succinic acid.

<5 1.4  103 [SA]0 0.07  103 [SA]0 [catalyst]

>5 7.5  103 0.37  103 [catalyst]

30 r experimental (micromol/L/min)

1

25 20

15

10 5

0 10

0

20

30

r modeling (micromol/L/min) Fig. 7. Experimental rate versus modelled rate of catalytic ozonation of succinic acid [catalyst] = 20.3 and 64.7 mg L1, [SA]0 = 1–10 mmol L1.

Under the continuous conditions, and as also noted in semi-continuous mode, the catalyst location had no influence on the amount of SA removed. For initial SA concentrations of 1.7 and 4 mmol L1 and 64.7 mg L1 catalyst, the removed concentrations were 1 and 1.6 mmol L1 irrespective of where the catalyst was located. Similar to the semi-continuous conditions, the amount of SA removed increased linearly with the initial SA concentration (up to about 5.5 mmol L1) and then levelled off at 2.2 mmol L1 for higher initial concentrations (Fig. 6, small symbols). Consequently, it was calculated that 22.5 mmol was the maximum amount of SA that could be removed per hour by 64.7 mg L1.

90 y = 1.8112x

80 Succinic acid removal (%)

[Catalyst] (mg L1)

R2 = 0.9071

70 60 50 40 30 20 10

3.4. Continuous experiments Additional experiments were performed in continuous mode with the configuration presented in Fig. 1. In these conditions, the flow rate of the feed solution was 10 L h1, corresponding to a hydraulic retention time of 102 min, while the gas flow rate was 160 Nm3 h1 and the recirculation flow rate was 120 L h1.

0 0

10

20 30 DOC removal (%)

40

50

Fig. 8. Comparison between succinic acid removal and mineralization during catalytic ozonation in semi continuous or continuous mode ([catalyst] = 64.7 mg L1).

1009

A. Audirac et al. / Chemical Engineering Journal 279 (2015) 1004–1009 Table 2 Rate expressions from the experimental data for [DOC]0 between 43 and 490 mg C L1 and [catalyst] equal to 64.7 mg L1. [Catalyst] (mg L1)

64.7

[DOC]0 (mg L1) Rate (mg C L1 min1)

<290 2.5  103 [DOC]0 0.037  103. [DOC]0 [catalyst]

3.5. Mineralization experiments DOC (dissolved organic carbon) removal was monitored over a time course for each experiment in continuous or semi-continuous mode. As can be seen in Fig. 8, there was a difference between the SA and DOC elimination yields. SA was not oxidized to CO2 upon its removal. The ratio between SA removal and DOC removal was 1.8, which means that by-products were formed and oxidized more slowly in CO2. 100% DOC removal was obtained for higher reaction times only. From the catalytic ozonation experiments with 64.7 mg L1 of catalyst, the results presented in Table 2 showed that the rate of DOC removal had the same profile as the rate of SA removal with the two zones. In the first area, between 43 and 290 mg C L1 of the initial DOC concentration (corresponding to 0.9–6 mmol L1 of SA, respectively), DOC removal was linear, with a slope of 2.5  103 min1. Above 290 mg C L1 of the initial DOC concentration, i.e. in the second area, the rate was constant and equal to 680  103 mg C L1 min1. 4. Conclusion Succinic acid is a small dicarboxylic acid refractory to single ozonation in aqueous solution. It was not found to adsorb to a significant extent on the surface of the catalyst used. Catalytic ozonation experiments were performed in a bench scale reactor in which the mixing efficiency was checked. Two catalyst locations were tested, i.e. within the reactor itself and in a recirculation loop. Similar SA removal yields were obtained in these two experimental configurations and SA concentrations decreased linearly with the reaction time. For the smallest initial SA concentrations (0.9– 5.6 mmol L1), the removal rate increased linearly with the initial SA concentration. Above 5.6 mmol L1 of the initial SA concentration and independently of the catalyst concentration, the rate of SA removal reached a constant value. This behaviour could indicate that successive reactions were involved to enable the catalytic ozonation to proceed, with a necessary step independently of the catalyst. However, the rates were always proportional to the catalyst concentration. Under the present experimental conditions, the maximum amount of succinic acid removed per hour for the catalyst concentration used, i.e. 20.3 and 64.7 mg L1, was found to be 7.5 and 23.5 mmol, respectively, and total mineralization did not occur upon SA removal. A 1.8 ratio between the rate of SA and DOC removal was observed.

>290 680  103 10.0  103 [catalyst]

These results obtained with a solid catalyst within the system will allow to envisage the catalytic ozonation process to be implemented at an industrial scale. Indeed, it would avoid a separation step for catalyst recovery at the end of the process and would allow the treatment of large flowrates. The modelization used in this study enable the forecast of the amounts of pollutant that could be removed and consequently the design of industrial treatment units. The data provided from this study were integrated in a patent filed by our research group [25]. References [1] R. Gracia, S. Cortés, J. Sarasa, P. Ormad, J.L. Ovelleiro, Ozone Sci. Eng. 22 (2000) 461–471. [2] T.S. Ping, L.W. Hua, Z.J. Qing, C.C. Nan, Ozone Sci. Eng. 24 (2002) 117–122. [3] Udrea, C. Bradu, Ozone Sci. Eng. 25 (2003) 335–343. [4] Y.X. Yang, J. Ma, J. Zhang, S.J. Wang, Q.D. Qin, Ozone Sci. Eng. 31 (2009) 45–52. [5] D.S. Pines, D.A. Reckhow, Ozone Sci. Eng. 25 (2003) 25–39. [6] V.S.R.R. Pullabhotla, C. Southway, S.B. Jonnalagadda, Catal. Commun. 9 (2008) 1902–1912. [7] L. Li, W. Ye, Q. Zhang, F. Sun, P. Lu, X. Li, J. Hazard. Mater. 170 (2009) 411–416. [8] R.C. Martins, R.M. Quinta-Ferreira, Ozone Sci. Eng. 31 (2009) 402–411. [9] B. Kasprzyk-Hordern, P. Andrzejewski, J. Nawrocki. 27 (2005) 301–310. [10] F.J. Beltràn, J.P. Pocostales, P.M. Alvarez, J. Jaramillo, J. Hazard. Mater. 169 (2009) 532–538. [11] T. Merle, J.S. Pic, M.H. Manero, H. Debellefontaine, Ozone Sci. Eng. 32 (2010) 391–398. [12] F.J. Beltràn, F.J. Rivas, R. Montero de Espinosa, Appl. Catal. B 39 (2002) 221– 231. [13] S. Baig, F. Petitpain, Proceedings of the International Conference on Ozone, 7–9 April, Berlin, 2003. [14] V. Fontanier, S. Baig, J. Albet, J. Molinier, Proceedings of the International Conference ‘‘Advances in Science and Engineering for Industrial Applications of Ozone and Related Oxidants’’, 10–12 March, Barcelona, Espagne, 2004, II.2.6-1–II.2.6-7. [15] V. Fontanier, V. Farines, J. Albet, S. Baig, J. Molinier, Ozone Sci. Eng. 27 (2005) 115–128. [16] P. Barratt, A. Baumgartl, N. Hannay, M. Vetter, F. Xiong, Water Sci. Technol. 35 (4) (1997) 347–352. [17] P. Barratt, F. Xiong, Proceedings of the 12th World Congress of the International Ozone Association, 15–18 May, 1995, 419–437. [18] J.P. Kaptijn, Ozone Sci. Eng. 19 (1997) 297–305. [19] J.P. Kaptijn, M.F.C. Plugge, J.H.J. Annee, Proceedings of the 12th World Congress of the International Ozone Association, 15–18 May, Lille, France, 1995. [20] M. Ernst, F. Lurot, J.C. Schrotter, Appl. Catal. B 47 (2004) 15–25. [21] N. Karpel Vel Leitner, F. Delanoë, B. Acedo, F. Papillaut, B. Legube, Proceedings of International Regional Conference, 23–25 September, Poitiers, France, 1998. [22] F. Delanoë, B. Acedo, N. Karpel Vel Leitner, B. Legube, Appl. Catal. B 29 (2001) 315–325. [23] H. Bader, J. Hoigné, Determination of ozone in water by Indigo method, Ozone Sci. Eng. 4 (1981) 169–176. [24] A.G. Gonçalves, J.L. Figueiredo, J.J.M. Orfão, M.F.R. Pereira, Carbon 48 (2010) 4369–4381. [25] F. Pontlevoy, A. Audirac, N. Karpel Vel Leitner, J.C. Vasse, F. Deliane, Patent PCT/ FR2013/051921.