JANUARY-MARCH 2016 Vol.22, Number 1, 1-126

ISSN 1451 - 9372(Print) ISSN 2217 - 7434(Online) JANUARY-MARCH 2016 Vol.22, Number 1, 1-126

www.ache.org.rs/ciceq Journal of the Association of Chemical Engineers of Serbia, Belgrade, Serbia

Vol. 22 Belgrade, January-March 2016 No. 1

Chemical Industry & Chemical Engineering CONTENTS Quarterly (ISSN 1451-9372) is published quarterly by the Association of Chemical Shenghua Zhu, Yonghui Bai, Lunjing Yan, Qiaoling Hao Engineers of Serbia, Kneza Miloša 9/I, 11000 Belgrade, Serbia Fan Li, Characteristics and synergistic effects of co- pyrolysis of Yining coal and poplar sawdust ...... 1 Editor: Jasmina Gubić, Jelena Tomić, Aleksandra Torbica, Mirela Vlada B. Veljković Iličić, Tatjana Tasić, Ljubiša Šarić, Sanja Popović, [email protected] Characterization of several milk proteins in domestic balkan donkey breed during lactation, using lab-on-a- Editorial Office: -chip capillary electrophoresis ...... 9 Kneza Miloša 9/I, 11000 Belgrade, Serbia Phone/Fax: +381 (0)11 3240 018 Seyed Ali Alavi Fazel, Goharshad Hosseyni, Experimental E-mail: [email protected] investigation on partial pool boiling heat transfer in www.ache.org.rs pure liquids ...... 17 Marija Ilić, Franz-Hubert Haegel, Vesna Pavelkić, Dragan For publisher: Tatijana Duduković Zlatanović, Snežana Nikolić-Mandić, Aleksandar Lolić, Zoran Nedić, The Influence of alkyl polyglucosides Secretary of the Editorial Office: (and highly ethoxylated alcohol boosters) on the phase Slavica Desnica behavior of a water/toluene/technical alkyl polyethoxylate microemulsion system ...... 27 Marketing and advertising: V. Sangeetha, V. Sivakumar, biogas production from AChE Marketing Office synthetic sago wastewater by anaerobic digestion: Kneza Miloša 9/I, 11000 Belgrade, Serbia Phone/Fax: +381 (0)11 3240 018 optimization and treatment ...... 33 Muhammad Imran Ahmad, Muhammad Sajjad, Irfan Ahmed Publication of this Journal is supported by the Khan, Amina Durrani, Ali Ahmed Durrani, Saeed Gul, Ministry of Education and Science of the Republic of Serbia Asmat Ullah, Sustainable production of blended cement in Pakistan through addition of natural Subscription and advertisements make payable pozzolana ...... 41 to the account of the Association of Chemical Yuehao Luo, Robert Smith, Lork Green, Exploring instant- Engineers of Serbia, Belgrade, No. 205-2172- aneous micro-imprinting technology on semi-cured 71, Komercijalna banka a.d., Beograd epoxy resin coating based on relationship between forming precision and curing degree ...... 47 Computer typeface and paging: Vladimir Panić Zorana Boltić, Mića Jovanović, Slobodan Petrović, Vojislav Božanić, Marina Mihajlović, Continuous improvement Printed by: concepts as a link between quality assurance and Faculty of Technology and Metallurgy, implementation of cleaner production – Case study in Research and Development Centre of Printing the generic pharmaceutical industry ...... 55 Technology, Karnegijeva 4, P. O. Box 3503, 11120 Belgrade, Serbia Aleksandar Golubović, Ivana Veljković, Maja Šćepanović, Mirjana Grujić-Brojčin, Nataša Tomić, Dušan Mijin, Abstracting/Indexing: Biljana Babić, Influence of some sol-gel synthesis Articles published in this Journal are indexed in parameters of mesoporous TiO2 on photocatalytic Thompson Reuters products: Science Citation degradation of pollutants ...... 65 TM Index - Expanded - access via Web of Mehdi Asadollahzadeh, Shahrokh Shahhosseini, Meisam Science®, part of ISI Web of KnowledgeSM Torab-Mostaedi, Ahad Ghaemi, The effects of operating parameters on stage efficiency in an oldshue- rushton column ...... 75

CONTENTS Continued Xiaolei Li, Chunying Zhu, Gas–liquid mass transfer with instantaneous chemical reaction in a slurry bubble column containing fine reactant particles ...... 85 Jelena Popović, Goran Radenković, Jovanka Gašić, Sla- voljub Živković, Aleksandar Mitić, Marija Nikolić, Rado- mir Barac, The examination of sensitivity to corrosion of nickel-titanium and stainless steel endodontic instruments in tooth root canal irrigating solutions ...... 95 O.S. Glavaški, S.D. Petrović, V.N. Rajaković-Ognjanović, T.M. Zeremski, A.M. Dugandžić, D.Ž. Mijin, Photodeg- radation of dimethenamid-P in deionised and ground water ...... 101 Sonja V. Smiljanić, Snežana R. Grujić, Mihajlo B. Tošić, Vladimir D. Živanović, Srđan D. Matijašević, Jelena D. Nikolić, Vladimir S. Topalović, Effect of La2O3 on the structure and the properties of strontium borate glasses ...... 111 Marija Šljivić-Ivanović, Aleksandra Milenković, Mihajlo Jović, Slavko Dimović, Ana Mraković, Ivana Smičiklas, NI(II) immobilization by bio-apatite materials: Appraisal of chemical, thermal and combined treatments ...... 117

Activities of the Association of Chemical Engineers of Serbia are supported by:

- Ministry of Education, Science and Technological Development, Republic of Serbia - Hemofarm Koncern AD, Vršac, Serbia - Faculty of Technology and Metallurgy, University of Belgrade, Belgrade, Serbia - Faculty of Technology, University of Novi Sad, Novi Sad, Serbia - Faculty of Technology, University of Niš, Leskovac, Serbia - Institute of Chemistry, Technology and Metallurgy, University of Belgrade, Belgrade, Serbia

Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 22 (1) 1−8 (2016) CI&CEQ

SHENGHUA ZHU CHARACTERISTICS AND SYNERGISTIC YONGHUI BAI EFFECTS OF CO-PYROLYSIS OF YINING LUNJING YAN COAL AND POPLAR SAWDUST QIAOLING HAO

FAN LI Article Highlights State Key Laboratory Breeding • Co-pyrolysis characteristics of a Chinese coal and poplar sawdust were studied • Base of Coal Science and Gas product yields of co-pyrolysis shows notable increase than that of separate pyrolysis Technology Co-founded by Shanxi • The synergistic effect is contributed mainly by the ash in the poplar sawdust Province and the Ministry of Science and Technology, Taiyuan Abstract University of Technology, Taiyuan, Co-processing of biomass and coal is perceived as a way to enhance the China energy utilization by virtue of the integrated and interactive effects between SCIENTIFIC PAPER different types of carbonaceous fuels. The purpose of this study was to investigate the co-pyrolysis characteristics of Yining coal and poplar sawdust, and to UDC 66.092-977-922(510):622.33 determine whether there is any synergistic effect in pyrolytic product yields. The coal was blended with sawdust at mass ratios of 9:1, 7:3, 5:5, 3:7 and 1:9. DOI 10.2298/CICEQ141125012Z The change of char yields, maximum weight loss rate and the corresponding temperature of different coal/sawdust blends during pyrolysis were compared by thermogravimetric analysis (TG). The total tar yields during pyrolysis of separate coal and sawdust, as well as their blends, were acquired from the low temperature aluminum retort distillation test. From the comparison of experimental and theoretical values of the char yields from TG and tar yields from carbonization test, it was observed that co-pyrolysis of coal/sawdust blends produced less char and tar than the total amount produced by separate coal and sawdust pyrolysis. The different product distribution suggested that there was synergy effect in gas product yields. The co-pyrolysis of demineralized and devolatilized sawdust with coal indicated that the ash in the sawdust was the main contributor to the synergistic effect. Keywords: co-pyrolysis, poplar sawdust, aluminum retort carbonization, synergies.

To face the severe situation of fossil fuel supply which is of vital importance to gasification reactivity. shortage and environmental pollution, biomass energy Therefore, it is necessary to understand the interac- has become a global interest and seen wide appli- tions between coal and biomass, the change in pyro- cation in partly replacing coal in gasification process. lysis characteristic and possible synergistic effects of As a renewable energy source, rational utilization of coal–biomass blending [6-10]. biomass can effectively solve the problem of energy Although co-processing has become a widely shortage [1-5]. Co-pyrolysis, as the preliminary pro- accepted practice all over the world, the research cess of co-gasification, plays a crucial role in deter- results published about synergies still do not have mining gas product distributions and char structure, agreeable conclusions. Some research results showed that the syner- Correspondence: F. Li, State Key Laboratory Breeding Base of gistic effect is dependent on the extent of contact Coal Science and Technology Co-founded by Shanxi Province between fuel particles, and the synergy is more likely and the Ministry of Science and Technology, Taiyuan University to happen when pyrolysis is carried out on a fixed-bed of Technology, Taiyuan 030024, China. reactor than on a fluidized-bed or drop tube reactor E-mail: [email protected] Paper received: 25 November, 2014 [11,12], while other scholars [13] revealed the lack of Paper revised: 27 April, 2015 synergistic effects on pyrolytic products yields as well Paper accepted: 29 April, 2015

1 S. ZHU et al.: CHARACTERISTICS AND SYNERGISTIC EFFECTS OF… Chem. Ind. Chem. Eng. Q. 22 (1) 1−8 (2016) as gas composition from pyrolysis of coal/sawdust Crushed samples were sieved and only particles blends under both low heating rate in a fixed-bed between the sizes of 0.074–0.154 mm were used to reactor and high heating rate in a drop-tube reactor. run the TG and aluminum retort carbonization test. Kastanaki et al. also confirmed that the co-pyrolysis of The specific particle diameter is consistent with those coal/biomass blends did not have substantial inter- produced in practical milling systems used in pulver- action in the solid phase [14]. Masnadi et al. per- ization units [22]. In our previous study [23], the par- formed switch-grass and coal co-pyrolysis in a ther- ticle size less than 0.125 mm was found to be able to mogravimetric analysis and no significant interaction eliminate the effects of mass and heat transfer limit- between coal and biomass were observed during co- ations. The mixture of YN and SA (SA-YN) was made -pyrolysis [4]. On the contrary, some researchers [15- by mechanically blending them together in different -18] found an obvious synergy on the overall weight proportions, aSA-bYN means that the mass ratio of loss yield and characteristics of pyrolysis products SA to YN is a:b. The addition of biomass to coal has during the co-pyrolysis of coal and other feedstock, important influence on the pyrolysis product distri- including oil residue and biomass. Krerkkaiwan et al. bution. In order to comprehensively investigate co- researched co-pyrolysis of coal and biomass using a pyrolysis characteristics, five blending rates ranging drop tube fixed reactor. The results showed that the from pure coal to pure biomass in 20 wt.% increments biomass has a significant influence not only on the were chosen in this study. Moreover, the blending magnitude of the synergetic effect during the co-pyro- rates are similar to those typically used in industry- lysis but also on the reactivity of the resultant chars scale co-firing trails. The ultimate and proximate anal- [19]. However, the authors did not consider the inter- yses of the samples were shown in Table 1. The ulti- actions between the volatile and coal char, so the mate analysis of the coal was determined following results they got has important dependency on the vol- the Chinese National Standard GB/T 476-2008 for atile. Previous study shows that the volatile-char inter- carbon, hydrogen and nitrogen, and GB/T 214-2007 actions can affect almost every aspect of low-rank for sulfur [24,25]. The proximate analysis of the coal fuel gasification and pyrolysis [20]. Therefore, the was measured following the Chinese National Stan- synergies in co-pyrolysis of biomass and coal are still dard GB/T 212-2008 for moisture (Mad), ash (Ad) and not clear. Further and more detailed researches are volatile matter (Vdaf) [26]. needed. It can be seen that the SA had high content of Considering that the type of blending fuels was volatile and low content of ash, while YN had a a major factor that triggers the synergy [21], a typical relatively low content of volatile and high content of Chinese coal and biomass, which have striking differ- ash. ence in volatile content and ash content, were selected as the experimental sample. In this study, the Table 1. Ultimate and proximate analysis data of YN and SA pyrolysis characteristics and the char yields of the (wt.%); ad: air-dried basis; daf: dry and ash-free basis; *: by dif- coal, sawdust and their mixtures were investigated ference using TG analyzer, and low temperature aluminum Sample Ultimate analysis Proximate analysis retort distillation test were carried out to compare the Cad Had O*ad Nad St,ad Mad Vdaf Ad tar yields of coal, sawdust and their mixture. The syn- YN 58.99 3.28 18.37 0.96 0.36 12.61 33.806.21 ergy during co-pyrolysis was examined by comparing SA 49.01 2.69 40.50 0.72 0.01 6.69 82.940.41 the theoretical and experimental data. To further understand the conditions that lead to synergy, Preparation of DASA and DVSA demineralized sawdust (DASA) and devolatilized sawdust (DVSA) were prepared, and then added to First, the pretreatment of raw sawdust in acid coal to appraisal whether one of them in biomass is was conducted by the ratio of 1 g:20 mL of HCl sol- the main contributor to the synergy. ution (37 wt.% HCl was diluted in 1:1 proportion), soaking for 24 h at room temperature and stirring EXPERIMENTAL continuously using a magnetic stirrer. Then, the HCl- -washed sample was blended with hydrofluoric acid Preparation of raw materials (HF) at a ratio of 1 g to 12.5 mL to prepare the demineralized sawdust sample, soaking for 24 h at room Chinese bituminous coal from Yining (YN) and temperature and stirring continuously using a mag- poplar sawdust (SA) were used in this study. The YN netic stirrer. Finally, the demineralized sawdust sample was drawn and manually crushed using a pestle and was obtained after oven drying to constant weight at a mortar, while the SA was crushed using a grinder. 60 °C for 14 h. As to the preparation of DVSA, the

2 S. ZHU et al.: CHARACTERISTICS AND SYNERGISTIC EFFECTS OF… Chem. Ind. Chem. Eng. Q. 22 (1) 1−8 (2016) sawdust was first pyrolyzed in a high-temperature sil- Data processing methods icon carbide furnace from room temperature to 800 Theoretical calculation of co-pyrolysis yields of ° ° C with a heating rate of 10 C/min in argon atmo- tar and char sphere and had a residence time of 20 min, then the In order to investigate whether interactions sample was cooled to room temperature in Ar atmo- existed between the coal and sawdust, the theoretical sphere, the residual solid was the DVSA. and experimental value of pyrolysis products were Pyrolysis experiments in TG compared. The theoretical value was given by the following equation: The TG experiments were performed in the thermogravimetric analyzer (NETZSCH STA449F3), MT = xMSA + (1-x)MYN (1) the maximum temperature error of the measurement where M is the tar or char yield during single saw- ° SA is ±1 C and the mass precision is 1 μg. Approxim- dust pyrolysis or aluminum retort carbonization test, ately 10 mg initial sample was fed into the Al2O3 plate M is the tar or char yield during single coal pyrolysis ° YN and heated from room temperature to 1000 C at a or aluminum retort carbonization test, M is the theo- ° T constant heating rate of 10 C/min under argon atmo- retical value of char yield during co-pyrolysis or tar sphere at a constant flow rate of 50 mL/min. At least 3 yield during aluminum retort carbonization test, and x repetitions were conducted to ensure the reprodu- is the mass fraction of sawdust in solid feed mixture. cibility of the experiments and accuracy of the data. The maximum mass loss standard deviation was 3%. RESULTS AND DISCUSSION Test of low temperature distillation by aluminum retort TG analysis of the pyrolysis of coal, sawdust and To obtain the tar yields of coal, sawdust and coal/sawdust blends their mixture during pyrolysis, tests of low tempera- Figure 1 shows the weight and the derivative ture distillation by aluminum retort were performed weight change profiles for coal and sawdust as a according to China Standard GB/T 480-2010, “Test of function of temperature. TG/DTG curves of them at a low temperature distillation of coal by aluminum heating rate of 10 °C/min suggest that both the ther- retort” [27]. First, 20 g coal sample was packed in mal decomposition and mass loss of poplar sawdust aluminum retort. Then, in the temperature range of and Yining coal have three steps during pyrolysis, but 260–510 °C, it was heated with a heating rate of the temperature of the maximum degradation rate 5 °C/min and it was held at the final temperature for (T ) of each step are rather different. 20 min. After the distillation experiment, the tar, pyro- max The first stage of YN pyrolysis occurs between lysis water, char and gas yields were measured. All 33 and 250 °C, while sawdust pyrolysis occurs the experiments were replicated at least three times between 33 and 185 °C. A minor mass decay was to make sure that the results were reproducible; the observed, this is due to the release of H O and some maximum standard deviation in tar yield was 5%. 2

100

0 80

-2 60 DTG%/min /

SA 40 -4

TG / % YN

20 -6

-8 0 200 400 600 800 1000

Temperature / oC

Figure 1. TG/DTG curves of SA and YN at a heating rate of 10 °C/min.

3 S. ZHU et al.: CHARACTERISTICS AND SYNERGISTIC EFFECTS OF… Chem. Ind. Chem. Eng. Q. 22 (1) 1−8 (2016)

absorbing gases such as CH4, CO2 and N2. The from 0.96 to 7.22%/min with increasing SA content in greatest fraction of mass loss of sawdust occurs in the SA-YN, which means the SA-YN has higher pyro- the second step, in the temperature range of 250- lysis reactivity and the addition of SA could promote 445 °C, which is attributed to the drastic thermal the overall evolution rate of volatile matters. decomposition of SA, the maximum devolatilization Char and tar yields rate is 7.22%/min and the Tmax is 378 °C. The third stage covers a wide temperature range from 445 °C Figure 3 shows the theoretical and experimental to the final temperature and a slight weight loss was value of char yields of five different mixtures with the observed, which was associated with the degradation YN:SA rates spanning from 9:1 to 1:9 in 20 wt.% of heavier chemical structures in the SA matrix [28]. increments during co-pyrolysis. With increasing SA Figure 1 shows that the profile trend of YN is addition from 10 to 90%, the experimental and pre- similar to that of SA, but different in some pyrolysis dicted char yields all had a remarkable decrease, the former was from 32.58 to 5.73% and the latter was characteristic parameters, especially the Tmax and the maximum mass loss rate. A more detailed analysis of from 45.87 to 8.96%, which was caused by the increase in the absolute amount of high volatile SA the parameters (see Figure 2) suggests that the Tmax and the maximum mass loss rates of SA-YN with (82.49, daf). Moreover, it is obvious that the experi- different ratio, YN and SA varied from each other. The mental char yield of any kind YN-SA is always lower than the predicted value, the maximum and minimum Tmax for SA-YN was lower than either of the two pure samples and the maximum weight loss rate shifted differences are 13.28 and 3.23%, respectively, and the addition of SA have inhibiting effect on pyrolysis

7 Maximum weight loss rate 420 T max 6 400 5

380 C o

4 ,

max 3 360 T

2 340

Maximum weight loss rate, %/min rate, loss weight Maximum 1 320 YN 9:1 7:3 5:5 3:7 SA Proportion of coal to poplar sawdust in the blending sample Figure 2. Pyrolysis characteristic parameters of different coal/sawdust blending.

100 Difference Theoretical value of char yield 80 Experimental value of char yield

40 Char yields(%)

0 9:1 7:3 5:5 3:7 1:9 Proportion of coal to poplar sawdust in the blending sample Figure 3. Theoretical and experimental value for char yields of different coal/sawdust blending during co-pyrolysis.

4 S. ZHU et al.: CHARACTERISTICS AND SYNERGISTIC EFFECTS OF… Chem. Ind. Chem. Eng. Q. 22 (1) 1−8 (2016) char yield and promotion effect on gas product form- Synergy analysis ation. Pyrolysis behavior of SA, DASA and DVSA As is known, the gas yield was the summation of Considering co-pyrolysis of SA and YN could un-condensed light gases and condensed volatile have synergistic effects on gas product yields, the matter (coal tar) at room temperature and pressure. contribution of ash and volatile in the SA on syner- Therefore, to have a more detailed analysis of SA gistic effect was investigated. The pyrolysis behavior addition on tar yields, low temperature aluminum of DASA and DVSA were analyzed first. Figure 5 retort distillation tests of YN, SA, and YN-SA were shows the comparison of the TG and DTG curves of conducted respectively. Figure 4 shows 7YN-3SA and SA, DASA and DVSA. It suggests that the DVSA did 9YN-1SA as examples for interpretation of the differ- not have an obvious mass loss until the pyrolysis ence of experimental and predicted values of tar temperature surpassed 750 °C. The final mass decay yields during distillation test. The tar yield of SA alone accounts for 39% of DVSA, far below the final value pyrolysis is much higher than that of YN pyrolysis, the of SA (92%). It reveals the differences of pyrolysis experimental values of tar during 7YN-3SA and 9YN- behavior of SA and DASA, the onset and final volatile -1SA distillation are 6.41 and 3.26%, and 0.43 and evolution temperature of DASA was shifted to the 1.38% lower than the predicted values, which means higher and lower temperature, respectively, the vola- that SA has an inhibiting effect on tar yields during co- tile release rate of DASA was much higher than that pyrolysis, i.e., co-pyrolysis of SA and YN could have of SA, and the release time of volatile was shortened. synergistic effects on gas product yields.

Tar yields / % / Tar yields 4

0 0246T YN SA M7:3 T7:3 M9:1 9:1 Figure 4. Theoretical and experimental value for tar yields of different samples during low temperature aluminum retort distillation (where T7:3 and M7:3 are the experimental and theoretical value of 7YN-3SA).

100

DASA 60 TG

SATG DVSA 40 TG TG/% 20

0 0

-4 DASADTG SA -8 DTG DVSA DTG -12 DTG/%/min -16

0 200 400 600 800 1000 o Temperature/ C Figure 5. TG and DTG profiles comparisons between SA, DVSA and DASA.

5 S. ZHU et al.: CHARACTERISTICS AND SYNERGISTIC EFFECTS OF… Chem. Ind. Chem. Eng. Q. 22 (1) 1−8 (2016)

It may suggest that the DASA has a larger surface °C, which, however, was not observed in the DTG area and more active sites, which could accelerate curve of 3SA-7YN. Table 2 illustrates the experimen- the releasing of volatile. In addition, the alkali/alkaline tal and predicted values of the final mass loss fraction earth metals in the SA have important influence on of 3SA-7YN, 3DVSA-7YN and 3DASA-7YN at term- the pyrolysis reactivity, which could make the pyro- inal pyrolysis temperature of 1000 °C. Since the expe- lysis reaction happen at lower temperature [28,29]. As rimental values are higher than the respective theo- shown in Figure 5, the maximum volatile release rate retical ones, it may be concluded that the synergy of DASA was 16%/min, while the max volatile release effect between YN and SA on gas product yields was rate of SA was 7%/min. observed. Compared with 3DVSA-7YN and 3DASA- Co-pyrolysis behavior of YN, DASA and DVSA -7YN, it is obvious that the ash in the SA could widen the difference between experimental and theoretical Figure 6 gives the comparison of co-pyrolysis value, suggesting that the ash in the SA was the main behaviors of YN, DASA and DVSA. The whole pyro- contributor to the synergy between SA and YN. The lysis process could be divided into three stages, the demineralized coal and coal-containing fuel blends first stage occurred between 26 and 200 °C, a faster along with the demineralized and devolatilized bio- H O release rate and larger amount of 3SA-7YN was 2 mass components will be tested in the future. observed than that of 3DASA-7YN due to that DASA was dried for 24 h after it was prepared. The second Table 2. Experimental and predicted values (%) of the final stage was in the temperature range of 200–400 °C mass loss fraction of 3SA-7YN, 3DVSA-7YN and 3DASA-7YN and the strongest weight loss observed was attributed under terminal pyrolysis temperature of 1000 °C to the active thermal decomposition of the raw feeds, Value 3SA-7YN 3DVSA-7YN 3DASA-7YN the maximum mass loss rate of 3DASA-7YN was 2.68%/min higher than that of 3SA-7YN. It may be Experimental 41.2 56.8 40.1 that the ash in the SA had obvious effects on the Predicted 26.1 40.4 33.8 synergy. The third stage appeared between 400–1000 °C. As it was also shown in Figure 1, the temperature CONCLUSIONS range of 400–600 °C was the most drastic mass loss In this research, experiments were conducted by interval of YN, but the DTG curve of 3SA-7YN in this TGA and low temperature aluminum retort distillation temperature range only had slight difference com- to study the changes in pyrolysis characteristic pared with that of 3DASA-7YN. parameters and possible synergistic effects of coal– Compared to the co-pyrolysis curves of 3SA- biomass blending during pyrolysis. The following con- 7YN and 3DVSA-7YN, it is clear that the pyrolysis clusions can be drawn from the results: behaviors of them are quite different. The DTG profile The degradation of poplar sawdust and Yining of 3DVSA-7YN had a remarkable mass loss peak coal had three stages in pyrolysis temperature rang- when the pyrolysis temperature was higher than 680 ing from room temperature to 1000 °C and the tem-

100

3SA-7YNTG

40 3DASA-7YNTG TG/% 3DVSA-7YNTG 20

0 0

-2 3SA-7YN DTG 3DASA-7YN -4 DTG 3DVSA-7YN

DTG/%/min DTG

-6 0 200 400 600 800 1000 o Temperature/ C Figure 6. Co-pyrolysis behavior of 3SA-7YN, 3DASA-7YN and 3DVSA-7YN.

6 S. ZHU et al.: CHARACTERISTICS AND SYNERGISTIC EFFECTS OF… Chem. Ind. Chem. Eng. Q. 22 (1) 1−8 (2016) peratures of the maximum weight loss rate of each [8] T. Sonobe, N. Worasuwannarak, S. Pipatmanomai, Fuel Process. Technol. 89 (2008) 1371-1378 stage were rather different. The Tmax for SA-YN was lower than either of the two pure samples and the [9] S.G. Sahu, P. Sarkar, N. Chakraborty, A.K. Adak, Fuel maximum weight loss rate shifted from 0.96 to Process. Technol. 91 (2010) 369-378 7.22%/min with the increase of the SA content in the [10] Y. Ninomiya, L. Zhang, T. Sakano, C. Kanaoka, M. SA-YN. It indicated that the SA-YN had higher pyro- Masui, Fuel 83 (2004) 751-764 lysis reactivity and the addition of SA could promote [11] A.G. Collot, Y. Zhuo, D.R. Dugwell, R. Kandiyoti, Fuel 78 (1999) 667–679 the overall evolution rate of the volatile matters. [12] J.M. Jones, M. Kubacki, K. Kubica, A.B. Ross, A. The addition of SA had inhibiting effect on char Williams, J. Anal. Appl. Pyrolysis 74 (2005) 502–511 and tar yields during pyrolysis and supportive effect [13] C. Meesri, B. Moghtaderi, Biomass Bioenergy 23 (2002) on gas product formation. Co-pyrolysis of SA and YN 55-66 had synergistic effects on gas product yields. [14] E. Kastanaki, D. Vamvuka, P. Grammelis, E. Kakaras, The volatile matter during SA pyrolysis only had Fuel Process. Technol. (2002) 77-78 slight influence on synergy, while the ash in the SA [15] I. Suelves, M.J. Lazaro, R. Moliner, J. Anal. Appl. Pyro- was the critical factor that led to the synergistic effects lysis 65 (2002) 197–206 on gas product yields. [16] L. Zhang, S.P. Xu, W. Zhao, S.Q. Liu, Fuel 86 (2007) Acknowledgements 353–359 [17] H. Haykiri-Acma, S. Yaman, Fuel 86 (2007) 373–380 The authors gratefully acknowledge the financial [18] J.X. Fei, J. Zhang, F.C. Wang, J. Wang, J. Anal. Appl. support from Shanxi Coal Based Key Scientific and Pyrolysis 95 (2012) 61-67 Technological Project (No. MH2014-02) and National [19] S. Krerkkaiwan, C. Fushimi, A. Tsutsumi, P. Natural Science Foundation (21176166). Kuchonthara, Fuel Process. Technol. 115 (2013) 11-18 [20] C.Z. Li, Fuel 112 (2013) 609-623 REFERENCES [21] D. Vamvuka, S.S. Troulino, E. Kastanaki, Fuel 85 (2006) 1763-1771 [1] S.W. Park, C.H. Jang, Energy 39 (2012) 187-195 [22] J.A. Kostamo, in: Sixteenth Annual International Pitts- [2] K. Srirangan, L. Akawi, M. Moo-Young, C.P. Chou, Appl. burgh Coal Conference. Pittsburgh, PA, USA, 1999 Energy 100 (2012) 172-186 [23] Y.H. Bai, Y.L. Wang, S.H. Zhu, L.J Yan, F. Li, K.C. Xie, [3] N. Kythreotou, S.A. Tassou, G. Florides, Energy 47 Fuel 126 (2014) 1-7 (2012) 253-261 [24] GB/T 476-2008. Standards Press of China, Beijing (2008) [4] M.S. Masnadi, R. Habibi, J. Kopyscinski, J. M. Hill, X. Bi, pp. 1-16 (in Chinese) C.J. Lim, N. Ellis, .J.R. Grace, Fuel 117 (2014) 1204- [25] GB/T 214-2007. Standards Press of China, Beijing (2007) –1214 pp. 1-10 (in Chinese) [5] R.Habibi, J.Kopyscinski, M.S. Masnadi, J.Lam, J.R. [26] GB/T 212-2008. Standards Press of China, Beijing (2008) Grace, C.A. Mims, and J.M. Hill, Energy Fuels 27 (2012) pp. 1-14 (in Chinese) 494-500

[6] T. Wu, M. Gong, E. Lester, P. Hall, Fuel 104 (2013) 194- [27] GB/T 480-2010. Standards Press of China, Beijing (2010) pp. 1-13 (in Chinese) -200 [7] K.D. Wang, Z.H.K.D. Wan, Z.H. Wang, Y. He, J. Xia, Z.J. [28] N. Tsubouchi, C.B. Xu, Y. Ohtsuka, Fuel Process. Technol. 85 (2004) 1039-1052 Zhou, J.H. Zhou, K.F. Cen, Fuel 139 (2015) 356-364 [29] X.W. Zou, J.Z. Yao, X.M. Yang, W.L. Song, W.G. Lin, Energy Fuels 21 (2007) 619-624.

7 S. ZHU et al.: CHARACTERISTICS AND SYNERGISTIC EFFECTS OF… Chem. Ind. Chem. Eng. Q. 22 (1) 1−8 (2016)

SHENGHUA ZHU KARAKTERISTIKE I SINERGIJSKI EFEKAT YONGHUI BAI PIROLIZE UGLJA YINING I PILJEVINE TOPOLE LUNJING YAN QIAOLING HAO Procesiranje biomase i uglja predstavlja način za poboljšanje korišćenja energije na FAN LI osnovu integrisanih i interaktivnih efekata između različitih vrsta ugljeničnih goriva. Cilj State Key Laboratory Breeding ovog istraživanja je analiza pirolitičkih karakteristika uglja iz Jininga (Yining, Kina) i pilje- Base of Coal Science and vine topole, kao i da se utvrdi postojanje sinergiijskog efekta u prinosima pirolitičkih proiz- Technology Co-founded by Shanxi voda. Ugalj je pomešan sa piljevinom u masenim odnosima 9:1, 7:3, 5:5, 3:7 i 1:9. Pro- Province and the Ministry of mena prinosa čađi, maksimalna brzina gubitka mase i odgovarajuća temperatura različitih Science and Technology, Taiyuan mešavina ugalj/piljevina tokom pirolize su upoređeni termogravimetrijskom analizom (TG). University of Technology, Taiyuan, Destilacijom u aluminijumskoj retorti na niskoj temperaturi određen isu ukupni prinosi China katrana za vreme pirolize uglja, piljevine i njihovih mešavina. Poređenjem eksperimentalnih i teorijskih vrednosti prinosa čađi iz TG analize i prinosa katrana iz ispitivanja karbo- NAUČNI RAD nizacije uočeno je da ko-piroliza mešavine ugalj/piljevina proizvodi manje čađi i katrana od ukupnih količina proizvedeni u odvojenim procesima pirolize uglja i piljevine. Različita dis- tribucija proizvoda ukazuje na sinergistički efekat u prinosu gasovitih proizvoda. Piroliza demineralizovane i devolatilizovane piljevine sa ugljem pokazuje da pepeo iz piljevine daje najveći doprinos sinergističkom efektu.

Ključne reči: Piroliza, piljevina topole, karbonizacija u aluminjimuskoj retorti, si- nergija.

8 Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 22 (1) 9−15 (2016) CI&CEQ

JASMINA GUBIĆ1 CHARACTERIZATION OF SEVERAL MILK JELENA TOMIĆ1 1 PROTEINS IN DOMESTIC BALKAN DONKEY ALEKSANDRA TORBICA BREED DURING LACTATION, USING LAB-ON- MIRELA ILIČIĆ2 TATJANA TASIĆ1 -A-CHIP CAPILLARY ELECTROPHORESIS LJUBIŠA ŠARIĆ1 1 Article Highlights SANJA POPOVIĆ • Protein profile of domestic Balkan donkey milk during lactation period to the 280th day 1Institute of Food Technology, were determined • Donkey’s milk protein profiles were determined by applying lab-on-a-chip electropho- University of Novi Sad, Novi Sad, resis Serbia • 2 Domestic Balkan donkey milk was found to be low in casein content Faculty of Technology, University • Lysozyme, lactoferrin and immunoglobulins were identified of Novi Sad, Novi Sad, Serbia • Balkan donkey milk represents a rich source of high nutritive components

SCIENTIFIC PAPER Abstract Domestic Balkan donkey (Equus asinus asinus) is a native donkey breed, pri- UDC 637.12’618(497):543.545.2 marily found in the northern and eastern regions of Serbia. The objective of the DOI 10.2298/CICEQ150105013G study was to analyze proteins of Domestic Balkan donkey milk during the lactation period (from the 45th to the 280th day) by applying lab-on-a-chip electrophoresis. The chip-based separations were performed on the Agilent 2100 Bioanalyzer in combination with the Protein 80 Plus lab chip kit. The protein content of domestic Balkan donkey milk during the lactation period of 280 days ranged from 1.40 to 1.92% and the content of αs1-casein, αs2-casein, β-casein, α-, β-lactoglobulin, lysozyme, lactoferrin and serum albumin was relatively quantified. Lysozyme (1040-2970 mg/L), α-lactalbumin 12 kDa (1990-2730 mg/L) and α-lactalbumin 17.7 kDa (2240-3090 mg/L) were found to be the proteins with the highest relative concentrations. Keywords: donkey milk, protein, lab-on-a-chip electrophoresis.

Over the past decades donkey milk has been human milk, but scarce information is available reg- less studied compared to ruminant milk, but in the last arding the use of donkey milk for this purpose. few years, interest in donkey milk has considerably Domestic Balkan donkey is a native breed, pri- increased among the scientific community of Europe. marily found in the northern and eastern regions of Donkey milk has been successfully used in clinical Serbia, with about 1000 subjects reared [6,7]. The studies, with children who suffer from cow’s milk pro- population of this breed is nowadays reduced to a tein allergy (CMPA), and has good palatability [1,2]. very low number. Therefore, it is very important to Its composition is more similar to human milk than preserve the breed and to increase the number of ruminant milk, however. It has a relatively low lipid animals, in order to achieve milk production in signi- content and adequate lipid integration is needed for ficant amounts. FAO – the organization of food and toddlers’ diet [3,4]. Other types of milk, such as agriculture – has initiated and recommended activities mare’s [5], have been proposed as a substitute for for the mentioned breed protection. Specific milk characteristics and parameters are the effect of the keeping conditions and pasture feeding, climate, as Correspondence: Ja. Gubić, Institute of Food Technology, well as the race [8]. Donkey milk has been tradition- University of Novi Sad, Bulevar cara Lazara 1, 21000 Novi Sad, ally used in Serbia as a natural remedy for the treat- Serbia. ment of asthma and bronchitis. Considering this fact, E-mail: [email protected] Paper received: 5 January, 2015 there has recently been a growing demand for donkey Paper revised: 29 April, 2015 milk in the Serbian market [9]. Paper accepted: 8 May, 2015

9 J. GUBIĆ et al.: CHARACTERIZATION OF SEVERAL MILK PROTEINS… Chem. Ind. Chem. Eng. Q. 22 (1) 9−15 (2016)

Donkey milk has a lower protein content than Polidori et al. [11] analyzed donkey milk proteins other ruminant milk ranging from 13 to 28 g/L, while using two-dimensional electrophoresis (2-DE) fol- proteomic profile is quite similar to human milk lowed by N-terminal analysis and found and deter- [10,11]. Protein content varies considerably among mined β-caseins with molecular weights ranging from species and is influenced by breed, stage of lactation, 33.10 to 33.74 kDa and from 31.15 to 32.15 kDa and feeding, climate, parity, season, and udder health lower. status [12]. The content of casein in donkey milk Since literature data on use of donkey milk in ranges from 6.4 to 10.3 g/kg of total protein content. human nutrition and its changes during lactation is Generally, casein present in different types of milk very limited, the main objective of this study was to consists of four genetic fractions: αs1-, αs2-, β- and characterize several of the proteins of domestic Bal- k-casein [13]. Guo et al. [14] reported that the content kan donkey milk during the lactation period from the of whey proteins in donkey milk is within the range 45th to the 280th day. Moreover, the aim of this study from 4.9 to 8.0 g/kg of total protein. According to the was to evaluate the nutritional value of domestic Bal- research by Cunsolo et al. [15], considerable differ- kan donkey milk from the protein point of view, by ences can exist between the primary structure of don- applying lab-on-a-chip electrophoresis. Deep know- key and bovine αs1-casein, which could be related to ledge of the protein composition and variability could the previously demonstrated low allergenic properties be beneficial for a more appropriate use in infant of donkey milk and could contribute to its better feeding. human tolerance. The basic whey proteins in donkey milk are MATERIALS AND METHODS β-lactoglobulin, α-lactalbumin, immunoglobulins, blood serum albumins, lactoferrin and lysozyme [16,17]. Sample collection The β-lactoglobulin is present in donkey milk as a The research on domestic Balkan donkeys, a monomer whereas this protein is a dimer in ruminant native breed, was conducted in the Special Nature milk [11] and has better digestibility in newborns due Reserve Zasavica [23]. Zasavica is located in the to higher digestibility and absorption of soluble mono- north-west region of Serbia and is currently home for mer proteins [18,19]. Furthermore, there is a possi- a herd of more than 150 female donkeys. Donkey milk bility of utilization of low-cost protein in formulations samples were individually collected from 10 female for infant feeding [20]. domestic Balkan donkeys, after parturition from spring Donkey milk contains several antimicrobial com- (April) to winter (January) season, on the 45th, 60th, ponents, including lactoferrin, lactoperoxidase and 80th, 100th, 125th, 150th, 170th, 200th, 230th and 280th lysozyme [14,21]. Šarić et al. [9] investigated the anti- day of lactation. From June to early October the ani- bacterial properties and the protein profile of raw milk mals in the grassland of Zasavica reservation were from the native Serbian donkey breed with an empha- reared outdoors on pasture, where they had the pos- sis on the lysozymes and lactoferrin. The average sibility to consume meadow plants. During other lysozyme content of 1.0 mg/mL determined is consi- months of the year, donkeys were reared indoors, in a derably higher compared to the milk of other species covered area, and they were fed with corn and corn [13,16]. stalks, while hay was available ad libitum. From April SDS-PAGE (sodium dodecyl sulphate polyacryl- to June the way of feeding changed substantially. amide gel) electrophoresis analysis is a commonly Corn and fresh water clover were given to the animals used method for protein separation, which is also before milking. Donkeys had access to water ad libi- widely applied to donkey milk analysis. Salimei et al. tum. [13] and Guo et al. [14] have, using this method, con- The animals were manually milked twice a day, cluded that whey proteins such as β-lactoglobulin, at 7:00 am and 3:00 pm, 120 min after separating lysozyme and α-lactalbumin are the most abundant in foals from their mothers. During milking, foals rem- donkey milk originating from Italian breeds, and lacto- ained in visual and tactile contact with their mothers. ferrin, serum albumin and immunoglobulins were Milk was completely removed from both udders. Each found to be minor protein components. individual raw milk sample was collected into glass Criscione et al. [22] used mass spectrometry flasks and stored in an ice box at 4 °C. For each day and high performance liquid chromatography to char- of sampling, which is 10 days in total, 10 individual acterize IEF patterns. The authors reported the abs- samples were collected twice (morning and evening) ence of αs1-casein in some individual cases and the for a total of 200 samples. presence of αs2-casein in all donkey milk samples.

10 J. GUBIĆ et al.: CHARACTERIZATION OF SEVERAL MILK PROTEINS… Chem. Ind. Chem. Eng. Q. 22 (1) 9−15 (2016)

Protein determination RESULTS AND DISCUSSION Total protein concentration was measured Protein content of domestic Balkan donkey milk through nitrogen determination. Total nitrogen was during the lactation period of 280 days is shown in determined by the application of ISO standard Table 1. method [24]. A nitrogen conversion factor of 6.38 was The protein content reached the highest value of used for the calculation of the protein content of milk 1.92% on the 60th day of the lactation stage. After- samples. wards, the concentration decreased until the end of Electrophoretic analysis the lactation period when it reached the value of The proteins of donkey milk were separated and 1.40%. The protein content in Domestic Balkan don- quantitated using lab-on-a-chip electrophoresis tech- key milk is in agreement with others studies on Italian nique based on their molecular mass in comparison donkey breeds – Martina Franca, Ragusana and with the marker protein ladder [25]. Sample prepar- Amiata [4,13,28,29]. ation was carried out according to Tidona et al. [26] Figure 1a and b shows the molecular weight (in with minor modifications. Milk samples were diluted in kDa) of the bands present at the beginning and at the 1:1.5 (V/V) ratio, sample: buffer (0.125 M Tris-HCl, end of the lactation period. The proteins were deter- 4% SDS, 2% glycerol, 2% β-mercaptoethanol, pH mined based on the literature data [11,21,22] by com- 6.8) and heated at 100 °C for 5 min. parison of molecular weights and relative concentra- The chip-based separations were performed tions. using Agilent 2100 bioanalyzer (Agilent Technologies, The bands of basic casein proteins that have ∼ Santa Clara, CA, USA) in combination with the Pro- been discovered have molecular weights 30.3 kDa α ∼ α tein 80 Plus lab chip kit and the dedicated Protein 80 ( s1-casein) and 26.7–27.0 kDa ( s2-casein). The software assay on 2100 expert software. Chips were findings showed two bands with 16-16.7 and 34.5- β prepared according to the protocol provided by the –35.0 kDa for -casein in donkey milk. The results of Protein 80 lab chip kit. The Protein 80 ladder (1.6, 3.5, electrophoresis showed a pattern similar to that rep- 6.5, 15, 28, 46, 63 and 95 kDa) and the internal orted in the literature [11,13,16]. The chip-based sep- markers were used as reference for sizing and rel- aration profiles of soluble proteins of donkey milk α ative quantification. According to Živančev et al. [27] quantified -lactalbumin with approximate molecular β values of LOD and LOQ for the proteins in the anal- weight of 12 and 17.7 kDa. -lactoglobulin and serum yzed solutions were 5.4 and 8.4 ng/μL, respectively. albumin molecular weights in whey protein fraction were around 19.6 and 66.0 kDa, respectively. Statistical analysis Major antimicrobial proteins determined in don- The one way ANOVA analysis and Duncan post key milk were immunoglobulin (Mr 37-38 kDa), lacto- hoc test were performed to assess data differences ferrin (Mr 74-78 kDa) and lysozyme (Mr 14.7-15.0 between various samples using Statistical software kDa). The results obtained in this research are similar version 12 (STAT SOFT Inc., 2013, USA). The data to those obtained by other authors [11,21,22]. means were considered to be significantly different at The concentrations of several protein fractions P < 0.05. in Balkan donkey milk during the lactation period are shown in Figure 2. The trend for αs1-casein content

Table 1. Protein content of domestic Balkan donkey milk (total n = 200) during the lactation period of 280 days; results are given as mean ± standard deviation;*p < 0.05; ns - not significant

Time of milk sampling, day Protein (nitrogen×6.38), % Significance level 45 1.83 ± 0.12 * 60 1.92 ± 0.20 * 80 1.73 ± 0.23 * 100 1.70 ± 0.28 * 125 1.64 ± 0.30 * 150 1.62 ± 0.19 * 170 1.49 ± 0.28 ns 200 1.50 ± 0.19 ns 230 1.45 ± 0.19 * 280 1.40 ± 0.20 *

11 J. GUBIĆ et al.: CHARACTERIZATION OF SEVERAL MILK PROTEINS… CI&CEQ 22 (1) 9−15 (2016)

(a)

(b)

Figure 1. a) Gel-like image of samples taken on the 45th and the 280th day; b) electropherogram of samples taken from the 45th until the280th day of the lactation stage. showed a high variability and ranged from 1160 to (P < 0.05). However, the β-casein content (P < 0.05) 730 mg/L, whereas αs2-casein content ranged from decreased significantly (84 to 13 mg/L) from the 110 to 74 mg/L. The content of αs1-casein began to beginning to the end of the lactation period. decline significantly from the 60th day until the 150th The content of α-lactalbumin was 2730 mg/L in day, compared to αs2-casein whose content started the early and 2240 mg/L in the late lactation stage, decreasing after the 150th day and continued until the which is very similar to the content found in human end of the lactation period. Trend variations for αs1- milk (2200 mg/L) [16]. The α-lactalbumin content casein content throughout the lactation period were showed a significant increase four months after par- related to the change in total protein content. During turition and reached values of 2450 to 3090 mg/L, the lactation period, the values of αs2-casein did not after which the content decreased and remained quite change significantly until the 150th day, and followed stable until the end of the lactation period. The con- the same pattern from the 170th to the 230th day centration of β-lactoglobulin varied from 139 to 263

12 J. GUBIĆ et al.: CHARACTERIZATION OF SEVERAL MILK PROTEINS… CI&CEQ 22 (1) 9−15 (2016)

Figure 2. Trend of the concentration (mg/L) of several proteins fractions in Balkan donkey milk during the lactation period.

mg/L. β-Lactoglobulin concentration decreased sig- growth of a large number of gram positive bacteria. nificantly (P < 0.05) after the 60th day and also after Šarić et al. [9] investigated the antibacterial properties the 150th day of lactation. Though β-lactoglobulin is and the protein profile of raw milk from the native generally resistant to gastro-intestinal enzymes, in a donkey Serbian breed with an emphasis on the lyso- simulated in vitro digestion of donkey milk 70% of the zyme and lactoferrin contents. Lysozyme and α-lac- β-lactoglobulin was digested, which is the amount talbumin showed high resistance to human gastric twice as high compared to the bovine counterpart and duodenal juices as already reported for raw [30,31]. Also, equine β-lactoglobulin was digested sig- equine, cow and human milk [30,31]. In our study, nificantly faster compared to bovine and caprine -lactalbumin was the dominant protein fraction of β-lactoglobulin [31]. The serum albumin content (113- donkey milk, while immunoglobulin and lactoferrin -238 mg/L) showed a tendency towards stabilization were minor components. The lactoferrin content (87– during the mid-lactation period. –13 mg/L) decreased significantly from the beginning High concentrations of lysozyme were quantified to the end of the lactation period, and showed similar in Balkan donkey milk, in which it ranged from 1040 to trend as β-casein. The immunoglobulin had an inc- 2970 mg/L. The lysozyme content was stable during reasing trend, reaching the climax on the 100th day different stages of lactation and significantly dec- (88.3 mg/L) and decreasing thereafter. The high con- reased (P < 0.05) after the 150th day. Donkey milk is tent of protective antimicrobial compounds in donkey known to be a rich source of lysozyme (1000 mg/L) milk taken from the early and middle lactation period [16] and has a significantly higher content of lyso- suggested its beneficial impact on gut health and zyme than human milk (400 mg/L) and bovine milk immune defense system. (130 mg/L), while being quite similar to that in equine milk (400-1000 mg/L) [29,32]. Lysozyme inhibits the

13 J. GUBIĆ et al.: CHARACTERIZATION OF SEVERAL MILK PROTEINS… Chem. Ind. Chem. Eng. Q. 22 (1) 9−15 (2016)

CONCLUSIONS [9] Lj. Šarić, B. Šarić, A. Mandić, A. Torbica, J. Tomić, D. Cvetković, Đ. Okanović, Int. Dairy J. 25 (2012) 142-146 Lab-on-a-chip capillary electrophoresis could be [10] E. D’Auria, C. Agostoni, M. Giovannini, E. Riva, R. Zetter- applied to the relative quantification of several milk ström, R. Fortin, G. F. Greppi, L. Bonizzi, P. Roncada, proteins. αs1-casein, αs2-casein, β-casein, α-lactal- Acta. Paediatr. 94 (2005) 1708–1713 bumin, β-lactoglobulin, lysozyme, lactoferrin and [11] P. Polidori, S. Vincenzetti, in Milk Protein, Ch. 8, InTech serum albumin were relatively quantified, with the Rijeka, 2012, p.p. 215-232 highest relative concentration of lysozyme and α-lac- [12] Y. Park, M. Juarez, M. Ramos, G. Heinlein, Small Rum talbumin. The concentration of all determined proteins Res. 68 (2007) 88-113 decreased during the lactation period. The minimum [13] E. Salimei, F. Fantuz, R. Coppola, B. Chiofalo, P. Poli- dori, G. Varisco, Anim. Res., A 53 (2004) 67–78 significant changes were observed for αs2-casein, [14] H.Y. Guo, K. Pang, X.Y. Zhang, L. Zhao, S.W. Chen, M.L. lysozyme and β-lactoglobulin. It can be concluded Dong, F. Z. Ren, J. Dairy Sci. 90 (2007) 1635-1643 that Balkan donkey milk represents a source of anti- [15] V. Cunsolo, E. Cairone, D. Fontanini, A. Criscione, V. bacterical proteins such as lysozyme and highly dig- Muccilli, R. Salettia, S. Fotia, J. Mass Spectrom. 44 α estible proteins such as whey protein, -lactalbumin (2009) 1742–1753 and lactoferrin. [16] S. Vincenzetti, P. Polidori, P. Mariani, N. Cammertoni, F. Acknowledgments M. Fantuz, A. Vita, Food Chem. 106 (2008) 640-649 [17] A. D’Alessandro, A. Scaloni, L. Zolla, J. Proteome Res. 9 This paper is part of the research work on the (2010) 3339-3373 project III46012 financed by the Ministry of Education, [18] B. Lönnerdal, Am. J. Clin. Nutr. 77 (2003) 1537-1543 Science and Technological Development of the Rep- [19] J. Barłowska, M. Szwajkowska, Z. Litwińczuk, J. Król, ublic of Serbia. We would like to thank Mr Slobodan Comp. Rev. Food Sci. Food Safety 10 (2011) 291-302 Simić and Nikola Nilić (Special Nature Reserve [20] M. Friedman, J. Agric. Food Chem. 44 (1996) 6-29 Zasavica, Serbia) for milk samples and great cooper- [21] F. Nazzaro, P. Orlando, F. Fratianni, R. Coppola, Open ation. Food Sci. J. 4 (2010) 43-47 [22] A. Criscione, V. Cunsolo, S. Bordonaro, A. M. Guastella, REFERENCES R. Saletti, A. Zuccaro, Int. Dairy J. 19 (2009) 190-197 [23] http://www.zasavica.org.rs/en/o-magarcima/ (accessed [1] G. Monti, E. Bertino, M.C. Muratore, A. Coscia, F. Cresi, 11 May 2011) L. Silvestro, C. Fabris, D. Fortunato, M. G. Giuffrida, A. [24] ISO Standard 8968-1:2001 (E) (2001) Conti, Pediatr Allergy Immunol. 18 (2007) 258-264 [25] S.G. Anema, Int. Dairy J. 19 (2009) 198-204 [2] I. Dello Iacono, M.G. Limongelli, Riv. Immunol. Allergol. Ped. 4 (2010)10-15 [26] F. Tidona, C. Sekse, A. Criscione, M. Jacobsen, S. Bor- donaro, D. Marletta. G.E. Vegarud, Int. Dairy J. 21 (2011) [3] E. D’Auria, M. Mandelli, P. Ballista, F. Di Dio, M. Gio- vannini, Case Rep. Ped. 2011 (2011) 1-4 158-165 [27] D.R. Živančev, B.G. Nikolovski, A.M. Torbica, J.S. [4] M. Martini, I. Altomonte, F. Salari, Ital J. Anim. Sci. 13 (2014) 123-126 Mastilović, N.H. Đukić, Chem. Ind. Chem. Eng. Q. 19 (2013) 553-561 [5] W.J. Park, H. Zhang, B. Zhang, L. Zhang, in Non-bovine [28] C. Giosuè, M. Alabiso, G. Russo, M.L. Alicata, C. Torrisi, Mammals, Y.W. Park, G.F.W. Heinlein, Eds., Blackwell Anim. 2 (2008) 1491-1495 Publishing Professional, Oxford, 2006, p. 275-296 [29] A. Alabiso, C. Giosuè, M.L. Alicata, F. Mazza, G. [6] FAO, DAD-IS Domestic Animal Diversity Information Sys- Iannolino, Anim. 3 (2008) 543-547 tem, http://www.fao.org/dad-is/ (accessed 25 February 2009) [30] R.A.E. Inglingstad, T.G. Devold, E.K. Eriksen, H. Holm, [7] W. Kugler, H.P. Grunenfelder, E. Broxham, Donkey M. Jacobsen, K.H. Liland, E.O. Rukke, G.E. Vegarud, Dairy Sci. Tech. 90 (2010) 549-560 breeds in Europe: inventory, description, need for action, conservation, report 2007/2008, Monitoring Institute for [31] F. Tidona, A. Criscione, T.G. Devold, S. Bordonaro, D.M. Rare Breeds and Seeds in Europe/SAVE Foundation: St Gerd, E. Vegarud, Int. Dairy J. 35 (2014) 57-62 Gallen, Switzerland. [32] R. Floris, I. Recio, B. Berkhout, S. Visser, Curr. Pharm. [8] S. Stojanović, PhD Thesis, University of Novi Sad, Novi Des. 9 (2003) 1257-127. Sad, 2012, p. 25 (in Serbian)

14 J. GUBIĆ et al.: CHARACTERIZATION OF SEVERAL MILK PROTEINS… Chem. Ind. Chem. Eng. Q. 22 (1) 9−15 (2016)

JASMINA GUBIĆ1 KARAKTERIZACIJA NEKOLIKO PROTEINA MLEKA JELENA TOMIĆ1 RASE DOMAĆI BALKANSKI MAGARAC TOKOM 1 ALEKSANDRA TORBICA LAKTACIJE, UPOTREBOM LAB-ON-A-CHIP MIRELA ILIČIĆ2 1 KAPILARNE ELEKTROFOREZE TATJANA TASIĆ LJUBIŠA ŠARIĆ1 Domaći balkanski magarac (Equus asinus asinus) je autohtona rasa, primarno nađena u SANJA POPOVIĆ1 severnom i istočnom regionu Srbije. Cilj rada je bio da se analiziraju proteini mleka 1Naučni institut za prehrambene domaćeg balkanskog magarca tokom laktacije primenom Lab-on-a-Chip elektroforeze. tehnologije, Univerzitet u Novom Sadu, Razdvajanje na čipu izvršeno je korišćenjem uređaja Agilent 2100 bioanalyzer u kombi- Bulevar cara Lazara 1, 21000 Novi naciji sa Protein 80 Plus Lab Chip kitom. Sadržaj proteina mleka domaćeg balkanskog Sad, Srbija magarca tokom laktacionog perioda od 280. dana kretao se od 1,40 do 1,92% i sadržaj 2 Tehnološki fakultet, Univerzitet u αs1-kazeina, αs2-kazeina, β-kazeina, α-laktalbumina, β-laktoglobulina, lizozima, lakto- Novom Sadu, Bulevar cara Lazara 1, ferina i serum albumina je relativno kvantifikovan. Lizozim (1040-2970 mg/L), α-laktal- 21000 Novi Sad, Srbija bumin 12 kDa (1990-2730 mg/L) i α-laktalbumin 17,7 kDa (2240-3090 mg/L) su proteini NAUČNI RAD koji su nađeni u relativno visokim koncentracijama. Ključne reči: magareće mleko, protein, lab-on-a-chip elektroforeza.

Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 22 (1) 17−26 (2016) CI&CEQ

SEYED ALI ALAVI FAZEL EXPERIMENTAL INVESTIGATION ON GOHARSHAD HOSSEYNI PARTIAL POOL BOILING HEAT TRANSFER IN Department of chemical PURE LIQUIDS engineering, college of chemistry and chemical engineering, Article Highlights Mahshahr Branch , Islamic Azad • Boiling heat transfer increases with increasing surface roughness University, Mahshahr, Iran • The bubble shape and oscillating characteristics determined the boiling heat transfer coefficient SCIENTIFIC PAPER • Eotvos and Roshko numbers are related to boiling heat transfer coefficient

UDC 66:544.355-145.13 Abstract Saturated partial pool boiling heat transfer on a horizontal rod heater was inves- DOI 10.2298/CICEQ150213014F tigated experimentally. The boiling liquids included water and ethanol. The heating section was made from various materials including SS316, copper, aluminum and brass. Experiments were performed at several degrees of surface roughness ranging between 30 and 360 µm average vertical deviation. The boiling heat transfer coefficient, bubble departing diameter and frequency, and nucleation site density were measured. The data have been compared to major existing correlations. It was found that experimental data do not match with major correlations in the entire range of experiments with acceptable accuracy. The boiling heat transfer area was divided in two complementary areas, the induced forced convection area and the boiling affected area. Based on two dimensionless groups, including Eötvös and Roshko numbers, a semi-empirical model is proposed for prediction of the boiling heat transfer coefficient. It is shown that the proposed model provides improved performance in prediction of the boiling heat transfer coefficient in comparison with to existing correlations. Keywords: induced force convection, pool boiling, surface roughness, heat transfer coefficient.

The nucleate pool boiling phenomenon is widely density. In addition, the structures of boiling heat applied in many engineering processes. The heat transfer surface are usually very complex and contain transfer mechanism from the surface to the boiling nucleation cavities with various shapes and sizes. fluid is known to be a very complicated phenomenon. This information is not completely available for every Design, operation and optimization of the involved given heating surface. equipment require precise prediction of the boiling In this investigation, the experimental data heat transfer coefficient. There has been a lot of covers a wide range of heating surfaces character- research on pool boiling over the past few decades. istics and liquids physical properties. Water and etha- However, the mechanism of pool boiling heat transfer nol have been selected as the boiling liquids. The is still not completely understood. This is because of cylindrical heating surfaces were made by various the intense complexity of three interconnected hetero- metals including SS316, copper, aluminum and brass. geneous parameters: 1) bubble departing diameter, Each surface has been sanded with several grades to 2) bubble departing frequency, and 3) nucleation site provide various degrees of roughness. Note that the roughness is defined as the arithmetic average of the Correspondence: S. Ali Alavi Fazel, Department of chemical vertical deviations of the surface. engineering, college of chemistry and chemical engineering, The experimental data have been compared to Mahshahr Branch , Islamic Azad University, Mahshahr, Iran. E-mail: [email protected] major existing correlations. It is shown that the exist- Paper received: 13 February, 2015 ing correlations cannot predict the boiling heat trans- Paper revised: 17 April, 2015 fer coefficient with a satisfactory accuracy. Some of Paper accepted: 8 May, 2015

17 S. ALI ALAVI FAZEL, G. HOSSEYNI: EXPERIMENTAL INVESTIGATION… Chem. Ind. Chem. Eng. Q. 22 (1) 17−26 (2016) the existing correlations may agree well with present correlation, many tuning parameters have been imp- experimental data in some limited degrees of rough- lemented and additionally the physical properties of ness; however the deviations between present exper- heating surface are totally ignored. Stephan and imental data and existing correlations exceed 50% Abdelsalam [4] proposed four specific correlations absolute average error (A.A.E.) in some other deg- applying a statistical multiple regression technique to rees of roughness. In this investigation, a new semi- the following liquid classes: water, organics, refriger- empirical model is presented to predict the boiling ants and cryogenics. In these correlations, the bubble heat transfer coefficient with A.A.E. of 11% at full diameter is estimated by Fritz [5] correlation. Cooper range of roughness degrees, which is much less than [6] proposed a new reduced pressure form of pool the A.A.E. of the existing correlations and is within boiling heat transfer correlation including the rough- maximum expected uncertainty of the experimental ness of the boiling surface. Gorenflo [7] has proposed procedure. an empirical correlation based on the reduced pres- McNelly [1] has proposed one of the first emp- sure of the boiling liquid. In this correlation, the sur- irical correlations for prediction of pool boiling heat face roughness is also included. Application of the transfer coefficient. In this correlation, the physical Gorenflo [7] correlation requires the specific reference characteristics of heating surface are not involved. heat flux, qo and also reference boiling heat transfer

Rohsenow [2] has proposed an empirical correlation coefficient, α0. Vinayak and Balakrishnan [8] and also based on the bubble agitation mechanism. In this Alavi Fazel, Jamialahmadi and Safekordi [9] have correlation, the boiling fluid is assumed to be single also a wide-ranging survey on some other correl- phase. In the Rohsenow [2] correlation, the Nusselt ations. In Table 1, the major existing correlations number is empirically correlated to Prandtl and Rey- have been summarized. nolds numbers. Mostinski [3] has ignored the surface Modeling of problems with stochastic roughness effects and applied the principle of corresponding is very difficult due to the complicated interactions states to pool boiling heat transfer. In this correlation, between bubbles and surface. This problem has been the experimental data are correlated to the reduced reviewed by McHale and Garimella [10]. Also the pressure and critical pressure of boiling liquid. In this problem of surface topography is investigated by

Table 1. A summary of major existing correlations

Author Correlation

0.69 McNelly [1]  0.31 ρ 0.33 qClll Pk  α = 0.225− 1 Ah σρ fgv

0.33 Rohsenow [2] 0.5 Δ  CTl qA/ σ =Csf Pr l h μ hg()ρρ− fg l fg l v

Mostinski [3] 0.7 0.17 1.2 10     α 0.69 qPPP  =bPc  1.8 +  + ; b=3.75E-5 SIUnit APPP     ccc  

0.371 0.35 −1.73 Stephan and Abdelsalam [4] 0.674 0.297 2 α k qd ρ hd 2lρ ()ρρ− α = 0.23 lvfg  llv  σ dAkTlsρα l 2 d ρ l  l 

− −0.55 0.67 Cooper [6] α 0.12 0.443Ra − −0.55 =P55rr() log P MwqA () / Gorenflo [7] n qA/ P αα=FFFF ; F=;0.90.3 n=− P0.3 ; F=1.2 P0.27 + 2.5 P+ r ; 0qPWRWM q ()qA/ r p rr1−P 0 r

2/15 1/ 4 R kllplρC kW F=a0 ; F=; qA=/20 WR R WM k ρ C 0 2 a 00p0 m

0.8 Nishikawa [24] 31.4P 0.2 0.2() 1−p pqA0.23 / α =Rc ()8 r r ; R=0.125 μm Mw0.1 T a ()− 0.9 p c0.9 10.99pr

0.33 Boyko and Kruzhiline [25] 0.7  0.5 k hqfg / Aρ TCσP σ α = 0.082 lvsl ; l=  − 2 ρρ− lgTkslρρ l v h ρ l g () fg v 2 lv

18 S. ALI ALAVI FAZEL, G. HOSSEYNI: EXPERIMENTAL INVESTIGATION… Chem. Ind. Chem. Eng. Q. 22 (1) 17−26 (2016)

Jabardo [11] and more recently by Moita, Teodori and of the rod heater is shown in Figure 1. Moreira [12]. The heating surface temperature was calculated by the integrated form of Fourier’s conduction equa- EXPERIMENTAL tion in cylindrical coordinates. In this investigation, several rod heaters have been produced from differ- Apparatus ent metals, including stainless steel 316 (SS316), The boiling vessel contained 35 L of test liquid. copper, aluminum and brass. These metals have This volume was sufficient to provide pool boiling been selected based on: 1) ease of metalworking, 2) conditions. The vessel was thermally insulated to providing a wide range of physical properties and 3) minimize the heat loss. The temperature of the sys- availability of the physical properties in the literature. tem was constantly monitored and regulated to satur- Because the bubble dynamics and boiling heat trans- ation point. The vessel was equipped with a rod fer coefficient are strongly affected by surface rough- heater, which includes four thermocouples, emb- ness, the surfaces of the heaters were sanded to edded parallel to the heating surface. The input AC provide various degrees of roughness. Roughness is electrical power to the rod heater was adjustable by a generally quantified by the vertical deviations of a real variable electrical transformer. This transformer con- surface from its ideal form. There are many definitions verts the input AC voltage of 220 V into any sel- for surface roughness. In this research, the rough- ectable voltage between 0-240 V. The electrical input ness Ra, is defined as the arithmetic average of the power to the rod heater is calculated by the product of vertical deviations. Figure 2 presents a typical electrical voltage, current and cosine of the difference measured value of surface roughness. between electrical voltage and current. A schematic

Figure 1. A schematic of the rod heater.

Figure 2. A typical value of surface roughness.

19 S. ALI ALAVI FAZEL, G. HOSSEYNI: EXPERIMENTAL INVESTIGATION… Chem. Ind. Chem. Eng. Q. 22 (1) 17−26 (2016)

In this investigation, water and ethanol have camera is 10.1 megapixels and the shutter speed is been chosen as the boiling liquids, based on: 1) the 1/2000 s. These values were sufficient to provide a availability of the physical properties, 2) covering a sharp and clear image from the heating surface. wide range of physical properties and 3) non-toxic A summary of surface characteristics and boiling properties. fluid at various degrees of roughness are presented in Table 2. The measured values of the cosine of con- Procedure tact angle between boiling liquid and surface, which Initially, the entire system was cleaned by cir- describes the wettability characteristics, are pre- culating and draining the boiling liquid through the sented in Table 3. Note that the contact angle slightly vessel, after which the test solution was introduced. varies upon variation of surface temperature; the The pressure of the system was kept at about 10 kPa mentioned table presents the average values. The (abs.) for an hour by a vacuum pump to degas the static contact angles are measured when droplet is boiling liquid. The temperature of the system was standing on the surface and the three-phase bound- then raised to the saturation temperature, and the ary is not moving. To measure the static contact electrical voltage was supplied to the rod heater up to angle, a droplet of specific liquid was placed on the the maximum value. After reaching steady state, the particular metallic surface. Then, the image of the surface temperature was recorded. Then, the elec- drop was captured using a digital camera, which was trical voltage was decreased in various intervals and equipped with a micro-lens to magnify the subject. the recordings were repeated in each interval after The experimental static contact angle was then steady-state accomplishment. Note that the dec- defined by fitting the tangent line on the liquid-solid reasing path of heat flux was to prevent a hysteresis contact point. Note that because of the hysteresis effect. Each experiment took about five minutes to effect, the static contact angle has a spectrum of reach a steady state at any specific condition. The contact angles ranging from advancing (maximal), to wall temperature was calculated based on the rec- the receding (minimal) contact angle. The equilibrium orded temperatures of the thermocouples inside the contact angle is somewhere between those values, rod heater. The distance between thermocouples loc- and was calculated by the Tadmor correlation [13]. ation and surface was 0.5 mm, which was introduced in the integrated form of the Fourier conduction law in Table 2. The experimented degrees of roughness at various cylindrical coordinates. In addition, the thermal con- boiling fluids and surfaces ductivities of individual heating materials were intro- Fluid Aluminum Brass Copper Stainless steel 316 duced in the Fourier conduction law. The arithmetic Water 1.9×10-4 m 3.0×10-5 m 3.0×10-5 m 2.5×10-6 m averages of four thermocouples were assigned to the 3.5×10-4 m 1.4×10-4 m actual wall temperature. Note that the measured tem- 3.6×10-4 m peratures from the four measuring points were appro- Ethanol 3.0×10-5 m 3.0×10-5 m 3.0×10-5 m Not tested ximately correspondent within ±0.2 K. Some runs 1.9×10-4 m 1.4×10-4 m were repeated twice to ensure the reproducibility of -4 -4 3.5×10 m 3.6×10 m the experiments. The physical properties of liquid and heating surface were evaluated at bulk and wall tem- Table 3. The measured values of the cosine of contact angle perature, respectively. between boiling liquid and surface To measure the bubble diameter, photographs of the heating surface have been captured at high Fluid Aluminum Brass Copper Stainless steel 316 speed at each heat flux. The diameters of 20 bubbles Water 0.738 0.623 0.435 0.813 were measured and the arithmetic averages were Ethanol 0.902 0.901 0.922 Not tested calculated. To measure the bubble departing frequency and nucleation site density, high speed video Experimental uncertainty recording (1000 fps) was performed at each con- The resolution of the voltmeter, ammeter and dition. The slow motion of the recordings was anal- the millivolt meter used in the present study was ±1 V, yzed and the nucleation sites were counted and ±0.1 A and ±0.01 mV respectively. The uncertainty in divided to project area. In addition, the bubble fre- the measurement of temperature is ±0.2 K. The pro- quencies of the nucleation sits were counted and the pagation error of the four parallel thermocouples, arithmetic averages were calculated. A Casio EX- which are provided to measure the surface tempera- -FH25 camera was used to record the visual inform- ture, was estimated to (4×0.2)/4 = 0.2 K, which is ation. The effective resolution of the mentioned equal to 0.2/100 = 0.002 or 0.2% at boiling point of

20 S. ALI ALAVI FAZEL, G. HOSSEYNI: EXPERIMENTAL INVESTIGATION… Chem. Ind. Chem. Eng. Q. 22 (1) 17−26 (2016) pure water. The maximum propagation error of heat The performances of major existing correlation flux in terms of fractional uncertainty is estimated to: are presented in Figure 4. The numerical comparisons show that for water/SS316, water/brass and Δ q  ethanol/SS316, the Mostinski [3] correlation has the A ΔΔIV0.1 1 ===+ + 0.11 best performance with 11, 24 and 11% absolute aver- Δ q IV 1.1 50  age error (A.A.E.), respectively, while the mentioned A best value correlation has about 56% A.A.E. for Water/Cu and Ethanol/Cu boiling systems. For water/Cu, ethanol/ Note that the heat flux is calculated by the /brass, ethanol/Al and ethanol/Cu, the correlation pro- products of electrical voltage and current divided to posed by Stephan and Abdelsalam [4] has the best the heating area of the rod heater. It is considered agreement with experimental data with 44, 16, 14 and that the heating area was accurate enough to ignore 34% absolute average error respectively, while the from error analysis. The standard deviation of mea- mention correlation has more than 60% A.A.E. for sured bubble diameters and bubble frequencies for water/SS316 boiling system at the average exper- water at 25 kW/m2 was typically equal to 0.00009 mm imented roughness and heat fluxes. These deviations and 2 Hz, respectively. This means that the uncer- are large because of the experimental basis of the tainty for measured bubble diameter and bubble fre- existing correlations. quency would be about 0.00009/0.002 = 4.5% for a To cross-check the validity of the derived model, bubble with 2 mm in diameter and 2/100 = 2% for an independent dataset was collected [14]. The men- bubble frequency of 100 Hz. The nucleation sites tioned dataset consists of water, acetone, ethyl ace- were visually counted by the slow motion playback of tate, 2-propanol, methanol and ethanol at the boiling the recorded videos without any significant degree of liquid. The heating section was pure copper with uncertainty. smooth texture. To quantify the impact of various Experimental results physical properties on boiling heat transfer coefficient, The raw numerical values of heat flux versus the sensitivity analysis was performed by arranging degree of superheat are presented in Figure 3. It can the following equation: be inferred that the boiling heat transfer coefficient αρρ= g0 g1 g2 g3 μσ g4 g5 g6 ρ g7 g8 g9 lvfgllhC k lsss Ck (1) increases with increasing the degree of superheat at any constant condition. In addition, the boiling heat By using the genetic algorithm, the vector G = transfer coefficient increases with increasing the sur- = [g0,g1,…,g9], which represents the exponents of Eq. face roughness at any constant condition. This is (1), is found equal to: because of the enhancement in nucleation site density. G =[0.46, −− 0.86, 0.39,0.78, − 0.16,0.24, −−− 0.32, 0.26, 0.10,0.37, − 0.10,0.38] (2)

Figure 3. The raw numerical values of heat flux versus degree of superheat.

21 S. ALI ALAVI FAZEL, G. HOSSEYNI: EXPERIMENTAL INVESTIGATION… CI&CEQ 22 (0) 000−000 (2016)

== + QqAqAqAcc bb (7)

By combining Eqs. (4), (5) and (7) the heat flow rate can be calculated as: ==αα Δ+Δ QqAATbb cc AT (8)

Assuming the affected areas by spherical bub-

bles are equal to the projected area of the bubbles, Ab can be calculated by: A N π b = β d 2 (9) AA4 where N/A is the nucleation site density. In the aforementioned equation, β is the ratio of area of influence to projected area of bubble at departure. Judd and Hwang [15] have matched their predicted Figure 3. The raw numerical values of heat flux versus degree heat fluxes with experimental data and reported that of superheat. β = 1.8. Some other investigators, such as Han and Griffith [16], postulated that β = 4. In this investigation, Modeling it is found that the parameter β depends on the shape According to Newton’s cooling law, the heat and the oscillating behavior of the departing bubble. transfer is proportional to the area and the thermal Clift, Grace and Weber [17] proposed that the bubble driving force, i.e.: shapes can be describes by the dimensionless Eöt- vös number. In addition, because the dimensionless qAAT=Δα (3) Roshko number describes the oscillating nature of the In the presence of bubbles on the heating sur- rising bubbles, here it is postulated that the parameter face, the heating area can be divided by two com- β should be a function of both Ro and Eö. By reg- plementary zones: 1) Ab, the area that is affected by ression analysis it is found that: bubbles and 2) A , the convective heat transfer area, c β = 22−4 Ro . Eö (10) as designated in Figure 5. Each zone has the individual magnitude of heat transfer, i.e.: where: =Δα qccAAT cc (4) ()ρρ− gd 2 Eö = lv (11) and: σ =Δα and: qbbAAT bb (5) fd 2 where the subscripts “c” and ”b” stand for “convection” Ro = (12) μρ and “bubble affected” areas, respectively. Openly, the ll/ following equation is already established: Figure 6 describes the relation between the A AA aforementioned dimensionless groups. Combining =+=cb1 (6) AAA Eqs. (8) and (9) yields:

N πd 2 αα=−+() α αβ (13) cbcA 4

where α is the total heat transfer coefficient, αc is the

convective heat transfer coefficient and αb is the heat transfer coefficient in the bubble affected area (all in W m-2 K-1). When a bubble is develops on a heating surface, the heat transfer coefficient through the bubble Figure 5. Dividing the boiling heat transfer area in the modeling. stem, αb can be predicted by the correlation proposed by Mikic and Rohsenow [18]. The heat transfer Summing up Eqs. (4) and (5) yields:

22 S. ALI ALAVI FAZEL, G. HOSSEYNI: EXPERIMENTAL INVESTIGATION… Chem. Ind. Chem. Eng. Q. 22 (1) 17−26 (2016) mechanism is substantiated to be transient conduct- age absolute error of 14% for the entire systems. The ion around nucleation sites. The heat transfer coef- Stephan correlation [21] is an experimental correl- ficient is calculated by: ation with the following mathematical form:

0.5 απ= 2 k ρCf (14) 2 bllpl Ja 100000 2σ d =+0.25 1 (18) Pr Ar g ()ρρ−    lv To predict the nucleation site density, the correlation proposed by Xiao, Jiang, Zheng, Chen and Liu [22] is recommended:

N − = 7.8125e-29() 1− cosφ R 6 (19) A c,min In the aforementioned correlation, the minimum cavity radius is calculated by:

 2  δζθθ4 c R =−−−−11ss2  (20) c,min  θθδθ  c1w w w   Figure 6. The ratio of area of influence to projected area of where: bubble at departure as a function of the products of Roshko number and Eötvös number. 2σT ζ = sat (21) ρ vfh g The convective heat transfer coefficient, αc can be calculated by the equation proposed by Churchill 1cos+ ϕ c = (22) and Bernstein [19], which is applicable to forced con- 1 sinϕ vection around the horizontal cylinders: =+ ϕ c2 1cos (23) Nu =+0.3 OD ϕ 4/5 where is the contact angle of the fluid and the 5/8 1/2 1/3  (15) ++0.62ReOD Pr Re OD heater material. The boundary layer thickness can be 1/ 4 1  2/3 282000 calculated by dividing the liquid thermal conductivity 10.4/+ ()Pr   to the natural convection heat transfer coefficient: where the dimensionless Reynolds number is cal- k δ = l (24) α culated based on the upward terminal velocity of NC bubbles, uT. The upward terminal velocity can be calculated by: To calculate the bubble departing frequency, the experimental correlation proposed by Zuber [23] is 4 d ρ − ρ recommended: ug= lv (16) T 3 C ρ dl σρρ− 0.25  g ()lv fd = 0.59   (25) and the drag coefficient, C , is already calculated by ρ 2 d  l  Ishii and Zuber [20]: Note that in the modeling of the present data, =+24 0.75 CRedd()10.1 (17) the experimental values of nucleation site density, Red bubble departing frequency and diameter are used. Note that to find the terminal velocity, Eqs. (16) MODEL VALIDATION and (17) should be calculated iteratively, because the Reynolds number is already a function of the terminal The performance of the new model is compared velocity. to experimental data, as presented in Figure 7. It is To predict the bubble departing diameter, many found that 95% of the data points are matching within correlations have been compared to experimental ±11% absolute average error with experimental data. data. It is found that the Stephan correlation [21] has Note that the value of ±11% is calculated as the max- the best agreement with experimental data with aver-

23 S. ALI ALAVI FAZEL, G. HOSSEYNI: EXPERIMENTAL INVESTIGATION… Chem. Ind. Chem. Eng. Q. 22 (1) 17−26 (2016) imum expected uncertainty. To revalidate the new CONCLUSION model, an independent dataset [14] have been compared to the new model as presented in Figure 8. It is In this investigation, saturated nucleate pool shown that about 90% of the data points are matching boiling heat transfer was studied experimentally for within ±20% absolute average error with experimental two boiling liquids – water and ethanol. Several heat- data. It is important to note that the new model is very ing elements were made and tested with different sensitive to three key parameter including nucleation metals including copper, aluminum, brass and SS316 site density, bubble departing frequency and diame- with various degrees of roughness. The measured ter. In evaluating the performance of the new model data includes boiling heat transfer coefficient, nuc- by the independent dataset [14], the three aforemen- leation site density, bubble departing diameter, as tioned parameters were estimated by existing correl- well as bubble departing frequency. The experimental -2 ation introduced by Eqs. (18), (19) and (25). In the heat fluxes were limited to 100 kW m to be able to mentioned reference [14], these values are not rep- measure the visual information of the boiling pheno- orted. menon. It was found that bubble departure frequency, diameter and nucleation site density are three key- -parameters in determining the boiling heat transfer coefficient. Furthermore, the ratio of area of influence to projected area of bubble at departure can be correlated to the products of two dimensionless groups, Roshko and Eötvös numbers. The Roshko number describes the oscillating nature of bubble dynamics, while the Eötvös number characterizes the shape of bubbles. It was shown that by dividing the heating surface to two complementary areas, one is directly influenced by bubbles and the other is free from bubbles effects; the boiling heat transfer is predictable. A new semi-empirical model was proposed to predict the boiling heat transfer coefficient.

Acknowledgment Figure 7. The predicted values of heat flux versus the experimental values of present study. The author is thankful to department of chemical engineering, college of chemistry and chemical engineering, Mahshahr Branch , Islamic Azad University, Mahshahr, Iran for financial support of research project entitled “Pool boiling heat transfer in pure liquids”. NOMENCLATURE A area, m-2 Ar Archimedes number b constant (see Mostinski correlation)

c1 constant, see Eq.(22)

c2 constant, see Eq.(23)

Cd drag coefficient C heat capacity, J kg-1 K-1

Csf constant, see Rohsenow correlation d bubble diameter, m Eö Eötvös number f bubble departing frequency, Hz

Figure 8. The predicted values of heat flux versus the Fp see Gorenflo correlation experimental values from independent investigation. Fq see Gorenflo correlation

FWM see Gorenflo correlation

FWR see Gorenflo correlation g acceleration of gravity, N kg-1

24 S. ALI ALAVI FAZEL, G. HOSSEYNI: EXPERIMENTAL INVESTIGATION… Chem. Ind. Chem. Eng. Q. 22 (1) 17−26 (2016)

-1 hfg specific heat of vaporization, J kg [2] W.M. Rohsenow, Trans. ASME 74 (1952) 969-976 Ja Jakob number [3] I.L.Mostinski, Teploenergetika 4 (1963) 66 k thermal conductivity, W m-1 K-1 [4] K. Stephan, K. Abdelsalam, Int. J. Heat Mass Transfer 23 Mw molecular weight, g mol-1 (1980) 73–87 N number of nucleation sites [5] W. Fritz, Phys. Z. 36 (1953) 379–384 P pressure, Pa [6] M.G. Cooper, Adv. Heat Transfer 16 (1984) 157-239 Pr Prandtl number [7] D. Gorenflo, VDI Heat Atlas. (1993)

Pr reduced pressure [8] R. Vinayak Rao, A.R. Balakrishnan, Exp. Therm. Fluid q heat flux, W/m2 Sci. 29 (2004) 87–103 Re Reynolds number [9] S.A. Alavi Fazel, M. Jamialahmadi, A.A. Safekordi, Iran. Ro Roshko number J. Chem. Chem. Eng. 27(3) (2008) 135–150

Ra absolute roughness, m [10] J.P. McHale, S.V. Garimella, Int. J. Multiphase Flow 36 T temperature, K (2010) 249-260 -1 [11] J.M. Saiz Jabardo, Open Transp. Phenom. J. 2 (2010) uT terminal velocity, m s 24-34 Subscripts [12] A.S. Moita, E.Teodori, A.L.N. Moreira, Int. J. Heat fluid 0 reference flow 52 (2015) 50-63 b boiling [13] R. Tadmor, Langmuir 20(18) (2004) 7659-7664 c convection or critical [14] S. Matthew, Ph.D. Thesis, Michigan State University, l liquid 1988 NC natural convection [15] R.L. Judd, K.S. Hwang, J. Heat Transfer 98 (1976) 623– OD outside diameter -629 s saturated or solid [16] C.Y. Han, P. Griffith, MIT Press Energy Lab. Ser. (1962) v vapor 1-76 w wall [17] R. Clift, M.E. Grace, Bubbles Drops and Particles, Aca- demic Press, New York, 1978, p. 26 Greek symbols [18] B.B. Mikic, W.M. Rohsenow, J. Heat Transfer 91 (1969) -2 -1 α heat transfer coefficient, W m K 245–250 α 2 -1 l thermal diffusivity, m s [19] S.W. Churchill, M. Bernstein, Trans. ASME 99 (1977) β the ratio of area of influence to projected 300–306 area of bubble at departure [20] M. Ishii, N. Zuber, AIChE J. 25 (1979) 843–855 δ boundary layer thickness, m [21] K. Stephan, Ph.D. Thesis, University of Auckland, ζ see Eq. (21) Auckland 1992

θs D-value of Ts-T∞ [22] B. Xiao, G. Jiang, D. Zheng, L. Chen, B. Liu, Res. J. Appl.

θw D-value of Tw-T∞ Sci., Eng. Technol. 6(4) (2013) 587-592 ρ density, kg m-3 [23] N. Zuber, Appl. Mech. Rev. 17 (1964) 663-672 σ surface tension, N m-1 [24] K. Nishikawa, Y. Fujita, H. Ohta, S. Hidaka, in Proceed- ϕ contact angle – see Eq. (19), (22) and (23) ings of the 7th International Heat Transfer Conference, München, Germany, 4, 1982, pp. 1–66 REFERENCES [25] L.D. Boyko, G.N. Kruzhilin, Int. J. Heat Mass Transfer 10 (1967) 361. [1] M.J. McNelly, J. Imp. Coll. Chem. Soc. 7 (1953) 18–34

25 S. ALI ALAVI FAZEL, G. HOSSEYNI: EXPERIMENTAL INVESTIGATION… Chem. Ind. Chem. Eng. Q. 22 (1) 17−26 (2016)

SEYED ALI ALAVI FAZEL EKSPERIMENTALNA ISTRAŽIVANJA PRENOSA GOHARSHAD HOSSEYNI TOPLOTE PRI DELIMIČNOM KLJUČANJU Department of chemical ZASIĆENIH ČISTIH TEČNOSTI engineering, college of chemistry and chemical engineering, Prenos toplote pri delimičnom zasićenom ključanju je eksperimentalno ispitivan u sudu sa Mahshahr Branch , Islamic Azad horizontalnim grejačem oblika šipke. Ispitivanje je uključilo vodu i etanol. Sekcija za gre- University, Mahshahr, Iran janje je napravljena od različitih materijala: SS316, bakar, aluminijum i mesing. Eksperi- NAUČNI RAD menti su izvršene sa površinama čiji stepen hrapavosti meren srednjom vertikalnom devi- jacijom u opsegu između 30 i 360 μm. Ispitivani su koeficijent prenosa toplote, prečnik i učestalost otkidanja mehura, kao i gustina nukleacionih mesta. Podaci su poređeni sa glavnim postojećim korelacijama. Pokazano je da se eksperimentalni podaci ne poklapaju sa prihvatljivom tačnošću sa glavnim korelacijama u celom opsegu eksperimentlnih uslova. U ovom radu, oblast prenosa toplote pri ključanju je podeljen u dve komplemen- tarne podoblasti: izazvane prinudne konvekcije i oblasti pod uticajem ključanja. Za koefi- jent prenosa toplote predložen je polu-empirijski model koji uključuje dva bezdimenziona kriterijuma, Etvesov i Roškov kriterijum. Pokazano je da predloženi model nudi poboljšane performanse u predviđanju koeficijenta prenosa toplote u poređenju sa postojećim korelacijama.

Ključne reči: indukovana prinudna konvekcija, ključanje zasićene tečnosti, hra- pavost površine, koeficijent prenosa toplote.

26 Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 22 (1) 27−32 (2016) CI&CEQ

MARIJA ILIĆ1 THE INFLUENCE OF ALKYL POLYGLUCO- FRANZ-HUBERT HAEGEL2 3 SIDES (AND HIGHLY ETHOXYLATED VESNA PAVELKIĆ ALCOHOL BOOSTERS) ON THE PHASE DRAGAN ZLATANOVIĆ4 SNEŽANA NIKOLIĆ-MANDIĆ5 BEHAVIOR OF A WATER/TOLUENE/ ALEKSANDAR LOLIĆ5 /TECHNICAL ALKYL POLYETHOXYLATE ZORAN NEDIĆ6 MICROEMULSION SYSTEM 1Faculty of Mining and Geology, Article Highlights University of Belgrade, Belgrade, • Addition of sugar surfactant to system of water/tolune/Lutensol ON 50 was investigated Serbia • Sugar surfactant shifts the phase behavior to lower temperature 2Forschungszentrum Jülich, Institut • Microemulsion of water/tolune/Lutensol ON 50 and alcohol ethoxylate C18E100 was für Bio- und Geowissenschaften, investigated

IBG-3 Agrosphäre, Jülich, • Strongly hydrophilic C18E100 shifted the one phase region to higher temperature Germany 3Institut of Chemistry, Technology Abstract and Metallurgy, University of The influence of additives (alkyl polyglucoside, Glucopon 600 CS UP and alco- Belgrade, Belgrade, Serbia hol ethoxylate C18E100) on the behavior of the water/toluene/Lutensol ON 50 4Innovation Center Faculty of (technical oxoalcohol, i-C10E5) microemulsion system as a function of tempera- Mechanical Engineering, University ture and composition has been investigated. The phase behavior of the micro- of Belgrade, Belgrade, Serbia emulsions was determined by vertical sections through the Gibbs phase prism 5Faculty of Chemistry, University of Belgrade, Belgrade, Serbia (fish-like phase diagrams). Alkyl polyglucoside shifts the one phase region to 6Faculty of Physical Chemistry, lower temperatures compared with water/toluene/Lutensol ON 50 mixtures. University of Belgrade, Belgrade, This is contrary to the expectation, considering the extreme hydrophilic nature Serbia of the sugar headgroup. The addition of hydrophilic alcohol ethoxylate (C18E100) to the water/toluene/Lutensol ON 50 system increases the solubiliz- SCIENTIFIC PAPER ation capacity of the surfactant, even if the co-surfactant is used in small quantities, and shifts the one-phase region to higher temperature by a few °C. UDC 547.533:66:544 Keywords: microemulsion, toluene, alkyl polyglucoside, oxoalcohol DOI 10.2298/CICEQ141105015I ethoxylate, efficiency booster, “fish” diagrams.

Microemulsions are thermodynamically stable The stability of microemulsions over a large tem- optically isotropic mixtures, consisting of two immis- perature range and low surfactant concentration is cible components, oil and water, made miscible by a required for technical applications. Some investigat- third component, the surfactant. They may contain ions [1,2] show that the use of surfactants with longer additives such as salt or alcohol. The properties of hydrophobic units reduces the amount of surfactant ternary nonionic surfactant/water/oil microemulsions needed for microemulsification due to the increasing are very interesting scientifically and technically. A efficiency of the surfactant. Adding suitable additives convenient way to study these systems is to measure such as sugar surfactants and nonionic alcohol the phase behavior at constant oil/water ratios as a ethoxylates received much attention in recent years. It function of temperature, T, and surfactant mass was found that the addition of a small amount of fraction, γ. polymer with amphiphilic properties to the microemulsion system increases the efficiency of the surfactant [3-10]. Alcohol ethoxylate surfactants are widely used Correspondence: M. Ilić, Faculty of Mining and Geology, Uni- for microemulsion applications. The phase behavior versity of Belgrade, Djušina 7, 11000 Belgrade, Serbia. with respect to the length of the hydrocarbon tails and E-mail: [email protected] the number of ethylene oxide (EO) units depends on Paper received: 5 November, 2014 Paper revised: 13 May, 2015 the purity of alcohol ethoxylates. The structure of the Paper accepted: 20 May, 2015

27 M. ILIĆ et al.: THE INFLUENCE OF ALKYL POLYGLUCOSIDES… Chem. Ind. Chem. Eng. Q. 22 (1) 27−32 (2016) hydrocarbon tail of the surfactant strongly influences ance (HLB) of 13.8, according to Griffin [29]. Lutensol the microemulsion behavior [11,12]. Technical grade ON 50 (C10 oxoalcohol polyethoxylate with an aver- alcohol ethoxylates usually contain mixtures of differ- age of 5 ethylene oxide units, i-C10E5) is a commercial ent alcohols and often exhibit a distribution over a nonionic surfactant of BASF AG, Ludwigshafen (Ger- large range of ethoxylation degrees. many). It has a HLB of 11.5 (technical information In recent years, alkyl polyglucosides, a class of sheet of BASF). sugar surfactants, have received considerable inter- Glucopon 600 CS UP alkyl polyglucoside (alkyl est as nonionic surfactants because of their excellent chain containing 10 to 16 carbon atoms, and an aver- biodegradability, ease of manufacture from renewable age number of glucose units of 1.4) is a commercial resources, such as sugar and vegetable oil feed- nonionic sugar surfactant containing 51% active mat- stocks [13], and potential use in a large number of ter and 49 mass% water (technical information sheet industrial applications [14-19]. One potential use of of Cognis, Monheim, Germany). The alcohol ethoxyl- sugar surfactants is in microemulsion formulations. ate C18E100 was synthesized under argon using a high Making microemulsions with alkyl polyglucosides is vacuum line by Frank et al. [28] in the laboratory of difficult owing to the low surfactant solubility in many JCNS-1 at Forschungszentrum, Jülich. Water was classes of oils. Fundamentally, it is of substantial deionized and twice distilled. interest to form microemulsions with sugar surfactant Phase diagram determination as co-surfactant or to form sugar surfactant-based microemulsions using co-surfactant. Kahlweit et al. [30,31] introduced a way of studying the phase behavior of ternary or quaternary mix- Alkyl polyglucosides, abbreviated as CmGn, represent complex mixtures [20] where m is the number tures. A procedure to obtain an overview of the of carbon atoms in the hydrocarbon chain and n is the phases is to draw the phase diagram at a constant average number of glucose units in the hydrophilic oil/water ratio as a function of temperature, T, and headgroup. The nonionic surfactants, n-alkyl poly- surfactant mass fraction, γ. The phase boundaries resemble the shape of a fish. Typically, temperature- glycol ethers (CiEj) are typically used. These surfactants contain i carbon atoms in the hydrophobic -composition phase diagrams obtained for a 1/1 mass alkyl chain and j ethoxy units in the hydrophilic head- ratio of oil and water show a one-phase microemul- group. The phase behavior of microemulsion systems sion at relatively high surfactant concentration. At containing alkyl polyglucoside has been studied by lower surfactant concentration a three-phase body some authors [21-23]. exists consisting of a middle-phase microemulsion in In the last decades, a lot of papers were pub- equilibrium with excess phases of oil and water, lished concerning many aspects of polymers in micro- surrounded by two-phase regions illustrated as 2Φ emulsions, such as solubilization efficiency boosting and 2Φ . When a surfactant is mainly dissolved in by amphiphilic polymers in microemulsions [24-26]. water and two phases consist of a surfactant-rich The influence of various polymers on the phase equi- water (lower) phase in equilibrium with an excess oil librium of microemulsions containing nonionic surfact- phase, the region is denoted as 2φ. At high tempe- ant, water and oil, has also been studied. It was found rature a nonionic surfactant is more soluble in oil and that water-soluble polymers expelled into the coexist- forms a surfactant-rich oil (upper) phase in equilibrium ent water phase cause the coexistence of the lamellar with an excess water phase denoted 2Φ (Figure 1). phase with water, which does not appear in the absence of polymer [27]. In this work, we studied the phase behavior of Lutensol ON 50-based microemulsions after addition of alkyl polyglucoside Glucopon 600 CS UP as a co- surfactant or a hydrophilic alcohol ethoxylate (C18E100) as an efficiency booster [28].

MATERIALS AND METHODS

Toluene (purity 99% by GC) was purchased from Merck Schuchardt (Germany). Octaethylene glycol decylether (C10E8, octaethyleneoxide decylether) Figure 1. Schematic “fish cut” phase diagram of a nonionic with purity higher than 98% (GC) was purchased from microemulsion with equal water to oil proportions as a function Fluka (Germany). It has a hydrophilic-lipophilic bal- of surfactant concentration.

28 M. ILIĆ et al.: THE INFLUENCE OF ALKYL POLYGLUCOSIDES… Chem. Ind. Chem. Eng. Q. 22 (1) 27−32 (2016)

The convenient variables are the temperature branched i-C10E5 (Lutensol ON 50) in order to find the and the following composition variables (pressure is composition and temperature of optimum solubil- always kept constant) – the mass fraction of oil in the ization (Figure 2). mixture of water and oil:

α = mB /(mA + mB) (1) The mass fraction of surfactant in the mixture of all three components in ternary mixtures:

γ = mC/(mA + mB + mC) (2) or:

γ = (mC + mD)/(mA + mB + mC + mD) (3) when two surfactants or mixtures of surfactant and co-surfactant are used. In this case, the mass fraction of one of the surface-active components (δ) in the mixture is defined as:

δ = mC/(mC + mD) (4) Figure 2. Phase diagrams of water/toluene/C E and where the capital indices A, B, C, D refer to the com- 10 8 water/toluene/Lutensol ON 50 at equal mass fractions ponents, water, oil, surfactant and co-surfactant, res- of water and toluene (α = 0.5). pectively. The so-called γ -point, where the three-phase The phase boundaries of the system with C10E8 body meets the one-phase region, defines the mini- meet at γ = 0.123, T = 24.9 °C. Beyond that point, a mum mass fraction of surfactant needed to solubilize single homogeneous phase appears, when the mass water and oil and is a measure for the efficiency of the fraction of surfactant γ is further increased. Thus, this  surfactant. The corresponding temperature, T , is a point where the three-phase (3φ) and one-phase measure for the phase inversion temperature (PIT). regions (1φ) meet represents the lowest surfactant The phase diagrams were recorded by success- concentration needed to solubilize equal masses of ively adding water and oil to the initial water-oil-surf- the two immiscible components, water and toluene. At actant mixture. The samples were prepared by weigh- lower temperatures the microemulsion coexists with ing appropriate amounts of components (1:1 ratio of excess oil (denoted by 2Φ). At higher temperatures oil to water) on 0.1 mg precision scales in the order the microemulsion coexists with excess water ( 2Φ ). surfactant, toluene, water to suppress intermediate This system exhibits a strong tendency to form liquid formation of liquid crystals. The mass fraction of the crystals. Its phase diagram shows a large area of surfactant (or surfactant/additive) is calculated with liquid crystals (LC) surrounded by microemulsion Eqs. (2) or (3) and the mass fraction of the additive in (regions 1φ and 1φ*). These results are very similar to the surfactant /additive mixture with Eq. (4). those reported for the same system with equal vol- Samples were weighed into test tubes, which ume fraction φ of water and toluene, i.e., lower mass were immediately sealed (glass stoppers) and put into fraction α of oil (α ≈ 0.465). In that case, the “fishtail a thermostated water bath with temperature control point” was found at γ = 0.114 and T = 24.34 °C [32]. ° up to 0.2 C. In the thermostated bath, the mixtures Owing to the lower density difference of toluene and were stirred with small magnetic stirrers to ensure water compared to alkane systems, the samples complete mixing of the components at the given tem- exhibit somewhat slower phase separation. The perature. After equilibrium was established, the bicontinuous microemulsion containing water, toluene occurring phases were characterized by visual ins- and C10E8 also shows some unusual behavior at low pection between crossed polarizers. temperatures. Shear-induced birefringence was observed between crossed polarizers in the region RESULTS AND DISCUSSION denoted 1φ* while stirring the sample. For the water/toluene/Lutensol ON 50 system, The pseudo-binary phase diagrams of micro- the determination of the point of optimum solubiliz- emulsion systems were determined at equal mass ation of equal masses of water and toluene failed. At fractions of water and toluene (α = 0.5) for two non- high γ values, a single homogeneous phase (1φ) was ionic surfactants, pure linear C10E8 and technical

29 M. ILIĆ et al.: THE INFLUENCE OF ALKYL POLYGLUCOSIDES… Chem. Ind. Chem. Eng. Q. 22 (1) 27−32 (2016) found in a temperature range similar to that of the The water/toluene/Lutensol ON 50 “fish” (δ = 0)

C10E8 system, but with considerably higher surfactant is shown for reference in Figure 3. As can be seen, content needed for mutual solubilization of both water with the addition of CmGn, the homogeneous micro- and toluene. The shape of the “fishtail” is also not emulsion region (“fish tail” with the phase sequence symmetrical with respect to temperature. This behavior 2Φ-1- 2Φ ) becomes wider and the efficiency of the is typical for technical surfactants with a distribution of surfactant mixture increases slightly. more or less hydrophobic components due to different With increasing δ, the “fish tail” unexpectedly degrees of ethoxylation [10]. In contrast to the system moves downward on the temperature scale despite with linear C10E8, no liquid crystals are found within the hydrophilic nature of the sugar surfactant. The the region of the bicontinuous microemulsion (1φ) efficiency of the surfactant mixture increases with indicating a less rigid structure of the surfactant layer increasing δ. In these quaternary systems the location at the interface between the oil and the water micro- of γ is not determined because, in contradiction to -phases. Approaching the composition of optimum many other microemulsion systems, equilibration for soulubilization of equal masses of water and toluene, the system water/toluene/Lutensol ON 50 was parti- samples with the branched technical surfactant cularly slow near the point of optimum solubilization needed very long and sometimes extremely long for the investigated ternary mixture (see Figure 1). times for equilibration after agitation or temperature Effect of hydrophilic alcohol ethoxylate changes. They show delayed visible phase separation and the phase boundaries to the two-phase The effect of small amounts of hydrophilic alco- regions determined visually are diverging. hol ethoxylate C18E100, on the location and width of Using the ternary mixtures water/toluene/Luten- the one-phase region was investigated. sol ON 50 as a base case, the role of added sub- Figure 4 shows phase diagrams for the stances, sugar surfactant and hydrophilic alcohol water/toluene/Lutensol ON 50 system with and ethoxylate on the phase behavior is explored. without addition of C18E100. In this presentation the oil mass fraction is α = 0.5 and the mass fraction of Effect of alkyl polyglucoside (Glucopon 600 CS UP): additive is δ = 0.01. Compared with the system δ Phase behavior as a function of without additive, the one-phase region with additive is Figure 3 shows the temperature-composition shifted to higher temperature. This effect can be phase diagram at mass fraction α = 0.5 for the quater- explained by the large hydrophilic moiety of C18E100. nary system water/toluene/Lutensol ON 50/Glucopon For higher content of C18E100, formation of liquid 600 CS UP for varying fractions of sugar surfactant, δ crystals was observed, which makes the system (0.05, 0.10 or 0.20). unsuitable for many applications (data not shown).

Figure 3. Temperature-composition phase diagram at water to Figure 4. Temperature-composition phase diagram at water to toluene mass fraction α = 0.5 for the system toluene mass fraction α = 0.5 of the water/toluene/Lutensol ON water/toluene/Lutensol ON 50 with added sugar surfactant, 50 system with C18E100 as additive (mass fraction δ = 0.01). Glucopon 600 CS UP , for varying mass fractions of sugar surfactant, δ (.05, 0.10 or 0.20).

30 M. ILIĆ et al.: THE INFLUENCE OF ALKYL POLYGLUCOSIDES… Chem. Ind. Chem. Eng. Q. 22 (1) 27−32 (2016)

CONCLUSIONS [2] K. Holmberg, B. Jönsson, B. Kronberg, B. Lindman, Surfactants and polymers in aqueous solution, John We have studied the phase behavior of water/ Wiley & Sons, Chichester, 2003 /toluene/Lutensol ON 50/Glucopon 600 CS UP and [3] A. Kabalnov, U. Olsson, K. Thuresson, H. Wennerström, Langmuir 10 (1994) 4509-4513 water/toluene/Lutensol ON 50/C18E100 mixtures as a function of temperature and composition. Both addi- [4] H. Endo, J. Allgaier, G. Gompper, B. Jakobs, M. Monken- tives lead to improved solubilization and shift the one- busch, D. Richter, T. Sottmann, R. Strey, Phys. Rev. Lett. phase region to lower surfactant content near to the 85 (2000) 102-105 [5] B. Jakobs, T. Sottmann, R. Strey, Tenside, Surfactants, values obtained for pure C10E8. Deterg. 37 (2000) 357-364 The addition of the alkyl polyglucoside, Gluco- [6] G. Gompper, D. Richter, R. Strey, J. Phys.: Condens. pon 600 CS UP, to the ternary water/toluene/Lutensol Matter 13 (2001) 9055-9074 ON 50 system additionally shifts the phase behavior [7] R. Strey, M. Brandt, B. Jakobs, T. Sottmann, in Studies in to lower temperature. This behavior is unexpected, Surface Science and Catalysis, Vol. 132, Y. Iwasawa, N. because the sugar surfactant is absolutely insoluble Oyama, H. Kunieda (Eds.), Elsevier Science, in toluene and should act as a hydrophilic surfactant. Amsterdam, 2001, pp. 39-44 It should rather increase the temperature for the one- [8] H. Endo, M. Mihailescu, M. Monkenbusch, J. Allgaier, G. phase region. No formation of liquid crystals was obs- Gompper, D. Richter, B. Jakobs, T. Sottmann, R. Strey, I. erved by the addition of Glucopon 600 CS UP. The Grillo, J. Chem. Phys. 115 (2001) 580-600 microemulsion consisting of water, toluene, Lutensol [9] B. Jakobs, T. Sottmann, R. Strey, J. Allgaier, L. Willner, D. Richter, Langmuir 15 (1999) 6707-6711 ON 50 and alcohol ethoxylate C18E100 as an additive also increases the efficiency of the surfactant system, [10] T. Sottmann, R. Strey, in Fundamentals of Interface and but shifts the one-phase region to higher temperature, Colloid Science, J. Lyklema (Ed.), Academic Press, New as expected. The formation of liquid crystals at higher York, 2005, pp. 5.1-5.96 [11] K.R. Wormuth, S. Zushma, Langmuir 7 (1991) 2048-2053 content of C18E100, however, makes this additive less suitable. These findings make the sugar surfactant a [12] C. Frank, H. Frielinghaus, J. Allgaier, H. Prast, Langmuir better choice for improving the ternary system water/ 23 (2007) 6526-6535 /toluene/Lutensol ON 50 with respect to applications. [13] K. Hill, in Alkyl Polyglucosides: Technology, Properties and Applications, K. Hill, W. von Rybinski, G. Stoll (Eds.), Alkyl polyglucosides are also preferable because of VCH, New York, 1997, pp. 1-7 their outstanding ecological and biological properties. [14] D. Geetha, R. Tyagi, Tenside, Surfactants Deterg. 49 Nomenclature (2012) 417-427 [15] M. Haeger, K. Holmberg, Tetrahedron Lett. 41 (2000) C E n-alkyl polyglycol ethers i j 1245-1248 C G alkyl polyglucosides m n [16] A. Stradner, B. Mayer, T. Sottmann, A. Hermetter, O. EO ethylene oxide units Glatter, J. Phys. Chem., B 103 (1999) 6680-6689 HLB hydrophile-lipophile balance [17] I.F. Uchegbu, S.P. Vyas, Int. J. Pharm. 172 (1998) 33-70 T temperature  [18] R. Schwering, D. Ghosh, R. Strey, T. Sottmann, J. Chem. T phase inversion temperature Eng. Data 60 (2015) 124-136 γ surfactant mass fraction [19] K. Hill, W. von Rybinski, G. Stoll, Alkyl Polyglucosides: Acknowledgments Technology, Properties and Applications, VCH, New York, 1997 The authors thank the Ministry of Education, [20] C. Stubenrauch, Curr. Opin. Colloid Interface Sci. 6 Science and Technological Development of the (2001) 160-170 Republic of Serbia for financial support under project [21] L.D. Ryan, E.W. Kaler, Colloids Surfaces, A 176 (2001) number 172051, D. Richter and J. Allgaier (JCNS-1), 69-83 Forschungszentrum, Jülich, for providing experimen- [22] K. Fukuda, U. Olsson, M. Ueno, Colloids Surfaces, B 20 tal equipment and the synthesis of C18E100. Lutensol (2001) 129-135 ON 50 was a gift from BASF. Glucopon 600 CS UP [23] J.L. Chai, Y.T. Wu , X.Q. Li, B. Yang, L.S. Chen, S.C. was a gift from Cognis. M. Ilić further thanks DAAD Shang, J.J. Lu, J. Chem. Eng. Data 56 (2011) 48-52 (Germany) for a scholarship. [24] T. Sottmann, Curr. Opin. Colloid Interface Sci. 7 (2002) 57-65 REFERENCES [25] D. Byelov, H. Frielinghaus, O. Holderer, J. Allgaier, D. Richter, Langmuir 20 (2004) 10433-10443 [1] I.D. Morrison, S. Ross, Colloidal Dispersions: Suspen- sions, Emulsions and Foams, Wiley, New York, 2002

31 M. ILIĆ et al.: THE INFLUENCE OF ALKYL POLYGLUCOSIDES… Chem. Ind. Chem. Eng. Q. 22 (1) 27−32 (2016)

[26] M. Mihailescu, M. Monkenbusch, H. Endo, J. Allgaier, G. [29] W.C. Griffin, J. Soc. Cosmet. Chem. 1(5) (1949) 311-326 Gompper, J. Stellbrink, D. Richter, B. Jakobs, T. Sott- [30] M. Kahlweit, R. Strey, Angew. Chem. Int. Ed. Engl. 24 mann, B. Farago, J. Chem. Phys. 115 (2001) 9563-9577 (1985) 654-668 [27] A. Kabalnov, U. Olsson, H. Wennerström, Langmuir 10 [31] M. Kahlweit, R. Strey, P. Firman, D. Haase, Langmuir 1 (1994) 2159-2169 (1985) 281-288 [28] C. Frank, H. Frielinghaus, J. Allgaier, D. Richter, Lang- [32] S. Burauer, T. Sottmann, R. Strey, Tenside, Surfactants muir 24 (2008) 6036-6043 Deterg. 37 (2000) 8-16.

MARIJA ILIĆ1 UTICAJ ALKILPOLIGLUKOSIDA (I ALKOHOLA FRANZ-HUBERT HAEGEL2 VISOKOG STEPENA ETOKSILACIJE U ULOZI 3 VESNA PAVELKIĆ POJAČIVAČA) NA FAZNO PONAŠANJE DRAGAN ZLATANOVIĆ4 SNEŽANA NIKOLIĆ-MANDIĆ5 MIKROEMULZIONOG SISTEMA ALEKSANDAR LOLIĆ5 VODA/TOLUOL/TEHNIČKI ALKIL POLIETOKSILAT ZORAN NEDIĆ6 Ispitivan je uticaj aditiva (alkil poliglukosida, Glukopon 600 CS UP, i alkohol-etoksilata 1Faculty of Mining and Geology, C18E100) na fazno ponašanje mikroemulzionog sistema voda/toluol/lutensol ON 50 (tehnički University of Belgrade, Belgrade, oksoalkohol, i-C10E5) u funkciji temperature i sastava sistema. Za određivanje faznog Serbia ponašanja u mikroemulziji korišćeni su vertikalni preseci Gibbs-ovih faznih prizmi (dija- 2Forschungszentrum Jülich, Institut grami oblika tela ribe). Jednofazni region sistema voda/toluol/Lutensol ON 50 dodatkom für Bio- und Geowissenschaften, IBG-3 Agrosphäre, Jülich, alkil poliglukozida se pomera ka nižim temperaturama. Ovakvo ponašanje je suprotno oče- Germany kivanom, polazeći od izrazito hidrofilne prirode glave molekula površinski aktivnog šećera. 3 Institut of Chemistry, Technology Dodatak malih količina hidrofilnog alcohol etoksilata (C18E100) sistemu voda/toluol/Lutensol and Metallurgy, University of ON 50, povecava kapacitet rastvaranja površinski aktivne supstance i pomera jednofazni Belgrade, Belgrade, Serbia region sistema ka višim temperaturama. 4Innovation Center Faculty of Ključne reči: mikroemulzija, toluol, alkil poliglukozid, oksoalkohol etoksilat, adi- Mechanical Engineering, tiv-pojačivač, “riba” dijagram. University of Belgrade, Belgrade, Serbia 5Faculty of Chemistry, University of Belgrade, Belgrade, Serbia 6Faculty of Physical Chemistry, University of Belgrade, Belgrade, Serbia

NAUČNI RAD

32 Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 22 (1) 33−39 (2016) CI&CEQ

V. SANGEETHA1 BIOGAS PRODUCTION FROM SYNTHETIC V. SIVAKUMAR2 SAGO WASTEWATER BY ANAEROBIC 1Department of Food Technology, DIGESTION: OPTIMIZATION AND Kongu Engineering College, TREATMENT Perundurai, Tamil Nadu, India 2Department of Chemical Engineering, Alagappa College of Article Highlights • Technology Campus, Anna Optimization for biogas yield was conducted using response surface methodology • Mixed culture from sago industry sludge can produce effective biogas University, Chennai, India • The optimum condition for biogas production and COD removal was at pH 7 and 32 °C SCIENTIFIC PAPER • Adequacy of the model shows R2 value for COD removal and biogas production was 0.9943 and 0.9880, respectively UDC 662.756.3:628.3 • Abstract DOI 10.2298/CICEQ140612016S Sago processing industries generate a voluminous amount of wastewater with extremely high concentration of organic pollutants, resulting in water pollution. Anaerobic digestion was employed for reduction of COD and maximization of biogas production using synthetic sago wastewater by batch process. Mixed culture obtained from sago industry sludge was used as a source for microorg- anisms. Response surface methodology was used to optimise the variables, such as pH, initial BOD, temperature and retention time. Statistical results were assessed with various descriptives, such as p-value, lack of fit (F-test), coefficient of R2 determination, and adequate precision values. Pareto analysis of variance revealed that the coefficients of determination value (R2) of COD and BOD removal and biogas production were 0.994, 0.993 and 0.988, respectively. The optimum condition in which maximum COD removal (81.85%), BOD removal (91.61%) and biogas production of 99.4 ml/day were achieved was at pH 7 with an initial BOD of 1374 mg/l, and with the retention time of 10 days at 32 °C. Keywords: anaerobic digestion, synthetic sago wastewater, biogas production, chemical oxygen demand, optimisation.

Sago, the common edible starch processed from a rotary drum with low chemical oxygen demand the tubers of cassava (Mannihotesculenta) is one of (COD) and the other from the extraction process the major tuber crops grown in more than 80 which owns a high contaminating load of COD and countries in the humid tropics. In the southern part of biochemical oxygen demand (BOD). Hence, large India, particularly in Tamil Nadu, there are about 800 quantities of processed water up to 15 m3/t of fresh small-scale units of sago industries discharging about cassava root are converted into wastewater, which 40,000 to 50,000 L of sago wastewater and 15 to 30 t must be treated before its release into the environ- of sludge per unit per day [1,2]. Sago processing ment. The amount of water used to produce one ton industries generates two types of wastewater; one of starch ranges from 10-30 m3 and repeated washing resulting from the washing and peeling of cassava in improves the starch quality [3]. Due to stringent environment protection regul- Correspondence: V. Sivakumar, Department of Chemical Eng- ations, it is necessary for the processing industry to ineering, Alagappa College of Technology Campus, Anna Uni- treat wastewater [4]. Hence, it has become manda- versity, Chennai, India. tory for these units to treat the wastewater for safe E-mail: [email protected] Paper received: 12 June, 2014 discharge. There is ample space for an effective and Paper revised: 13 April, 2015 complete treatment system which will ensure a safe Paper accepted: 25 May, 2015

33 V. SANGEETHA, V. SIVAKUMAR: BIOGAS PRODUCTION… Chem. Ind. Chem. Eng. Q. 22 (1) 33−39 (2016) effluent standard limit and hidden energy recovery in MATERIALS AND METHODS the form of biogas before the disposal [1]. Viewing the socio-economic profile of small-scale industrial farm- Sago wastewater ing operations, it is necessary to develop a suitable Preparation of synthetic sago wastewater was low-cost treatment method for treatment of sago reported elsewhere [27] and the physicochemical wastewater [3]. Physical and chemical methods of characteristics of the synthetic sago wastewater were treating the sago wastewater have been unpredict- analysed as per standards of American Public Health able due to the problem of sludge disposal. Biological Association (APHA) [28]. The characteristics of the methods are classified into two types: aerobic pro- wastewater were pH: 6.8, COD: 2286 mg/L, BOD: cesses [5-7], which have limited applicability due to 840 mg/L, TDS: 1237 mg/L, TSS: 537 mg/L, VS: 610 aeration cost [8,9], and anaerobic processes at high mg/L and VSS: 1015 mg/L. treatment rate such as anaerobic filter [10], hybrid Experimental setup and procedure UASB [11], anaerobic rotating biological contactor [12] and fluidized bed [1] systems. The experiment was carried out in a batch reac- From previous literature, it is found that most of tor of 1 L capacity (Figure 1) for different time inter- the researchers successfully used anaerobic pro- vals (4-12 days). Mixed sludge from sago industry cesses for treatment of sago wastewater [11-13]. Var- was used as inoculum (10 vol.%) containing metha- ious anaerobic treatment techniques, including con- nogenic bacteria of Methanosarcina, Methanococ- ventional method, pave way for sustainable environ- coides, Methanoplanus and Methanospirillum. The ment [14-18]. Anaerobic treatment has an advantage pH was adjusted by 1 M HCl or 1 M NaOH using a pH of degrading concentrated waste and producing sig- meter (1283286 Eutech Instruments, Singapore). nificantly less sludge [4]. Because of variations in Initial BOD was varied from 798 to 1702 mg/L by process variables, anaerobic treatment processes are adding sago powder and temperature was adjusted rare at the industrial scale and easily unstable under from 26 to 34 °C with the help of a water bath. The certain circumstances. Therefore, the model has been samples were taken for analysis of COD by the open developed to optimize the treatment process for COD reflux method and for BOD by the standard dilution removal, BOD removal and biogas production, as technique according to APHA [28] and also for biogas functions of the following operating variables: pH, ini- production [13]. tial BOD, temperature and retention time. Response surface methodology (RSM), a mathematical (statistical) technique, is commonly used for developing, analysing, optimizing, and understanding performance of complex variables in an efficient mode. Rec- ently, it has been successfully applied to different wastewater treatment for achieving optimization using experimental designs [19-25]. The advantage of using RSM is the reduction in the number of experiments, compared to a full experimental design at the same level [26]. The objective of the present work is to study the treatment of COD and BOD removal and biogas pro- Figure 1. Experimental setup. duction in anaerobic digestion of synthetic sago wastewater, and also optimizing the effect of the pro- Experimental design cess variables such as pH (4-8), initial BOD concentration (798–1702 mg/L), temperature (26–34 Four factors and five levels of rotatable CCD °C) and retention time (4–12 days) using RSM. A full were carried out with 30 experimental runs. Twenty factorial Central Composite Design (CCD) was emp- four experiments were augmented with six replicates loyed for the optimisation of process variables. at the design centre to evaluate the pure error. Each variable was varied from 5 levels and the relationship between the coded and actual values are described as follows:

()XX− x = i o (1) i Δ X i

34 V. SANGEETHA, V. SIVAKUMAR: BIOGAS PRODUCTION… Chem. Ind. Chem. Eng. Q. 22 (1) 33−39 (2016)

=− + + where xi and Xi are the dimensionless and actual COD Removal (YX11 ) 513.15 15.10 values of the independent variable i, X is the actual ++ + + o 0.061XX23 23.152 10.289 XXX 4120 .009 value of the independent variable at the centre point +−+− 2.539 XX131.062 XX 14 0.003 X 2 X 3 (3) and ΔXi is the step change of Xi corresponding to the −+−−0.0007XX 0.165 XX 6.94 X2 unit variation of the dimensionless value. The vari- 24 34 1 −−−222 ables and its levels are designated as –2, –1, 0, +1 and 0.00008XXX234 0.669 0.322 +2. The second order polynomial equation was used BOD Removal (YX )=− 1418.08 + 55.88 + to describe the effect of independent variables in 21 +++ terms of linear, quadratic and interactions: 0.21XXXXX23412 67.62 23.68 0.007 +−1.52XX 0.115 XX + 0.0004 X X + (4) Y =+ββXXX + β + β + β X + β XX + 13 14 2 3 o112233441212 +−−−−22 ++ββ +++ β β + 0.003XX24 0.46 XX 34 8.14 X 1 0.001 X 2 13XX 1 3 14 XX 1 4 23 X 2 X 3 24 X 2 X 4 (2) −−22 +++++βββββ2222 1.82XX34 0.66 34XXXXXX 3 4 11 1 22 2 33 3 44 4 =− + + Biogas production (YX31 ) 2497.74 91.60 where Y is predicted response, βo is constant coef- ++0.45XXXXX 117.33 + 35.29 + 0.01 ficient, β1, β2, β3 and β4 are linear coefficients, β11, β22, 23412 β β β β β β ++ − 33 and 44 are quadratic coefficients, 12, 13, 14, 23, 1.02XX13 0.52 XX 14 - 0.0007 X 2 X 3 (5) β24 and β34 are cross-products coefficients, and X1, X2, −−−−22 0.002XX24 0.20 XX 34 9.91 X 1 0.0018 X 2 X and X are input variables (pH, initial BOD, 3 4 −−1.96XX22 1.64 temperature and retention time).The data obtained 34 from the response surface methodology on COD To check the estimated regression equation for removal and BOD removal and biogas production the goodness of fit, Fishers F-test was employed and was subjected to the ANOVA. the multiple correlation coefficients R2 was calculated The quality of the fit polynomial model was [21]. The ANOVA results showed the significant res- 2 stated by the coefficient of determination (R ), ponse models with highest (p < 0.05) R2 value of 2 adjusted R , and its statistical significance was deter- 0.994, 0.993 and 0.988 for removal of COD, BOD and mined by F test. The individual effect of each variable biogas production, respectively. The two different as well as the effect of the interaction were deter- tests, such as sequential model sum of squares and mined, and numerical optimisation was performed to model summary statistics are used to decide the ade- determine the optimal solution (maximum COD rem- quacy of various models. prob > F values for the oval, BOD removal and biogas production). quadratic model were less than 0.0001, while the maximum adjusted R2 value and predicted R2 value RESULTS AND DISCUSSION were found to be 0.989 and 0.970 for COD removal. Even though the cubic model was found to be ali- Statistical analysis and fitting of second order ased, prob > F values were greater than 0.05. There- polynomial equation fore, the quadratic model was chosen for further anal- Several factors influence the removal of COD ysis. Adeq Precision measures the signal-to-noise and biogas production from the synthetic sago waste- ratio; typically a ratio greater than 4 is desirable. water, but initial BOD, pH, temperature and retention Thus, signal-to-noise ratios of 56.422, 47.407 and time play important roles. 31.701 for removal of COD, BOD and biogas pro- The response COD, BOD and biogas were mea- duction, respectively, indicate an adequate signal, sured for different runs according to the design matrix and this model can be used to navigate the design carried out based on the design of experiment and the space. The result indicates that the process variables values for random runs are shown in Table 1. CCD are significant factors that affect the response vari- seeks to minimise the integral of the prediction variables. The interacting terms significant for removal of able across the design space. Experimental results COD, BOD and biogas production are shown in Table 2. were analysed, approximating the function of COD and BOD removal and biogas production. The regres- Effect of independent variables on % COD and % sion equations (3)–(5) shown below are obtained after BOD removal the ANOVA: The polynomial equation framed for the above analysis was expressed as three-dimensional surface plots to visualise individual and interactive outcome of factors on the response within the design range.

According to the quadratic model X1, X2, X3 and X4

35 V. SANGEETHA, V. SIVAKUMAR: BIOGAS PRODUCTION… CI&CEQ 22 (1) 33−39 (2016)

Table 1. The design of experiment and response for random runs of anaerobic digestion

Run pH Int. BOD T t COD Removal, % BOD Removal, % Biogas production, ml/day

X1 X2 X3 X4 Yexp Ypre Yexp Ypre Yexp Ypre 1 7 1024 28 10 45.38 47.93 64.62 66.62 72.3 72 2 6 1702 30 8 50.83 49.73 60.81 60.81 54.1 56.26 3 6 798 30 8 43.83 44.22 54.55 54.87 42.6 45.75 4 5 1476 28 6 25.27 26.57 35.27 35.38 26.5 25.54 5 7 1024 28 6 43.24 41.49 55.65 54.54 55.9 52.32 6 6 1250 26 8 38.64 37.8 52.24 51.69 46 45.5 7 6 1250 30 8 63.55 63.95 78.79 81.31 84.6 88.9 8 5 1024 28 6 29.61 30.34 38.08 39.41 21.6 22.37 9 6 1250 30 4 46.02 47.41 56.92 58.4 43.6 48.51 10 8 1250 30 8 56.9 57.23 70.9 72.79 83.1 90 11 4 1250 30 8 16.1 15.06 26.26 24.69 10.1 8.51 12 6 1250 34 8 68.52 68.66 72.23 73.1 63.6 69.41 13 7 1024 32 6 62.33 62.62 76.9 74.66 73.6 70.72 14 5 1476 32 10 48.72 49.96 56 56.36 43.6 42.94 15 7 1476 28 6 46.6 45.78 58.33 57.07 65.3 64.67 16 6 1250 30 8 65.21 63.95 80.41 81.31 89.6 88.9 17 5 1476 32 6 34.98 33.65 45.63 44.06 35.1 34.33 18 6 1250 30 8 64.35 63.95 82.56 81.31 90.1 88.9 19 5 1024 28 10 46.36 45.28 54.4 52.41 37.4 37.87 20 7 1024 32 10 73.52 71.71 80.2 79.34 90.4 87.12 21 7 1476 28 10 50.4 50.94 76.8 75.83 82.6 80.72 22 5 1024 32 6 32.2 31.15 47.1 47.32 34.9 32.54 23 6 1250 30 8 64.2 63.95 81.54 81.31 90 88.9 24 6 1250 30 12 72.25 70.16 83.94 82.78 76.4 76.8 25 5 1024 32 10 46.7 48.74 51.25 52.94 45.2 44.77 26 6 1250 30 8 62.5 63.95 82.54 81.31 88.6 88.9 27 7 1476 32 6 72.6 73.17 76.69 77.94 86.4 81.69 28 7 1476 32 10 80.5 80.99 90.21 89.31 96.3 94.47 29 5 1476 28 10 39.3 40.23 52.4 55.07 35.6 37.42 30 6 1250 30 8 63.86 63.95 82.01 81.31 90.5 88.9

Table 2. ANOVA of the second order polynomial equation for COD and BOD removal and biogas production

Source df COD Removal, % BOD Removal, % Biogas production, ml/day Coefficient p-Value Coefficient p-Value Coefficient p-Value estimate Prob > F estimate Prob > F estimate Prob > F Model 14 7184.55 < 0.0001 8093.31 < 0.0001 18091.68 < 0.0001

X1 1 2667.67 < 0.0001 3469.21 < 0.0001 9959.30 < 0.0001

X2 1 45.46 0.0010 52.96 0.0018 165.90 0.0043

X3 1 1428.36 < 0.0001 687.05 < 0.0001 858.01 < 0.0001

X4 1 776.46 < 0.0001 891.45 < 0.0001 1199.92 < 0.0001

X1×X2 1 64.92 0.0002 42.87 0.0040 84.18 0.0303

X1×X3 1 412.80 < 0.0001 148.66 < 0.0001 67.65 0.0487

X1×X4 1 72.21 0.0001 0.86 0.6373 17.43 0.2934

X2×X3 1 39.28 0.0018 0.57 0.6999 1.89 0.7249

X2×X4 1 1.63 0.4515 44.72 0.0034 13.14 0.3595

X3×X4 1 7.04 0.1293 54.58 0.0016 10.73 0.4065 2 X1 1 1324.67 < 0.0001 1818.38 < 0.0001 2694.50 < 0.0001

36 V. SANGEETHA, V. SIVAKUMAR: BIOGAS PRODUCTION… Chem. Ind. Chem. Eng. Q. 22 (1) 33−39 (2016)

Table 2. Continued

COD Removal, % BOD Removal, % Biogas production, ml/day Source df Coefficient p-Value Coefficient p-Value Coefficient p-Value estimate Prob > F estimate Prob > F estimate Prob > F 2 X2 1 493.56 < 0.0001 944.20 < 0.0001 2461.88 < 0.0001 2 X3 1 196.93 < 0.0001 613.25 < 0.0001 1695.16 < 0.0001 2 X4 1 45.70 0.0010 196.96 < 0.0001 1180.88 < 0.0001 Residual 15 40.96 55.74 220.55 Lack of fit 10 36.87 0.0549 44.96 0.2159 196.27 0.0683 Pure error 5 4.08 10.78 24.28

are the important factors determining Y1 and Y2. The results shown in Figures 2 and 3 indicate that at pH 7, COD and BOD removal are 81.85 and 91.16%, respectively. In mixed sludge, methane producing bacteria are sensitive to mesosphilic temperature range; the graph shows that at 32 °C removal of COD and BOD were achieved at a maximum. Further increase in temperature is not significant in COD removal and also the production of bio gas decreases. Retention time less than 4 days is insufficient for a stable digestion because initial volatile fatty acid concentration was high in the wastewater. After 8-10 days there is a decrease in volatile fatty acid which leads to high COD removal [29]. Therefore, increase in retention time increases the COD removal [1,13]. Similarly, increase in retention time increases the BOD removal due to reduction in organic content of the wastewater Figure 3. Effect of temperature and retention time on BOD removal (%) at optimum pH and initial BOD. caused by anaerobic digestion. Effect of independent variables on biogas production Biological decomposition of organic wastes results in biogas production. The variation in parameters such as pH, initial BOD, temperature and retention time are significant factors affecting the growth of microbes during anaerobic digestion. From Figure 4 it is observed clearly that an increase in retention time proportionately increases the biogas production, which further indicates that a maximum of 99.4 ml/day of the biogas was recovered at optimum condition. Anaerobic digestion can take place at either mesophilic or thermophilic temperatures. Even small changes in temperature from 32–34 °C have been shown to reduce the biogas production rate. Hence, mixed sludge was suitable for biogas recovery in the mesophilic temperature in which anaerobes are active at 32 °C. pH is an important parameter for Figure 2. Effect of temperature and retention time on COD removal (%) at optimum pH and initial BOD. anaerobic digestion. The suitable pH range for methane producing bacteria is 6.8–7.2. The pH range of 5.5–6.5 is suitable for acetogenic bacteria. The pH is maintained with a methanogenic range to prevent the predominance of the acid forming bacteria [4].

37 V. SANGEETHA, V. SIVAKUMAR: BIOGAS PRODUCTION… Chem. Ind. Chem. Eng. Q. 22 (1) 33−39 (2016)

From the results, it was found that the optimum pH is thetic sago wastewater using anaerobic digestion is 7 for biogas production. very effective and the operating variables highly influence the response variables. Hence, this study was a unique attempt to optimise the treatment and production of effective biogas using anaerobic digestion treatment. The RSM model helped to identify the most significant operating factors and the optimum levels with minimum effort and time.

REFERENCES

[1] R. Saravanane, D.V.S. Murthy, K. Krishnaiah, Water Air Soil Pollut. 127 (2001) 15–30 [2] S. Savitha, S. Sadhasivam, K. Swaminathan, Feng Huei Lin, J. Cleaner Prod. 17 (2009) 1363–1372 [3] X. Colin, J.L. Farinet, O. Rojas, D. Alazarda, Bioresour. Technol. 98 (2007) 1602–1607 [4] K. Gurdal, S. Arslan, Environ. Model Assess. 14 (2009) 607–614 Figure 4. Effect of temperature and retention time on biogas [5] P.M. Ayyasamy, R. Banuregha, G. Vivekanandhan, S. production at optimum pH and initial BOD. Rajakumar, R. Yasodha, S. Lee, P. Lakshmanaperumal- samy, World J. Microbiol. Biotechnol. 24 (2008) 2677– Optimization of experimental conditions –2684 The optimum region was identified by consider- [6] M. Rajasimman, C. Karthikeyan, Degradation of starch industry effluent in an inverse fluidized bed reactor, In: ing the maximum removal of COD, BOD and biogas th Proceedings of the 58 Annual meeting of the Indian Ins- production. The optimized process conditions obtained titute of Chemical Engineers, Mumbai, India, 2004, pp. at pH 7, initial BOD 1374 mg/L, temperature 32 °C 27-30 and a retention time of 10 days showed a maximum [7] M. Rajasimman, C. Karthikeyan, Int. J. Environ. Res. 3 COD and BOD removal of 81.85 and 91.16%, res- (2009) 569-574 pectively, and maximum biogas production of 99.4 [8] R.C. Leitao, A.C. Van Haandel, G. Zeeman, G. Lettinga, ml/day with a desirability of 0.991. The results Bioresour. Technol. 97 (2004) 1105–1118 obtained at 10 days of retention time show higher BOD [9] H. Spanjers, J. B. van Lier, Water Sci. Technol. 53 (2006) removal, COD removal and biogas production [2]. 63–76 [10] P. Khageshan, V.S. Govindan, Anaerobic filter for treat- th CONCLUSION ment of sago wastewater, In: Proceedings of 4 National Symposium on Environment, Chennai, India, 1995, pp. In the present study, anaerobic digestion metho- 248–252 dology has been employed for reduction of COD and [11] B. Rajesh, J.S. Kaliappan, D. Beck, Int. J. Environ. Sci. biogas production under optimal condition. The RSM Tech. 3 (2006) 69–77 based CCD was shown to be useful for the design of [12] P. Sujana, T.K. Ramanujam, Bioprocess Eng. 16 (1997) experiments to investigate the effect of the four 163-168 estimated parameters (pH, initial BOD, temperature [13] R. Parthiban, P.V.R. Iyer, G. Sekaran, J. Environ. Sci. 19 (2007) 1416–1423 and retention time) on the response parameters (COD and BOD removal and biogas production). The [14] C.A. Sastry, R. Murahari, K. Saroja, Studies on the treatment of sago mill wastewater, In: Seminar on Environ- results showed good agreement between experimen- mental Pollution, Kerala, India, 1964 tal and predicted values. Based on the ANOVA table, 2 [15] K. Saroja, C.A. Sastry, Report on treatment of Sago the coefficient of determination (R ) values of 0.994 wastes, National Environmental Engineering Research and 0.988 indicate the adequacy of the model for Institute, India,1972 COD removal and biogas production, respectively. [16] C. Tongkasane, Water Waste Treat. 13 (1970) 392–398 Maximum COD reduction and biogas production was [17] M.B. Pescod, N.C. Thanh, Water Sci. Technol. 9 (1977) achieved at a pH value of 7, initial BOD of 1374 mg/L, 563–574 temperature of 32 °C and with the retention time of 10 [18] G. Lettinga, Water Sci. Technol. 52 (2005) 1–11 days. It was identified from this study that COD reduction and biogas production in the treatment of syn-

38 V. SANGEETHA, V. SIVAKUMAR: BIOGAS PRODUCTION… Chem. Ind. Chem. Eng. Q. 22 (1) 33−39 (2016)

[19] A. Danion, C. Bordes, J. Disdier, J. C. Gauvrit, C. Guil- [25] H.C. Yatmaz, A. Akyol, M. Bayramoglu, Ind. Eng. Chem. lard, P. Lanteri, J. Photochem. Photobiol. A: Chemo- Res. 43 (2004) 6035–6039 sphere 168 (2004) 161–167 [26] R.H. Myers, D.C. Montgomery, Response Surface [20] B.K. Korbahti, N. Aktas, A. Tanyolac, J Hazard. Mater. Methodology: Process and Product Optimization Using 148 (2007) 83–90 Design Experiments, John Wiley and Sons, New York, [21] M. Muthukumar, D. Sargunamania, N. Selvakumara, J. 2002 VenkataRao, Dyes Pigments 63 (2004) 127–134 [27] V. Sangeetha, V. Sivakumar, A. Sudha, K.S. Priyenka [22] K. Rajkumar, M. Muthukumar, Environ. Sci. Pollut. Res. Devi, Int. J. Eng. Technol. 6 (2014) 1053-1058 19 (2012) 148–160 [28] APHA, Standard Methods for the Examination of Water [23] R. Sridhar, V. Sivakumar, V. P. Immanuel, J.P. Maran, and Wastewater, American Public Health Association, Environ. Prog. Sustainable Energy 31 (2011) 558-565 New York, 2005 [24] B. Subha, M. Muthukumar, Sci. World J. (2012), Article ID [29] L. Appels, J. Baeyens, J. Degrève, R. Dewil, Prog. 239271, doi: 10.1100/2012/239271 Energy Combust. Sci. 34 (2008) 755-781.

V. SANGEETHA1 PROIZVODNJA BIOGASA IZ VEŠTAČKIH V. SIVAKUMAR2 OTPADNIH VODA IZ PROIZVODNJE SKROBNOG 1Department of Food Technology, BRAŠNA PALME SAGO ANAEROBNOM Kongu Engineering College, DIGESTIJOM: OPTIMIZACIJA I TRETMAN Perundurai, Tamil Nadu, India 2Department of Chemical Engineering, Alagappa College of Prerađivačka industrija skrobnog brašna palme sago stvara veliku količinu otpadnih voda Technology Campus, Anna sa izuzetno visokom koncentracijom organskih zagađivača, što dovodi do zagađenja vode. University, Chennai, India Za redukciju HPK i maksimalnu proizvodnju biogasa korišćena je anaerobna digestija veštačkih otpadnih voda iz industrije skrobnog brašna šaržnim postupkom. Mešana kultura NAUČNI RAD dobijena iz mulja industrije skrobnog brašna je korišćeno kao izvor mikroorganizma. Metodologija odzivne površine je korišćena za optimizaciju faktora procesa, kao što su: pH, početna vrednost BPK, temperatura i vreme zadržavanja. Statistički rezultati su oce- njeni preko p vrednosti, odstupanja (F-test), koeficijenta determinacije R2 i adekvatne preciznosti. Pareto analiza varijansi je pokazala da koeficijenti determinacije (R2) za sma- njenje HKP i BPK i proizvodnju biogasa iznose 0,994, 0,993 i 0,988, redom. Optimalni uslovi pri kojima je postignuto maksimalno uklanjanje HPK (81,85%) i BOD uklanjanje (91,61%) i proizvodnju biogasa (99,4 ml po danu) su: pH 7, početni BPK 1374 mg/l, vreme zadržavanja 10 dana i temperatura 32 °C.

Ključne reči: anaerobna digestija, veštačka optadna voda iz proizvodnje skrobnog brašna, produkcija biogasa, hemijska potrošnja kiseonika, optimizacija.

Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 22 (1) 41−45 (2016) CI&CEQ

MUHAMMAD IMRAN AHMAD1 SUSTAINABLE PRODUCTION OF BLENDED MUHAMMAD SAJJAD1 2 CEMENT IN PAKISTAN THROUGH ADDITION IRFAN AHMED KHAN OF NATURAL POZZOLANA AMINA DURRANI2 1 ALI AHMED DURRANI Article Highlights 1 SAEED GUL • Ordinary Portland cement is partially substituted with rhyolite to reduce cost ASMAT ULLAH1 • Blended cements employing rhyolite are demonstrated to possess satisfactory compressive strength 1 Department of Chemical • Inter-grinding of rhyolite and clinker to produce blended cement shows reduced Engineering, University of energy consumption Engineering and Technology, Peshawar, Pakistan Abstract 2Qadir Enterprises, Peshawar, In this work, pozzolana deposits of district Swabi, Pakistan were investigated Pakistan for partial substitution of Portland cement along with limestone filler. The cement samples were mixed in different proportions and tested for compressive SCIENTIFIC PAPER strength at 7 and 28 days. The strength activity index (SAI) for 10% pozzolana, UDC 666.94(549.1) and 5% limestone blend at 7 and 28 days was 75.5 and 85.0% satisfying the minimum SAI limit of ASTM C618. 22% natural pozzolana and 5% limestone DOI 10.2298/CICEQ141012017A were interground with clinker and gypsum in a laboratory ball mill to compare the power consumption with ordinary Portland cement (OPC) (95% clinker and 5% gypsum). The ternary blended cement took less time to reach the same fineness level as OPC due to soft pozzolana and high grade lime stone, indicating that intergrinding may reduce overall power consumption. Blended cement production using natural pozzolana and limestone may reduce the energy consumption and greenhouse gas emissions. Keywords: ternary blended cement, natural pozzolana, limestone filler, cement production.

Natural pozzolans have been employed in civil cementitious materials in the process resulting in red- works since ancient times [1]. The addition of natural uction in fuel consumption required for clinker form- volcanic rocks to cement or to concrete mixes results ation, CO2 emissions, as well as enhanced durability in improving chemical and physical properties such as and life cycle performance of the concrete structures reduction in heat release when mixed with water, [3]. good ultimate compressive strength, low permeability, The addition of natural pozzolans to form blended high resistance to sulphates and chloride attacks, and cements has been investigated extensively by reduced alkali-silica reaction [2]. Addition of limestone researchers previously demonstrating benefits in red- as a filler increases the early strength development in uction of energy consumption, green house gas emis- concrete; however, chloride ion diffusion may also sions, and cost [4-7]. The addition of natural pozzo- increase depending upon the blending ratio. A careful lans is constrained due to increase in hydration choice of additives and their blending ratios may yield requirements and decrease in early strength develop- cements with enhanced performances. Cement pro- ment [8]. Blending of cement with natural pozzolans duction may become more sustainable by addition of and others additives offers the advantage of exploit- ing characteristic of various materials while compen- Correspondence: M. Imran Ahmad, Department of Chemical sating for disadvantageous features [9-14]. Blended Engineering, University of Engineering and Technology, Pesha- cements are also produced on a commercial scale, war, Pakistan. for example in Algeria, using natural pozzolana and E-mail: [email protected] Paper received: 12 October, 2014 limestone [15]. Paper revised: 12 October, 2014 Paper accepted: 1 June, 2015

41 M. IMRAN AHMAD et al.: SUSTAINABLE PRODUCTION OF BLENDED CEMENT… Chem. Ind. Chem. Eng. Q. 22 (1) 41−45 (2016)

Natural pozzolans are known to react with the these extruded to the ground surface. The estimated calcium hydroxide formed during the reaction of ordi- quantity of deposit above ground level is 9.2 million nary Portland cement with water. The reaction of tons, while the quantity below ground level needs to silica component of pozzolana with calcium hydroxide be estimated after proper drilling. The pozzolana is relatively slow, and produces calcium silicate hyd- deposits of Swabi are whitish in color without any rates. The addition of pozzolana also results in inc- significant variation in size and composition [21]. rease of cementitious aluminates resulting from the reaction of alumina component of pozzolana with MATERIAL AND METHODS calcium hydroxide and sulphate ions [16-19]. This research work attempts to explore the pro- Pozzolana samples were collected and tested duction of blended cements in Pakistan through addi- for chemical, mineralogical composition, using XRF, tion of natural pozzolana for sustainable growth of the XRD, and other properties essential to determine cement, and construction sector. Natural pozzolana feasibility of use as cementitious material. Ordinary deposits are available in different areas of KPK, Portland cement (OPC) was used with natural pozzo- Pakistan such as in Karak, Mohmand agency, Swabi lana from Swabi, Pakistan and high grade limestone and Swat. Bentonite deposits of Karak district have (consisting of more than 95% calcium carbonate) been investigated for partial substitution of ordinary from the quarry of Askari Cement, Nizampur, Pakis- Portland cement in mortars and concrete [20]. tan. The chemical composition of OPC, natural poz- In this paper the natural pozzolana deposits of zolana, and limestone employed in this work are Swabi are investigated for production of ternary shown in Table 1. It may be observed from Table 1 blended cement. Pozzolana deposits are located in that the minimum requirement of oxides as per ASTM Gohatee, on both sides of Swabi-Mardan road as C618, i.e., the sum of silica, alumina, and iron oxides extrusive rocks, i.e., during geological transformation content should be greater than 70%, for natural poz-

Table 1. Chemical composition (%) of the cement, pozzolana and limestone employed in experiments

Material SiO2 Al2O3 Fe2O3 CaO MgO K2O N2O SO3 Cement 20.5 4.89 4.49 61.41 1.65 0.95 0.22 3.59 Pozzolana 70.61 11.97 0.69 1.95 0.61 4.06 0.0 0.09 Limestone 5.25 1.4 1.2 53.0 0.8 0.05 0.03 0.01

Figure 1. X-ray diffractogram of natural pozzolana.

42 M. IMRAN AHMAD et al.: SUSTAINABLE PRODUCTION OF BLENDED CEMENT… Chem. Ind. Chem. Eng. Q. 22 (1) 41−45 (2016) zolana is satisfied. The mineralogical composition of while total quantity of grinding media was of 96 kg. natural pozzolana is shown in Figure 1. The mineral- Media sizes were 72 (12.56 kg), 63 (21.7 kg), 49 ogical composition as determined by X-ray diffraction (22.06 kg) and 39 mm (10.64 kg). The dimensions of bears similarity with the mineralogical composition of the cylinders were 27 mm, length 37 mm (17.17 kg), a natural pozzolana reported previously [15]. and diameter 25 mm, length 31 mm (11.87 kg). The Pozzolana sample was also tested for loss on feed quantity was 5 kg. The ball mill was drained at ignition using BS–FLS–2011–04 standard. The loss on regular time intervals for sieve analysis using 600, 90 ignition was 1.15%, satisfying the maximum of 10% and 45 μm mesh as well as for Blaine fineness. Insol- specification of ASTM C618. It was concluded based uble residue (IR) was determined by the BS–FLS– on the loss on ignition that natural pozzolana under 20051–04 standard. consideration could be mixed with clinker or cement without any drying through external heat source. RESULTS AND DISCUSSION The formulation of blended cement was varied by substitution of ordinary Portland cement with poz- The chemical composition of ordinary Portland zolana ranging from 5 to 22%, while the limestone cement, pozzolana, and limestone employed in this content was maintained constant at 5%. Ordinary work is shown in Table 1. Portland cement used was from Askari Cement Ltd., It may be observed from Table 1 that the mini- Nizampur, Pakistan, with fineness of 289.3 m2/kg and mum requirement of oxides as per ASTM C618, i.e., residue of 10% on 45 μm. Pozzolana and limestone the sum of silica, alumina, and iron oxides content were separately ground to 370 m2/kg and then mixed should be greater than 70%, for natural pozzolana is with ordinary Portland cement in specified ratios, as satisfied. The compressive strength (MPa) at 7 and shown in Table 2. Mortar cubes were casted and 28 days of various blends is shown in Table 2. tested for compressive strength at 7 and 28 days. Mortar cubes were prepared using 1:3 ratio of cement Table 2. Compressive strength of tested composite cement mortars and sand, taking 200 g of cement and 600 g of sand. Cement Pozzolana Limestone Compressive strength, MPa Cube dimensions were 70.1 mm×70.1 mm×70.1 mm. % % % 7 days 28 days ° Curing of cubes was carried out at 27±2 C water 100 0 0 54.7 61.8 temperature in curing tank until the day of testing. 90 5 5 44.3 58.3 Table 3 presents the composition, loss on ignition, 85 10 5 41.3 52.5 specific surface area, i.e., Blaine and residue of var- 80 15 5 40.0 49.6 ious blends. 73 22 5 37.1 48.3 Strength activity index (SAI) was calculated for all the blends to test for minimum specification of 75% The strength activity index calculated using Eq. as per ASTM C618. The strength activity index is (1) for various blends at 7 and 28 days is shown in defined as [22]: Figure 2. 100A It may be observed from Table 2 and Figure 2 SAI = (1) B that increasing the weight percentage of pozzolana above 10% while maintaining limestone percentage where A = average compressive strength of the fixed at 5% resulted in violation of the ASTM C618 blended cement mortar cubes and B = average com- specification, i.e., below the specified limit of 75%. pressive strength of the cement mortar cubes without Alternatively, it may be noted that substitution of more any substitution. than 15% of ordinary Portland cement, using pozzo-

lana and limestone, in cement mortars resulted in sig- The effect of pozzolana substitution on power nificant loss of compressive strength at 7 days. The consumption was investigated by grinding clinker, two blends (80/15/5 and 73/22/5) with significant loss pozzolana, high grade limestone and gypsum mix, of compressive strength at 7 days show a recovery of and compared with grinding of clinker and gypsum in strength between 7 and 28 days, as shown by the a laboratory ball mill. The ball mill consisted of a satisfactory SAI value at 28 days. All the cement com- single chamber manufactured by Wuxi Building Mat- posite mortars showed an increase in compressive erial Instrument & Machinery Co, China. The labor- strength after 7 days indicating that between 7 and 28 atory ball mill had a diameter of 560 mm and length of days both the OPC hydration and pozzolanic hydra- 520 mm. The feed size was less than 30 mm as per tion reactions contributed to strength development. the mill requirement. The installed motor was 1.5 kW The reduction in early strength development is char-

43 M. IMRAN AHMAD et al.: SUSTAINABLE PRODUCTION OF BLENDED CEMENT… Chem. Ind. Chem. Eng. Q. 22 (1) 41−45 (2016)

Figure 2. Strength activity indices for composite cement mortars after 7 and 28 days. acteristic to use of natural pozzolans due to the need loying a mix of clinker 68%, pozzolana 22%, lime- for longer duration of moist-curing. However, this stone 5% and gypsum 5% and compared with the shortcoming is partially addressed by addition of lime- grinding to produce OPC, i.e., clinker 95% and gyp- stone filler, which accelerates the reactions in cement sum 5%. The specific surface area obtained after pastes and mortars [15]. regular time intervals for the ternary blended cement The chemical composition and physical pro- and OPC are shown in Table 4. perties such as Blaine fineness and residue are shown in Table 3. Table 4 Grinding test for OPC and blended cement

Blaine fineness, m2/kg Table 3. Chemical composition (%) and physical properties of Grinding time, min OPC Blended cement blended cement composites 10 163.3 245.8 OPC/Pozzolana/Limestone blend Component 15 183.8 301.9 100/0/0 90/5/5 85/10/5 80/15/5 73/22/5 20 239.5 363.4 SiO2 20.5 22.84 25.65 28.22 32.48 25 275.3 408.5

Al2O3 4.89 5.21 5.79 6.3 7.15

Fe2O3 4.49 4.43 4.5 4.56 4.68 It may be observed from Table 4 that blended CaO 61.41 58.56 55.17 51.39 45.88 cement was easier to grind compared to OPC due to MgO 1.65 1.47 1.22 1.00 0.62 reduced percentage of clinker in blended cement. The

K2O 0.95 1.06 1.20 1.32 1.51 grinding tests indicate that inter-grinding of pozzolana

N2O 0.22 0.22 0.23 0.23 0.23 and limestone with clinker and gypsum in the cement

SO3 3.59 3.26 3.09 2.88 2.54 mill may reduce the power consumption required to LOI 2.12 3.74 3.52 3.67 3.69 achieve the specified fineness for cement. Physical properties Blaine fine- 289.3 330.6 335.6 342.2 361.2 CONCLUSIONS ness, m2/kg Residue, % 8.50 10.80 12.60 14.40 15.40 In this paper, partial substitution of ordinary Portland cement with natural pozzolana from the Swabi district and limestone was investigated for pro- It may be observed from Table 2 that the sub- duction of ternary blended cement in Pakistan. Poz- stitution of OPC with pozzolana and limestone results zolana percentage was varied from 5 to 22%, to in reduction in early strength development. However, substitute OPC, while maintaining a fixed percentage the specific surface area, i.e., Blaine fineness inc- of limestone. It is concluded based on the results of reases, indicating that pozzolana is a relatively soft compressive strength at 7 and 28 days that up to 15% additive. The partial substitution of OPC with pozzo- OPC may be substituted with 10% pozzolana and 5% lana and limestone may also result in further saving in limestone in cement composites. The use of ternary energy consumption through reduction in power con- blended cement composites in civil works would need sumption requirement in the cement mill for grinding further investigation to determine the long term strength of clinker and gypsum. The grinding test for blended development, i.e., after 28 days, as well durability cement was carried out in a laboratory ball mill emp- through sulphate resistance, and permeability tests.

44 M. IMRAN AHMAD et al.: SUSTAINABLE PRODUCTION OF BLENDED CEMENT… Chem. Ind. Chem. Eng. Q. 22 (1) 41−45 (2016)

The power consumption required in the cement [9] J. Bai, B. Sabir, S. Wild, J.M. Kinuthia, Mag. Concr. Res. mill was inferred through investigation on a laboratory 52(2) (2000) 153-162 ball mill indicating that intergrinding of pozzolana and [10] R. Bleszynski, R.D. Hooton, M.D.A. Thomas, C.A. limestone with clinker and gypsum may reduce the Rogers, ACI Mater. J. 99(5) (2002) 499-508 overall power consumption of cement production. [11] M.F. Carrasco, G. Menendez, V. Bonavetti, E.F. Irassar, Cem. Concr. Res. 35(7) (2005) 1324-1331 Acknowledgement [12] J.M. Khatib, J.J. Hibbert, Constr. Build. Mater. 19(6) The support of Askari Cement, Nizampur, Pakis- (2005) 460-472 tan is acknowledged for providing assistance in test- [13] Z. Li, Z. Ding, Cem. Concr. Res. 33(4) (2003) 579-584 ing of composition and properties of samples. [14] G. Menendez, V. Bonavetti, E.F. Irassar, Cem. Concr. Compos. 25(1) (2003) 61-67 REFERENCES [15] M. Ghrici, S. Kenai, M. Said-Mansour, Cem. Concr. Compos. 29 (2007) 542-549 [1] S.H. Kosmatka, B. Kerkhoff, W.C. Panarese, N.F. Mac- [16] V.L. Bonavetti, V.F. Rahhal, E.F. Irassar, Cem. Concr. leod, R.J. McGrath, Design and Control of Concrete Mix- Res. 31(6) (2001), 853-859 th ture, 7 ed., Portland Cement Association, Skokie, IL, [17] M. Heikal, H. El-Didamony, M.S. Morsy, Cem. Concr. 2002 Res. 30(12) (2000) 1827-1834 [2] ACI, ACI Mater. J. 91(4) (1994), 410-426 [18] G. Kakali, S. Tsivilis, E. Aggeli, M. Bati, Cem. Concr. Res. [3] P.K. Mehta, ACI SP-178, Farmington Hills, MI, 1998, pp. 30(7) (2000) 1073-1077 1-25 [19] A.M. Neville, Properties of Concrete, 3rd ed.,. Pearson, [4] P.K. Mehta, Cem. Concr. Res. 11 (1981) 507-518 Harlow, 1981 [5] F. Massazza, Cem. Concr. Compos. 15(4) (1993) 185-–214 [20] J. Mirza, M. Riaz, A. Naseer, F. Rehman, A.N. Khan, Q. [6] A. Tagnit-Hamou, N. Pertove, K. Luk, ACI Mater. J. Ali, Appl. Clay Sci. 45 (2009), 220-226 100(1) (2003) 73-78 [21] A.H. Kazmi, S.G. Abbas, Metallogeny and Mineral [7] L. Turanli, B. Uzal, F. Bektas, Cem. Concr. Res. 35 Deposits of Pakistan, Orient Petroleum Inc., Islamabad, (2005) 1106-1111 2001 [8] B. Uzal, L. Turanli, Cem. Concr. Compos. 34 (2012) 101- [22] ASTM, Standard Specification for Coal Fly Ash and Raw –109 or Calcined Natural Pozzolan for Use in Concrete. C618- -08a. ASTM International, West Conshohocken, PA, 2008.

MUHAMMAD IMRAN AHMAD1 ODRŽIVA PROIZVODNJA CEMENTNE MEŠAVINE MUHAMMAD SAJJAD1 IZ PAKISTANA UZ DODATAK PRIRODNOG 2 IRFAN AHMED KHAN PUCOLANA AMINA DURRANI2 1 ALI AHMED DURRANI U ovom radu istraživana su nalazišta pucolana u oblasti Svabi (Swabi, Pakistan), radi 1 SAEED GUL parcijalne zamene portland cementa uz dodatak krečnjaka kao punioca. Komponente su 1 ASMAT ULLAH mešane u različitim odnosima, a uzorci betona su testirani na pritisnu čvrstoću posle 7 i 28 1Department of Chemical dana. Indeks pritisne čvrstoće (SAI) za mešavine od 10% pucolana i 5% krečnjaka posle 7 Engineering, University of i 28 dana bio je 75,5 i 85,0%, redom, što zadovoljava minimalnu SAI granicu prema ASTM Engineering and Technology, C618. Smeša sa 22% prirodnog pucolana i 5% krečnjaka je samlevena sa klinkerom i Peshawar, Pakistan gipsom u laboratorijskom kugličnom mlinu radi poređenja potrošnje energije sa običnim 2 Qadir Enterprises, Peshawar, portland cementom (OPC) (95% klinkera i 5% gipsa). Trojnoj cementnoj mešavini je tre- Pakistan balo manje vremena da se postigne ista finoća kao OPC zbog prisustva mekih faza pucolana i visokog udela krečnjaka, što pokazuje da je moguće smanjiti ukupnu potrošnju ener- NAUČNI RAD gije. Evidentno da proizvodnja cemente mešavine, koristeći prirodni pucolan i krečnjak, može da smanji potrošnju energije i emisiju gasova staklene bašte.

Ključne reči: trojna cementna mešavina, prirodni pucolan, krečnjački punioc, proizvodnja cementa.

Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 22 (1) 47−53 (2016) CI&CEQ

YUEHAO LUO1 EXPLORING INSTANTANEOUS MICRO- ROBERT SMITH2 3 -IMPRINTING TECHNOLOGY ON SEMI- LORK GREEN -CURED EPOXY RESIN COATING BASED 1School of Engineering and ON RELATIONSHIP BETWEEN FORMING Applied Science, The George Washington University, PRECISION AND CURING DEGREE Washington D.C., USA 2College of Applied Sciences and Article Highlights Technology, Ball State University, • Micro-dimple imprinting method is explored and adopted Muncie, IN, USA • Relationship between plastic deformation capacity and curing degree is investigated • 3 School of Engineering, Boston Instantaneous micro-imprinting method is exploited • Forming precision of instantaneous micro-imprinting can surpass 90% University, Boston, MA, USA

Abstract SCIENTIFIC PAPER Nano/micro-imprinting technology based on polymer material coating has UDC 678.686:544:66 attracted increased attention throughout the world in the past several decades, and it is at present progressively developing into a hot topic, in which how to DOI 10.2298/CICEQ150119018L improve the manufacturing efficiency is becoming the urgent issue to be resolved. Polymer’s curing process is exactly complicated and sophisticated, which involves simultaneously performing physical and chemical changes, when the curing reaction reaches certain level, the system will abruptly transform into insoluble, non-melting gel with rapidly increased viscosity and rigidity, which can generate fixed deformation under persistent external pressure. In this paper, the plastic deformation capacity of epoxy resin in the curing process is investigated by the micro-dimple imprinting experiment, and the relationship between forming precision and curing degree is ascertained adopting the DSC (differential scanning calorimetry) method. In addition, the instantaneous micro-imprinting technology based on the micro-grooves is explored, and the experimental results indicate that the forming precision can surpass 90%. The paper will establish a novel avenue for application of the nano/micro imprinting technology into practical engineering. Keywords: polymer material, semi-cured, DSC, micro-imprinting,micro grooves, forming precision.

Entering the 21st century, with the increasingly fluid engineering applications, such as natural gas serious energy crisis and climate warming, saving pipelines, navigation, agriculture, industry, airplanes, energy and reducing emission of greenhouse gases everyday life, etc. [1-3]. It has been illustrated that has turned into an important issue, in which lowering bio-inspired micro-structured morphology has the the friction force on the contact surfaces between apparent drag-reducing effect in turbulence with moving objects can take a critical role. For protecting smooth skin as baseline [4, 5]. Expanding the appli- surfaces from corrosion and obtaining the smooth- cations of bio-inspired drag-reducing technology to ness, polymer coatings have been widely exploited in coating surfaces is an effective way to reduce friction force and conserve resources. Traditional imprinting methods with micro-textures have been explored and Correspondence: Y. Luo, School of Engineering and Applied investigated comprehensively; however, for the pur- Science, The George Washington University, Washington D. C., 20052, USA. pose of holding the perfect forming effect, long dur- E-mail: [email protected]; [email protected] ation with persistent external pressure is imperative Paper received: 19 January, 2015 [6]. This can lower the manufacturing efficiency and Paper revised: 18 May, 2015 Paper accepted: 13 June, 2015 restrict the practical engineering applications; there-

47 Y. LUO et al.: EXPLORING INSTANTANEOUS MICRO-IMPRINTING… Chem. Ind. Chem. Eng. Q. 22 (1) 47−53 (2016) fore, new and feasible manufacturing ways should be MATERIALS AND METHODS researched further. For the current nano/micro imprinting methods, Curing of epoxy resin the time-points exerting pressure on semi-cured coat- The curing and cross-linking mechanism of ther- ings are not concentrated into the short time-zone [7- mosetting epoxy resin is very complex and intricate. -9], and the consequence is that the duration should The chemical kinetics and physical interactions both be very long to maintain the high forming precision. exist in the curing process [10-12], which has so far Reducing the duration for improving the manufact- not been understood thoroughly. The phenomeno- uring efficiency has thus developed into an urgent logical method is the most popular and commonly problem. If the duration could be condensed into the adopted in studying epoxy resins. The following semi- time of polymer padding into the hollows of the mold, empirical formula, which has laid the basic foundation while obtaining satisfactory forming precision after to investigate the curing mechanism of epoxy resin, curing, the manufacturing efficiency will improve has been applied widely [13]: greatly with good machining quality. Therefore, ascer- α taining the most appropriate time-points to exert ins- d =−−E a α n A0 exp(1 ) (1) tantaneous pressure on the semi-cured coating in the dtRT curing process is a critical issue. In this article, the where: α is the curing degree, t is the duration time, relationship between forming precision and curing A0 is the frequency factor, Ea is the activation energy, degree is investigated by the micro-dimple imprinting R is the universal gas constant, T is the absolute experiment, and the instantaneous micro imprinting temperature, and n is the reaction cardinal number. technology is preliminarily explored adopting the The epoxy resin used in this paper (AW-01 micro-grooved mol. The results show that when the epoxy resin) can be cured completely at temperature curing degree is located into 0.8-0.9, the coating has higher than 30 °C. SEM images of the epoxy resin the best plastic deformation capacity, and forming coating at curing temperatures 40-100 °C are shown precision can be more than 90%. in Figure 1a-d. It can be seen that the internal structure of the epoxy resin cured at 45 °C is more meti-

Figure 1. SEM images and wear scars of epoxy resin coating curing at different temperatures.

48 Y. LUO et al.: EXPLORING INSTANTANEOUS MICRO-IMPRINTING… Chem. Ind. Chem. Eng. Q. 22 (1) 47−53 (2016) culous than that cured at 95 °C. In order to validate Micro-dimple imprinting the wear property, wearing experiments were con- To ascertain the plastic deformation capacity of ducted on an SRV machine (produced in Germany). epoxy resin at different time-points, the micro-dimple The schematic drawing and key parts of the tester are imprinting experiment is designed and as shown in shown in Figures 2 and 3. In the experiment, the force Figure 4. The system was comprised of parallel light, exerted on the smooth steel ball was 25 N, the stroke coating sample, substrate, convex lens imaging sys- of the rectilinear reciprocating movement was set as tem, displaying screen, rigid probe, etc. 1.6 mm, the frequency was 15 Hz, and the wear scars The basic steps of the micro-dimple imprinting corresponding with the internal structures are illus- experiment are as follows: 1) exerting instantaneous trated in Figures 1a-1d. It can be concluded that with displacement on the semi-cured coating surface by decreasing curing temperature, the width of wear the rigid probe; 2) epoxy resin coating gradually scars drops from 2.51 to 1.87 mm, which implies that becomes cured; 3) measuring the depth of micro- the anti-wear property of epoxy resin is gradually imp- dimple by the precise scanning method. The forming roved. Additionally, finer structures can enable better precision of the imprinting experiment is defined as mechanical and chemical properties, such as anti-cor- the ratio of depth of micro-dimple and instantaneous rosion, tensile strength, anti-acid, hardness, rough- displacement. The feeding displacement of rigid ness, etc. [14-16]. Therefore, the curing temperature probe can be exactly exerted by the screw rotation ° of 45 C was used in this paper. with the precision of 1 μm. In the experiment, the epoxy resin is covered on the plat aluminum substrate by the airless spraying method, and the depth of liquid film is 300 µm, which can be measured by the liquid film gauge, as shown in Figure 5. The three- dimensional morphology and parameters on cured epoxy resin coating can be obtained by the highly precise scanning, as shown in Figure 6a-c.

Figure 2. Schematic drawing of wearing testing experiment.

Figure 5. Measuring the depth of epoxy resin liquid film.

The obtained results from experiments in which the curing temperature is fixed at 45 °C and duration is varied are displayed in Table 1. It can be concluded

that with increasing duration, the forming precision Figure 3. Key parts of SRV wearing tester.

Figure 4. Schematic diagram of micro-dimple imprinting method.

49 Y. LUO et al.: EXPLORING INSTANTANEOUS MICRO-IMPRINTING… Chem. Ind. Chem. Eng. Q. 22 (1) 47−53 (2016)

Figure 6. Three-dimensional morphologies on cured epoxy resin coating.

Table 1. Results of micro-dimple imprinting experiments; T = 45 °C

Duration, min Depth of dimple, μm Forming precision, % Duration, min Depth of dimple, μm Forming precision, % 100 145.2 72.6 114 180.8 90.4 102 156.2 78.1 115 180.6 90.3 105 172.2 86.1 117 180.4 90.2 107 174.2 87.1 118 180.2 90.1 110 179.8 89.9 119 180.0 90.0 111 180.4 90.2 120 179.0 89.5 112 181.2 90.6 121 176.6 88.3 113 182.2 91.1 122 176.0 88.0

first increases and then decreases gradually. More- capacity. The relationship between forming precision, over, the maximum of forming precision can be curing degree and heat flow is shown in Figure 7, greater than 90%. indicating that the best plastic deformation corresponds to the curing degree (α) range of 0.8-0.9. Differential scanning calorimetry (DSC) Furthermore, with increasing curing degree, the form- For studying the properties and characteristics ing precision first increases and reaches a maximum, of epoxy resin system more comprehensively, the and then decreases. The measured hardness of kinetics of the curing reaction were monitored by a semi-cured coating fitting to exert the external pres- differential scanning calorimeter (DSC) and inves- sure varied from 68-72 HA (Figure 8). tigated by the constant heating method, which is an important basis of composite materials forming tech- Exploration on instantaneous micro-imprinting nology and can supply the basic theoretical found- technology ation [18-22]. Bio-inspired micro-structured surfaces have an DSC is the most popular phenomenological apparent drag reduction effect in turbulent flowing method to explore and investigate the curing kine- conditions [25-29], and have extensively been put into matics of epoxy resin, which mainly involves two application in fluid engineering [30, 31]. However, the operation models: isothermal DSC and dynamic DSC practical applications are limited by the manufacturing [23,24]. The isothermal DSC can obtain the reaction efficiency to some extent. The instantaneous micro- heat at some certain temperature, the dynamic DSC imprinting technology with micro-grooved mold is pre- can obtain the reaction heat (ΔHtotal) or the residual liminarily exploited to improve the manufacturing effi- reaction heat (ΔHres). In addition, the glass transition ciency according the aforementioned conclusion. If temperature can be measured by the shift of the the operating time is not enough, the polymer will not baseline. The isothermal DSC experiment was con- fill the hollows of the mold completely [32]. Mean- ducted at 45 °C. while, if the operating time is too long, the imprinting efficiency will be affected and restricted. The equation RESULTS AND DISCUSION for shortest duration required for filling the hollows completely has been derived by Heyderman [33]: Relationship between different parameters η S 2 11 When the forming precision is greater than 90%, t =−0 () (2) f 2P hh22 it can be regarded as the best plastic deformation f0

50 Y. LUO et al.: EXPLORING INSTANTANEOUS MICRO-IMPRINTING… Chem. Ind. Chem. Eng. Q. 22 (1) 47−53 (2016)

Figure 7. Relationship between heat flow rate, forming precision and curing degree.

where tf is the duration, η0 is the viscosity of polymer, In the experiments, the operating time was

S is area of pattern, P is the pressure, h0 is the height about 1 s, which was greater than the filling time and of the pattern before imprinting, and hf is the height of much less than that of available imprinting. The pattern after imprinting. When the curing degree of images of micro-grooved mold and cured micro-struc- epoxy resin is 0.8-0.9, the duration tf is 0.25-0.35 s. tured coating morphology are shown in Figure 9a and b, and the cross section curves are illustrated in Figure 10a and b. It can be seen that the depth and semi-width of micro-grooved mold were 41.4 and 95.3 µm, and those of the coating surface were 38.3 and 95.3 µm, respectively. From the forming precision analysis (Figure 11), it can be concluded that the forming errors only exist in the vertical direction (forming precision can be more than 90%, 38.3/414×100 = = 92.5%), and there is no error in the horizontal direction.

Figure 8. Measuring the hardness of semi-cured epoxy resin.

Figure 9. Three-dimensional morphology of micro-grooved model and imprinted coating.

51 Y. LUO et al.: EXPLORING INSTANTANEOUS MICRO-IMPRINTING… Chem. Ind. Chem. Eng. Q. 22 (1) 47−53 (2016)

Figure 10. Cross section curves of micro-grooves and morphology on imprinted coating.

Figure 11. Forming precision analysis of instantaneous micro-imprinting technology.

CONCLUSIONS REFERENCES

In this paper, the relationship between plastic [1] J. Wu, G. Huang, Q. Pan, J. Zheng, Y. Zhu, B. Wang, deformation capacity and curing degree was inves- Polym.48 (2007) 7653-7659 tigated by the micro-dimple imprinting experiment, [2] K. Formela, J.T. Haponiuk, Iran. Polym. J. 23 (2014) 185- and the following conclusions can be drawn: –194 1) When the curing degree is located in the [3] S.B. Mortazavi, Y. Rasoulzadeh, A.A. Yousefi, A. range of 0.8-0.9, the forming precision can be greater Khavanin, Iran. Polym. J. 19 (2010) 197-205 than 90% and the epoxy resin has the best plastic [4] L. Yuehao, Z. Deyuan, L. Yufei, Oil Gas-Eur. Mag. 40 (2014) 96-97 deformation capacity. [5] D.Y. Zhang, Y.H. Luo, H.W. Chen, Oil Gas-Eur. Mag. 37 2) The relationships between forming precision, (2011) 85-90 curing degree and heat flow rate are ascertained for [6] Y. Luo, Y. Liu, Adv. Mater. Res. 884–885 (2014) 378–381 the first time. With increasing curing degree, the form- [7] J. Du, Z.Y. Wei, S.Z. Li, Y.P. Tang, J. Micromech. ing precision first increases and reaches a maximum, Microeng. 23 (2013) 05526 and then decreases. [8] H.D. Rowland, A.C. Sun, P.R. Schunk, W.P. King, J. 3) The instantaneous micro-imprinting techno- Micromech. Microeng. 15 (2005) 2414 logy adopting the micro-grooved template was preli- [9] L.L. Gomes, Y. Kawano, J. Appl. Polym. Sci. 79 (2001) minarily explored, and the experimental results showed 910-919 that it has good forming effect and quality. The form- [10] W. Wu, J. Zhang, Iran. Polym. J. 21 (2012) 763-769 ing precision in the horizontal direction is about 100%, [11] F. Fenouillot F, P. Cassagnau, J. C. Majeste, Polym. 50 and the forming precision in the vertical direction is (2009) 1333-1350 greater than 90%.

52 Y. LUO et al.: EXPLORING INSTANTANEOUS MICRO-IMPRINTING… Chem. Ind. Chem. Eng. Q. 22 (1) 47−53 (2016)

[12] L. Yuehao, L. Yufei, Z. Deyuan, Eur. J. Envir. Civil. Eng. [23] H. Ismail, N.F. Omar, N. Othman, J. Appl. Polym. Sci. 32 (2015) 25-31 121 (2011) 1143-1150 [13] A. Campanella, L. M. Bonnaillie, R. P. Wool, J. Appl. [24] R. Scaffaro, D. Tzankova, M. Nocilla, L. Mantia, Polym. Polym. Sci. 112 (2009) 2567-2578 Degrad. Stabil. 90 (2005) 281-287 [14] Y. Luo, D. Zhang, H. Chen, Adv. Sci. Lett. 5 (2012) 49–55 [25] Y. Luo, Y. Liu, D. Zhang, E.Y. K. Ng, J. Mech. Med. Biol. [15] B. Adhikari, D. De, S. Maiti, 25 (2000) 909-948 14 (2014) 1450029 [16] Z. Deyuan, L. Yuehao, C. Huawei, Pipeline Gas J. 238 [26] L. Carli, O. Bianchi, R. S. Mauler, J. S. Crespo, Polym. (2011) 58–60 Bull. 67(2011)1621-1631 [17] P. Nevatia, T.S. Banerjee, B. Dutta, A. Jha, A.K. Naskar, [27] L. Yuehao, Z. Deyuan, Appl. Surf. Sci. 282 (2013) 370– A.K. Bhowmick, J. Appl. Polym. Sci. 83 (2002) 2035-2042 –375 [18] T. Chaubey, H. Arastoopour, J. Appl. Polym. Sci. 119 [28] Y.H. Luo, J. Mech. Med. Biol. 15 (2015) 1530002 (2011) 1075-1083 [29] K. Tan, L. Chunxi, H. Meng, Z. Wang, Polym. Test. 28 [19] R. Sonnier, E. Leroy, L. Clerc, A. Bergeret, J. M. Lopez- (2009) 2-7 -Cuesta, A.S. Bretelle, P. Lenny, Polym. Test. 27 (2008) [30] Y. Luo, D. Zhang, Appl. Mech. Mater. 44–47 (2011) 151– 901-907 –157 [20] L. Yuehao, Z. Deyuan, Oil Gas-Eur. Mag. 38 (2012) 213– [31] L. Yuehao, L.Z. Deyuan Z.C. Huawei, Adv. Mater. Res. –214 189–193 (2011) 9–12 [21] H. Ismail, R. Nordin, A. M. Noor, Iran. Polym. J. 12 (2003) [32] D. Zhang, Y. Luo, L. Xiang, C. Huawei, J. Hydrodyn. 23 373-380 (2011) 204–211 [22] V.V. Rajan, K. Dierkes, R. Joseph, J. Noordermeer, Prog. [33] L.J. Heydermana, H. Schift, C. Davida, J. Gobrechta, T. Polym. Sci. 31 (2006) 811-834 Schweizerb, Microelectron. Eng. 54 (2000) 229-245.

YUEHAO LUO1 ISTRAŽIVANJE TEHNOLOGIJE TRENTUNOG ROBERT SMITH2 MIKRO-ŠTAMPANJA NA POLUOČVRSLIM 3 LORK GREEN PREMAZIMA EPOKSI SMOLA ZASNOVANIM 1School of Engineering and NA ZAVISNOSTI IZMEĐU PRECIZNOSTI Applied Science, The George Washington University, OBLIKOVANJA I STEPENA OČVTŠĆAVANJA Washington D.C., USA 2College of Applied Sciences and Tehnologija nano/mikro-štampanja baziranim na premazima polimernih materijala privlači Technology, Ball State University, sve veću pažnju širom sveta u poslednjih nekoliko decenija, da bi danas postala top tema Muncie, IN, USA u kojoj je kako poboljšati efikasnost proizvodnje urgentan problem koji treba rešiti. Proces 3 School of Engineering, Boston očvršćavanja polimera je komplikovan i sofisticiran postupak, koji uključuje istovremeno University, Boston, MA, USA obavljanje fizičkih i hemijskih promena. Kada reakcija očvršćavanja dostigne određeni nivo, sistem se naglo transformiše u nerastvorljiv i netopljiv gel sa naglo povećanim visko- NAUČNI RAD zitetom i rigidnošću, koji mogu da dovedu do stalne deformacije pod stalnim spoljašnim pritiskom. U ovom radu, kapacitet plastične deformacije epoksidne smole u procesu očvrš- ćavanja je praćen utiskivanjem mikro-otisaka, a odnos između preciznosti oblikovanja i stepena očvršćavanja je utvrđen metodom diferencijalne skaning kalorimetrije. Pored toga, proučavana je i tehnologija trenutnog mikro-štampanja zasnovana na pravljenju mikro-otisaka, a eksperimentalni rezultati ukazuju da preciznost oblikovanja može biti preko 90%. Rad utvrđuje novi put primene tehnologije nano/mikro štampanja u praktičnom inženjer- stvu.

Ključne reči: polimerni material, semi-čvršćavanje, diferencijalna skaning kalori- metrija; mikri-štampanje, mikro-otisci, preciznost oblikovanja

Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 22 (1) 55−64 (2016) CI&CEQ

ZORANA BOLTIĆ1 CONTINUOUS IMPROVEMENT CONCEPTS AS MIĆA JOVANOVIĆ2 2 A LINK BETWEEN QUALITY ASSURANCE AND SLOBODAN PETROVIĆ IMPLEMENTATION OF CLEANER VOJISLAV BOŽANIĆ3 MARINA MIHAJLOVIĆ4 PRODUCTION – CASE STUDY IN THE GENERIC PHARMACEUTICAL INDUSTRY 1Hemofarm A.D, Vršac, Serbia 2 Faculty of Technology and Article Highlights Metallurgy, University of Belgrade, • CI as a relationship between QA and CP implementation in the generic pharmaceutial Belgrade, Serbia 3 industry Faculty of Organizational • Application of Lean and Six Sigma tools for process improvement and link to other Sciences, Belgrade, Serbia known concepts 4 Innovation Centre of the Faculty of • Evaluation of the production systems in terms of CI, considering both quality and effi- Technology and Metallurgy, ciency University of Belgrade, Belgrade, Serbia Abstract The subject and the research objective presented in this article is establishing SCIENTIFIC PAPER of the relationship between quality assurance and implementation of cleaner production in the generic pharmaceutical industry through the comprehensive UDC 615:66:338.45 concept of continuous improvement. This is mostly related to application of DOI 10.2298/CICEQ150430019B Lean and Six Sigma tools and techniques for process improvement and their link to other known concepts used in the industrial environment, especially manufacturing of generic pharmaceutical products from which two represent- ative case studies were selected for comparative analysis, also considering relevant regulatory requirements in the field of quality management, as well as appropriate quality standards. Although the methodology discussed in this conceptual and practice oriented article is strongly related to chemical engineering, the focus is mainly on process industry, i.e., production systems, rather than any specific technological process itself. The scope of this research is an engineering approach to evaluation of the production systems in terms of continuous improvement concepts application, considering both quality aspects and efficiency of such systems. Keywords: quality assurance, cleaner production, pharmaceutical industry, continuous improvement, lean, six sigma.

The present article is based on the application of process of the products under consideration, but also specific continuous improvement techniques on to quality systems relevant for production of pharma- relevant performance measures (PM) of different pro- ceuticals. This implies optimization techniques of processes in the pharmaceutical industry and evaluation cess design versus treatment processes (including of their effectiveness in the actual industrial environ- techno-economic analysis) with the aim of introducing ment [1,2]. The processes that are studied in this cleaner process technologies and how it relates to the work are related to design and operation of the equip- improvement of the quality systems as well. Conti- ment, as well as to the flow of the manufacturing nuous improvement techniques also represent a part of chemical engineering, but in the sense of process Correspondence: M. Jovanović, Faculty of Technology and Met- industry, i.e. in the area of production systems, rather allurgy, University of Belgrade, Karnegijeva 4, 11000 Belgrade, than any individual technological process. An original Serbia. E-mail: [email protected] and novel approach is applied to the evaluation of Paper received: 30 April, 2015 industrial processes, as well as a new industrial eng- Paper revised: 17 June, 2015 ineering methodology and its application in practice is Paper accepted: 15 November, 2015

55 Z. BOLTIĆ et al.: CONTINUOUS IMPROVEMENT CONCEPTS… Chem. Ind. Chem. Eng. Q. 22 (1) 55−64 (2016) studied through the evaluation of the production sys- effectively make decisions. Selection of the appropri- tems, which are determined both by their capacities ate key performance indicators, as well as measure- and quality aspects. The scope is therefore the pro- ment methodologies, represents a critical factor for cesses in the generic pharmaceutical industry, which, the success of measurement and analysis proce- as such, has certain specificities that have to be con- dures. sidered when evaluating the applicability of certain In the new philosophy of sustainability, the con- techniques in the area of continuous improvement. cept of sustainable development is replaced with the Pharmaceutical industry is involved in develop- term sustainable success, as a result of the orga- ment, production and marketing of drugs, i.e., phar- nization's ability to achieve and maintain long-term maceutical products approved by relevant regulatory goals, and this is the main novelty in revised standard authorities. Pharmaceutical companies can have their ISO 9004: 2009 [3] in relation to the second revision own research and development – originators, i.e., of this standard [4]. ISO 9001 [5] specifies the requi- innovators, or produce generic drugs – products bio- rements for the quality management system and is equivalent to the innovator’s, i.e., comprising the focused on its effectiveness in complying with custo- same medicinal or therapeutic substances as the ori- mer requirements. ISO 9004 gives additional guid- ginal drug. Pharmaceutical companies based on their ance for the organizations willing to move further than own development made a significant progress in these requirements, to solving the needs and expect- treatment of numerous diseases, but increasing ations of all interested parties and their satisfaction healthcare costs over time also resulted in the inc- through systematic and continuous performance imp- reased use of generic drugs with certain advantages rovement. Therefore, it represents a powerful tool for over more expensive original drugs, especially among the management, and the sustainable success of the the poorer population of patients. Because of this, organization is developed through its capability to generic pharmaceutical production represents a signi- satisfy needs and expectations imposed by its custo- ficant portion of the world pharmaceutical industry. mers and other interested parties in a balanced way Either based on its own development or generic, and over a long period of time. Self-assessment is pharmaceutical industry is regulated through numer- used to identify areas for improvement and inno- ous laws and regulations in the area of patenting, vation, establish priorities and develop action plans testing and ensuring safety and efficiency of drugs, aimed at sustainable success. The results of the i.e., product release to the market. organization's evaluation according to ISO 9004 may The aim of this work is to evaluate the relation- represent a valuable input for management review as ship between process performance measurement required by ISO 9001 (i.e., for information review from with key indicators of success on the company level monitoring, measurement and analysis, as stipulated and the concept of introducing cleaner production in ISO 9004:2009), but this self-assessment process principles in the generic pharmaceutical industry, also has a certain potential to be a learning tool linked to the continuous improvement programs. enabling the improved involvement of interested parties whose needs and expectations have to be pro- THEORETICAL PART perly understood, as the key element of the organization's maturity model. It is a well known fact that pharmaceutical pro- Cleaner production and eco-efficiency are part duction is one of the most regulated industrial sectors of the consideration in the area of the environment, as and that quality is the key factor that determines each prescribed by the social responsibility standard. manufacturing system, including product character- These are the strategies for satisfaction of human istics, its appearance, duration, maintenance, but also needs through more efficient utilization of resources its supply and relevant documentation. For the pur- and producing less pollution and waste. Significant pose of achieving sustainable success, the organiza- focus is to introduce the improvements at the source tions regularly measure, analyse and review their instead at the end of process or activity. Cleaner and performance, including evaluating the progress in safer production, as well as eco-efficiency approaches achieving planned results compared to their mission, include also the improvement of the sustainable prac- vision, policies, strategy and goals on all levels and in tice, introducing new technologies or processes with all relevant processes and functions in the organi- lower consumption of material and energy (i.e., util- zation. Measurement and process analysis is used to ization of its renewable resources) and rationalization track this progress, gather and have all the necessary of water consumption. Cleaner production assumes information available to evaluate performance and elimination or safe management of toxic and hazard-

56 Z. BOLTIĆ et al.: CONTINUOUS IMPROVEMENT CONCEPTS… Chem. Ind. Chem. Eng. Q. 22 (1) 55−64 (2016) ous materials/wastes, as well as improvement of the accordance with Lean philosophy in the area of products and services projects. Cleaner production in manufacturing and supply chain [11]. There are seven general represents a contemporary approach in pre- common forms of waste, namely: transport, inventory, venting the creation of pollution that provided the movement, waiting, over-production, over-processing greatest contributions in the production sector, espe- and defects, and one of the tools introduced by the cially in industry [6]. Lean methodology is Kaizen (Ky = change and Zen = Six Sigma approach to continuous improvement = for the better, generally being translated as con- uses the methodology known as DMAIC (Define, tinuous improvement through solving problems). Measure, Analyze, Improve, Control) for process imp- Kaizen is a quick, intensive look at the process with rovement from beginning to end [7]. In each of these the aim of improvement. It gathers customers, sup- phases, appropriate tools are being used, such as pliers, support and people performing the work, where project plan, SIPOC (Supplier, Input, Process, Output, the latter is key for its success. The actual process is Customer), process mapping and different team man- observed and waste identified and eliminated through agement techniques in the define phase, measure- establishing a new process but as a continuous effort ment system analysis, histogram and Pareto in mea- rather than reaching perfection in one step. Produc- sure, FMEA (Failure Mode and Effects Analysis), tion system established in Toyota is the starting basis “5 whys” and fishbone diagram for cause-and-effect for numerous literature sources in this field, including analysis, statistic process control and control charts in The Machine That Changed the World: The Story of the control phase of the implemented improvements. Lean Production [12] and Lean Thinking [13]. According to well known statistical principles, The international conference on harmonisation sigma represents a Greek letter for standard deviation of technical requirements for registration of pharma- showing how much the measured results are distant ceuticals for human use (ICH) in its guidance Q10 from the average of the observed set of data, i.e., [14] describes one comprehensive model for an representing a measure of process variability and its effective pharmaceutical quality system that is based capability to work without errors and as little variation on International Organization for Standardization as possible. Measures are necessary in order to (ISO) quality concepts, includes applicable good determine whether the process of interest is stable manufacturing practice (GMP) regulations, and com- and predictable, as well as how much variation is pre- plements ICH “Q8 Pharmaceutical Development” and sent [8]. Six Sigma means that the interval between ICH “Q9 Quality Risk Management.” ICH Q10 is a both upper and lower limit of the process specification model for a pharmaceutical quality system that can be and the average of the results obtained from that implemented throughout the different stages of a pro- actual process is 6 standard deviations. This number duct lifecycle. Much of the content of ICH Q10 appli- of standard deviations is inversely proportional to the cable to manufacturing sites is currently specified by probability of defects and in fact illustrates how much regional GMP requirements. Implementation of ICH of the obtained results is within the interval required Q10 throughout the product lifecycle should facilitate by the customer, i.e. increasing the sigma level of the innovation and continual improvement and strengthen process decreases the cost and increases product- the link between pharmaceutical development and ivity and customer satisfaction. In order to state that a manufacturing activities. process is “sigma” it is not allowed to have more than In this work, the selected criteria used to deter- 3.4 defects per million opportunities [9], therefore six mine the existence of either Lean or Six Sigma appro- sigma virtually represents a measure of quality prac- ach (or both) in the case studies subject to evaluation tically aiming for perfection. ISO/TR 10017:2003 also are as follows: 1) data based approach/performance provides guidance on the selection of appropriate sta- measurement, 2) link to the customer, 3) proactive tistical techniques that may be useful to an organiz- thinking and 4) tools and techniques. Additionally, the ation in developing, implementing, maintaining and link between the processes subject to relevant case improving a quality management system in compli- studies (quality assurance – QA and cleaner product- ance with ISO 9001 [10]. ion – CP) was analyzed against relevant regulatory Lean, on the other hand, is a concept originating requirements and quality standards. from Toyota in the 1950s and representing a common sense and practical approach to solving problems CASE STUDY with the focus on identifying and eliminating waste from the process. Over the years, Toyota started a Starting hypotheses are based on the impor- global transformation in almost all industry sectors in tance of measurement as the key element to control,

57 Z. BOLTIĆ et al.: CONTINUOUS IMPROVEMENT CONCEPTS… Chem. Ind. Chem. Eng. Q. 22 (1) 55−64 (2016) manage and improve processes on one side and the implementation of cleaner production in the generic link of continuous improvement to the cleaner pro- pharmaceutical industry [2], as the case study CS II. duction concept in the industrial environment on the Description of the analyzed case studies other. Performance measurement represents the basis of most continuous improvement programs, enabling CS I in the area of quality assurance was aimed for example the implementation of cleaner production to evaluate the implementation of the modern step by step, using appropriate tools and techniques. approach based on measures and key performance This is especially true for generic pharmaceutical indicators in a pharmaceutical company using the industry where it is necessary to change technological example of the delivery time improvement through procedures in a highly regulated environment in con- decreasing the number of deviations and time spent trast to minimizing harmful effects of the production on unnecessary investigations. Problems subject to processes in the end-of-pipe approach. the analysis were identified in both cases based on Therefore, the scope of this study is to evaluate relevant information and data available in the indus- the following: trial information system related to actual processes, • Performance measurement system repre- and appropriate corrective actions and suggested sents the comparison of the current values with the improvements were implemented through the des- predefined objectives and enables feedback to the cribed improvement projects. Results were discussed participants in the process - this approach should relative to the previously established objectives: in the result in the improvement of the quality management area of quality assurance, significant decrease of total system and continual adjusting of the performance number of deviations was shown – more than 50% [1]. measures. On the other hand, as CS II for introducing the • Implementation of the continuous improve- cleaner production principles into the processes ment program based on process performance mea- within the generic pharmaceutical industry, the case surement leads to decreased costs related to different of tablets coating was selected, as one of the most forms of waste, i.e., redundant engagement of common and widely used operation in the pharma- resources. ceutical production in general. The conclusion was • Efficient elimination of waste from the pro- made that the option of preventing pollution through cesses with positive effects on quality, environment, modifying the formulation has a significant advantage working conditions and social responsibility at the both considering financial benefits and minimization same time, can be accomplished through a unique of waste, i.e. negative impact on the environment [2]. approach of the continuous improvement program Data based approach/Performance measurement implementation, introducing step by step improve- In CS I, methods used for gathering the inform- ments in individual areas. ation related to key performance indicators (KPI) are This evaluation is performed using the examples selected as feasible and appropriate for the organiz- of quality assurance processes [1] as the case study ation, which is also one of the requirements in ISO CS I, and dealing with volatile organic compounds 9004:2009 (Figure 1). On the other hand, the KPI is (VOCs) emissions, as one of the challenges for the

Figure 1. Main characteristics of the key indicators, performance measures and data analysis performed.

58 Z. BOLTIĆ et al.: CONTINUOUS IMPROVEMENT CONCEPTS… Chem. Ind. Chem. Eng. Q. 22 (1) 55−64 (2016) selected to enable its quantification and make fea- Organization performance measurement thus sible to the organization to set measurable goals, represents an important source of data for a system- identify, monitor and predict trends and implement atic approach to evaluation of the available inform- corrective, preventive and improvement measures, as ation to assure that this information serves as a basis needed. It is important to ensure that measurable and for making important decisions. Improvement, inno- reliable information are available for the implement- vations and learning can be applied on products, pro- ation of these corrective actions when the perform- cesses and technology, as well as on organizational ance is not in compliance with the previously estab- structures, management systems, infrastructure and lished goals. work environment and the basis for this is the cap- As efficient utilization of resources is also one of ability of the organization to draw conclusions based the requirements that need to be assured by the on relevant data analysis. One of the key benefits of management system, it is shown in the CS II that the continuous improvement, besides improving perform- processes are established to monitor and optimize ance through enhanced capability of the organization these resources in order to ensure their effective and and harmonizing the improvement activity at all efficient use. Therefore, the organization continually levels, in accordance with its strategic orientation, is measures their current utilization to identify oppor- the adaptability, i.e., flexibility to respond fast enough tunities for the improvement in this area. In addition, to opportunities, mostly in terms of increased compe- the environmental impact was measured as part of titiveness on the market. continual monitoring to enable the organization to Security management within the supply chain for identify and implement the appropriate risk manage- which the KPI is selected for evaluation in CS I is ment in this area. related to numerous other business aspects, and relevant requirements covering these management sys- Link to the customer tems are defined in ISO 28001:2007 [15]. According In the CS I the selected KPI is decomposed as a to this standard, the supply chain represents the inter- performance indicator in relevant functions and levels linked ensemble of resources and processes begin- in the organization to support reaching the higher ning with the source of raw materials, through product level objectives in line with the strategy and corporate and/or services supplied to the end user by means of policies (Figure 2). different kinds of transport. The system of supply chain managements in the CS I is also subject to con- Strategic goals tinuous improvement. Cleaner production as a concept evaluated in CS II can be indirectly linked to the population of Management Review Business pharmaceutical industry customers through sus- Strategy tainable development taking care of limited environmental capacity to receive a specific quantity of waste, mostly related to industrial pollution. The relationship between the elements of procurement, production and the consumers in a broader sense is shown implying the need to develop preventive acti- Key Performance vities through the product life cycle. This is addition- Indicators ally supported by the fact that cleaner production represents an application of the comprehensive preventive strategy of environmental protection on the production processes, products and services with the aim to increase overall efficiency and decrease health and environmental risks (UNEP), meaning preser- Processes vation of resources, water and energy, reduced appli- Continuous Improvement cation of toxic and hazardous raw materials and reduced quantities and toxicity of all emissions and wastes at the source of the production process ins- Performance tead of the End-of-Pipe technologies (Figure 3). measures Cleaner production does require significant Figure 2. Link between strategic goals, key indicators and changes in the organization and its processes. This performance measures.

59 Z. BOLTIĆ et al.: CONTINUOUS IMPROVEMENT CONCEPTS… Chem. Ind. Chem. Eng. Q. 22 (1) 55−64 (2016)

Figure 3. Example of EOP – a typical adsorption process for VOCs [16]. can be accomplished through an approach that can identified risk in the field of environmental manage- bring benefits to all interested parties, which is also ment. Furthermore, cleaner production is by its def- shown based on the techno-economic analysis per- inition a proactive approach to dealing with the envi- formed as part of CS II. ronmental impact of the processes in all industries. Using prevention in formulating environmental pro- Proactive thinking tection policy is also required according to ISO Processes and their relationships, as shown in 140001 [17] and cleaner production completely sup- CS I, are regularly reviewed and appropriate actions ports this concept complying with the organization's are taken for their improvement. In the course of practices for environmental management systems in planning and management of these processes, the achieving the common objectives to continually imple- organization's environment was considered and anal- ment the improvements. Živković et al. have done a yzed, mostly taking into account relevant regulatory case study for the Oil Refinery Belgrade that con- and other requirements. The planning process firmed an improvement of environmental perform- according CS I considers the established needs of the ances using the ISO 14001 standard [18]. organization to develop or apply new process characteristics as an added value, which at the same time Tools and techniques represents one of the requirements of ISO 9004:2009. Pareto diagrams were used in CS I as a tool to In CS II, the focus is on process optimization focus attention to problems offering the greatest and new technologies. When it comes to the infra- potential for improvement. This technique is based on structure which is planned, enabled and managed by the rule that 20% of causes lead to 80% of problems the organization in an efficient and effective manner, (Pareto principle). Fishbone diagram or root cause appropriate attention is also given to safety and pro- analysis was also applied in this study (CS 1) rep- tection, the elements of this infrastructure linked to resenting a visual description of individual contribu- production processes, as well as its impact on the tions to a certain problem. The fish head represents a working environment, overall efficiency, costs and problem to be solved while the bones serve to picture capacities. At the same time, working environment is the root causes classified in 4–6 main categories: mat- evaluated in accordance with applicable laws and erials, methods, people, machines, the environment other regulations in the field of environmental, health and measurement. Additional categories or further and safety management. Therefore, technological classifications within individual categories are also opportunities were considered in CS II to improve org- possible depending on their importance. anization performance in different areas, including Suitable working environment, as a combination product realization and interaction with interested of human and physical factors, assumes also max- parties. It is shown that the organization is also imum efficiency and minimization of waste, which is considering the integration of the environmental asp- one of the key principles of Lean manufacturing and ects in the design and product development, as well the basis of the CS II at the same time. Furthermore, as developing the specific processes to minimize the cleaner production focuses on the root causes of

60 Z. BOLTIĆ et al.: CONTINUOUS IMPROVEMENT CONCEPTS… Chem. Ind. Chem. Eng. Q. 22 (1) 55−64 (2016) problems related to the environment and not the financial benefits. As a result, a natural link is deve- consequences, which is one of the main goals in the loped between the environmental goals and improve- six sigma analyze phase of the improvement process. ment projects initiated to increase productivity, achieve better yields, implement savings in materials RESULTS AND DISCUSSION and decrease the cost of waste management. There- fore, cleaner production becomes an important ele- Results of the two case studies analysis against ment of the comprehensive strategy of performance pre-defined criteria are summarized in Table 1. improvement and efficiency increase through the Additionally linked to CI through regulations and application of production concepts in accordance with quality standards (a–e are referred in Table 1): Lean principles. a. ISO 9004 requirements; Based on the studied facts, as well as numerous b. ICH Q10 and GMP requirements; literature findings in this area, but also with regards to c. System of ecological management; practical experience in management of the quality d. ISO 140001 focus on prevention; assurance processes and pharmaceutical production e. Similar to PDCA in ISO 9001. in general, it is obvious that the continuous improve- Linking the objectives in the field of environmen- ment programs can be considered as the link tal protection with improved productivity, material between establishing sustainable process perform- savings and decreased cost of handling and waste ance measurement systems and implementation of management, cleaner production is imposed as an cleaner production in the pharmaceutical industry inseparable part of the overall strategy to improve (Figure 4). performances and increase total efficiency. Environ- Requirements for the Pharmaceutical Quality mental protection aspects may be regarded as an System are stipulated in ICH Q10 [14] and related to important motivation factor to come to innovative sol- quality assurance processes as described in Good utions leading to both safety increase and significant Manufacturing Practice. Performance measures eva-

Table 1. Comparative analysis against the suggested Continuous Improvement Criteria

Continuous Improvement (CI) CS I related to QA processes CS II related to CP Criteria Data based approach/ data gathered to support the selected KPIa Monitoring and optimization of resources Performance measurement PMs established in relevant functionsb Measuring the environmental impact Customer orientation KPI and PMs link to strategy to achieve flexibility in Indirectly through sustainable developmentc terms of market requirements through KPI Techno-economic analysis selection for supply chain management Proactive thinking Regular review of processes and measures for Focus on optimization and new technologies their improvement Evaluation against applicable laws and regulations Integration of environmental aspects in design and Product developmentd Tools and techniques Pareto Lean manufacturing Fishbone Kaizen philosophye

Figure 4. Continuous improvements as the link between quality assurance and cleaner production principles.

61 Z. BOLTIĆ et al.: CONTINUOUS IMPROVEMENT CONCEPTS… Chem. Ind. Chem. Eng. Q. 22 (1) 55−64 (2016) luation in terms of quality management processes is represents a logical path adopted by an organization one of the key activities required for achieving the that has successfully implemented the basics continuous improvement as one of the most important required by ISO 9001, as it provides broader focus on requirements of the Pharmaceutical Quality System the quality management system through a wider given in ICH Q10. At the same time, system of eco- model based on processes. It relates to needs and logical management, as one of the cleaner production expectations of all interested parties and gives ins- elements, is also an instrument to recognize and tructions for systematic and continual improvement of solve environmental problems based on the conti- overall performance of an organization. Additionally, nuous improvement concepts. Implementation of the ISO 9004 also comprises broader requirements for Q10 model should facilitate continual improvement, in management of the resources and their efficient use, order to identify and implement appropriate product which again brings performance measurement within quality improvements, process improvements, vari- the quality system in relation to cleaner production ability reduction, innovations, and pharmaceutical principles through the concept of continuous impro- quality system enhancements, thereby increasing the vement (Figure 5). ability to fulfill a pharmaceutical manufacturer’s own It is clear that in both case studies the focus was quality needs consistently. Quality risk management on the contemporary requirements of the ICH Q10

Figure 5. The link between performance measurement and cleaner production through the concept of CI. can be useful for identifying and prioritizing areas for related to establishing the process performance continual improvement. indicators in critical areas within the Pharmaceutical The results of the performed study may also be Quality System and implementation of the continuous summarized are follows: improvement concept in general, leading to evalu- • single methodology for improvement of qua- ation of the overall organization performance through lity management and industrial processes through ISO 9004 and enhanced model based on processes, continual monitoring and control of relevant perform- as well as taking into consideration the needs and ance measures; expectations of all interested parties, which defines • correlation between different regulatory the CI as a common term both for quality manage- requirements for pharmaceutical industry to enable ment and cleaner production principles. compliance through implementation of a single concept of continuous improvement; CONCLUSION • dissemination of continuous improvement philosophy and knowledge in the area of integrated Measures and KPIs represent an important environmental and quality management, as well as element of the CI concept, which on the other hand safety, all in accordance with the main principles and plays the key role in the modern Quality Management key aspects of social responsibility. System (QMS) of the pharmaceutical company. The Performance measurement plays the key role in appropriate application of the process performance the area of quality management, providing insight in measurement system actually means measuring the parts of the process where change and improvement current values of the specific parameters against the is required, necessary feedback as the basis for objectives and providing the feedback to relevant further improvements, as well as relevant information participants in the process. This approach should lead for analysis and evaluation of the achieved perform- to the continuous improvement of the QMS, as well as ance. When it comes to models based on Total Qua- performance measures in various processes, includ- lity Management (TQM), process management and ing sustainable environmental protection. customer orientation are regarded as the key factors Applying the appropriate Lean and Six Sigma for implementation. On the other hand, new ISO 9004 tools and techniques, further significant problems

62 Z. BOLTIĆ et al.: CONTINUOUS IMPROVEMENT CONCEPTS… Chem. Ind. Chem. Eng. Q. 22 (1) 55−64 (2016) categories can be identified to be gradually solved [7] M. Brassard, L. Finn, D. Ginn, D. Ritter, The Six Sigma leading to completely eliminating, e.g., an entire class Memory JoggerTM II, A Pocket Guide of Tools for Six of deviations or environmental impacts when it comes Sigma Improvement Teams, GOAL/QPC, 1994 to harmful emissions for example. Additionally, links [8] A.W. Roberts, D.E. Varberg, Faces of Mathematics: An Introductory Course for College Students, Harper and and precise correlations can be determined between Row, New York, 1982 such performance measures managed locally and [9] C. Gyigi, N. DeCarlo, B. Williams, Six Sigma for Dum- higher level objectives and the analysis can be exp- mies, Wiley Publishing, Inc., 2005 anded to other processes supporting these object- [10] International Organization for Standardization (2003), ives. In this regard, the approach and concept of ISO/TR 10017:(2003), Guidance on statistical techniques application of appropriate tools and analysis methods for ISO 9001:2000 shown and developed in this work may be readily [11] J. Liker, The Toyota Way: 14 Management Principles used in all similar improvement projects through dif- from the World's Greatest Manufacturer, McGraw-Hill, ferent areas of the quality system and generic phar- New York, 2004 maceutical production. [12] J.P. Womack, D.T. Jones, D. Roos, The Machine That Changed the World: The Story of Lean Production, Free Acknowledgements Press, A Division of Simon&Schuster, Inc., New York, The authors are grateful to the Ministry of Sci- 1990 ence and Technological Development of the Republic [13] J.P. Womack, D.T. Jones, Lean Thinking. Free Press, A of Serbia for the support (project TR 34009). Division of Simon&Schuster, Inc., New York, 1996 [14] ICH Harmonized Tripartite Guideline, Pharmaceutical REFERENCES Quality System Q10. EMEA, 2008 [15] International Organization for Standardization, ISO [1] Z. Boltić, N. Ružić, M. Jovanović, S. Petrović, Accredit. 28001:(2007), Security management systems for the Qual. Assur. 15 (2010) 629-636 supply chain - Best practices for implementing supply [2] Z. Boltić, N. Ruzić, M. Jovanović, M. Savić, J. Jovanović, chain security, assessments and plans - Requirements S. Petrović, J. Cleaner Prod. 44 (2013) 123–132 and guidance [3] International Organization for Standardization, ISO [16] European Commission (2011), Best Available Tech- 9004:(2009), Managing for the sustained success of an niques (BAT) Reference Document for Common Waste organization - A quality management approach. Geneva, Water and Waste Gas Treatment/Management Systems ISO. in the Chemical Sector. Industrial Emissions Directive 2010/75/EU (Integrated Pollution Prevention and [4] International Organization for Standardization, ISO Control). Draft 2, 20 July 2011 9004:(2000), Quality management systems - Guidelines for performance improvements. Geneva, ISO. [17] International Organization for Standardization, ISO 140001:(2014), Environmental Management [5] International Organization for Standardization, ISO 9001:(2008), Quality management systems – require- [18] S. Živković, Lj. Takić, N. Živković, Chem. Ind. Chem. Eng. ments. Geneva, ISO. Q. 19 (2013) 541−552. [6] Official Gazette of the Republic of Serbia, No. 17/(2009), Strategy for Implementation of Cleaner Production in the Republic of Serbia

63 Z. BOLTIĆ et al.: CONTINUOUS IMPROVEMENT CONCEPTS… Chem. Ind. Chem. Eng. Q. 22 (1) 55−64 (2016)

ZORANA BOLTIĆ1 KONCEPTI KONTINUIRANOG UNAPREĐENJA KAO MIĆA JOVANOVIĆ2 VEZA IZMEĐU OBEZBEĐENJA KVALITETA I 2 SLOBODAN PETROVIĆ UVOĐENJA ČISTIJE PROIZVODNJE – STUDIJA VOJISLAV BOŽANIĆ3 4 SLUČAJA U GENERIČKOJ FARMACEUTSKOJ MARINA MIHAJLOVIĆ INDUSTRIJI 1Hemofarm A.D, Beogradski put b.b, 26300 Vršac, Srbija Predmet i cilj istraživanja koje je predstavljeno u ovom radu jeste uspostavljanje veze 2Tehnoško-metalurški fakultet, Univerzitet u Beogradu, između obezbeđenja kvaliteta i uvođenja čistije proizvodnje u generičkoj farmaceutskoj Karnegijeva 4, 11000 Beograd, industriji kroz sveobuhvatni koncept kontinuiranog unapređenja. Ovo se u najvećoj meri Srbija odnosi na primenu “lean” i “šest sigma” alata i tehnika za unapređenje procesa i njihovu 3Fakultet organizacionih nauka, povezanost sa drugim poznatim konceptima koji se koriste u industrijskom okruženju, a Jove Ilića 154, 11000 Beograd, posebno proizvodnji generičkih farmaceutskih proizvoda, gde su za potrebe komparativne Srbija analize odabrane dve reprezentativne studije slučaja, uzimajući u obzir i relevantne regu- 4 Inovacioni centar Tehnološko- latorne zahteve u oblasti menadžmenta kvalitetom, kao i odgovarajuće standarde kvaliteta. -metalurškog fakulteta, Univerzitet Iako je metodologija razmatrana u ovoj konceptualnoj i praktičnoj studiji usko povezana sa u Beogradu, Karnegijeva 4, 11000 hemijskim inženjerstvom, akcenat je u najvećoj meri stavljen na procesnu industriju, Beograd, Srbija odnosno proizvodne sisteme, pre nego na pojedinačne tehnološke procese. U tom smislu, NAUČNI RAD predmet ovog istraživanja jeste inženjerski pristup evaluaciji proizvodnih sistema u po- gledu primene koncepta kontinuiranog unapređenja, uzimajući u obzir kako aspekte kvaliteta, tako i efikasnost tih sistema.

Ključne reči: obezbeđenje kvaliteta, čistija proizvodnja, farmaceutska industrija, kontinuirano unapređenje, “lean”, “šest sigma”.

64 Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 22 (1) 65−73 (2016) CI&CEQ

ALEKSANDAR GOLUBOVIĆ1 INFLUENCE OF SOME SOL-GEL SYNTHESIS IVANA VELJKOVIĆ2 1 PARAMETERS OF MESOPOROUS TIO2 ON MAJA ŠĆEPANOVIĆ PHOTOCATALYTIC DEGRADATION OF MIRJANA GRUJIĆ-BROJČIN1 NATAŠA TOMIĆ1 POLLUTANTS DUŠAN MIJIN3 4 Article Highlights BILJANA BABIĆ • Anatase nanopowders were synthesized by sol–gel method using tetrabutyl titanate as 1Institute of Physics, University of precursor • XRPD data showed slight growth of crystallites in synthesized samples (from 24 to 35 Belgrade, Belgrade, Serbia nm) 2Institute for Multidisciplinary • Raman scattering data confirmed the anatase as dominant TiO2 phase Research, University of Belgrade, • The BET showed that specific surface area was greater at the lower temperature of Belgrade, Serbia calcination 3 Faculty of Technology and • Photodegradations were comparable with Degussa P25 for C.I. Reactive Orange 16 Metallurgy, University of Belgrade, and phenol Belgrade, Serbia 4Institute of Nuclear Sciences Abstract

“Vinča”, University of Belgrade, Titanium dioxide (TiO2) nanopowders were produced by sol-gel technique from Belgrade, Serbia tetrabutyl titanate as a precursor by varying some parameters of the sol-gel synthesis, such as temperature (500 and 550 °C) and the duration of SCIENTIFIC PAPER calcination (1.5, 2 and 2.5 h). X-ray powder diffraction (XRPD) results have UDC 54:66:544.526.5 shown that all synthesized nanopowders were dominantly in the anatase phase, with the presence of a small amount of rutile in samples calcined at 550 DOI 10.2298/CICEQ150110020G °C. According to the results obtained by the Williamson-Hall method, the anatase crystallite size was increased with the duration of the calcination (from 24 to 29 nm in samples calcined at lower temperature, and from 30 to 35 nm in samples calcined at higher temperature). The analysis of the shift and line-

width of the most intensive anatase Eg Raman mode confirmed the XRPD results. The analysis of pore structure from nitrogen sorption experimental data described all samples as mesoporous, with mean pore diameters in the range of 5-8 nm. Nanopowder properties have been related to the photocatalytic activity, tested in degradation of the textile dye (C.I. Reactive Orange 16), carbofuran and phenol. Keywords: nanostructures, anatase, X-ray diffraction, Raman scattering.

Photocatalysis is a well-known process mostly bility. Nanocrystalline TiO2 is essentially a cheap and employed to degrade or transform organic and inorg- biocompatible wide band-gap semiconductor with an anic compounds, and the kinetics depend on catalyst involving photogenerated holes and photocatalytic surface area, availability of active sites, pore sizes, capabilities for organic pollutants [1-3]. Namely, many number and nature of trapped sites, as well as on organic compounds can be decomposed in aqueous adsorption/desorption characteristics. TiO2 is an solution in the presence of TiO2 powders or coatings important photocatalyst mainly because of its strong illuminated with near ultraviolet (UV) or visible light. oxidizing power, non-toxicity and long-term photosta- The structural, morphological, optical and photocatalytic properties of TiO nanocrystals are strongly 2 dependent on the synthesis process [4,5]. Among the Correspondence: A. Golubović, Institute of Physics, University of Belgrade, Pregrevica 118, 11080 Belgrade, Serbia. various synthesis methods, the sol-gel method has E-mail: [email protected] recently attracted a lot of attention, since it is simple Paper received: 10 January, 2015 and cost-effective way of producing nanostructured Paper revised: 7 April, 2015 Paper accepted: 30 June, 2015 anatase TiO2 with tailored properties.

65 A. GOLUBOVIĆ et al.: INFLUENCE OF SOME SOL-GEL SYNTHESIS… Chem. Ind. Chem. Eng. Q. 22 (1) 65−73 (2016)

Many factors influence photocatalytic reactivity a phenol [27,28]. The commercial TiO2 (Degussa of TiO2 which is documented by numerous public- P25) was applied in a number of photodegradation ations in the last decades [6-11]. Generally, anatase processes of pollutants, and we wanted to synthesize is considered a desirable phase for photocatalysis TiO2 nanopowders using various parameters of syn- application as it shows higher activity then rutile [8,12- thesis and to compare the photocatalytic properties of –13]. However, a mixture of anatase and rutile with a such prepared catalysts. The mechanism of the sintered interface, like commercial TiO2 (Degusa photodegradation process is not completely defined, P25), is claimed to be more active then pure anatase as many parameters are involved. According to this, [14-17]. In order to obtain the highest performance, our manuscript is a contribution in understanding of the main challenge is the synthesis of preferably such a complex process. To the best of our know- nanocrystalline anatase TiO2 that enables a balance ledge, this study is original and it was not found in the between major influencing parameters: crystal struc- literature. ture, surface hydroxylation and crystallinity. Several methods of characterization, such as The sol-gel process represents a flexible chem- XRPD, Brunauer-Emmett-Teller (BET) measure- ical route to synthesize various high-performance ments, and Raman scattering were employed in this nanostructured ceramic materials with controlled study to correlate structural and morphological pro- internal morphology and chemistry. Materials with perties of synthesized TiO2 nanopowders and their designed internal nanostructure (entirely intercon- photocatalytic activity under UV light irradiation. nected open nanoporosity, hierarchical, fractal or nanocrystalline solid network) and various possible EXPERIMENTAL chemical compositions (from organic to inorganic) can be obtained in large range of shapes (finely divided Synthesis nanopowders, nanoparticles, thin and thick films, The TiO2 nanocrystals were prepared by a sol- fibers, granular beds and monolithic materials). The gel method. All of the reagents were of analytical sol-gel process is a solution-based technique, where grade and were obtained from commercial sources the material structure is created through chemical and used without further purification. Tetrabutyl reactions in the liquid state, giving the high flexibility titanate (99%, Acros Organics, Belgium) was used as of the process for easy application. the precursor of titania, hydrochloride acid (36.2%,

The photocatalytic efficiency of TiO2 powder Zorka, Serbia) as the catalyst, ethanol (96%, denat- heavily depends on its microstructure and physical ured, Carlo Erba, Italy) as the solvent, and distilled properties, which are in turn determined by the pre- water for hydrolysis. pH of the solution was 7. The paration conditions. Among these, the presence of reagent molar ratio was Ti(OBu)4:HCl:EtOH:H2O = mesopores gives rise to a large surface area, which = 1:0.3:15:4 according to [29], which enabled obtain- offers abundant interaction sites with external mole- ing a stable gel. The process of gelation was carried cules [18]. The photocatalytic process involves the out at 4 °C, where appropriate amounts of Ti(OBu)4, separation of the electron-hole charge pair, their HCl and EtOH were stirred one hour by magnetic stir- transport and trapping to/at the surface, and, finally, rer. After that, an appropriate amount of distilled water their reaction with the desired molecules. These pro- was added in the mixture due to hydrolysis and form- cesses always compete with the charge pair recom- ation of the gel. This gel was “aged” (the process of bination. The nanostructure significantly affects these polycondensation) for two hours, the wet gels were elemental processes based on several reasons. Apart dried at 80 °C, and then calcinated at 500 and 550 °C from a high surface-to-volume ratio, which must be for 1.5, 2 and 2.5 h, to obtain TiO2 nanocrystals. The beneficial for all chemical processes, the first factor is heating and the cooling rates were 135 °C/h. Accord- the quantum confinement and improved reduction/ ing to the calcination conditions (various temperature /oxidation power. The second factor is the practical of calcinations and duration of the calcinations), syn- absence of band bending and the consequent easier thesized samples were labeled as: T500/1.5, T500/2, access of both charged particles to the surface [19]. T500/2.5, T550/1.5, T550/2 and T550/2.5. TiO nanopowders are very efficient compounds 2 Characterization methods for the photodegradation of many pollutants [20,21]. In our investigations, we made a focus on degrad- Generally, instrumental broadening is negligible ation of organic pollutants having different chemical in the case of low crystallinity samples. Broadening of structure. Namely, a textile dye (C.I. Reactive Orange the peaks because of low crystallinity is dominant. 16) [22,23], an insecticide (carbofuran [2,24-26]) and These are fundamentals of X-ray powder analysis.

66 A. GOLUBOVIĆ et al.: INFLUENCE OF SOME SOL-GEL SYNTHESIS… Chem. Ind. Chem. Eng. Q. 22 (1) 65−73 (2016)

Structural analysis of prepared samples was done by Orange 16), 6.86×10-4 M (carbofuran), 4×10-4M (phe- XRPD on an Ital Structures APD2000 diffractometer, nol), respectively. Upon preparation of the solution, using CuKα radiation (λ = 1.5406 Å), angular range: agitation was applied in dark by continuous stirring 20°< 2θ < 90°. Data were collected at every 0.01° in (magnetic stirrer) at 400 rpm to keep the suspension the 20-90° 2θ using a counting time of 80 s/step. MDI homogenous for 90 min. Then, the lamp was Jade 5.0 software was used for calculation of the switched on and the suspension sampled after appro- structural and microstructural parameters. The Wil- priate times of irradiation. The concentration of pollu- liamson-Hall method [30] was applied for the deter- tants was determined after centrifugation of a sample mination of average microstrain and the mean crys- on Mini Spin Eppendorf at 12000 rpm by a UV-Vis tallite sizes, , of the prepared samples. The spectrophotometer (Shimadzu 1700) at appropriate obtained values were compared to the mean crys- wavelength. tallite sizes calculated by the Scherrer formula [31]. The Scheerer formula is an estimate of crystallite size RESULTS AND DISCUSSION calculated from FWHM of all diffractions collected during measurement. XRPD measurements Raman scattering measurements was per- The XRPD measurement confirmed that sol-gel formed in the backscattering geometry at room tem- synthesis resulted with preparation of anatase modi- perature in air, using Jobin-Yvon T64000 triple spec- fication of TiO2, which is clearly indicated with the trometer, equipped with a confocal microscope and a main anatase reflection at 2θ ≈ 25° (JCPDS card no. nitrogen-cooled coupled device detector. The spectra, 21-1272). The samples calcinated at 500 °C were excited by 514.5 nm line of Ar+/Kr+ laser with output found to be phase-pure anatase (Figure 1), with crys- power less than 5 mW to avoid local heating due to tallite sizes growing with increasing calcination time laser irradiation, was recorded with high spectral (Table 1), while the samples calcinated at 550 °C resolution of about 0.7 cm-1. have small amount of rutile impurities, which are The porous structure of anatase samples is confirmed by small peaks at 2θ ≈ 27° in Figure 1 evaluated from adsorption/desorption isotherms of N2 (JCPDS, card no. 21-1276). The presence of rutile in at –196 °C, using the gravimetric McBain method. The calcined anatase samples can be caused both by pH main parameters of the porosity, such as specific value and the temperature of the calcination [34]. In surface area and pore volume, have been estimated our case, the small amount of rutile in samples by BET method from αs-plot [32]. The pore size dis- calcined at 550 °C is caused by the temperature of tribution was estimated from hysteresis sorption data the calcination as pH value is the same (pH 7). by the Barret-Joyner-Halenda (BJH) method [33]. Measurements of photocatalytic activity UV irradiation of a suspension (an appropriate amounts of pollutant and TiO2 powder as the catalyst) was performed in an open flask (100 ml volume) with ® an Osram Ultra-Vitalux 300 W (UV-A) lamp placed 50 cm from the surface of the solution. The light intensity was 40 mW cm–2, and it was measured on the Amprobe Solar-100, Solar Power meter, Beha- Amprobe, GmbH. The textile dye, C.I. Reactive Orange 16, was obtained from the company Bezema, Switzerland, as a gift (commercial name Bezaktiv Orange V-3R) and used without futher purification. Carbofuran (99.2 %) was obtained from FMC, USA. Phenol, p.a. grade, was purchased from Fluka. The photodegradation of organic pollutants was studied by preparing a solution containing known concentration of organic and appropriate amount of TiO2. In a typical experiment, 25 ml of a solution was used, the quantity of TiO2 was 50 mg, whereas the pollutants × -5 solution molarities were 8.1 10 M (C.I. Reactive Figure 1. The XRPD patterns of TiO2 samples, where rutile diffraction is denoted by “R”.

67 A. GOLUBOVIĆ et al.: INFLUENCE OF SOME SOL-GEL SYNTHESIS… Chem. Ind. Chem. Eng. Q. 22 (1) 65−73 (2016)

Table 1. The unit cell parameters and unit cell volume, together with average crystallite size, , of anatase and microstrain obtained by Scherrer and Williamson-Hall methods

Sample Calcination conditions Unit cell parameters Scherrer method Williamson-Hall method 3 Temperature, °C Time, h a and c in Å, V in Å / nm / nm Microstrain, %

T500/1.5 500 1.5 a = 3.784(3) 15 24 0.301 c = 9.53(0) V = 136.4(8)

T500/2 500 2.0 a = 3.789(9) 18 28 0.231 c = 9.52(1) V = 136.7(5)

T500/2.5 500 2.5 a = 3.789(2) 19 29 0.247 c = 9.50(3) V = 136.4(5)

T550/1.5 550 1.5 a = 3.789(1) 24 30 0.108 c = 9.51(5) V = 136.6(1)

T550/2 550 2.0 a = 3.788(7) 28 33 0.077 c = 9.51(4) V = 136.5(7)

T550/2.5 550 2.5 a = 3.789(1) 30 35 0.085 c = 9.53(4) V = 136.8(9)

According to the Scherrer formula, the crystallite size for samples calcinated at lower temperature has been estimated in the range from 15 to 19 nm, while the samples calcinated at higher temperature have higher crystallinity, with crystallite size in the range from 24 to 30 nm, while for these estimated by Wil- liamson-Hall method were in the range from 24 to 29 nm for lower temperature and from 30 to 35 nm for higher temperature. The analysis of XRPD data by the Williamson-Hall method has shown higher microstrain value in the samples calcinated at 500 °C compared to the samples calcinated at 550 °C. In all futher discussion, values of crystalline size evaluated by the Williamson-Hall method were used. Raman scattering measurements The Raman spectra of all synthesized nanopowders are dominated by anatase Raman modes −1 −1 [35,36]: Eg(1) (∼143 cm ), Eg(2) (∼199 cm ), B1g (∼399 − − − cm 1), A +B (∼518 cm 1), and E (∼639 cm 1), as 1g 1g g(3) can be seen from the spectrum of two chosen samples shown in Figure 2. The most intensive Figure 2. The Raman spectra of samples T500/1.5 and T550/2.5. The experimental spectra (circles) are fitted by the sum of Raman Eg(1) mode is positioned between 142.8 and -1 -1 Lorentzians (thin lines). Anatas modes are denoted by “A” 143.5 cm , with linewidths from 9 to 11.5 cm . The and rutile by “R”. dependence of Raman shift on linewidth of this mode is shown in Figure 3. The Eg(1) Raman modes in the by XRPD (24-35 nm), slight shift and broadening samples T500/1.5, T500/2 and T500/2.5, calcined at lower relative to bulk anatase [33] may rather be ascribed to ° temperature (500 C), are more shifted and more defects and disorder in anatase crystal structure, than broadened then the mode in samples calcined at to the phonon confinement effects. The smaller line- ° higher temperature (550 C). Having in mind the rel- width and the Raman shift of Degussa P25 compared atively large crystallite size in all samples registered

68 A. GOLUBOVIĆ et al.: INFLUENCE OF SOME SOL-GEL SYNTHESIS… Chem. Ind. Chem. Eng. Q. 22 (1) 65−73 (2016)

to the series of obtained samples can be explained as T550/2 the porosity was very small (the pore concentra- the least deffective and disordered anatase structure. tion is within experimental error). The mean pore diameters obtained by BET and BJH method are in good agreement. The pore size distribution for synthesized anatase samples and Degussa P25, obtained by BJH method, are shown in Figure 4. It

could be seen that in the rows T500/1.5, T500/2, T500/2.5

and T550/1.5, T550/2, T550/2.5 value of specific surface area had the highest value for the first member, lowest for the second and close to the first for the third member. The explanation for this tendency lies in the fact that the pores tranformed during the time of calcination. The tendency of microstrain in Table 1 was in accordance with the tendency of a pore evolution. Also, the pores in the samples calcined at 500 °C (mean pore diameter around 5-6 nm) are smaller than those in the

samples calcined at higher temperature (7-8 nm), as Figure 3. The experimental dependence of Raman shift on can be seen in Table 2. From Figure 4 can be also linewidth for the most intensive Eg(1) mode of synthesized anatase samples and Degussa P25.

Some additional Raman features, detected in the sample T550/2.5 shown in Figure 2, can be ascribed -1 to the rutile modes [37] Eg (∼445 cm ) and A1g (∼609 cm-1). The Raman modes related to the brookite phase [38] in the synthesized samples were not detected. Porosity To investigate the effects of synthesis conditions parameters on the adsorption abilities and pore structure of TiO2 samples, the nitrogen sorption isotherms measurements have been carried out. The specific surface area, pore volume and mean pore diameter calculated from both BET and BJH are listed in Table

2. The samples calcined at 500 °C (samples T500/1.5,

T500/2 and T500/2.5) are obviously more porous than those calcined at 550 °C (samples T550/1.5 and T550/2.5). Note that the parameters of porosity, determined from the αs-plots [16,39], suggest that the samples are fully mesoporous (Smeso = SBET), whereas in the sample Figure 4. The pore size distribution for synthesized anatase sample and Degussa P25 obtained by BJH method.

Table 2. The porous properties of synthesized anatase samples, as well as Degussa P25: specific surface areas (SBET, SBJH), pore volumes (Vp, Vt), and mean pore diameters ( DBET , DBJH ) obtained by BET and BJH methods, respectively

2 3 –1 2 –1 3 –1 Sample SBET = Smeso, in m /g Vp / cm g DBET / nm SBJH / m g Vt / cm g DBJH / nm

T500/1.5 52 0.1063 5.3 52.0 0.1025 5.1

T500/2 33 0.0757 5.9 34.2 0.0777 5.9

T500/2.5 45 0.0922 5.3 45.9 0.0903 5.1

T550/1.5 18 0.0580 8.3 18.5 0.0599 8.3

T550/2 2 – – – – –

T550/2.5 17 0.0454 6.9 18.3 0.0504 7.1 Degussa P25 13 0.0244 7.5 11.6 0.0214 7.4

69 A. GOLUBOVIĆ et al.: INFLUENCE OF SOME SOL-GEL SYNTHESIS… Chem. Ind. Chem. Eng. Q. 22 (1) 65−73 (2016) seen that the pore distribution of Degussa 25 was ution and c is the concentration after irradiation with almost uniform (except for the largest value of about 9 UV light. The efficiencies of the studied TiO2 catalysts nm) and that could be the crucial fact why Degussa as well as the observed pseudo first reaction rate 25 was the powerful tool for the photocatalytic deg- constants were presented in Table 3. The main dif- radation. ference between Degussa P25 and synthesized catalysts is the reaction rate as a result of pore distri- Photocatalytic activity bution uniformity. The photocatalytic activity of synthesized cat- In case of carbofuran, Degussa P25 showed alyst was studied using three representatives of org- higher photodegradation efficiency than all syn- anic pollutants: C.I. Reactive Orange 16 (textile dye), thesized samples (98% of carbofuran was photodeg- carbofuran (pesticide) and phenol. The samples were raded after 90 min). After 150 min of UV irradiation, (after mixting with pollutants, sorption and UV irradi- the highest photodegradation efficiency was obtained ation) withdrawn and analyzed on a UV-Vis spectro- by samples T500/1.5 (75%), whereas the lowest effi- photometer at 492.5 nm for C.I. Reactive Orange 16, ciency were observed for the samples T550/1.5 and 277 nm for carbofuran and 270 nm for phenol. The T550/2.5 (both 49%). The obtained results are in accord- time after the agitation 90 min in dark is denoted as 0, ance with the pore distribution influence on the reac- and these concentrations are denoted as c0. The tion rate. Photocatalitic degradation of carbofuran reactions were performed using Degussa P25 TiO 2 using synthesized of TiO2 series and Degussa P25 as for comparison. The results are shown in Figure 5. catalysts are presented in Figure 5b. Here, one can In Figure 5a, the effectiveness of synthesized observe that the photocatalytic reaction rate is highest TiO2 catalysts in photodegradation reaction of (C.I. when Degussa P25 is used, while the differences Reactive Orange 16) is presented. In comparison to between synthesized catalysts are much less pro- Degussa P25, the catalyst T500/1.5 showed almost the nounced. As given above, the observed reaction rate same photodegradation effectiveness (99 and 98% might be the result of the mean pore diameter range, after 90 min of UV irradiation, respectively), while the and the combination of specific surface area and others samples, except T550/1.5, showed similarly good mean pore diameter. photodegradation effectiveness. The photodegrad- Phenol [40] was also subjected to photodeg- ation efficiency can be determined as: radation using synthesized catalysts and the results cc− are shown in Figure 5c. It appears that TiO2 (both Efficiency= 100 0 c synthesized and Degussa P25) is able to remove 0 phenol too, but it requires more time, since the con- where c0 is the initial concentration of pollutant sol- centrations continuously decrease. After 150 min of

Figure 5. The kinetics of degradation of: a) C.I. Reactive Orange 16, b) carbofuran and c) phenol, under UV irradiation monitored in the

presence of synthesized TiO2 samples and Degussa P25.

70 A. GOLUBOVIĆ et al.: INFLUENCE OF SOME SOL-GEL SYNTHESIS… Chem. Ind. Chem. Eng. Q. 22 (1) 65−73 (2016)

Table 3. The efficiency, %, of the studied TiO2 catalysts as well as the observed pseudo-first reaction rate constants

Catalyst Time, min P-25 T500/1.5 T500/2 T500/2.5 T550/1.5 T550/2 T550/2.5 RO16 0 0 0 0 0 0 0 0 30 72 40 34 25 33 42 32 60 98 61 62 63 60 71 63 90 99 96 98 95 85 97 95 k / min-1 0.0546 0.0287 0.0335 0.0271 0.0189 0.0322 0.0271 Carbofuran 0 0 0 0 0 0 0 0 30 38 1 1 1 1 1 1 60 82 4 4 3 2 4 2 90 98 17 10 10 9 15 5 120 48 29 32 21 36 19 150 75 63 68 49 68 49 k / min-1 0.0372 0.0062 0.0041 0.0027 0.0028 0.0049 0.0027 Phenol 0 0 0 0 0 0 0 0 30 14 1 6 5 4 2 0 60 24 10 10 19 14 13 1 90 32 23 23 23 14 19 16 120 41 38 38 33 23 34 32 150 54 52 52 45 46 44 34 k / min-1 0.0048 0.004 0.004 0.0035 0.003 0.0033 0.0025

UV irradiation, Degussa P25 degraded 54%, almost CONCLUSIONS the same value as for T500/1.5 and T500/2 (52%). Other The structural and morphological properties of samples (T500/2.5, T550/1.5 and T550/2) degraded about TiO powders were intentionally varied by the tem- 45% and T550/2.5 showed the worst result (34%). 2 Obtained photodegradation result for Degussa P25 is perature and duration of the calcination. The analysis in accordance with results from de la Cruz Romero et of XRPD data showed that rising of temperature and al. [3] where phenol was not 100% photodegraded extending the duration of the calcination caused slight even with the UV irradiation of 10 h (only 60% under growth of crystallites in synthesized samples (from 24 similar experimental conditions). to 35 nm), which was confirmed by Raman scattering. One would expect that the smaller molecule, It was also noticed that the most intensive Raman Eg phenol, can easily access the internal surface of mode in the samples calcined at higher temperature Degussa sample giving higher degradation rate in (550 °C) is less broadened and blueshifted than in the comparison to other two organic pollutants. Namely, if samples calcined at 500 °C, pointing to less deffective only size of the molecule is important, than the and disordered anatase structure. The BET analysis reaction rate order would be: phenol > carbofuran > showed that the greatest specific surface area was in ° > RO16. On the contrary, the rate order is inverse, the sample calcined for 1.5 h at 550 C (T550/1.5). The RO16 being most reactive. The main reason for such samples calcined at 500 °C displayed higher photo- reaction rate is due to the different mechanisms of catalytic activity in the degradation in comparison with degradation and different part of molecules involved. the samples calcined at 550 °C. The results of photo- If only one molecule is concerned, then the influence degradation of C.I. Reactive Orange 16 for the ° of the catalyst is more complex. Not only the mean sample calcined 2 h at 500 C (sample T500/2) was pore diameter is important, but also the combination comparable with Degussa P25. The samples calcined of specific surface area and mean pore diameter, for 1.5 and 2 h at the same temperature (samples giving Degussa an advantage when voluminous T500/1.5 and T500/2) showed comparable efficiency with molecules are concerned. Degussa P25 in photodegradation of phenol, while in

71 A. GOLUBOVIĆ et al.: INFLUENCE OF SOME SOL-GEL SYNTHESIS… Chem. Ind. Chem. Eng. Q. 22 (1) 65−73 (2016) photodegradation of carbofuran Degussa P25 showed [18] W. Li, X. Guo, Y. Zhu, Y. Hui, K. Kanamori, K. Nakanishi, superior photocatalytic properties. J. Sol-Gel Sci. Technol. 67 (2013) 639-645 [19] M. Fernández-García, A. Martínez-Arias, J. C. Hanson, Acknowledgement J. A. Rodriguez, Chem. Rev. 104 (2004) 4063-4104 This work was financially supported by the [20] S. Ahmed, M.G. Rasul, W.N. Martens, R. Brown, M.A. Serbian Ministry of Education, Science and Techno- Hashib, Desalination 261 (2010) 3-18 logical Development, Projects No. III45018 and [21] S. Ahmed, M.G. Rasul, W.N. Martens, R. Brown, M.A. ON171032, as well as SASA project F-134. Hashib, Water Air Soil Pollut. 215 (2011) 3-29 [22] C.-Y. Chen, Water Air Soil Pollut. 202 (2009) 335-342 REFERENCES [23] D. Mijin, M. Radulović, D. Zlatić, P. Jovančić, Chem. Ind. Chem. Eng. Q. 13 (2007) 179-185 [1] R. Pourata, А. R. Khataee, S. Aber, N. Daneshvar, [24] B. Lopez-Alvarez, R.A. Torres-Palma, G. Peñuela, J. Desalination 249 (2009) 301-307 Hazard. Mater. 191 (2011) 196-203 [2] M. Mahalakshmi, B. Arabindoo, M. Palanichamy, V. [25] J. Fenoll, P. Hellín, P. Flores, C. M. Martínez, S. Navarro, Murugesan, J. Hazard. Mater. 143 (2007) 240-245 J. Photochem. Photobiol., A 251 (2013) 33-40 [3] D. de la Cruz Romero, G. Torres Torres, J. C. Arévalo, R. [26] F. Javier Benitez, J.L. Acero, F.J. Real, J. Hazard. Mater. Gomez, A. Aguilar-Elguezabal, J. Sol-Gel Sci. Technol. 89 (2002) 51-65 56 (2010) 219-226 [27] M. Jesus, S. Benito, O. Aaron O, A.H. de Lasa, Chem. [4] Y. Wang, A. Zhou, Z. Yang, Mater. Lett. 62 (2008) 1930- Eng. Sci. 78 (2012) 186-203 –1932 [28] K. Majeda, W. Lijun, A.H. Al-Muhtaseb, A.B. Albadarin, [5] Hari-Bala,Y. Guo, X. Zhao, J Zhao, W. Fu, X. Ding, Y. G.M. Walker, Chem. Eng. J. 213 (2012) 125-134 Jiang, K. Yu, X. Lv, Z. Wang, Mater. Lett. 60 (2006) 494– [29] Y.L. Du, Y. Deng, M.S. Zhang, J. Phys. Chem. Solids 67 –498 (2006) 2405-2408 [6] W. Dong, Y. Sun, C. W. Lee, W. Hua, X. Lu, Y. Shi, S. [30] G.K. Williamson, W.H. Hall, Acta Metall. 1 (1953) 22-31 Zhang, J. Chen, D. Zhao, J. Am. Chem. Soc. 129 (2007) [31] H.P. Klug, L.E. Alexander, X-Ray Diffraction Procedures: 13894-13904 For Polycrystalline and Amorphous Materials, 2nd ed., [7] E. Stathatos, D. Papoulis, C. A. Aggelopoulos, D. Pana- Wiley-VCH, New York, 1974, p. 687 giotaras, A. Nikolopoulou, J. Hazard. Mater. 211-212 [32] K. Kaneko, C. Ishii, H. Kanoh, Y. Hanzawa, N. Setoyama, (2012) 68-76 T. Suzuki, Adv. Colloid Interfac. Sci. 76-77 (1998) 295- [8] T. Ohno, K. Sarukawa, M. Matsumura, New J. Chem. 26 –320 (2002) 1167-1170 [33] E.P. Barrett, L.G. Joyner, P.P. Halenda, J. Am. Chem. [9] M. Inagaki, R. Nonaka, B. Tryba, A. W. Morawski, Soc. 73 (1951) 373-381 Chemosphere 64 (2006) 437-445 [34] W. Zhang, S. Chen, S. Yu, Y. Yin, J. Cryst. Growth 308 [10] A. Sclafani, L. Palmisano, M. Schiavello, J. Phys. Chem. (2007) 122–129 94 (1990) 829-832 [35] T. Ohsaka, F. Izumi, Y. Fujiki, J. Raman Spectrosc. 7 [11] N. Serpone, D. Lawless, R. Khairutdinov, E. Pelizzetti, J. (1978) 321-324 Phys. Chem. 99 (1995) 16655-16661 [36] M.J. Šćepanović, M. Grujić-Brojčin, Z. Dohčević-Mitrović, [12] T. Ohno, K. Sarukawa, M. Matsumura, J. Phys. Chem., B Z.V. Popović, Appl. Phys., A 86 (2007) 365-371 105 (2001) 2417-2420 [37] X. Wang, J. Shen, Q. Pan, J. Raman Spectrosc. 42 [13] K. Tanaka, M. F. V. Capule, T. Hisanaga Chem. Phys. (2011) 1578-1582 Lett. 187 (1991) 73-76 [38] Y.-H. Zhang, C. K. Chan, J. F. Porter, W. Guo, J. Mater. [14] D.C. Hurum, K.A. Gray, T. Rajh, M.C. Thurnauer, J. Res. 13 (1998) 2602-2609 Phys. Chem., B 109 (2004) 977-980 [39] A. Golubović, B. Abramović, M. Šćepanović, M. Grujić- [15] J. Zhang, Q. Xu, Z. Feng, M. Li, C. Li, Angew. Chem. Int. –Brojčin, S. Armaković, I. Veljković, B. Babić, Z. Doh- Edit. 47 (2008) 1766-1769 ćević-Mitrović, Z. V. Popović, Mater. Res. Bull. 48 (2013) [16] T. Kawahara, Y. Konishi, H. Tada, N. Tohge, J. Nishii, S. 1363-1371 Ito, Angew. Chem. Int. Edit. 41 (2002) 2811-2813 [40] J. Moreira, B. Serrano, A. Ortiz, H. de Lasa, Chem. Eng. [17] A. Zachariah, K. V. Baiju, S. Shukla, K. S. Deepa, J. Sci. 78 (2012) 186-203. James, K.G.K. Warrier, J. Phys. Chem., C 112 (2008) 11345-11356

72 A. GOLUBOVIĆ et al.: INFLUENCE OF SOME SOL-GEL SYNTHESIS… Chem. Ind. Chem. Eng. Q. 22 (1) 65−73 (2016)

ALEKSANDAR GOLUBOVIĆ1 UTICAJ NEKIH PARAMETARA SOL-GEL SINTEZE IVANA VELJKOVIĆ2 MEZOPOROZNOG TIO NA FOTOKATALITIČKU 1 2 MAJA ŠĆEPANOVIĆ DEGRADACIJU ZAGAĐIVAČA MIRJANA GRUJIĆ-BROJČIN1 1 NATAŠA TOMIĆ Nanoprahovi titan-dioksida (TiO ) su proizvedeni sol-gel tehnikom iz tetrabutil-titanata kao 3 2 DUŠAN MIJIN prekursora, varirajući neke parametre sol-gel sinteze kao što su temperatura kalcinacije 4 BILJANA BABIĆ (500 i 550 °C) i dužina kalcinacije (1,5; 2 i 2,5 h). XRPD rezultati su pokazali da su svi 1Institut za fiziku, Univerzitet u sintetizovani nanoprahovi dominantno u anataz fazi sa prisustvom malih količina rutilne Beogradu, Pregrevica 118, 11080 faze u uzorcima kalcinisanim na 550 °C. Saglasno rezultatima dobijenim Williamson-Hall Beograd, Srbija metodom, kristaliti anataza rastu sa vremenom kalcinacije (od 24 do 29 nm u uzorcima 2Institut za multidisciplinarna kalcinisanim na nižoj temperaturi, i od 30 do 35 nm u uzorcima kalcinisanim na višoj tem- istraživanja, Univerzite u peraturi). Analize pomeraja i poluširine najintenzivnijeg Eg Ramanskog moda anataza su Beogradu, Kneza Višeslava 1, potvrdile XRPD rezultate. Parametri veličine pora dobijeni pomoću eksperimentalnih poda- 11000 Beograd, Srbija taka sorpcije azota su ukazali na to da su svi uzorci mezoporozni, sa srednjom veličinom 3 Tehnološko-metalurški fakultet, pora u opsegu 5-8 nm. Fotokatalitička aktivnost dobijenih nanoprahova je testirana na deg- Univerzitet u Beogradu, radaciji tekstilne boje (C.I. Reactive Orange 16), karbofurana i fenola. Karnegijeva 4, 11000 Beograd, Srbija Ključne reči: nanostrukture, anataz, difrakcija X-zraka na prahu, rasipanje. 4Institut za nuklearne nauke „Vinča”, Univerzitet u Beogradu, 11001 Beograd, Srbija

NAUČNI RAD

Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 22 (1) 75−83 (2016) CI&CEQ

MEHDI ASADOLLAHZADEH1,2 THE EFFECTS OF OPERATING PARAMETERS SHAHROKH SHAHHOSSEINI1 2 ON STAGE EFFICIENCY IN AN MEISAM TORAB-MOSTAEDI OLDSHUE-RUSHTON COLUMN AHAD GHAEMI1

1Department of Chemical Article Highlights Engineering, Iran University of • Stage efficiency of the investigated column is high in comparison with other extractors Science and Technology (IUST), • Stage efficiency is strongly dependent on the agitation rate and interfacial tension Tehran, Iran • Stage efficiency is better when the mass transfer direction is from continuous to 2Nuclear Fuel Cycle Research dispersed phase • Empirical correlation is derived for prediction of stage efficiency School, Nuclear Science and

Technology Research Institute, Abstract Tehran, Iran In this research, stage efficiency was measured in a 113 mm Oldshue-Rushton SCIENTIFIC PAPER column for two systems including toluene-acetone-water and n-butyl acetate- -acetone-water. The experiments were performed in two directions of mass UDC 54:66.063 transfer. The effects of different parameters such as rotor speed, dispersed and continuous phase velocities and direction of mass transfer on the stage DOI 10.2298/CICEQ150426022A efficiency were investigated. The experimental data show that the stage efficiency is strongly dependent on the agitation rate and interfacial tension, but only slightly dependent on phase velocities. It was observed that the stage efficiency is better when the mass transfer direction of acetone is from the continuous to the dispersed phase in comparison to opposite direction due to the presence of oscillations created by surface tension gradient. The investigated column is one of the extraction columns with high stage efficiency. An empirical correlation is proposed to describe the stage efficiency in terms of Rey- nolds and Froude numbers. The predictions of the equation had good agreement with the experimental data. Keywords: Oldshue-Rushton column, stage efficiency, axial mixing, throughput.

Solvent extraction is one of the key unit operation speed, the relative velocity between the dis- ations in the processes including the petrochemical, persed phase and continuous phase decreases like- pharmaceutical, hydrometallurgical, and environmen- wise, which lowers the throughput. In addition, the tal industries. Among various types of solvent ext- agitation can increase the axial mixing and reduce the raction units, the extraction column is emerging as extraction efficiency by decreasing solute concen- one of the best choices because of a high throughput tration gradients and as a consequence the mass and stage efficiency [1]. transfer rate. Neglecting the effect of axial mixing The droplet size and the degree of turbulence when designing an extraction column can lead to are dependent on the mechanical agitation in the overestimation of mass transfer efficiency of about extraction column. Mixing can intensify the stage effi- 30% or more [4]. Thus, the mechanical agitation can ciency due to the large interfacial area with small dis- be used to control the droplet size, dispersed phase persed drops [2,3]. As drop size decreases with agit- holdup, stage efficiency and, consequently, the performance of the extraction columns [5]. Correspondence: S. Shahhosseini, Department of Chemical In several studies, a number of authors have Engineering, Iran University of Science and Technology (IUST), reported different methods to decrease axial mixing P.O. Box 16765-163, Tehran, Iran. by coalescing small drops in the section between E-mail: [email protected] Paper received: 26 April, 2015 stages. Internal column geometry reduces axial mix- Paper revised: 24 June, 2015 ing, increases droplet coalescence and breakage Paper accepted: 2 July, 2015

75 M. ASADOLLAHZADEH et al.: THE EFFECTS OF OPERATING… Chem. Ind. Chem. Eng. Q. 22 (1) 75−83 (2016) rates resulting in increased mass transfer rates, and rate mixing operation [14]. Experimental work in Old- affects the mean residence time of the dispersed shue-Rushton columns is limited and the studies phase, which allows the handling of large loads with about stage efficiencies in the column have rarely small differences of interfacial tension and density, been referred to in the literature. improving the hydrodynamic performance of the col- The objective of the present work is to inves- umn and, subsequently, the extraction efficiency [6,7]. tigate the influence of operating parameters such as The experimental set up with the mixing part rotor speed and velocity of dispersed and continuous and the packing part alternately to promote drop phase on the stage efficiency for mass transfer dir- coalescence in the packing part has been reported by ections as well as the two systems. An empirical Scheibel [8]. For the same objective, a three-dimen- correlation for prediction of stage efficiency is recom- sional lattice as a partition of the mixing stages was mended in terms of physical properties of liquid sys- investigated by Steiner et al. [9]. tems and operating conditions. The coalescence-dispersion pulsed-sieve-plate extraction column (CDPSEC) is a modified pulsed- EXPERIMENTAL sieve-plate extraction column (PSEC). It was reported that the CDPSEC with 50 mm in the plate spacing A pilot plant Oldshue-Rushton extraction column was of 120% overall mass transfer efficiency over the is used in these experiments. The column built in a standard PSEC [10]. However, when the plate cylindrical glass section was equipped with impellers spacing of the CDPSEC was reduced to 25 mm, it with accurate speed control and the internal parts was reported that the mass transfer efficiency of the were constructed from stainless steel; a schematic CDPSEC was only about 50% that of the standard diagram of the Oldshue-Rushton column used in this PSEC, although the interface renewal frequency was study is presented in Figure 1. The specifications of doubled [11]. this column and range of operating variables are Horvath and Hartland achieved the high stage listed in Table 2. efficiency with a mixer-settler extraction column in In normal operation, two types of immiscible which the inter-stage mixing was extremely small, liquids with different densities flow counter-currently similar to the throughput of the column [12]. Schwei- through the apparatus. One of them is in large quan- tzer reported a rectangular mixer-settler tower with tity (continuous phase), while the other, being in horizontal arrangement of the mixer and settler in minute quantity (few percent), is dispersed as drops. each stage. The arrangement between stages can Two flow meters are employed to supply and monitor reduce the axial mixing and result in the enhance- the fixed flow rates of continuous and dispersed ment of separation efficiency [13]. The comparison of phases. The inlet and outlet of the column are con- performance of various columns is shown in Table 1. nected to four tanks, each of 85 L capacity. The inter- The Oldshue-Rushton column manufactured by face is maintained at the required level by using an the mixing equipment company and commonly known optical sensor as previously described. as the Mixco column was developed in 1940 thanks to Two chemical systems for instance toluene- the best endeavors of Rushton and Oldshue. The unit acetone-water (high interfacial tension), and n-butyl consists of an outer shell in which horizontal stage acetate-acetone-water (medium interfacial tension) separators are constructed to form the desired num- are examined on the extraction column for both mass ber of processing stages, each equivalent to a sepa- transfer directions. The European Federation of

Table 1. Comparison of performance of various columns [12]

Diameter Stage height Stage efficiency NTS/m Throughput Mass transfer Column m m % 1/m m3 m–2 h–1 direction Enhanced coalescing column 0.072 0.060 45 3-7.5 10-60 d→c Kühni column 0.060 0.350 100/130 2.9-3.7 4-10 d→c Packed column 0.070 - - 1.8-2.5 15-30 d→c Pulsed packed column 0.070 - - 3.8-5.8 18-20 d→c Pulsed sieve plate column 0.050 0.100 60 3.5-6.0 20-30 d→c Karr reciprocating column 0.050 0.025 15 3.5-6.0 30-40 d→c Rotating disc column 0.070 - - 2.8-3.5 15-35 d→c Mixer settler extraction column 0.152 0.150 97 6.5 2-6 c→d Mixer settler extraction column 0.152 0.150 170 11.3 2-4 d→c

76 M. ASADOLLAHZADEH et al.: THE EFFECTS OF OPERATING… Chem. Ind. Chem. Eng. Q. 22 (1) 75−83 (2016)

Figure 1. Schematic flow diagram of Oldshue-Rushton column.

Chemical Engineering (EFCE) has adopted these Table 3 [15]. In the present work, the values of phys- systems as recommended systems. ical properties have been assumed to correspond to the arithmetic-mean concentrations of the continuous Table 2. Technical description of the Oldshue-Rushton column and dispersed phases at the inlet and outlet of the

Parameter Value Unit column. Column height (H) 700 mm Table 3. Physical properties of liquid systems at 20 °C [15] Column internal diameter (Dc) 113 mm Physical n-Butyl acetate/ Diameter of the rotor 50 mm Toluene/acetone/water Settler diameter 169 mm property /acetone/water –3 No. number of stages 9 - ρc / kg m 994.4-995.7 994.3-995.8 –3 Height of the stages 67 mm ρd / kg m 864.4-865.2 879.6-881.4 Fractional free cross section area 25 % μc / mPa s 1.059-1.075 1.075-1.088 Continuous phase flow rate 18-36 l/h μd / mPa s 0.574-0.584 0.723-0.738 –1 Dispersed phase flow rate 18-36 l/h σ / mN m 27.5-30.1 12.4-13.2 2 –1 × -9 × -9 Rotor speed 100-240 rpm Dc / m s 1.09-1.14 10 1.01-1.06 10 2 –1 -9 -9 Dd / m s 2.7-2.8×10 2.16-2.18×10 All experiments are carried out far from flooding conditions. Conditions became steady, as evidenced The drops were photographed by a very high- by a constant interface level, after three or four col- -resolution Nikon D5000 camera. Next, droplet dimen- umn volume of operation depending on the phase sions were compared with the thickness of stators as flow rates and rotor speed. At the end of each expe- a reference. It is found that the curved surface of the riment, the average hold-up of the column was mea- glass extraction column and significant differences sured by using the shutdown procedure (interface between air and the glass refractive indices leads to a position changes). parallax deformation of the objects photographed in In all experiments, dilute solutions were inves- the extraction column. In order to omit this pheno- tigated with approximately 3.5 wt.% acetone in the menon, a container filled with water was attached to organic phase. The acetone content of the aqueous the extraction column and the photographic approach and organic stream was measured by UV-Vis was used to calculate the size of stator thickness spectroscopy. The physical properties of the liquid– served as the reference for drop size measurements. liquid systems used in these experiments are listed in Consequently, digital image analysis software was

77 M. ASADOLLAHZADEH et al.: THE EFFECTS OF OPERATING… Chem. Ind. Chem. Eng. Q. 22 (1) 75−83 (2016)

applied in order to investigate the taken high quality stage efficiency. The stage efficiency, Eoy, based on photograph. A minimum of 1000 drops was analyzed the concentration of organic phase is illustrated as for each experimental condition in order to guarantee follows: the statistical significance of the determined size − = ()yynn−1 distributions. E oy (3) ()yy* − In the case of non-spherical droplets, the major nn−1 and minor axes, d and d , were measured and the 1 2 where y * = mx is the organic phase concentration in equivalent diameter, d , was calculated from Eq. (1): n n e equilibrium with the aqueous phase of nth stage, the 1/ 3 ddd= ()2 (1) value of m is 0.68 and 0.91 for toluene-water and e12 n-butyl acetate-water, respectively. Figure 2 illus- The Sauter mean diameter was then calculated trates the typical concentration profile for the organic according to the following equation: and continues phase for two systems and two directions of mass transfer. The experimental results N 3 obtained in these experiments are given in Tables A.1 ndii = i =1 and A.2 in Appendix and the pictures of drop sizes for d 32 N (2) 2 two systems is shown in Figure 3. ndii i =1 Effect of agitation speed where ni is the number of droplets of the mean Figure 4a shows the effect of changing the agi- diameter di within a narrow size range i. tation speed on the stage efficiency for both systems from dispersed to continuous phase mass transfer. It RESULTS AND DISCUSSION was observed that the stage efficiency in both systems is heavily dependent on the agitation speed. At The performance of an extraction column with low speeds, the stage efficiency is low due to inade- well-defined stages can be expressed in terms of quate mixing, resulting in low holdup and large drops

Figure 2. Typical concentration profile along the column (N=140 rpm, Vd= Vc=0.66 mm/s).

78 M. ASADOLLAHZADEH et al.: THE EFFECTS OF OPERATING… Chem. Ind. Chem. Eng. Q. 22 (1) 75−83 (2016)

Figure 3. Variation of drop sizes with rotor speed and interfacial tension for toluene-acetone-water: a) 140, b) 160, c) 180 rpm, and for n-butyl acetate-acetone-water: d) 140, e) 160, f) 180 rpm.

(low interfacial area) that is also observed in Figure 3. Having reached its maximum, the stage effi- The stage efficiency increases with an increase in the ciency falls to further increasing at agitation speed. A agitation speed and reaches a maximum of 58% for decrease in the stage efficiency could contribute to toluene-acetone-water system at a speed of 220 rpm the a significant decrease in mass transfer rates due and a maximum of 59% for n-butyl acetate-acetone- to small droplets behaving as rigid spheres, in which -water system at a speed of 180 rpm. case molecular diffusion would govern mass transfer in the system. The effect of rotor speed on the values of the stage efficiencies in the water-acetone-n-butyl acetate test system (medium interfacial tension) is greater than that of the water-acetone-toluene test system (high interfacial tension). The size of the droplets in higher interfacial tension test systems is larger than the droplet size in the lower interfacial tension test systems (Figure 3), which results in a decrease in their residence time in the column. Finally, the slip velocities increase and, consequently, the value of the dispersed phase holdup and stage efficiency will decrease; consequently, the column will operate in a more-stable manner. Effect of mass transfer direction The effect of the mass transfer direction on the stage efficiency is shown in Figure 4b. It is found from this figure that the mass transfer direction has a considerable effect on the stage efficiency. The stage efficiency in the continuous to dispersed phase transfer is lower than that in the opposite direction. This is due to the interfacial tension gradients that leads to the smaller drop sizes in continuous to dispersed phase transfer and larger drop sizes in the opposite direction. Therefore, the higher values of the stage efficiency in the case of the dispersed to continuous phase transfer are resulted from the increased mass transfer rates in drops of bigger sizes due to the pre- Figure 4. Effect of rotor speed on the stage efficiency: a) surface sence of oscillations created by coalescence between tension and b) direction of mass transfer (Vc= Vd= 0.66 mm/s). the droplets enhanced by the Marangoni effect [16].

79 M. ASADOLLAHZADEH et al.: THE EFFECTS OF OPERATING… Chem. Ind. Chem. Eng. Q. 22 (1) 75−83 (2016)

Effect of dispersed phase velocity leads to an increment in the holdup due to the red- As shown in Figure 5a, the stage efficiencies uction of the relative velocity between the drops and increase with an increase in dispersed phase velocity continuous phase, but it is not appreciable on the for mass transfer direction from the dispersed to the drop sizes. Therefore, the interfacial area increases continuous phase. This observation could be attri- with the positive effect of the holdup. An increase in buted to an increase in mean drop sizes because of drag forces arising from the relative velocity between an increase in drop formation and higher coalescence the continuous and dispersed phases leads to the frequency. The increment of the number of dispersed circulation in a drop and consequently, overall mass droplets leads to an increase in the dispersed phase transfer coefficient increases with an increase in Vc. holdup. It is observed that the effect of the holdup on The stage efficiency increases with both increase in the interfacial area is larger than that of mean drop overall mass transfer coefficient and interfacial area. size, i.e., the interfacial area increases with an inc- As mentioned earlier, it is observed from Figures 5b rease in the dispersed phase velocity; albeit an inc- and 6b that the stage efficiency in the dispersed to rease in the dispersed phase velocity leads to the continuous phase transfer is higher than that in the reduction of mass transfer coefficient, a decrease is opposite direction. more predominant when the increase in the interfacial area is considered. Therefore, the stage efficiency decreases along the column.

Figure 6. Effect of continuous phase velocity on the stage efficiency: a) surface tension and b) direction of mass transfer

(Vd= 0.66 mm/s). Figure 5. Effect of dispersed phase velocity on the stage efficiency: a) surface tension and b) direction of mass transfer Comparison of other type of extractors with present

(Vc= 0.66 mm/s). column A comparison of the separation performance of Effect of continuous phase velocity the Oldshue-Rushton column with some other type of The effect of the continuous phase velocity on extraction extractors is described in Figure 7. The pat- the stage efficiency is shown in Figure 6a. This effect tern is, as proposed by Pratt and Stevens, the number

80 M. ASADOLLAHZADEH et al.: THE EFFECTS OF OPERATING… Chem. Ind. Chem. Eng. Q. 22 (1) 75−83 (2016) of theoretical stages per unit of length against total ation can predict the stage efficiency of the column volumetric throughput of both phases [5]. This plot is accurately. of value in facilitating the comparison of the relative areas of application of various extractor types, despite being based on the data for a single system, viz. toluene-acetone-water in a phase ratio of 1.5. The present Oldshue-Rushton column reached values of between 5.14 and 6.55 NTS/m at low total through- puts. Therefore, it can be concluded that the present column has high stage efficiency while its throughput is low.

Figure 8. Comparison between experimental data and the proposed correlation.

CONCLUSION

Stage efficiency was measured in a 113 mm Oldshue-Rushton column for two systems. It is shown in this work that the performance of the column depends largely on the rotor speed. The stage efficiency increased with agitation speed and reached a Figure 7. Comparison of extractor performance; toluene- maximum, but after having reached its maximum, it –acetone-water system, Vd/ Vc = 1.5. fell to further increase in agitation speed. The com- Proposed correlation for stage efficiencies parison between the stage efficiencies for the two drops under the the same conditions of the two sys- There is no correlation for prediction of stage tems shows that the drop in n-butyl acetate-acetone- efficiency in the Oldshue-Rushton column. The expe- water system with a lower value of interfacial tension rimental data on the stage efficiency are correlated in has a higher value of E . It was observed that the terms of dimensionless numbers Re and Fr for both oy stage efficiency is higher when the mass transfer mass transfer directions as well as the two systems direction is from the continuous to the dispersed by using the least square method, as follows: phase. The comparison of Oldshue-Rushton column = −−0.203 0.169 EReFroy 1.399 (4) with some other types of extractors revealed that the stage efficiency is high in this column. where: Nomenclature = g Fr 2 (5) d32 Sauter mean drop diameter (m) dN 2 R D molecular diffusivity (m /s) ρ dV Dc column diameter (m) Re = c32s (6) μ dR rotor diameter (m) c Eoy stage efficiency 2 The experimental data are compared with the g acceleration due to gravity (m/s ) calculated results from the above equation in Figure N rotor speed (1/s) 8. The stage efficiency calculated according to this NTS number of stage efficiency correlation reproduces the experimental data with an Re Reynolds number average error of 4.64%. Thus, the proposed correl- Fr Froude number m distribution ratio

81 M. ASADOLLAHZADEH et al.: THE EFFECTS OF OPERATING… Chem. Ind. Chem. Eng. Q. 22 (1) 75−83 (2016) t time (s) [4] M.S.A. Nabli, P. Guiraud, C. Gourdon, Chem. Eng. Res. V superficial velocity (m/s) Des. 76 (1998) 951-960 [5] H.R.C. Pratt, G.W. Stevens, in Science and Practice in Vs slip velocity (m/s) Liquid–Liquid Extraction, Oxford University Press, Oxford, yn mass fraction of acetone in dispersed phase 1992, pp. 491-502 x mass fraction of acetone in continuous phase [6] W. Batey, J.D. Thornton, Ind. Eng. Chem. Res. 76 (1989) Greek letters 1096-1101 ρ density (kg/m3) [7] M. Jaradat, M. Attarakih, H. J. Bart, Ind. Eng. Chem. Res. σ interfacial tension (N/m) 50 (2011) 14121-14135 μ viscosity (Pa s) [8] E.G. Scheibel, Chem. Eng. Prog. 44 (1948) 681-690 φ dispersed phase holdup [9] L. Steiner, E.V. Fisher, S. Hartland, AIChE Symp. Ser. 80 (1984) 130-138 Subscripts [10] H.B. Li, G.S. Luo, W.Y. Fei, J.D. Wang, Chem. Eng. J. 78 c continuous phase (2000) 225-229 d dispersed phase [11] X.J. Tang, G.S. Luo, H.B. Li, J.D. Wang, Pet. Technol. 32 o overall value (2003) 1046-1050 Superscripts [12] M. Horvoth, S. Hartland, Ind. Eng. Chem. Process Des. Dev. 24 (1985) 1220-1225 * equilibrium value [13] P.A. Schweitzer, Hanson Mixer-Settler Handbook of Separation Techniques for Chemical Engineering, 3rd ed., REFERENCES McGraw-Hill, New York, 1997, p. 230 [14] J.H. Rushton, S. Nagata, T.B. Rooney, AIChE J. 10 [1] J. Rydberg, C. Musikas, G.R. Choppin, Solvent extraction (1964) 298-302 principles and practice, CRC Press, New York, 2004, p. 15 [15] T. Míšek, R. Berger, J. Schroter, EFCE Publ. Ser. 46 (1985) [2] J.C. Godfrey, M.J. Slater, Liquid-Liquid Extraction Equip- ment, Wiley, New York, 1995, p. 40 [16] M. Wegener, J. Grünig, J. Stüber, A.R. Paschedag, M. Kraume, Chem. Eng. Sci. 62 (2007) 2967-2978. [3] G.M. Ritcey, A.W. Ashbrook, Solvent extraction: principles and applications to process metallurgy, Vol. 1, Elsevier, New York, 1984, p.100

APPENDIX

Table A.1. Experimental data obtained in the experiments for toluene-acetone-water system

–1 –1 d to c transfer c to d transfer Qd / l h Qc / l h rpm φ d32 / mm φ d32 / mm 24 24 140 0.0687 2.48 0.072 2.44 24 24 160 0.0755 2.22 0.0792 2.09 24 24 180 0.089 1.9 0.0945 1.805 24 24 200 0.111 1.48 0.116 1.41 24 24 220 0.115 1.35 0.125 1.282 24 24 240 0.128 1.12 0.134 1.02 24 18 160 0.0703 2.23 0.0751 2.1 24 30 160 0.0768 2.19 0.08391 2.06 24 36 160 0.0805 2.17 0.0876 2.01 24 18 200 0.108 1.5 0.111 1.45 24 30 200 0.116 1.47 0.1205 1.37 24 36 200 0.119 1.46 0.1264 1.34 18 24 160 0.0671 2.16 0.0716 2.04 30 24 160 0.0818 2.28 0.0879 2.14 36 24 160 0.0893 2.39 0.0966 2.17 18 24 200 0.1045 1.42 0.1073 1.38 30 24 200 0.1212 1.53 0.1248 1.45 36 24 200 0.1331 1.63 0.1373 1.52

82 M. ASADOLLAHZADEH et al.: THE EFFECTS OF OPERATING… Chem. Ind. Chem. Eng. Q. 22 (1) 75−83 (2016)

Table A.2. Experimental data obtained in the experiments for n-butyl acetate-acetone-water system

–1 –1 Qd / l h Qc / l h rpm φ d32 / mm 24 24 100 0.0748 2.02 24 24 120 0.0893 1.7092 24 24 140 0.0978 1.4191 24 24 160 0.119 1.25 24 24 180 0.129 1.08 24 24 200 0.14 0.95 24 18 120 0.0848 1.7186 24 30 120 0.0943 1.6882 24 36 120 0.098 1.66 24 18 160 0.1145 1.263 24 30 160 0.1231 1.242 24 36 160 0.1262 1.22 18 24 120 0.0828 1.6512 30 24 120 0.0998 1.7412 36 24 120 0.1086 1.8092 18 24 160 0.1125 1.224 30 24 160 0.1256 1.309 36 24 160 0.1337 1.367

MEHDI ASADOLLAHZADEH1,2 EFEKTI RADNIH PARAMETARA NA EFIKASNOST SHAHROKH SHAHHOSSEINI1 STUPNJA OLDŠUE-RUŠTONOVE KOLONE MEISAM TORAB-MOSTAEDI2 1 AHAD GHAEMI U ovom istraživanju, efikasnost stupnja je ispitivana u Oldšue-Ruštonovoj koloni, prečnika 1Department of Chemical 113 mm, za dva sistema: toluen-aceton-voda i n-butil acetat-aceton- voda. Eksperimenti Engineering, Iran University of su uključili oba pravca prenosa mase. Ispitivan je uticaj različitih parametara, kao što su: Science and Technology (IUST), brzina mešanja, brzine strujanja dispergovane i kontinualne faze i pravac prenosa mase, Tehran, Iran na efikasnost stupnja. Eksperimentalni podaci pokazuju da efikasnost stupnja jako zavisi 2 Nuclear Fuel Cycle Research od brzine mešanja i međufaznog napona, a malo od brzine strujanja faza. Primećeno je da School, Nuclear Science and je efikasnost stupnja bolja kada je smer prenosa mase acetona od kontinualne prema dis- Technology Research Institute, pergovanoj fazi u odnosu na suprotan smer zbog prisustva oscilacija stvorenih gradijentom Tehran, Iran površinskog napona. Ispitivana kolona je jedna od ekstrakcionih kolona sa visokom efikas- nošću stupnja. Predložena je empirijska korelacija koja povezuje efikasnost stupnja sa NAUČNI RAD Rejnoldsovim i Frudovim brojem. Predviđanja jednačine se dobro slažu sa eksperimen- talnim podacima.

Ključne reči: Oldšue-Ruštonova kolona, efikasnost stupnja, aksijalna mešanje, kapacitet.

Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 22 (1) 85−93 (2016) CI&CEQ

XIAOLEI LI GAS–LIQUID MASS TRANSFER WITH CHUNYING ZHU INSTANTANEOUS CHEMICAL REACTION IN A School of Chemical Engineering SLURRY BUBBLE COLUMN CONTAINING and Technology, State Key FINE REACTANT PARTICLES Laboratory of Chemical Engineering, Tianjin University, Tianjin, China Article Highlights • A mass transfer model based on penetration theory was developed • SCIENTIFIC PAPER The effect of particle dissolution near the gas-liquid interface was considered in the model • UDC 66.06/.07:66.021.3:54 The absorption of SO2 into Mg(OH)2/water slurry was experimentally investigated in a bubble column DOI 10.2298/CICEQ141001021L Abstract In this study, the mass transfer accompanied by an instantaneous irreversible chemical reaction in a slurry bubble column containing sparingly soluble fine reactant particles has been analyzed theoretically. Based on penetration theory combined with the cell model, a one-dimensional mass transfer model was developed. In the model, the effects of particle size and particle dissolution near the gas-liquid interface on mass transfer were taken into account. The mass transfer model was solved and an analytical expression of the time-

-mean mass transfer coefficient was attained. Reactive absorption of SO2 from gas mixtures into Mg(OH)2/water slurry was investigated experimentally in a bubble column reactor to validate the mass transfer model. The results indicate that the present model has good predicting performance and could be used to predict mass transfer coefficient for the complicated gas-liquid-solid three- phase system with an instantaneous irreversible chemical reaction. Keywords: mass transfer, desulfurization, bubble column, instantaneous reaction, slurry.

Gas absorption accompanied by chemical reac- a steady-state homogeneous phase model. Sub- tions in slurries is widely employed in the chemical sequently, many studies on the gas absorption industry. Many chemical reactions involved in the accompanied by instantaneous chemical reactions process could be regarded as instantaneous when into slurries containing sparingly soluble reactant par- their rates are much greater than the rates of the ticles have been reported, and many theoretical molecular diffusion. Some typical examples are the models were developed. Ramachandran [2] has sum- removal of SO2 by means of the aqueous Mg(OH)2 or marized the gas absorption in slurries containing fine

Ca(OH)2 solution, the absorption of CO2 or H2S in the reactant particles till 2007. The models of gas-liquid aqueous Mg(OH)2 or Ca(OH)2 solution, etc. Rama- mass transfer accompanied by instantaneous chem- chandran et al. [1] theoretically analyzed gas absorp- ical reactions could be mainly classified into two tion into slurries by reactant particles using film theory categories: steady-state models and unsteady-state for the first time, and proposed and solved analytically models. Uchida et al. [3], Dagaonkar et al. [4,5], Scala [6] and Juvekar [7] developed and solved analytically the steady-state models in terms of film theory. The Correspondence: C. Zhu, School of Chemical Engineering and steady-state models based on film theory of mass Technology, State Key Laboratory of Chemical Engineering, transfer are well adapted for actual situations when Tianjin University, Tianjin 300072, China. E-mail: [email protected] the processes are steady-state and homogeneous Paper received: 1 October, 2014 phase, but they are unreasonable and could bring Paper revised: 27 April, 2015 great errors for the unsteady-state processes. By Paper accepted: 1 July, 2015

85 X. LI, C. ZHU: GAS–LIQUID MASS TRANSFER WITH INSTANTANEOUS… Chem. Ind. Chem. Eng. Q. 22 (1) 85−93 (2016) comparison, penetration theory and surface renewal ration theory by taking into account the effect of par- theory are more reasonable for describing the ticle size and particle dissolution near the gas-liquid unsteady-state processes. Mehra [8] studied the interface on the mass transfer. Mg(OH)2 has drawn effect of particle size distribution on the gas absorp- widespread attention as an absorbent for SO2 rem- tion, established and solved numerically an unsteady- oval in recent years because of the advantages of state model based on the penetration model using a high removal efficiency, recycling, no secondary pol- population balance approach to track the evolving lutants, etc. Thus, the mass transfer process of the particle size distributions. Kakaraniya et al. [9,10] ext- fast reactive absorption of SO2 into aqueous Mg(OH)2 ended the model of Mehra [8] to the SO2-Mg(OH)2- slurry in a bubble column was experimentally inves-

–MgSO3 system and CO2-Ca(OH)2-CaCO3 system. tigated to validate the proposed mass transfer model. Akbar et al. [11] studied the three-phase mass transfer in a spray scrubber with dissolving reactive par- EXPERIMENTAL ticles, proposed and solved numerically an unsteady- state model based on the penetration model. The The schematic diagram of the experimental studies on the gas-side mass-transfer coefficient of setup is shown in Figure 1, which consisted of the bubbles have been conducted. Patoczka [12], Mehta mixing gas generation unit, exhaust detection devices [13], Filla [14], Cho [15], Rocha [16], Guedes [17], and a bubble column. The bubble column was made Ricardo [18] and Sada [19] have reported the gas- of stainless steel with 20 cm in diameter and 45 cm in side mass-transfer coefficient for the bubble columns. height. The experimental conditions in this work are similar to Firstly, the column was filled with Mg(OH)2/water that of Sada [19], thus the estimation of mass transfer slurry, the air was fed into the bubble column until the proposed by Sada would be employed directly. Up to system reached steady-state. Then, SO2 with volume now, the effect of particle size on the mass transfer of fraction 0.98 (Tianjin Kermel Chemical Reagent Co., a gas-liquid-solid system with an instantaneous irre- Ltd.) supplied from a cylinder was mixed in the gas versible chemical reaction has not been theoretically mixer with the air from the air compressor (Shanghai analyzed and discussed in detail, and the particle size Jiebao Compressor Manufacture Co., Ltd.), and the and the particle dissolution near the gas-liquid inter- air-SO2 mixture was continuously fed at the bottom of face have a great influence on the mass transfer in the column. After SO2 reacted with Mg(OH)2 (particle the actual process. Therefore, in this paper, a theo- size: 1-10 μm, Tianjin Kermel Chemical Reagent Co., retical analysis of gas absorption accompanied by an Ltd.) in the liquid phase, the concentration of SO2 in instantaneous chemical reaction in slurries containing the exhaust was detected by a flue gas analyzer sparingly soluble fine reactant particles is presented, (Qingdao Minhope electronic instrument Co., Ltd) with -3 and a mathematical model of mass transfer is the accuracy of 0.1 mg m . The gas flow rates were developed and solved analytically based on penet- controlled by a rotameter (LZB-type, Tianjin flow Ins-

Figure 1. Schematic diagram of the experimental setup: 1) sulfur dioxide cylinder; 2) air compressor; 3) valve; 4) pressure reducing valve; 5) manometer; 6) rotor flow meter; 7) gas mixer; 8) bubble column; 9) feed tank; 10) flue gas detector; 11) manometer.

86 X. LI, C. ZHU: GAS–LIQUID MASS TRANSFER WITH INSTANTANEOUS… Chem. Ind. Chem. Eng. Q. 22 (1) 85−93 (2016)

trument Co., Ltd.) with a measuring range of 0-100 ml The Henry coefficient of SO2 was obtained by -1 -1 min for SO2 and 0-100 L min for air; the accuracy of [21]: rotameters was ±1.5%. For each experimental con- HTT=−exp() 55788 − 8.7615ln + 68.48 (5) dition, at least triplicate independent experiments were conducted to obtain the average value of the Theory volume mass transfer coefficient. The experiments were carried out at 298.15 K and atmospheric pres- The mass transfer accompanied by an instant- sure. aneous irreversible chemical reaction in a slurry bub- A number of methods have been developed to ble column reactor containing sparingly soluble fine measure the volumetric mass transfer coefficient, reactant particles is schematized in Figure 2. The they could be divided into two main categories: 1) solute diffuses from the gas phase into the liquid based on the measure of the concentration of the phase and reacts immediately and completely with solute gas in the liquid phase; 2) based on the the reactant present in the slurry, and then a sharp measure of the gas concentration in the gas phase. reaction plane parallel to the gas-liquid interface is The gas balance method is more adequate for the formed. In the zone between the interface and the three-phase system, because it is very difficult to reaction plane (0 < x < Λ) only the reactant A exists. measure the concentration of the solute gas in the Beyond the reaction plane (x > Λ) only the reactant B liquid phase [20]. When the reaction is rapid and exists. The following simplifying assumptions are instantaneous, the concentration of the solute gas A made for the modeling of gas absorption enhanced by in the bulk liquid phase can be regarded as zero. In sparingly soluble fine reactant particles: 1) the solid this experiment, low superficial gas velocities were particles are spherical and uniform in size, smaller adopted to form a bubble flow regime, thus the mass than the scale of the diffusion length; 2) there is no surface kinetic resistance to particle dissolution, and balance of SO2 is: the solid reactant has a low solubility in the liquid −= − QCin in Q out C out K tol aVC L() avg CL (1) phase and dissolves slowly, thus that particles shrinkage can be neglected [2-6]. where Cin and Cout are inlet and outlet bulk gas-phase concentrations respectively; Cavg is the logarithmic mean liquid phase equilibrium concentration of SO2 and CL bulk liquid-phase concentration. In our experiment, the inlet volume fraction of

SO2 in the gas mixture is very low, thus the variation of the gas phase volume is negligible. For an instantaneous reaction, the concentration of the gas reactant in liquid could be regarded as zero. Therefore, the volume mass transfer coefficient could be calculated by:

QC()− C = in out Katol (2) Cavg Figure 2. Schematic of mass transfer with instantaneous reaction in the slurry. The equilibrium concentration of SO2 in liquid phase could be calculated by Henry’s law: It is supposed that the particle and the sur- CPH= / (3) SO22 SO rounding liquid establish a micro-cell. If rp is the solid particle radius and εp is the solid hold-up of the The logarithmic mean liquid phase concen- particle, the radius of each micro-cell is [7]: tration of SO2 in liquid, Cavg could be calculated: rr= 3 ε (6) ()PHP//− H Cp p = in out Cavg (4) PH/ In the zone between the interface and the ln in  reaction plane (0 < x < Λ), the mass transfer inside PHout / each micro-cell could be considered as a steady-state where Pin is the inlet partial pressure of SO2 and Pout process due to the very small particle size. In the is the outlet partial pressure of SO2. zone between the interface and the reaction plane

87 X. LI, C. ZHU: GAS–LIQUID MASS TRANSFER WITH INSTANTANEOUS… Chem. Ind. Chem. Eng. Q. 22 (1) 85−93 (2016)

(0 < x < Λ), each micro-cell is divided into two parts by The consumption rate of reactant B per unit vol- the micro-spherical reaction plane rλ. The mass ume of slurry can be obtained: transfer process inside each micro-cell may be stated =−ππ23dCB = as follows: RrDBCB443r () r C dr C (17) ddC 2 2 B = =−3[1]DB() − C r() rr − r 0 (7) Bs BLCCp ddrr According to Figure 2, the mass transfer pro- B.C.: r = r , C = B ; r = r , C = 0. p B s λ B cess accompanied by an instantaneous irreversible chemical reaction in a slurry containing sparingly dd2 CA = r 0 (8) ddrr soluble fine reactant particles is divided into two parts by the reaction plane (x = rλ): A-only region from the B.C.: r = rλ, CA = 0; r = rC, CA = CAL. interface to the reaction plane (0 < x < Λ) and B-only The solutions of Eqs. (7) and (8) are respect- region beyond the reaction plane (x > Λ). The reac- ively: tant A and the reactant B react in the reaction plane, and the concentrations of the reactant A and the C=− Brr()λλ11() rr − (9) Bs p reactant B can be regarded as zero. The material =− − balance in the liquid phase before and after the reac- CA CAL()11 rrλλ() rr C (10) tion plane is given: In position of the micro-spherical reaction plane ∂∂2 CCAL=−ε AL − rλ, the following equation can be gotten: ()1 pAD ∂t ∂x 2 (18) 2 ddCCBA −+() −<< DD=− (11) 3[1]DCAAL DB Bs r C() r C r p (0 x xΛ ) BAddrrrrλλ ∂∂CC2 Substitution of Eqs. (9) and (10) into Eq. (11) BL=−()ε BL + 1 pBD 2 yields: ∂t ∂x (19) +−3[1]DB() C r2 () rr −< () x x =+ + Bs BLCCp Λ rλ rpBsAALpBSCAAL()11 DB DC() rDB rDC (12)

I.C.: t = 0, x > 0, CAL = 0, CBL = Bs Combining the Eqs. (10) and (12), the rate of B.C.: x = 0, C = C ∗ AL A ∂∂CC mass transfer in the interface of the micro-cell is: x = x , C = C = 0, DDALBL=− Λ AL BL AB∂∂ ∞ xx dC x = , CBL = Bs −=+DDCDBrrrA ()[1]() − (13) AAALBsCCpdr rC To solve Eqs. (18) and (19), the equal diffus- ivities condition and the concept of negative concen- The consumption rate of reactant A per unit tration of the solute are introduced [22]. And it is volume of slurry can be obtained: assumed that: dC =−ππ23A = DA = DB; CBL = –CAL (20) RrDACA443r () r C dr C (14) The following dimensionless variables are =+()2 − 3[1]DCAAL DB Bs r C() r C r p defined: ∗ ∗ − − In the zone beyond the reaction plane (x > xΛ), A = CAL/CA , B = CBL/CA , D = (1 εp)DA = (1 εp)DB, the mass transfer process inside each micro-cell may =−2 kDrrr3[ACCp() 1] (21) be represented as: By putting the concentration of B in the range of ddC r 2 B = 0 (15) x > Λ to B = −A and A' = A+q = 1+B /C ∗ , Eqs. (18) ddrr B s A  and (19) become identical:

B.C.: r = rp, CB = Bs; r = rC, CB = CBL. ∂∂'2' AA' The solution of Eq. (15) is: =−DkA (22) ∂t ∂x 2

=−()() −+ ' CBsC Brr11 rr Cp I.C.: t = 0, x > 0, A =0 (16) +− − ' ∗ CrrrrBL()11 p() p C B.C.: x = 0, A =1+qB = 1+Bs /C A

' x = ∞, A =0

88 X. LI, C. ZHU: GAS–LIQUID MASS TRANSFER WITH INSTANTANEOUS… Chem. Ind. Chem. Eng. Q. 22 (1) 85−93 (2016)

Solving Eq. (22) yields: The gas-liquid specific interfacial area [23]: x ad= 6ε (31) Aq' ()10.5+=(exp()() − xkDerfc + g B 4tD− kt (23) The gas hold-up [24]: + x exp(xkDerfc ) ( )) 0.25 4tD+ kt 32 εε()1−= 0.086 ug ρηρρ()() −  (32) g g single L L g  The mass transfer rate of A can be obtained: The average bubble rising velocity [24]: ∂ CAL ∗ −=+−×DqCDk= ()11()ε 2 A0BA∂x xAp =−−εε53 (24) uub single()11 g() g (33) ×+(erf() kt e−kt π kt ) The single bubble rising velocity:

The exposure time can be estimated by the 0.25 udgd=+()2.14σρ 0.505 (34) bubble diameter and bubble rising velocity: single L L θ = The bubble diameter [24]: 0bdu (25) = π 0.3333 Eq, (24) can be integrated: dV23() 4 (35)

θ 0 ∂ − CAL The bubble volume: ()DdtA0x = 0 ∂x N = (26) 1.2 0.6 A θ V = 0.976()QN g (36) 0

Substitution of Eq. (24) into Eq. (26) using the At 298.15 K, the diffusion coefficient of SO2 in the liquid phase, D , is 1.49×10-9 m2·s-1; the solubility relationship between D and DA yields: A -3 of Mg(OH)2 in water [4], Bs, is 0.46 mol·m . =+ −εθ + × NqAB()11(1)(0.5 p() k 0 (27) − θ ∗ RESULTS AND DISCUSSION ×+θπθk 0 πθ erf() k00 k e)() D A0A C The experimental data are the average values of From Eq. (27) the mass transfer coefficient can the volume mass transfer coefficient from the begin- be calculated by: ning to the SO2 concentration detected by the flue gas -3 =+ −εθ + × analyzer reaching 40 mg·m . KqLB()11(1)(0.5 p() k 0 (28) The calculated values according to the Eqs. − θ ×+θπθk 0 πθ erf() k00 k e)() D A0 (28)-(30) by MATLAB are shown in Figures 3–6.

=−2 -1 Effect of the solid hold-up where kDrrr3[ACCp() 1] (Eq. (21), unit: s ), θ0 is the average exposure time (unit: s). Then the terms The effect of the solid hold-up in the Mg(OH)2 in square brackets is dimensionless. Thus the unit of slurry on the volume mass transfer coefficient of SO2 πθ is presented in Figure 3, indicating that with inc- KL is the same as DA / ()0 . For the gas absorption accompanied by ins- reasing the solid hold-up in the Mg(OH)2 slurry, the tantaneous chemical reactions in slurry bubble col- volume mass transfer coefficient of SO2 increases. In umns containing sparingly soluble fine reactant par- the previous work, four mechanisms of mass transfer ticles, as the partial pressure of the gas reactant is enhancement were introduced, including the shuttling low, the gas-side mass transfer resistance could not mechanism, the boundary layer mixing mechanism, be ignored. The total mass transfer coefficient can be the coalescence inhibition mechanism and boundary expressed as: layer reaction mechanism [25]. In this work, for gas absorbed chemically into a slurry, the fine particles 11 = (29) could provide reactants into the liquid film to enhance K 1 H tol ()+ the mass transfer process, the enhance of mass KK g L transfer could be explained through the boundary layer reaction mechanism. Alper [26-28] introduced The gas-side volume mass transfer coefficient is the concept of effective film thickness to explain the [19]: effect of the catalyst concentration on the mass = 0.73 Kagb170 u (30) transfer enhancement for the gas absorption in cat-

89 X. LI, C. ZHU: GAS–LIQUID MASS TRANSFER WITH INSTANTANEOUS… Chem. Ind. Chem. Eng. Q. 22 (1) 85−93 (2016) alytic slurry reactors. Alper believed that the effective fies that the mass transfer model proposed in this film thickness would decrease with the increase of the paper is reasonable and acceptable in accuracy. solid concentration. Increasing the solid hold-up would lead to an increase of the number of particles per unit volume of slurry in the liquid film, making the reaction plane shift closer to the interface, which leads to decrease of the effective film thickness and intensify the mass transfer process. The higher the solid hold-up is, the closer the reaction plane shifts to the interface, which provides a higher value of the liquid-side mass transfer coefficient, KL. Thus the total volume mass transfer coefficient of SO2 increases with Mg(OH)2 solid hold-up in the slurry.

Figure 4. Effect of the gas flow rate on the volume mass transfer -3 coefficient, Mg(OH)2 solid hold-up: 0.6356×10 ; inlet partial

pressure of SO2: 0.2 kPa.

Effect of the particle size and the particle dissolution near the gas-liquid interface When the solid hold-up is constant, the effect of the particle size near the gas-liquid interface on the

volume mass transfer coefficient of SO2 is shown in Figure 5. For a given solid hold-up, when the radius of

particle is smaller than 5 μm, the change of the par- Figure 3. Effect of the Mg(OH)2 solid hold-up on the volume ticle size is found to have notable effect on the vol- mass transfer coefficient, gas flow rate: 20 L·min-1; inlet partial ume mass transfer coefficient. With the decrease of pressure of SO2: 0.2 kPa. particle size, the particle number near the gas-liquid Effect of the gas flow rate interface, especially in the zone between the interface and the reaction plane (0 < x < Λ), increases more Figure 4 presents the effect of the gas flow rate and more greatly because the volume of the particle on the volume mass transfer coefficient of SO2. It is proportional to the cube of the particle radius. Dec- could be found from Figure 4 that, when the gas flow reasing the particle size would lead to an increase of rate increases, the volume mass transfer coefficient the solid-liquid interface area, which would result in a of SO2 increases. An increase in the gas flow rate marked increase of the particle dissolution rate could increase the gas hold-up and provide more gas- (according to Eqs. (9) and (10)). Alper [28] believed liquid interfacial area. And increasing the gas flow that only the size of particle is smaller than the rate also speeds up the bubble rising velocity, which effective thickness could increase the absorption rate promotes the turbulence in the liquid, and then leads for the gas absorption in catalytic slurry reactors. As to the increase of the liquid-side mass transfer coef- the particle size increases, it becomes increasingly ficient. As a result, the volume mass transfer coef- closer to the effective film thickness, and then the ficient of SO2 increases with the increase of the gas intensification of mass transfer is weakened. In addi- flow rate. tion, with the increase of particle size (e.g., the par- Figures 3 and 4 show the comparison of experi- ticle radius is larger than 7 μm), the influences of the mental mass transfer coefficients with the predicted particle size variation on the particle number become values; the average deviation of present model is 1%. small, and the solid-liquid interface area decreases It can also be seen clearly that the calculated results obviously, leading to the decrease of the particle dis- agree well with the experimental values, which veri- solution rate and the reaction rate. Thus, the intensification of mass transfer is weakened and the vol-

90 X. LI, C. ZHU: GAS–LIQUID MASS TRANSFER WITH INSTANTANEOUS… Chem. Ind. Chem. Eng. Q. 22 (1) 85−93 (2016) ume mass transfer coefficient is remarkably dec- CONCLUSION reased. Therefore, the influence of the variation of the particle size on the volume mass transfer coefficient Mass transfer accompanied by an instantaneous becomes more and more significant with the dec- chemical reaction in a slurry bubble columns con- rease of the particle size. taining sparingly soluble fine reactant particles has been studied theoretically and experimentally. A model has been developed and analytically solved based on the penetration model. The analytical expression of the time-average mass transfer coefficient

has been derived. The fast reactive absorption of SO2

into aqueous Mg(OH)2 slurry in a bubble column was experimentally investigated. The volume mass trans-

fer coefficient of SO2 increases with the increase of

the solid hold-up in Mg(OH)2 slurry, the gas flow rate and the solubility of the particle reactant in the liquid

phase. For constant solid hold-up of Mg(OH)2, when the particle is large (> 7 μm), the variation of the particle size has little influence on the volume mass transfer coefficient; when the particle is small (< 5 μm), the particle size has notable effect on the vol- Figure 5. Effect of the radius of particles on the volume mass ume mass transfer coefficient. The calculated value × -3 transfer coefficient, Mg(OH)2 solid hold-up: 0.4237 10 ; by the present model agrees well with the experimen- inlet partial pressure of SO2: 0.2 kPa. tal data, which validates the proposed mass transfer model. Figure 6 demonstrates the effect of the solubility of the particle reactant in the liquid phase on the Acknowledgement volume mass transfer coefficient of SO2. The increase The authors gratefully acknowledge the financial of solubility of the particle reactant in the liquid phase support of the National Natural Science Foundation of enhances the rate of the particles dissolution rate and China (No. 21306127). the driving force of reactant B transferring from the bulk liquid to the reaction region. Therefore, the con- Nomenclature centration of reactant B in reaction region increases, a gas-liquid specific interfacial area (m2 m-3) and as a result, the reaction rate is accelerated, the A dimensionless concentration of A, defined by reaction plane moves close to the gas-liquid interface, Eq. (21) ' and the effective thickness of mass transfer is red- A defined by A+qB uced, thus the mass transfer between two phases is B dimensionless concentration of B, defined by obviously intensified [29]. Eq. (21) * -1 CA interfacial concentration of A (mol·L ) -1 Bs solubility of B (mol·L )

CA-B concentration of reactant A or B in the micro cell (mol·L-1) -1 CAL-BL liquid concentration of reactant A or B (mol·L )

Cavg average equilibrium concentration of SO2 in the liquid phase (mol·L-1)

Cin inlet concentration of SO2 in the gas phase (mol m-3)

CL concentration of SO2 in liquid phase

Cout outlet concentration of SO2 in the gas phase (mol m-3) C equilibrium concentration of SO in the liquid SO2 2 -1 phase (mol·L ) D defined by Eq.(21) Figure 6. Effect of the solubility of B on the total volumetric 2 -1 -3 DA-B molecular diffusivity of A or B (m s ) mass transfer, Mg(OH)2 solid hold-up: 0.8475×10 ; inlet partial d bubble diameter (m) pressure of SO2: 0.2 kPa; radius of particles:10 μm.

91 X. LI, C. ZHU: GAS–LIQUID MASS TRANSFER WITH INSTANTANEOUS… Chem. Ind. Chem. Eng. Q. 22 (1) 85−93 (2016) g gravitational acceleration (m·s-2) [3] S. Uchida, K. Koide, W.Y. Shindo, Chem. Eng. Sci. 30 H Henry coefficient (kmol·m-3·atm-1) (1975) 644-646 k defined by Eq.(21) [4] M.V. Dagaonkar, A.A.C.M. Beenackers, V.G. Pangarkar, -1 Chem. Eng. Sci. 56 (2001) 1095-1101 Kg gas-side mass transfer coefficient (m·s ) [5] M.V. Dagaonkar, A.A.C.M. Beenackers, V.G. Pangarkar, KL liquid-side physical mass transfer coefficient (m·s-1) Chem. Eng. J. 81 (2001) 203-212 -1 [6] F. Scala, Ind. Eng. Chem. Res. 41 (2002) 5187-5195 Ktol total mass transfer coefficient (m·s ) N number of holes on the gas distributor plate [7] V.A. Juvekar, A.A. Joshi, R. Thaokar, Ind. Eng. Chem. Res. 46 (2007) 3283-3295 NA average flux of A relative to a phase boundary (mol·m-2·s-1) [8] A. Mehra, Chem. Eng. Sci. 51 (1996) 461-477 [9] S. Kakaraniya, C. Kari, R. Verma, A. Mehra, Ind. Eng. PSO2 partial pressure of SO2, Pa Chem. Res. 46 (2007) 1904-1913 Pin inlet partial pressure of SO2, Pa [10] S. Kakaraniya, C. Kari, R. Verma, A. Mehra, Ind. Eng. Pout outlet partial pressure of SO2, Pa ∗ Chem. Res. 46 (2007) 3170-3179 qB defined by Bs/C A [11] M.K. Akbar, S.M. Ghiaasiaan, Chem. Eng. Sci. 59 (2004) Q flow rate of gas (m3·s-1) 967-976 r radial coordinate from the particle (m) [12] J. Patoczka, D.J. Wilson, Sep. Sci. Technol. 19 (1984) rc radius of the micro cell (m) 77–93 r radius of the particle (m) p [13] V.D. Mehta, M.M. Sharma, Chem. Eng. Sci. 21 (1966) rλ radius coordinate of the reaction plane in 361–365 micro-cell (m) [14] M. Filla, J.F. Davidson, J.F. Bates, M.A. Eccles, Chem. RA-B consumption rate of A or B per unite volume Eng. Sci. 31 (1976) 359–367 -3 -1 of slurry (mol·m ·s ) [15] J.S. Cho, N. Wakao, J. Chem. Eng. Jpn. 21 (1988) 576– R average total flux of A in the bubble column –581 SO2 -1 (mol·s ) [16] F.A.N. Rocha, J.R.F. Guedes de Carvalho, Chem. Eng. t time (s) Res. Des. 62 (1984) 303–320 T temperature (K) [17] J.R.F. Guedes de Carvalho, F.A.N. Rocha, M.I. Vascon- -1 ub superficial gas velocity (m·s ) celos, M.C.M. Silva, F.A.R. Oliveira, Chem. Eng. Sci. 41 -1 (1986) 1987–1994 usingle rise velocity of a single bubble (m·s ) V average volume of the bubble (m3) [18] R.C. Rodrigues, C.P. Ribeiro, P.L.C. Lage, Chem. Eng. J. 137 (2008) 282–293 VL volume of the liquid in the bubble column (m3) [19] E. Sada, H. Kumazawa, C. Lee, N. Fujlwara, Ind. Eng. x distance from the surface (m) Chem. Process. Des. Dev. 24 (1985) 255-261 [20] L. Poughon, D. Duchez, J.F. Cornet, C.G. Dussap, Bio- xΛ reaction plane distance from the interface (m) process Biosyst. Eng. 25 (2003) 341-348 Greek symbols [21] S. Ebrahimi, C. Picioreanu, Chem. Eng. Sci. 58 (2003)

εg gas hold-up 3589-3600

εp the solid hold-up of the particles [22] R. Garg, S. Nair, A.N. Bhaskarwar, Chem. Eng. J. 76 (2000) 89-98 θ0 exposure time (s) η effective slurry viscosity (kg·m-1·s-1) [23] N. Kantarci, F. Borak, K.O. Ulgen, Process Biochem. 40 λ radius distance of the reaction plane from the (2005) 2263-2283 particle (m) [24] S.K. Jana, A.N. Bhaskarwar, Chem. Eng. Sci. 65 (2010) Λ reaction plane distance from the interface (m) 3649-3659 -3 [25] K.C. Ruthiya, J. Van der Schaaf, B.F.M. Kuster, J.C. ρL liquid density (kg·m ) -3 Schouten, Chem. Eng. J. 96 (2003) 55-69 ρg gas density (kg·m ) -1 [26] E. Alper, W.D. Deckwer, Chem. Eng. Sci. 36 (1981) σL surface tension of liquid (N·m ) 1097-1099 [27] W.D. Deckwer, E. Alper, Chemie Ing. Tech. 52 (1980) REFERENCES 219-228 [1] P.A. Ramachandran, M.M. Sharma, Chem. Eng. Sci. 24 [28] E. Alper, B. Wichtendahl, W.D. Deckwer, Chem. Eng. Sci. (1969) 1681-1686 35 (1980) 217-222 [2] P.A. Ramachandran, Ind. Eng. Chem. Res. 46 (2007) [29] X.Q. Gao, Y.G. Ma, C.Y. Zhu, G.C. Yu, Chinese J. Chem. 3137-3152 Eng. 14 (2006) 158-163.

92 X. LI, C. ZHU: GAS–LIQUID MASS TRANSFER WITH INSTANTANEOUS… Chem. Ind. Chem. Eng. Q. 22 (1) 85−93 (2016)

XIAOLEI LI PRENOS MASE GAS–TEČNOST PRAĆEN CHUNYING ZHU TRENUTNOM HEMIJSKOM REAKCIJOM U School of Chemical Engineering BARBOTAŽNOJ KOLONI U PRISUSTVU FINIH and Technology, State Key ČESTICA REAKTANTA Laboratory of Chemical Engineering, Tianjin University, Tianjin, China U ovom radu analiziran je teorijski prenos mase praćen trenutnom nepovratnom hemijskom reakcijom u barbotažnoj koloni u prisustvu slabo rastvornih finih čestica reaktanta. NAUČNI RAD Na osnovu teorije penetracije, u kombinaciji sa modelom ćelija, razvijen je jedno-dimen- zionalni model prenosa mase. Ovaj model uzima u obzir uticaj veličine čestica i brzine rastvaranja blizu kontaktne površine gas-tečnost na prenos mase. Model prenosa mase je rešen, tako da je dobijen analitički izraz za koeficijent prenosa mase u funkciji vremena.

Reaktivna apsorpcija SO2 iz gasne smeše u suspenziji Mg(OH)2 u vodi je eksperimentalno ispitana u reaktoru tipa barbotažne kolone radi validacije modela prenosa mase. Rezultati pokazuju da razvijeni model dobro predviđa koeficijent prenosa mase u komplikovanom trofaznom sistemu gasno-tečno-čvrsto sa trenutnom ireverzibilnom hemijskom reakcijom.

Ključne reči: prenos mase, desumporizacija, barbotažna kolona, trenutna reakcija, suspenzija.

Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 22 (1) 95−100 (2016) CI&CEQ

JELENA POPOVIĆ1 THE EXAMINATION OF SENSITIVITY TO GORAN RADENKOVIĆ2 1 CORROSION OF NICKEL-TITANIUM AND JOVANKA GAŠIĆ STAINLESS STEEL ENDODONTIC SLAVOLJUB ŽIVKOVIĆ3 1 INSTRUMENTS IN TOOTH ROOT CANAL ALEKSANDAR MITIĆ • MARIJA NIKOLIĆ1 IRRIGATING SOLUTIONS RADOMIR BARAC1 Article Highlights 1 Department of Restorative • Corrosion of Ni-Ti and stainless steel endodontic files in irrigating solutions was exam- Dentistry and Endodontics, Clinic ined of Dentistry, Medical Faculty, • Testing of sensitivity to corrosion was performed by dynamic potentiometric method University of Niš, Niš, Serbia • Measurements were performed in 5.25% NaOCl, 0.2% CHX and 17% EDTA 2Department of Production Engineering, Faculty of Mechanical Abstract Engineering, University of Niš, Niš, The application of irrigating solutions is essential in chemomechanical treat- Serbia ment of tooth root canal. However, chemical and electrochemical aggressive- 3Department of Restorative ness of the solutions, which directly act on the instruments, may damage their Dentistry and Endodontics, Faculty surface. The aim of this study was to investigate the sensitivity of the nickel- of Dentistry, University of Belgrade, titanium (Ni-Ti) and stainless steel endodontic files to corrosive action of the Belgrade, Serbia sodium hypochlorite (NaOCl), chlorhexidine gluconate (CHX) and ethylenediamine tetraacetic acid (EDTA). Testing of sensitivity to corrosion of the SCIENTIFIC PAPER instruments was performed by dynamic potentiometric method. Measurements UDC 669.245:669.14.018.8:616.314-08 were made in 5.25% NaOCl, 0.2% CHX and 17% EDTA. Ni-Ti instruments immersed in 5.25% NaOCl showed the most intensive corrosive changes and DOI 10.2298/CICEQ150103023P the lowest value of pitting potential of 1.1 V. Stainless steel instruments immersed in 5.25% NaOCl showed higher value of pitting potential of 1.5 V. Stainless steel instruments immersed in 0.2% CHX showed lower corrosive surface changes and higher value of pitting potential of 1.6 V, whereas Ni-Ti instruments immersed in 0.2% CHX showed the pitting potential of 1.9 V. The corrosion was not observed in both types of instruments after immersion in 17% EDTA. The use of 5.25% NaOCl and 0.2% CHX may cause severe surface corrosion of Ni-Ti and stainless steel endodontic files. Keywords: corrosion, irrigating solutions, nickel-titanium, stainless steel, endodontic instruments.

Chemomechanical root canal preparation is tapering funnel shape, flowing with the original canal essential during endodontic treatment and involves from the coronal access to the apex. The functions of procedures of cleaning and shaping with endodontic the irrigants are to act as media for removing debris, instruments and irrigating solutions. The purpose of as lubricants, to dissolve smear layer from dentinal mechanical instrumentation is to obtain a continuous walls and to promote root canal sterility [1]. Many solutions, such as sodium hypochlorite (NaOCl), hydro-

gen peroxide (H O ), citric acid (C H O ), ethylene- Correspondence: J. Popović, Department of Restorative Dentis- 2 2 6 8 7 try and Endodontics, Clinic of Dentistry, Medical Faculty, Uni- diamine tetraacetic acid (EDTA), chlorhexidine glu- versity of Niš, Blv. Dr Zorana Djindjica 52, 18000 Niš, Serbia. conate (CHX) and physiological saline, have been E-mail: [email protected] used for root canal irrigation [2]. Even though the Paper received: 3 January, 2015 Paper revised: 11 June, 2015 benefits of irrigating solutions are essential for che- Paper accepted: 5 July, 2015 momechanical preparation, chemical and electro- • The paper was given as poster presentation at the Rosov pin chemical aggressiveness of these solutions may 2014, the second regional roundtable: Refractory, process industry and nanotechnology. damage the surface of the instruments [3].

95 J. POPOVIĆ et al.: THE EXAMINATION OF SENSITIVITY TO CORROSION… Chem. Ind. Chem. Eng. Q. 22 (1) 95−100 (2016)

There are many literature data about suscep- before the start of the potential rise and this time was tibility to corrosion of endodontic instruments in irri- set by the program. Anodic E–I polarization curves gating solutions [2,4]. The corrosion process could be were recorded by using software Par Stat by means activated during chemomechanical treatment, chem- of the linear sweep technique (sweep rate 0.2 mV/s) ical disinfection of the instruments and sterilization in an air atmosphere at room temperature of 23±3 °C. [5]. Corrosion adversely affects the metallic surfaces Potential value that showed sharp rise of the current by causing pitting and porosity, and decreases the was assigned as pitting potential. The sharp increase cutting efficiency of endodontic files [6]. Several stu- of the current was a result of local dissolution of the dies [7,8] have shown that corrosion of the endo- metal and forming of the pits. The measurements dontic files can degrade the mechanical properties were repeated six times for each solution and the and suddenly cause undesirable cracks that occur each type of the file, and the results were given as during root canal preparations. mean values. Statistical analysis was carried out The purpose of this study was to evaluate sen- using Student’s t-test and Mann-Whitney U test sitivity to corrosion of nickel-titanium and stainless- (SigmaStat statistical software). Electrochemical test- steel endodontic files in most commonly used root ings were performed at Department of Production canal irrigating solutions, NaOCl, CHX and EDTA. Engineering, Faculty of Mechanical Engineering, Uni- versity of Niš, and Department of Physical Chemistry EXPERIMENTAL and Electrochemistry, Faculty of Technology and Metallurgy, University of Belgrade. The study included 36 hand endodontic files divided according their type; 18 nickel-titanium (I- RESULTS AND DISCUSSION -FLEX, IMD, USA) and 18 stainless-steel (NTI-Kahla GmbH, Germany). To remove all debris received from The results of the study are shown in Table 1. the manufacturers, the files were cleaned in an ultra- The corrosion resistance was the lowest in the group sonic bath (JUS-S01, JEOL) with distilled water for 15 of Ni-Ti instruments immersed in 5.25% NaOCl. The min at the frequency of 28 kHz immediately after pitting potential was recorded at 1.1 V (Figure 1). taking them from the original packages. Each type of Higher resistance to corrosion was observed in Ni-Ti the instrument was divided into three groups accord- instruments tested in 0.2% CHX. The measurements ing to the irrigant solutions examined in the study, so showed that the pitting potential was 1.9 V (Figure 2). each group consisted of six files. Based on the obtained results it can be stated that Ni- Measurements were performed in 5.25% NaOCl -Ti instruments immersed in 5.25% NaOCl and 0.2% (prepared in the laboratory), 0.2% CHX (R4, Septo- CHX showed current increases and hence the tend- dont, France, diluted to 0.2%) and 17% EDTA (pre- ency to pitting corrosion (Figures 1 and 2). Comparing pared in the laboratory). All solutions used in this the behavior of Ni-Ti instruments in 5.25% NaOCl and study were freshly prepared, and stored in adequate 0.2% CHX we can notice that NaOCl caused higher conditions. The corrosion behaviour was assessed current increase that means less corrosion resist- using potentiodynamic method. The experiments ance. Statistical analysis showed that this difference were carried out in an ordinary, three-compartment was statistically significant (P < 0.001). On the cont- cylindrical glass cell. The counter electrode was a Pt rary, Ni-Ti instruments immersed in 17% EDTA showed foil and the reference electrode was a saturated calo- the highest resistance to corrosion. The rise of the mel electrode (SCE). All potentials were referred to current was not observed in the whole range of exam- SCE. The working electrode – endodontic instrument – ined potentials and the value remained approximately was placed into the cell in such a way that only the constant (Figure 3). working part of the instrument was immersed in the Similar behavior was observed in the group of solution, whereas the base and the hand were above stainless steel instruments. The increase of current the solution. The instruments were immersed 15 s density was also high in 5.25% NaOCl (1.5 V) and

Table 1. Pitting potential values of the Ni-Ti and stainless steel instruments in tested irrigant solutions

Instrument Irrigants Mean±SD Std. Error C.I. of Mean Max-Min Median Ni-Ti 5.25% NaOCl 1.1±0.089 0.037 ±0.094 1.2-1.0 1.1 0.2% CHX 1.9±0.141 0.058 ±0.148 2.1-1.7 1.9 Stainless steel 5.25% NaOCl 1.5±0.141 0.058 ±0.148 1.7-1.3 1.5 0.2% CHX 1.6±0.063 0.026 ±0.066 1.7-1.5 1.6

96 J. POPOVIĆ et al.: THE EXAMINATION OF SENSITIVITY TO CORROSION… Chem. Ind. Chem. Eng. Q. 22 (1) 95−100 (2016)

0.2% CHX (1.6 V), but the difference was not statistic- rosion and deterioration of the endodontic instruments ally significant (Figures 4 and 5). No significant inc- [9-11]. Corrosion is a deterioration of a metal by che- rease of current in wide range of examined potentials mical or an electrochemical reaction with its environ- was observed after immersion of stainless steel ins- ment, and a technique that evaluates the electroche- truments in 17% EDTA (Figure 6). mical properties of the instrument-irrigating solutions system would seem most appropriate in studying cor- 1,5x10-2 rosion [12]. Electrochemical techniques that are

2 Ni-Ti based on the electrode potential-current character- NaOCl istics define the susceptibility of a metal to react with 1,0x10-2 its environment [13].

Ni-Ti 5,0x10-3 1,0x10-2 EDTA 2 Current density, A/cm density, Current

0,0

1,01,21,41,6 Ep 0,0 Potential, V (SCE) Figure 1. Potentiodynamic polarisation curve of the Ni-Ti file in

5.25% NaOCl. Current density, A/cm

-2 -1,0x10 -3 1,5x10 1,01,21,41,61,82,0

2 Potential, V (SCE) Ni-Ti CHX -3 Figure 3. Potentiodynamic polarisation curve of the Ni-Ti file in 1,0x10 17% EDTA.

8,0x10-2 -4 5,0x10 SS NaOCl

2 -2 6,0x10 Current density, A/cm

0,0 -2 Ep 4,0x10

1,01,21,41,61,82,0 2,0x10-2 Potential, V (SCE)

A/cm Currentdensity,

Figure 2. Potentiodynamic polarisation curve of the Ni-Ti file in 0,0 Ep 0.2% CHX. 0,60,81,01,21,41,61,8 According to the examined potentials in both Potential, V (SCE) types of the instruments, after immersion in 5.25% Figure 4. Potentiodynamic polarisation curve of the stainless NaOCl Ni-Ti instruments showed less corrosion steel file in 5.25% NaOCl. resistance compared to the stainless steel instruments and this difference was statistically significant During endodontic therapy, the most frequently (P < 0.001). After immersion in 0.2% CHX, Ni-Ti ins- used irrigant is sodium hypochlorite (NaOCl) in a con- truments showed higher resistance to corrosion com- centration range of 0.5–6% [14]. It is an agent with pared to the stainless steel instruments, and the dif- wide spectrum of antimicrobial action and tissue dis- ference was statistically significant (P < 0.01). solution capacity [15], which is also used as a pre- The chemical mechanisms that occur either soaking solution in cleaning procedures of endodontic during instrumentation and irrigation of the root canal instruments after clinical use [9]. However, it is highly system, or after instrumentation (in procedures of ins- corrosive to metals and could cause corrosion of the trument disinfection and sterilization), may cause cor- endodontic files. Corrosion pattern involves pitting

97 J. POPOVIĆ et al.: THE EXAMINATION OF SENSITIVITY TO CORROSION… Chem. Ind. Chem. Eng. Q. 22 (1) 95−100 (2016) and potentially weakening of the structure of the ins- endodontic instruments [2]. The results of this study truments [16]. NaOCl contains active Cl-, and it is confirmed that intensive surface corrosion can occur well-known that Cl- is an aggressive ion that generally after immersion of the files in 0.2% CHX. According to increases corrosion rates [17]. This study showed that the Matamala [23], this high rate of corrosive changes the corrosion rate of the endodontic files was high in may depend on its acidic pH (5.72), as the acidic 5.25% NaOCl. These results are in accordance with environment increases the corrosion rate. the results of earlier studies and confirm that the corrosion of endodontic files in NaOCl is possible. 1,5x10-3 NaOCl is corrosive to many metals and selectively removes nickel from the Ni-Ti alloy [18]. Busslinger et 1,0x10-3 al. [19] found measurable release of titanium when 2 SS EDTA Lightspeed Ni-Ti files were immersed in NaOCl sol- -4 5,0x10 ution for 30 and 60 min. In the study of Stokes et al. [6] corrosion was visually observed on endodontic 0,0 files after immersion in 5.25% NaOCl, there was significant difference in corrosion frequency between -5,0x10-4 brands, but there was no difference between stainless A/cm density, Current Ep steel and Ni-Ti instruments. Oztan et al. [2] revealed severe corrosion on the surface of the stainless steel -1,0x10-3 endodontic instruments after immersion in 5.25% 0,8 1,0 1,2 1,4 1,6 1,8 Potential, V(SCE) NaOCl, in accordance to O’Hoy et al. [9] who have shown evident signs of corrosion after overnight Figure 6. Potentiodynamic polarisation curve of the stainless immersion of endodontic instruments in NaOCl. The steel file in 17% EDTA. fact that chloride and fluoride ions have negative effects on the corrosion resistance of stainless steel Ethylenediamine tetraacetic acid (EDTA) is the and Ni-Ti alloys is used in few investigations to pro- chelating irrigant with inorganic tissue dissolution cap- mote electrochemical dissolution and removing endo- acity, and is used due to ability to lubricate and facil- dontic instruments in cases where they are fractured itate root canal instrumentation especially in prepar- in the root canal system [20,21]. ation of narrow and curved root canals. In endodontics it is used as 15-17% solution [24]. The results -3 1,5x10 of potentiodynamic test in this study did not reveal corrosion of endodontic files after immersion in 17%

2 -3 SS EDTA, and it was in accordance with literature data 1,0x10 CHX [2,4,25]. Öztan et al. [2] have reported the lowest corrosion rate of stainless steel endodontic files in 5,0x10-4 17% EDTA. They have stated that EDTA forms com- plexes with metal ions (Fe, Ni, Cr, Co, etc.) at pH values < 4. EDTA’s ability to protect and passivate 0,0 instruments is due to its ability to complex with iron to Current density, A/cm Ep form an inhibiting barrier to oxidation and corrosion -5,0x10-4 [26]. According to Darabara et al. [4], large molecules 0,81,01,21,41,61,82,0 of R-EDTA have greater difficulty in concentrating and Potential, V (SCE) orienting the pit so as to increase the acidity to ade-

quate values for trigger corrosion. Atomic force micro- Figure 5. Potentiodynamic polarisation curve of the stainless scopic evaluation of Fayyad and Mahran [25] showed steel file in 0.2% CHX. that immersion in 17% EDTA did not affect the surface roughness of the Ni-Ti endodontic instruments. Chlorhexidine gluconat (CHX) at concentrations Endodontic files and reamers are generally 0.1-2% is a broad spectrum antimicrobial agent that is accepted as reusable instruments. In purpose to eli- used during root canal irrigation. Its cationic structure minate the risk of infection transmission, these instru- provides a unique property, named substantivity. This ments need to be cleaned and sterilized thoroughly prolonged antimicrobial activity in the root canal may after clinical use [27]. However, these procedures last up to 12 weeks [22]. However, the literature data could potentiate surface corrosion in irrigating sol- revealed that CHX can cause severe corrosion of

98 J. POPOVIĆ et al.: THE EXAMINATION OF SENSITIVITY TO CORROSION… Chem. Ind. Chem. Eng. Q. 22 (1) 95−100 (2016) utions [28]. Casella and Rosalbino [29] confirmed that [7] Y. Shen, M, Haapasalo, G.S. Cheung, B. Peng, J. Endod. sterilization process had negative influence on the 35 (2009) 129-132 corrosion behavior of endodontic instruments, and the [8] Y. Shen, G.S. Cheung, B. Peng, M. Haapasalo, J. Endod. effect appears to be more dramatic for longer steriliz- 35 (2009) 133-136 ation treatment periods. The presence of protein deb- [9] P.Y.Z. O’Hoy, H.H. Messer, J.E.A. Palamara, Int. Endod. ris in form of ground tooth structure or collagen, with J. 36 (2003) 724-732 NaOCl solutions, could increase the severity of the [10] A.A. Yahya, K.W. Majida, AL-Hashimi, J. Bagh. College. Dentistry. 21 (2009) 53-59 surface attack on the instrument [13]. Stokes et al. [6] [11] G. Spagnuolo, G. Ametrano, D. D’Antò, C. Rengo, M. evaluated the corrosive effect of 5.25% NaOCl on Simeone, M. Riccitiello, M. Amato, Int. Endod. J. 45 stainless steel and Ni-TI files using five commercial (2012) 1148–1155 brands. They reported that both the corroding and [12] G. Radenković, S.K. Zečević, Z. Cvijović, D.M. Dražić, J. non-corroding files were present in the same pack- Serb. Chem. Soc. 60 (1995) 51-59 ages. Those results showed that the severity of cor- [13] H.J. Mueller, J. Endod. 8 (1982) 246-252 rosive changes could also depend on manufacturing [14] I. Heling, I. Rotstein, T. Dinur, Y. Szwec-Levine, D. Stein- process and quality control. berg, J. Endod. 27 (2001) 278-280 [15] S. Stojicic, S. Zivkovic, W. Qian, H. Zhang, M. Haapa- CONCLUSION salo, J. Endod. 36 (2010) 1558-1562 [16] E. Berutti, E. Angelini, M. Rigolone, G. Migliaretti, D. The results of this study indicated that 5.25% Pasqualini, Int. Endod. J. 39 (2006) 693-699 NaOCl and 0.2% CHX, used as root canal irrigants, [17] H. Katayama, M. Yamamoto, T. Kodama, Corros. Eng. 49 cause severe corrosion on the surface of the Ni-Ti (2000) 41-44 and stainless steel endodontic files. The use of EDTA [18] N.K. Sarkar, W. Redmond, B. Schwaninger, A.J. Gold- did not cause corrosion of the surface of both types of berg, J. Oral. Rehabil. 10 (1983) 121-128 instruments. Due to the possibility of corrosion acting [19] A. Busslinger, B. Sener, F. Barbakow, Int. Endod. J. 31 to deteriorate endodontic instruments, irrigants should (1998) 290-294 be rinsed from files immediately after use and files [20] L.R.L. Aboud, F. Ormiga, J.A.C.P. Gomes, Int. Endod. J. should be replaced frequently. 47 (2014) 155-162 [21] C.C.F. Amaral, F. Ormiga, J.A.C.P. Gomes, Int Endod. J. Acknowledgement 48 (2015) 137-144 This work has been supported by the grant No. [22] J. Gasic, J. Popovic, S. Zivkovic, A. Petrovic, R. Barac, 175102 of the Serbian Ministry Education, Science M. Nikolic, Microsc. Res. Tech. 75 (2012) 1099-1103 and Technological Development. [23] G.R. Matamala, Corrosion. 43 (1987) 97-100 [24] M. Hülsmann, M. Heckendorff, A. Lennon, Int. Endod. J. REFERENCES 36 (2003) 810-830 [25] D.M. Fayyad, A.H. Mahran, Int. Endod. J. 47 (2014) 567- [1] J. Walcott, V.T. Himel, J. Endod. 23 (1997) 217-224 –573 [2] D.M. Öztan, A.A. Akman, L. Zaimoglu, S. Bilgiç, Int. [26] G. Reinhard, M. Radtke, U. Rammelt, Corros. Sci. 33 Endod. J. 35 (2002) 655-659 (1992) 307-313 [3] B.C. Sağlam, S. Koçak, M.M. Koçak, Ö. Topuz, Microsc. [27] M.A. Saghiri, F. Garcia-Godoy, M. Lotfi, P. Mehrvazfar, Res. Tech. 75 (2012) 1534-1538 M. Aminsobhani, S. Rezaie, K. Asgar, Scanning. 34 [4] M. Darabara, L. Bourithis, S. Zinelis, G.D. Papadimitriou, (2012) 309-315 Int. Endod. J. 37 (2004) 705–710 [28] X.R. Nóvoa, B. Martin-Biedma, P. Varela-Patiño, A. Col- [5] K. Sood, B. Mohan, L. Lakshminarayanan, Endodonto- lazo, A. Macías-Luaces, G. Cantatore, M.C. Pérez, F. logy 18 (2006) 34-41 Magán-Muñoz, Int. Endod. J. 40 (2007) 36–44 [6] W.O. Stokes, M.P. Di Fiore, T.J. Barss, A. Koerber, L.J. [29] G. Casella, F. Rosalbino, Corros. Eng. Sci. Technol. 46 Gilbert, P.E. Lautenschlager, J. Endod. 25 (1999) 17-20 (2011) 521-523.

99 J. POPOVIĆ et al.: THE EXAMINATION OF SENSITIVITY TO CORROSION… Chem. Ind. Chem. Eng. Q. 22 (1) 95−100 (2016)

JELENA POPOVIĆ1 ISPITIVANJE OSETLJIVOSTI ENDODONTSKIH GORAN RADENKOVIĆ2 INSTRUMENATA OD NIKL-TITANIJUMA I 1 JOVANKA GAŠIĆ NERĐAJUĆEG ČELIKA NA KOROZIJU U SLAVOLJUB ŽIVKOVIĆ3 1 RASTVORIMA ZA IRIGACIJU KANALA ALEKSANDAR MITIĆ MARIJA NIKOLIĆ1 KORENA ZUBA RADOMIR BARAC1 Primena sredstava za irigaciju kanala korena zuba je od suštinskog značaja u endo- 1 Odeljenje za bolesti zuba i dontskoj terapiji. Međutim, hemijska i elektrohemijska agresivnost ovih rastvora, koji endodonciju, Klinika za direktno deluju na instrumente, može oštetiti njihovu površinu. Cilj istraživanja je bilo stomatologiju, Medicinski fakultet, ispitivanje osetljivosti endodontskih turpija od nerđajućeg čelika i nikl-titanijuma (Ni-Ti) na Univerzitet u Nišu, Niš, Srbija koroziono delovanje natrijum-hipohlorita (NaOCl), hlorheksidin-glukonata (CHX) i etilen- 2Katedra za proizvodno-infor- macione tehnologije i diamin tetrasirćetne kiseline (EDTA). Ispitivanje otpornosti instrumenata na koroziju je menadžment, Mašinski fakultet, izvedeno potenciodinamičkom metodom. Merenje je izvedeno u rastvorima 5,25% NaOCl, Univerzitet u Nišu, Niš, Srbija 0,2% CHX i 17% EDTA. Najintenzivnije korozione promene i najnižu vrednost piting 3Klinika za bolesti zuba i potencijala od 1.1 V su pokazali Ni-Ti instrumenti potapani u 5,25% NaOCl. Višu vrednost endodonciju, Stomatološki fakultet, piting potencijala od 1.5 V su pokazali instrumenti od nerđajućeg čelika posle potapanja u Univerzitet u Beogradu, Beograd, 5,25% NaOCl. Manji intenzitet korozionih promena i piting potencijal od 1.6 V pokazali su Srbija instrumenti od nerđajućeg čelika potapani u 0,2% CHX, dok su Ni-Ti instrumenti potapani u 0,2% CHX pokazali vrednost piting potencijala od 1.9 V. Korozija nije zapažena kod obe NAUČNI RAD vrste instrumenata nakon potapanja u 17% EDTA. Primena 5,25% NaOCl i 0,2% CHX može izazvati ozbiljnu koroziju površina endodontskih turpija od nerđajućeg čelika i nikl-titanijuma.

Ključne reči: korozija, irigacioni rastvori, nikl-titanijum, nerđajući čelik, endodont- ski instrumenti.

100 Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 22 (1) 101−110 (2016) CI&CEQ

O.S. GLAVAŠKI1 PHOTODEGRADATION OF DIMETHENAMID-P S.D. PETROVIĆ2 IN DEIONISED AND GROUND WATER V.N. 3 RAJAKOVIĆ-OGNJANOVIĆ Article Highlights 1 T.M. ZEREMSKI • Photocatalytic degradation of dimethenamid-P herbicide is presented A.M. DUGANDŽIĆ2 • Degradation was studied in deionised and ground water under different conditions 2 • Photocatalytic degradation of dimethenamid-P is much faster in ground water D.Ž. MIJIN • HPLC showed almost complete removal of herbicide after 90 min in both water • 1Institute of Field and Vegetable TOC showed herbicide was mineralized 64% in deionised and 50% in ground water

Crops, Novi Sad, Serbia Abstract 2Faculty of Technology and Metallurgy, University of Belgrade, The study of photodegradation of dimethenamid-P herbicide was performed in Belgrade, Serbia deionised and ground water using TiO2 as a catalyst under UV light. The effect 3 • Faculty of Civil Engineering, of electron acceptor (H2O2), scavenger of OH radicals (C2H5OH) and scav- University of Belgrade, Bulevar enger of holes (NaCl and Na2SO4) as well as solution pH was analyzed. The Belgrade, Serbia photodegradation of dimethenamid-P was followed by HPLC. The formation of transformation products was followed using high performance liquid chromato- SCIENTIFIC PAPER graphy-electrospray mass spectrometry. Ion chromatography and total organic UDC 543.544:632.954:66 carbon measurements were used for the determination of the mineralization level. HPLC analysis showed the almost complete removal of herbicide after DOI 10.2298/CICEQ150608025G 90 min in deionised and ground water, while total organic carbon analysis showed that dimethenamid-P was mineralized 64 and 50% in deionised and ground water, respectively. The ion chromatography results showed that the mineralization process leads to the formation of chloride, sulphate and nitrate anions during the process. Transformation products were identified and the degradation mechanism was proposed. Keywords: salt effect; ion chromatography; liquid chromatography-electrospray mass spectrometry; photocatalysis; titanium dioxide.

Modern agricultural production in the last dec- levels of pesticides in drinking water should not ade involves the use of pesticides to a large extent. exceed 0.1 mg dm-3 for the individual components, Dimethenamid-P (2-chloro-N-(2,4-dimethyl-3-thienyl)- i.e., for some of their transformation products concen- -N-(2-methoxy-1-methylethyl) acetamide, DMA-P) tration should not exceed 0.5 mg dm-3 [6]. Within the belongs by its chemical properties and structure to strategy of protection of environmental resources, the group of chloroacetamides and plays an important heterogeneous photocatalysis has proved to be one role in the crop protection of broadleaf weeds and of the most effective techniques for the degradation of annual grasses in row crops [1], primarily in corn, organic pollutants [7]. It involves fotoinduction reac- soybean and sorghum [2]. These components include tion accelerated by a solid catalyst [8]. highly toxic and persistent substances and due to TiO2 as a photocatalytic semiconductor is the exceptional reactivity threaten to jeopardize the aqua- most suitable chemical compound for removal of tic environment through agricultural circle and wash- harmful substances from the environment by photo- ing [3-5]. The European Union has stipulated that the catalytic process. Its chemical inertness, stability to the photo and chemical corrosion, as well as low price

are its advantages as a catalyst [9]. Photocatalytic Correspondence: D.Ž. Mijin, Faculty of Technology and Metal- lurgy, University of Belgrade, Karnegijeva 4, 11120 Belgrade, degradation is based on the irradiation of UV light, Serbia. which results in the generation of oxidative species E-mail: [email protected] that are characterized by high and non-selective reac- Paper received: 8 June, 2015 Paper revised: 6 July, 2015 tivity, so they can easily attack and decompose the Paper accepted: 8 July, 2015

101 O.S. GLAVAŠKI et al.: PHOTODEGRADATION OF DIMETHENAMID-P… Chem. Ind. Chem. Eng. Q. 22 (1) 101−110 (2016)

molecules of organic pollutants. Photon has energy NaCl, and Na2SO4 as well as pH value of water sol- which is greater or equal to the band gap energy of ution in two different types of water (deionised and – semiconductor (TiO2). In that way electron (e ) from ground water) were studied. HPLC/MS (high perform- the valence band (VB) excitates to the conduction ance liquid chromatography-electrospray mass spec- band (CB), leaving a positive hole (h+) behind. The trometry) were applied for qualitative identification of energy level at the bottom of the valence zone effect- transformation products. ively reduces the potential of photoelectrons, while the peak energy of the valence zone creates its ability EXPERIMENTAL to oxidize. Electrons and cavities migrate to the surface of the catalyst and reduce species present on its Materials surface. Photogenerated cavities may oxidize organic DMA-P (purity higher than 99%) was supplied − molecules or react with OH , or H2O, oxidising them by Riedel de-Haen (Seelze-Hannover, Germany). • to OH. Photogenerated electrons can also react with Titanium dioxide (TiO2) labeled as P25 supplied by −• oxygen, translating it into superoxide anion O2 radi- Evonik was used in experimental part of the work. All • cal. This reaction leads to additional formation of OH. other chemicals were p.a. or higher grade. Deionized These radicals as very strong oxidative agents having water (DW) was obtained from a Millipore water puri- the ability to oxidize organic pollutants adsorbed on fication system. Ground water (GW) was obtained the surface of TiO2 to mineral products [10]. Redox from public-utility company Water supply and sewage reactions including photons can be presented by Eqs. treatment in Novi Sad, as alluvium of Danube. The -3 − 2– (1)-(5): ground water contains 269.6 mg dm of HCO3 /CO3 , -3 2− -3 – CB +VB 57.5 mg dm of SO4 , 20 mg dm of Cl , 1.496 mg TiO2 + hν(UV) → TiO2(e− + h ) (1) dm-3 of NO –, 0.343 mg dm-3 of Mn2+, 0.600 mg dm-3 • 3 TiO (h+VB) + H O → TiO + H+ + OH (2) -3 2+ -3 2+ 2 2 2 of NH3, 84 mg dm of Ca , 17.5 mg dm of Mg and −CB •− 2.810 mg dm-3 of Fe (total). Conductivity of deionized TiO2(e ) + O2 → TiO2 + O2 (3) –1 • • and ground water was 0.55 and 58.5 µS cm while O − + H+ → HO (4) 2 2 pH was 5.9 and 7.20, respectively. • HO + H+ + TiO (e−CB) → H O + TiO (5) 2 2 2 2 2 Photocatalytic experiment Conclusively, photocatalytic degradation of pes- The photodegradation of DMA-P was inves- ticides can be presented in simplified form by Eqs. tigated in two different types of water, in the deionised (6)-(8): and ground water, with pesticide concentration of Pesticide + h+VB → oxidation products (6) 34.5 mg dm-3. All the reactions were performed in an open reactor, thermostated at 25 °C [11]. For the Pesticide + e−CB → reduction products (7) irradiation an Osram Ultra Vitalux® 300 W lamp was • Pesticide + OH → transformation products (8) used, with ratio of UV-A and UV-B lights 13.6:3. The In order to get higher yield, agriculture relies on position of lamp was 40 cm from the surface of the the application of pesticides. The negative effects of reaction mixture. The temperature of solution changed for 2 °C after 90 min of irradiation. For every pesticides on the quality of ground and surface water 3 are well known [5]. The environmental issues con- experimental cycle 25 cm of the solution was placed cerning pesticides comprehend: inadequate control of into the reactor and stirred for 30 min in the dark. the usage (excessive concentrations of pesticides), Continuous stirring was maintained during the reac- non-biodegradability, long decomposition time and tion. The aliquots were taken at defined time intervals high mobility in different eco-systems. Dimethenamid- (after 10, 20, 30, 60 and 90 min from the beginning of -P belongs to the group of chloroacetamides which the reaction). All the aliquots were filtrated by 0.45 μm are persistent organic pollutants. Its specific feature is Cronus 13 mm Nylon Syringe filters, in order to rem- the migration from the soil to the ground and ground- ove the suspended TiO2 particles before the analysis. water [2]. All the experiments were done in triplicate. In this work, the study on the photocatalytic Analytical procedures behavior of DMA-P in aquatic environment is pre- During 90 min of irradiation time, the samples sented for the first time. the influence of various para- were taken from the suspension. The concentration of meters on the photocatalytic process, such as the ini- herbicide was determined by HPLC (high perform- tial concentration of catalyst, initial DMA-P concen- ance liquid chromatography) analysis. All analysis tration, the concentration of added H O , C H OH, 2 2 2 5 were performed at room temperature.

102 O.S. GLAVAŠKI et al.: PHOTODEGRADATION OF DIMETHENAMID-P… Chem. Ind. Chem. Eng. Q. 22 (1) 101−110 (2016)

The HPLC determinations were carried out with b) investigation of DMA-P degradation under

HPLC instrument Agilent 1100 Series equipped with UV light, in the absence of TiO2 (photolysis) (C(DMA-P) 0: Zorbax Eclipse XDB-C18 (Agilent). The analyses 34.5 mg dm-3, V: 25 cm3, T: 25 °C), were performed in isocratic mode using water/meth- c) heterogeneous photocatalysis of DMA-P sol- -3 anol/acetic acid (200:300:5 V/V/V), the mobile phase ution under UV light and a catalyst (Ckat: 2.0 g dm , 3 -1 -3 3 had flow rate of 0.8 cm min and the column tempe- C(DMA-P) 0: 34.5 mg dm , V: 25 cm , T: 25 °C). rature was 25 °C. The injection volume was 5 μL and DMA-P concentrations have not changed sig- UV detection was carried out at 240.4 nm. nificantly in the case of adsorption and photolysis. The pH value of the samples was adjusted by The results of the adsorption experiment show slight the addition of 0.1 mol dm-3 NaOH or HCl and the decrease in herbicide concentration (less than 5%) determination of pH value was performed on pH during a period of 90 min, indicating only slight ads- metar Inolab pH 730 (Germany). orption on the TiO2 surface. Photolysis results show The chromatographic separations were followed insignificant fall of the initial concentration of DMA-P. by an MS analyzer, Hypersil Gold Thermo Scientific On the other hand, photocatalysis shows almost com- (Bremen, Germany) (50 mm×2.1 mm, 3 mm particle plete destruction of this active substance (more than size) termostated at 25 °C using a Thermo survey 99% determined by HPLC) both in deionised and (USA) HPLC instrument. Injection volume was 50 μL ground water. The comparison of photocatalysis and and flow rate was 0.2 cm3 min-1. The mobile phases adsorption processes implies that stirring of suspen- were: A (0.10% acetic acid/99.9% water) and B sion for 30 min in the dark prior to the photocatalysis (0.10% acetic acid/99.9% acetonitrile). The analyses process is an important step for reaching the adsorp- were performed in isocratic mode. A LCQ Deca mass tion equilibrium [9]. spectrometer equipped with an atmospheric pressure To determine the optimum concentration of TiO2 interface and an ESI ion source was used as a det- for the photodegradation of DMA-P, experiments ector. The LC column effluent was delivered into the were conducted by varying the initial concentration of –3 ion source using nitrogen as sheath and auxiliary gas. TiO2 from 0.5 to 3.0 g dm , while keeping other para- The tuning parameters adopted for ESI source were: meters constant. It was found out that the maximum capillary voltage 45 V, capillary temperature 275 °C, removal efficiency of the chloroacetamide has been spray voltage 6 kV and gas flow was 20 arbitrary achieved with the catalyst concentration of 2.0 g dm–3, units. The analysis was performed in positive ion and, therefore, this concentration of TiO2 has been mode. Mass spectra were recorded across the range selected as the optimum one. Further increase of the 100–400 m/z. catalyst concentration decreases the rate of photo- Ion chromatographic (IC) analysis was per- degradation, and reduces the efficiency of degrad- formed on a Dionex DX-300 ion chromatograph at ation process. Theoretically, the increase of the cat- ambient temperature (25 °C) with a suppressed con- alyst concentration above an optimum value should ductivity detector. Ion chromatograph was equipped not have effect on the photodegradation rate since all with a Dionex IonPac AS14 column. the light available is already utilized. However, higher

Total organic carbon (TOC) was measured mass concentrations of TiO2 Evonik P25 led to the using a Zellweger LabTOC 2100 instrument. aggregation of its particles and thus to a decrease of contact surface between the substrate and the photo- RESULTS AND DISCUSSION catalyst. This caused a decrease in the number of active sites and a lower rate of photodegradation. Preliminary experiments When the concentration of catalyst is exceeded, a In the beginning of the photocatalytic study, part of the UV light is not utilized because of the three different experiments have been carried out in increased turbidity of solution and increased light aqueous environmental matrices with the aim to scattering by the photocatalyst particles, and there- evaluate adsorption and photolysis of the studied fore the overall performance decreases [12]. DMA-P. These experiments were conducted under Initial concentration of DMA-P affects the rate of the following conditions: its photocatalytic degradation. The increase of the a) investigation of DMA-P adsorption on the initial substrate concentration on the catalyst surface -3 • TiO2 in the dark (Ckat: 2.0 g dm , C(DMA-P) 0: 34.5 mg number of molecules/ions that react with OH inc- dm-3, V: 25 cm3, T: 25 °C), reases and the rate of degradation decreases. The increase of the initial substrate concentration above an optimum value leads to the decrease of the effi-

103 O.S. GLAVAŠKI et al.: PHOTODEGRADATION OF DIMETHENAMID-P… Chem. Ind. Chem. Eng. Q. 22 (1) 101−110 (2016) ciency of the photocatalytic process. the substrate The difference in DMA-P degradation rate molecules may adsorb on the catalyst surface instead caused by the change of pH value was more sig- of OH– and water molecules, which then result in the nificant in ground than in the deionised water. The • generation of fewer OH [7,9]. But since very small reaction rate in ground water is decreased in neutral adsorption was observed, the reduced photoactivity of and in the alkaline medium and pH has an insig- semiconductor may be due to the absorbtion of light nificant effect on the rate of degradation in deionised by organic molecules [13]. water. The explanation for the effect of pH on the photocatalytic reaction might be in the influence of The effect of the pH value electrically charged surface of the TiO2 on the phys- The influence of pH value on the photocatalytic ical and chemical properties of the parent molecule. degradation can be explained by electrostatic inter- Although DMA-P molecule is electrically neutral it can actions between the surface of TiO2, solvent mole- be repulsed from the negatively charged surface of cules, substrate and electrically charged radicals the photocatalyst at pH value above 6.8 due to formed during the process. At pH values above the unequal distribution of electron density in the sub- value of point of zero charge (PZC) of TiO2 (6.8), the strate molecule when electronegative atom like Cl is surface will remain negatively charged. For pH < PZC present [4]. the surface will remain positively charged [14]. For the The disappearance rate in the process of photo- influence of pH value on the photocatalytic degrad- catalytic degradation can be described by a pseudo- ation five different pH values were analyzed. The first kinetic order, as shown by Eqs. (9) and (10): adjustments of acidic medium (pH 2.0 and 4.0) and alkaline medium (pH 9.0 and 11.0) were made with C0 = lnkt (9) diluted HCl or NaOH. Before any adjustments the pH C value of pesticide solution in DW and GW was mea- CCe= −kt (10) sured and the obtained values were 6.33 and 7.34 for 0 DW and GW, respectively. The photocatalytic degrad- where C is the concentration of DMA-P at irradiation ation rate of DMA-P in DW and GW as a function of time t, and C0 is the initial concentration of DMA–P. pH value is shown in Figure 1. Lower degradation rate for both aqueous media was at pH values near to The effect of the addition of electron acceptor

PZC. The possible explanation for this phenomenon When TiO2 is used as a photocatalyst, one of – + is the fact that the TiO2 particles tend to agglomerate the problems that arise is the recombination of the e h and thus decrease the yield of degradation. As rep- pair. This problem is particularly apparent in the abs- orted, at pH values equal to the PZC, aggregate par- ence of appropriate electron acceptors which also ticles are larger, and number of active sites on the reduces the efficiency of photocatalytic reaction [17]. catalyst surface is decreased and degradation rate In order to enhance the formation of •OH and inhibit – + reduced [15,16]. e h pair recombination, the effect of addition of H2O2 as an electron acceptor on the efficiency of photodegradation has been investigated in a number of experiments conducted in both deionised and ground water [17,18]. In this study, a series of experiments has been carried out in both media. The obtained

results have shown that for H2O2 concentration of up to 0.005 mol dm-3 the reaction time is increased by 2 times in deionised water and by 1.6 times in ground water (as shown in Figure 2). This may be due to the increased concentration of •OH. Faster degradation

rate in the presence of H2O2 may be attributed to the generation of •OH and OH– in the presence of UV radiation, and not to the formation of less powerful •- O2 oxidant by the reduction of O2 [19]. At higher -3 concentrations (above 0.005 mol dm ), H2O2 acts as a “scavenger” of •OH and holes on the catalyst • Figure 1. The effect of pH value on the photocatalytic surface, leading to the formation of HO2 that react degradation rate of DMA-P (34.5 mg dm-3) in deionised and with •OH to generate oxygen and water as illustrated -3 ground water (ccat 2.0 g dm ).

104 O.S. GLAVAŠKI et al.: PHOTODEGRADATION OF DIMETHENAMID-P… Chem. Ind. Chem. Eng. Q. 22 (1) 101−110 (2016) by Eqs. (11) and (12). As the result, the efficiency of the inhibition of catalytic degradation is more pro- photocatalytic degradation is reduced. nounced in ground water where gradual addition of -3 • • ethanol (up to 0.8 mol dm ) decreases the reaction H O + OH → HO + H O (11) 2 2 2 2 rate 15 times, while in deionised water the reaction + → • + H2O2 + h HO2 + H (12) rate decreases 3 times under the same conditions. According to some researchers, if reaction products, such as oxygen and H2O2 are not present near the surface of TiO2, electron-hole pairs recombine and adsorbed energy is dissipated as heat. H2O2 added in smaller concentrations is able to prevent this reaction [4].

Figure 3. The effect of the added of C2H5OH on the photocatalytic degradation rate of DMA-P (34.5 mg dm-3) in -3 deionised and ground water (ccat 2.0 g dm ).

The effect of the inorganic ions When comparing the results of the rate of photocatalytic degradation of DMA-P in two types of water, Figure 2. The effect of the added H2O2 on the photocatalytic degradation rate of DPA-P (34.5 mg dm-3) in deionised and the great difference can be noticed. The rate of -3 ground water (ccat 2.0 g dm ). decomposition of DMA-P in ground water is two times faster than in deionised water. This could be ascribed The effect of the •OH scavenger to the presence of nitrate ions. Reaction of nitrate ions with photons ends with hydroxyl radicals accord- To confirm and to prove if heterogeneous photo- • ing to the Eqs (13)-(15) [21]: catalysis is taking place through OH, the effect of – – ethanol added in the reaction mixture containing NO3 + hν → NO2 + O (13) DMA-P and TiO Evonik P25 on the reaction rate has – • •– 2 NO + hν → NO + O (14) been investigated in both deionised and ground water 3 2 •– → • • (Figure 3). It was determined that as the ethanol con- O + H2O OH + OH (15) centration increases, the degradation rate decreases, Taking into account the fact that groundwater compared with the same reaction without addition of from the Danube alluvium is slightly alkaline it can be this solvent. The obtained result is in agreement with expected that hydroxyl radicals formed together with previous research from this field [20]. In the same photogenerated oxidative species generated with the research the effect of various solvents on the photo- irradiated TiO2, have higher degradation rates of catalytic degradation of benzidene yellow was stu- – DMA-P. Chen et al. [21] reported that NO3 as died. The results showed that the degradation effi- constituents found in natural waters absorb solar ciency decreases with the addition of solvents in the radiation in UV range less than 350 nm with max- following order: hexane < acetonitrile < 2-propanol < imum at 302 nm. Photolysis of these anions leads to < 1-butanol < 2-methyl-2-propanol. The obtained • • formation of OH under influence of UV radiation, as results confirm that alcohols are good OH scaven- shown by Eqs (13)-(15). gers and the products of reaction are weaker oxidants Ground water also contains dissolved metal (alkoxy-radicals) that react with the substrate. The ions, such as Fe3+, Mn2+, Ca2+ and Mg2+. Wei et al. results obtained in the present study also show that indicated another possible explanation for the differ-

105 O.S. GLAVAŠKI et al.: PHOTODEGRADATION OF DIMETHENAMID-P… Chem. Ind. Chem. Eng. Q. 22 (1) 101−110 (2016)

• ences observed in the kinetics of the process taking Cl– + h+ → Cl (18) place in these two aqueous media [22]. The effect of Cl– + •OH → ClOH•– (19) Fe3+ on the photodegradation efficiency of metami- • • 3+ – + → dophos was studied by varying amount of Fe from ClOH + H Cl + H2O (20) 0.001 to 0.8 mmol dm-3. The results showed that • – •– Cl + Cl → Cl2 (21) 3+ when higher Fe concentrations were added (up to 2– + •2– -3 SO + h → SO (22) 0.5 mmol dm ) its photodegradation efficiency inc- 4 4 2– • •– - reased rapidly (from 37.3 to 55.0%). When the con- SO4 + OH → SO4 + OH (23) centration of this cation exceeds this value, photodeg- • SO – + e – → SO 2– (24) radation efficiency is greatly reduced. It has been 4 CB 4 concluded that positively charged Fe3+ absorbed on 3+ – surface of the TiO2 are more easily reduced (Fe + e → Fe2+) thus decreasing electron-hole pair recombi- • 2– nation. This favours the formation of OH and O2 on the surface of the TiO2. The following reactions occur at the same time:

2+ + 3+ • Fe + H2O2 + H → Fe + OH + H2O (16) 2+ • + 3+ Fe + HO2 + H → Fe + H2O2 (17) It has also been shown that the presence of Na+, K+, Ca2+ and Mg2+ has no effect on the photodegradation rate. This is explained by the fact that these ions are in their most stable oxidation state and as such they do not show affinity for bonding photogenerated electrons and holes [23]. Physicochemical composition of groundwater – indicates the presence of HCO3 . When investigating – the previous studies the effects of HCO3 on the rate of photocatalytic degradation, showed that concentrations above 0.1 mol dm-3 lead to reduced photodegradation efficiency due to the formation of greater •- number of CO3 radical-ions which are less reactive than •OH [24]. However, in ground water where pH is - slightly alkaline, HCO3 are present to a greater extent 2- than CO3 and their concentration in this medium is below 0.05 mol dm-3, being, according to the findings of Lair et al., the most probable explanation for increased efficiency of herbicide degradation [25]. The effect of the added salts In addition, the salt effect on the reaction rate (Figure 4) was studied, using NaCl and Na2SO4. The salts were used at concentrations of 20 and 200 mmol dm-3, for each of the added salts. As can be seen from Figure 4. Effect of salt on the photocatalytic degradation of DMA-P (34.5 mg dm-3) in the presence of TiO (c 2.0 g dm-3) Figure 4, the photodegradation reaction is slower in the 2 cat in deionized water (A) and ground water (B). presence of salts in deionized water. Sodium chloride proved to be the stronger inhibitor than sodium sul- Chloride ions inhibit photocatalytic degradation fate. While chloride ions have hole scavenging pro- in both of analyzed waters, deionized and ground perties, sulfate anions react with positive holes and water. The inhibitor effect of these anions can be exp- hydroxyl radicals [26,27]. There is also a competitive lained through electrostatic interactions between sur- adsorption between DMA-P and chlorides and/or sul- face of photocatalyst and anions. In acidic solution fates [24]. These influences can be described by the surface of photocatalyst is positively charged and Eqs. (18)-(24) [7,26–29]: attracts anions, which has influence on the reduced

106 O.S. GLAVAŠKI et al.: PHOTODEGRADATION OF DIMETHENAMID-P… Chem. Ind. Chem. Eng. Q. 22 (1) 101−110 (2016) adsorption of molecules of DMA-P and intermediates The results lead to the conclusion that the form- – 2– – and therefore reduced rate of degradation process ation rates of Cl , SO4 and NO3 are slower compared and mineralization. In alkaline solutions such adsorp- to the degradation rate of DMA-P, and indicate the tion would be unlikely because of repulsive electro- formation of intermediates that contain chlorine, sul- static forces [24]. Sulfate ions inhibit photocatalytic phur and nitrogen. degradation in deionized water (at mildly acidic sol- The mineralization of DMA-P was studied by the ution, for pH 6.33), which can be explained by the total organic carbon analysis. For the period of 90 same inhibitor effect as for chloride ions. In ground min, the TOC elimination was 64% in deionised water water (in mild alkaline solution, for pH 7.34) these and 50% in ground water (Figure 6). This indicates anions increase rate of the degradation process that the TOC removal rate was not proportional to the which can be explained by oxidative ability of sulfate rate of DMA–P photodegradation and also confirms anions radicals. Although the sulfate anion radical is the formation of the organic intermediates. less reactive than •OH, it may oxidize the DMA-P mol- •− • ecule. At mild alkaline pH, both SO4 and OH are responsible for the degradation of DMA-P [24]. Results of total organic carbon elimination and ion chromatography The results of ion chromatography (shown in Figure 5) and TOC analysis have been used for the determination of the mineralisation level of DMA-P. Considering that the molecule of DMA-P contains one - 2– atom of chlorine, nitrogen and sulphur, Cl , SO4 and – NO3 may be separated after their complete mineralization. The degradation results in deionised water show that after 90 min of degradation 95% of chlorine is converted into chloride, while during the same period of irradiation 23% of total nitrogen is converted – 2– into NO3 and 8% of sulphur is converted into SO4 ions. However, mineralisation of the DMA-P in ground Figure 6. Time dependence of TOC concentration during the photocatalytic degradation of DMA-P (34.5 mg dm-3) in water is almost unnoticeable. -3 deionised and ground water (ccat 2.0 g dm ).

Figure 5. Time dependence of inorganic ions (nitrates, sulphates and chlorides) concentration during the photocatalytic degradation of -3 -3 DMA-P (34.5 mg dm ) in deionised and ground water (ccat 2.0 g dm ).

107 O.S. GLAVAŠKI et al.: PHOTODEGRADATION OF DIMETHENAMID-P… Chem. Ind. Chem. Eng. Q. 22 (1) 101−110 (2016)

Results of HPLC/MS analysis At the end of the photodegradation reaction (90 In order to identify intermediates formed during min) only traces of DMA-P were detected. When the photocatalytic process, a qualitative analysis of DMA-P was electrochemically degradated the major aliquot samples taken at various periods of degrad- degradation product was formed by C-N bond cleav- ation process was carried out. From the analysis of age, while Cl elimination produced minor degradation the obtained peaks identified by the value of m/z product [1]. ratios, based on the molecular weight and the nature of the chemical bond in the molecule of DMA-P, the CONCLUSION occurrence of transformation products can be con- The elimination of DMA-P from water with the firmed, which is illustrated in Figure 7. The efficiency mediation of TiO has been studied for the first time in of LC-MS hyphenated techniques for the character- 2 this study. Under optimal conditions almost complete ization of various photodegradation [30,31] products disappearance of 34.5 mg dm-3 of herbicide (deter- has been recently reported for chloracetamide herbi- mined by HPLC) and 50% TOC removal, occurred cide acetochlor, which has similar structure as the within 90 min in deionised and ground water, while investigated dimethenamide-P.

Figure 7. The degradation mechanism proposed for DMA-P photocatalytic degradation.

HPLC/MS analysis of the reaction mixture after TOC analysis showed that DMA-P was mineralized 10 minutes of the photocatalytic degradation, rev- 64 and 50% in deionised and ground water, respect- ealed the presence of one degradation product with ively. The ion chromatography results showed that m/z of 110.5. From the structural analysis of DMA-P it the mineralization process leads to the formation of can be assumed that the cleavage of C-N bond occur- chloride, sulphate and nitrate anions during the pro- red (path I) and that the dimethylthiophenyl cation has cess. DMA-P degradation products were identified by been formed, as shown in Figure 7. HPLC/MS analysis. They were formed by: C-N bond

The existence of ion fragment with m/z 262.1 cleavage (m/z 110.5), loss of CH3 group (m/z 262.1), may be explained by ordinary loss of CH3 group in elimination of chlorine and ketene, followed by rear- side chain of a parent molecule (II). rangement and cyclization (m/z 148.5). Another abundant degradation product was det- Acknowledgement ected after 30 min of the photocatalytic degradation with m/z 148.5 (III). the formation of this structure can This work has been financially supported by be explained by the following subsequent processes: Ministry of Education, Science and Technological elimination of chlorine and then loss of ketene, fol- Development, Republic of Serbia, under Grant No. lowed by rearrangement and cyclization to obtain a 172013. bicyclic product (Figure 7).

108 O.S. GLAVAŠKI et al.: PHOTODEGRADATION OF DIMETHENAMID-P… Chem. Ind. Chem. Eng. Q. 22 (1) 101−110 (2016)

REFERENCES [15] C. Minero, D. Vione, Appl. Catal., B 67 (2006) 257-269 [16] J. Wang, W. Sun, Z. Zhang, X. Zhang, R. Li, T. Ma, P. [1] O. Glavaški, S. Petrović, D. Mijin, M. Jovanović, A. Zhang, Y. Li, J. Mol. Catal., A: Chem. 272 (2007) 84-90 Dugandžić, T. Zeremski, M. Avramov Ivić, Electroanalysis [17] M. Qamar, M. Muneer, D. Bahnemann, J. Environ. Man- 26 (2014) 1877-1880 age. 80 (2006) 99-106 [2] R.A. Yokley, L.C. Mayer, S.B. Huang, J.D. Vargo, Anal. [18] S. Chen, Y. Liu, Chemosphere 67 (2007) 1010-1017 Chem. 74 (2002) 3754–3759 [19] W. Chu, W.K. Choy, T.Y.So, J. Hazard. Mater. 141 (2007) [3] J. Fenoll, P. Hellín, C-M. Martínez, P. Flores, S. Navarro, 86-91 J. Photochem. Photobiol., A 238 (2012) 81–87 [20] M.R. Sohrabi, M. Davallo, M. Miri, Int. J. ChemTech. Res. [4] V.A. Sakkas, P. Calza, A.D. Vlachou, C. Medana, C. Min- 1 (2009) 446-451 ero, T. Albanis, Appl. Catal., B 110 (2011) 238– 250 [21] Y. Chen, K. Zhang, Y. Zuo, Sci. Total Environ. 463–464 [5] E.M. Thurman, E.A. Scribner, in The Triazine Herbicides (2013) 802–809 50 Years Revolutionizing Agriculture, H.M. LeBaron, J.M. [22] L. Wei, C. Shifu, Z. Wei, Z. Sujuan, J. Hazard. Mater. 164 McFarland, O.C. Burnside Eds., Elsevier, San Diego, CA, (2009) 154–160 2008, pp. 451-475 [23] C. Wang, L. Zhu, M. Wei, P. Chen, G. Shan, Water Res. [6] Directive 2000/60/EC, EU Official Journal, 2000. L327, 46 (2012) 845-853 Council Directive 2000/60/EC establishing a framework for Community action in the field of water policy (Water [24] D.E. Santiago, J. Araήa, O. González-Díaz, M.E. Alemán- Framework Directive), pp. 1–73. Dominguez, A.C. Acosta-Dacal, C. Fernandez-Rodrí- guez, J. Pérez-Peήa, M. José, J.M. Doήa-Rodríguez, [7] S. Ahmed, M.G. Rasul, R. Brown, M.A. Hashib, J. Appl. Catal., B 156–157 (2014) 284–292 Environ. Manage. 92 (2011) 311–330 [25] A. Lair, C. Ferronato, J.M. Chovelon, J-M. Herrmann, J. [8] N.A. Laoufi, F. Bentahar, Desalin. Water Treat. 52 (2014) Photochem. Photobiol., A 193 (2008) 193-203 1947-1955 [26] Y. Wang, K. Lu, C. Feng, J. Rare Earths 31 (2013) 360- [9] D.A. Lambropoulou, I.K. Konstantinou, T.A. Albanis, A.R. 365 Fernández-Alba, Chemosphere 83 (2011) 367-378 [27] B. Neppolian, H.C. Choi, S. Sakthivel, B. Arabindoo, V. [10] A. Verma, M. Sheoran, A.P. Toor, Indian J. Chem. Murugesan, Chemosphere 46 (2002) 1173-1181 Technol. 20 (2013) 46-51 [28] C. Hu, J.C. Yu, Z. Hao, P.K. Wong, Appl. Catal., B 46 [11] D. Mijin, M. Savić, S. Perović, A. Smiljanić, O. Glavaški, (2003) 35-37 M. Jovanović, S.D. Petrović, Desalination 249 (2009) 286-292 [29] N. Kashif, F. Ouyang, J. Environ. Sci. 21 (2009) 527–533 [12] B. Abramović, S. Kler, D. Šojić, M. Laušević, T. Radović, [30] S. Bouchonnet, S. Bourcier, Y. Souissi, C.Genty, M. Sab- D.Vione, J. Hazard. Mater. 198 (2011) 123-132 lier, P. Roche, V. Boireau, V. Ingrand, J. Mass Spectrom. 47 (2012) 439-452 [13] A. Nezamzadeh-Ejhieh, S. Moeinirad, Desalination 273 (2011) 248-257 [31] Y. Souissi, S. Bourcier, S. Ait-Aissa, E. Maillot-Maréchal, S. Bouchonnet, C. Genty, M. Sablier, J. Chromatogr., A [14] W.Y. Wang, Y. Ku, Colloids Surfaces, A 302 (2007) 261– 1310 (2013) 98–112. –268

109 O.S. GLAVAŠKI et al.: PHOTODEGRADATION OF DIMETHENAMID-P… Chem. Ind. Chem. Eng. Q. 22 (1) 101−110 (2016)

O.S. GLAVAŠKI1 FOTODEGRADACIJA DIMETANAMIDA-P U S.D. PETROVIĆ2 DEJONIZOVANOJ I PODZEMNOJ VODI V.N. RAJAKOVIĆ-OGNJANOVIĆ3 1 T.M. ZEREMSKI Proučavanje reakcije fotodegradacije herbicida dimetanamida-P, izvršeno je u prisustvu 2 A.M. DUGANDŽIĆ TiO2 kao katalizatora i pod dejstvom UV zračenja, u dejonizovanoj i podzemnoj vodi. Ispi- 2 • D.Ž. MIJIN tan je uticaj koncentracije elektron-akceptora (H2O2), ″hvatača″ OH radikala (C2H5OH) i šupljina (NaCl and Na SO ), kao i uticaj pH sredine na brzinu reakcije fotodegradacije. 1Institut za ratarstvo i povrtarstvo, 2 4 Maksima Gorkog 30, 21000 Novi Promena koncentracije dimetanamida-P praćena je pomoću HPLC. Nastajanje degrada- Sad, Srbija cionih proizvoda analizirano je pomoću HPLC/MS. Jonska hromatografija kao i metoda 2Tehnološko-metalurški fakultet određivanja ukupnog organskog ugljenika primenjene su u cilju određivanja nivoa minera- Univerziteta u Beogradu, lizacije herbicida. HPLC analiza je pokazala da se u toku 90 min herbicid skoro potpuno Karnegijeva 4, 11120 Beograd, uklanja u dejonizovanoj i podzemnoj vodi. Metodom određivanja ukupnog organskog Srbija ugljenika utvrđeno je da se dimetanamid-P mineralizije 64% u dejonizovanoj, a 50% u 3 Građevinski fakultet Univerziteta u podzemnoj vodi. Jonska hromatografija je pokazala da pri degradaciji ispitivanog molekula Beogradu, Bulevar Kralja nastaju hloridni, sulfatni i nitratni anjoni. HPLC/MS analiza ukazala je da pri degradaciji Aleksandra 73, 11 000 Beograd, dolazi do raskidanja C-N veze (m/z 110,5), gubitka CH3 grupe (m/z 262,1), kao i do elimi- Srbija nacije hlora i ketena, praćene premeštanjem i ciklizacijom (m/z 148,5).

NAUČNI RAD Ključne reči: uticaj soli; jonska hromatografija; tečna hromatografija-elektron- sprej masena spektrometrija; fotokataliza; titan-dioksid.

110 Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 22 (1) 111−115 (2016) CI&CEQ

1 SONJA V. SMILJANIĆ EFFECT OF La2O3 ON THE STRUCTURE AND SNEŽANA R. GRUJIĆ1 2 THE PROPERTIES OF STRONTIUM BORATE MIHAJLO B. TOŠIĆ GLASSES VLADIMIR D. ŽIVANOVIĆ2 2 SRĐAN D. MATIJAŠEVIĆ Article Highlights 2 JELENA D. NIKOLIĆ • Selected lanthanum-strontium-borate glasses were prepared by conventional melt- VLADIMIR S. TOPALOVIĆ2 -quenching technique • The density and the molar volume were increasing with increasing La2O3 content 1 University of Belgrade, Faculty of • Oxygen molar volume values were increasing opposite to oxygen packing density Technology and Metallurgy, values Belgrade, Serbia • The HSM results were employed for obtaining the viscosity curves using VFT equa- 2Institute for the Technology of tion and GS Nuclear and Other Raw Mineral Materials, Belgrade, Serbia Abstract The selected lanthanum-strontium borate glasses were prepared by a conven- SCIENTIFIC PAPER tional melt-quenching technique. The compositions of the investigated glasses

were chosen to be: 5.7, 9.5, 14.3, 19.1 mol% for La2O3, 22.9, 19.1, 14.3, 9.5 UDC 666.1.031:66 mol% for SrO and 71.4 mol% for B2O3. The density, molar volume, oxygen molar DOI 10.2298/CICEQ150213031S volume, oxygen packing density, oxygen/boron ratios and structural transform- ations in the glass network were investigated according to the substitution of SrO

by La2O3. The density and the molar volume increased in parallel with La2O3 content increase. Simultaneously, oxygen molar volume values increased while the oxygen packing density values decreased. A hot stage microscope (HSM) and a differential thermal analysis (DTA) were used to determine the characteristic temperatures. By increasing the content of lanthanum, the glass transition temperatures, changed with the same trend as the molar volume. Glass stability parameters were calculated from the temperatures obtained by DTA and HSM. The HSM results were used to obtain the viscosity curves by applying the Vogel- –Fulcher-Tamman (VFT) equation. Keywords: glass, DTA, HSM, glass viscosity.

A growing interest recently has been focused on a formation of non-bridging oxygen (NBO) caused by alkaline earth-borate glasses due to their applications additional network modifier [2]. Therefore in borate as laser hosts, nonlinear optical and other photonic glasses, the main structural units are both [BO3] tri- devices [1]. The structure of vitreous B2O3 consists of angles and [BO4] tetrahedra forming boroxol rings and a random network of [BO3] triangles connected by chains with different number of NBO [2]. bridging oxygen at all three corners to form The present study aims to characterize the completely linked network. The addition of a network physical and structural properties of the selected lan- modifier in B2O3 glass could induce the conversion of thanum-strontium borate glasses. Glass compositions

[BO3] triangles to [BO4] tetrahedra. This conversion of for the synthesis were selected within the glass form- boron from 3- to 4-fold coordination occurs only until ing range of increasing content of La2O3, decreasing the network reaches some critical concentration of content of SrO and constant content of B2O3. Also, tetrahedral coordinated boron, and is then followed by the goal of this work was to determine glass stability and viscosity behavior of the selected glasses. The Correspondence: S.V. Smiljanić, University of Belgrade, Faculty physical and structural properties of the glasses were of Technology and Metallurgy, Karnegijeva 4, Belgrade, Serbia. investigated by measuring the densities of the glass E-mail: [email protected] samples, and calculating the molar volume, oxygen Paper received: 13 February, 2015 Paper revised: 27 June, 2015 molar volume, oxygen packing density values and Paper accepted: 8 July, 2015

111 S.V. SMILJANIĆ et al.: EFFECT OF La2O3 ON THE STRUCTURE… Chem. Ind. Chem. Eng. Q. 22 (1) 111−115 (2016) ratios of oxygen to boron atoms [2,3]. Differential grinding the bulk glass in an agate mortar and sieving thermal analysis (DTA) was applied to determine the to grain size < 0.048 mm. The DTA curves were glass transition temperature, Tg, the crystallization recorded by a Netzch STA 409 EP instrument at the -1 onset, Tx, and the crystallization peak temperature, heating rate 10 °C min , using Al2O3 as a reference

Tp. Hot stage microscope (HSM) was acquired for material. estimation the temperatures of: the first shrinkage HSM analysis was performed on the previously

(TFS), the maximum shrinkage (TMS), the deformation prepared glass powder pressed into cylinders, which

(TD), the sphere TS, the half-ball temperature (THB) were placed on a platinum plate, on an alumina sup- and the flow temperature (TF). The glass stability (GS) port, contacted with a (Pt/Rh/Pt) thermocouple. The parameters were calculated on the basis of these heating rate was 10 °C min–1. With temperature inc- characteristic temperatures. Viscosity curves of the rease, the geometric shape of the samples changed. glasses were set based on the results from the HSM Micrographs obtained were used to determine the using the Vogel-Fulcher-Tamman (VFT) relation [4]. temperatures corresponding to the typical glass viscosity points [5-7]. Combination of the DTA and HSM EXPERIMENTAL methods enabled determination of the GS parameters. The HSM results were applied to obtain viscosity The glasses with nominal composition yLa2O3- curves using the VFT relation [4,5]. –xSrO-(100-x-y)B2O3, where y = 5.7, 9.5, 14.3 and 19.1 and x = 22.9, 19.1, 14.3 and 9.5 (Table 1), were RESULTS AND DISCUSSION melted in a covered platinum crucible in an electric furnace and melted at 1200 °C for 30 min. The The densities (ρ) of the glass samples deter- reagent grade of H3BO3, SrCO3 and La2(CO3)3 were mined in the present study are shown in Table 1. The used as raw materials, mixed and homogenized in an molar volume (Vm) of the glass samples was calcul- agate mortar. Covered crucible and relatively short ated using the relative molecular mass (M) and den- melting time at relatively low temperature were sity (ρ) by the following relation [8]: applied in order to minimize boron evaporation. The M melt was cast and cooled on a stainless steel plate in V = (1) m ρ air at room temperature. The measurements of the weight loss due to the melting indicated that the glas- These values are included in Table 1 together ses were within 1–2 wt.% of the desired compositions. with the values of oxygen molar volume (Vo) and The obtained glasses were transparent without visible oxygen packing density (OPD), calculated using the bubbles. following relations:

Table 1. The glass compositions, density, molar volume, oxygen = 1 V0mV (2) molar volume and oxygen packing density n

Glass sample ρ Parameter OPD= 1000 n (3) 1 2 3 4 M La O , mol% 5.7 9.5 14.3 19.1 2 3 where n is the number of oxygen atoms per formula SrO, mol% 22.9 19.1 14.3 9.5 unit. B2O3, mol% 71.4 71.4 71.4 71.4 The following equation, based on the glass stoi- -3 ρ, g cm 3.21 3.25 3.56 3.66 chiometry, was used for the calculated number of 3 -1 Vm, cm mol 28.65 30.90 31.21 33.26 oxygen: 3 -1 Vo, cm mol 11.28 11.81 11.50 11.84 Number of oxygen=+xy 3 + 71.4 (4) OPD, mol dm-3 88.68 84.70 86.95 84.45 O/B 1.78 1.83 1.90 1.97 where x is mol% of the SrO, y is mol% of the La2O3

and 71.4 is constant content of the B2O3 in the The densities of the glasses were determined by glasses. using the pycnometer method, with uncertainty ±0.01. The density and the molar volume of the glasses

A hot-stage microscope (E. Leitz Wetzlar) equip- increased in parallel with La2O3 content increase in ped with a Cannon camera, and differential thermal the glasses. The increase of the density could be analysis (DTA) were used to determine the character- explained considering the higher relative molecular istic temperatures during the heating of the glass mass of lanthanum oxide as compared to the relative powder. The samples were prepared by crushing and molecular mass of strontium oxide. With the increase

112 S.V. SMILJANIĆ et al.: EFFECT OF La2O3 ON THE STRUCTURE… Chem. Ind. Chem. Eng. Q. 22 (1) 111−115 (2016)

of La2O3 content in the glasses, the oxygen content maximum shrinkage (TMS) is the temperature where rises as well, increasing the molar volume of the the sample shrinks to the maximum possible level, glass. The oxygen molar volume increases opposite but still has sharp edges, before softening, at to the oxygen packing density with the increasing viscosity log η = 7.8±0.1. The TD is the point of log η =

La2O3 content in glasses, indicating a less tight pack- = 6.3±0.1 when the first signs of softening could be ing of the glass network and more open glass network observed and the edges of the samples are rounded.

[9]. The O/B ratios increased together with La2O3 The TS is the temperature at which the sample content increase. The ratios of the O/B indicated the becomes spherical, at log η = 5.4±0.1, whereas the presence of the metaborate structures, so the both half-ball temperature (THB), at log η = 4.1±0.1, is the

[BO3] triangles and [BO4] tetrahedra units are present temperature at which the observed section of the in the glass systems. sample forms a semicircle. The flow temperature (TF) The characteristic temperatures obtained by HSM is the temperature at which the height of the drop of and DTA measurements are summarized in Table 2. molten glass corresponds to a unit on the microscopic

Тhe Tg exhibits the same trend of the changes as of scale, at log η = 3.4±0.1 [5].

Vo. The increase in the Tg could be attributed to the greater bond strength of the La-O (244 kJ mol-1) bond in comparison with the Sr-O bond (134 kJ mol-1). The addition of the La2O3 increased the Tg, which can be explained by higher field strength of La3+ (0.52 Å-2) with respect to Sr 2+ (0.32 Å-2) [10]. The decline in the Tg for the sample 3 could be explained by stoichiometry composition of this glass [11].

Table 2. Characteristic temperatures (°C) obtained by HSM and DTA

Glass sample Temperature 1 2 3 4

TFS 600 600 600 680

TMS 680 719 739 740

TD 700 720 760 760 Figure 1. The photomicrographs obtained by HSM with the

TS 740 760 800 800 graphs of the shrinkage.

THB 840 900 1000 1050

TF 890 950 1020 1060 These temperatures obtained by DTA and HSM

Tg 622 640 638 644 were used to determine the GS parameters, the Hruby parameter, K , the Weinberg, K , and the Tx 735 763 723 765 H W parameter K proposed by Lu and Liu [12]. Within Tp 809 792 749 792 LL this work, the used TF was determined by HSM. The photomicrographs for the glass sample 3, The glass stability parameters are defined by obtained by HSM, with the graphs of the shrinkage the equations: are shown in Figure 1 [11]. The shrinkage of the TT− K = x g (5) samples is determined by the ratios of A/A0 and H/H0, H − TTFx where A0 is the initial area of the sample whereas A is the area at the temperature T, H0 is the initial height TT− K = p g (6) and H is the height at the temperature T. W TF The temperatures corresponding to the typical viscosity points were determined from the photomic- T K = p (7) LL + rographs, by observing the geometric shape of the TTg F specimens, obtained by HSM (Figure 1). The temperature of the first shrinkage (TFS) is the temperature at The calculated parameters for the glasses are the typical viscosity, log η = 9.1±0.1, where η is in shown in Table 3. The resistance of a given glass dPa·s. At this temperature the sample shrinks to against crystallization upon reheating defines GS. about 3-5% of its initial height. The temperature of the High values of the parameters indicate high glass stability, higher stability of the glass with respect to

113 S.V. SMILJANIĆ et al.: EFFECT OF La2O3 ON THE STRUCTURE… Chem. Ind. Chem. Eng. Q. 22 (1) 111−115 (2016) devitrification. The lowest values of the parameters for the glasses, log η = f(T) as shown in Figure 2 in are related to the sample 3. This glass shows the the upper left corner. The activation energy of the smallest GS and the highest tendency toward crystal- viscous flow is obtained from the Arrhenius equitation, lization [11]. Increase of lanthanum content in the from the slope of the log η = f (1/T) curves (Figure 2, glasses was followed with decrease of the GS. Table 4).

Table 3. Glass stability parameters Table 4. VFT parameters and the activation energies of the viscous flow Glass sample KH KW KLL 1 0.73 0.21 0.53 Glass sample Parameter 2 0.66 0.16 0.50 1 2 3 4 3 0.29 0.11 0.45 A -0.315 -3.30 -6.12 1.22 4 0.41 0.14 0.46 B 1780 5103 11603 808

T0 / K 684 462 135 1142 -1 The viscosity values of the glasses and VFT Ea / kJ mol 365 340 297 543 parameters were determined based on the photomicrographs and the typical temperatures, obtained by CONCLUSION HSM: The investigation of the physical properties of B logη =+A (8) the glasses showed that the substitution of SrO by TT− 0 La2O3 increased the density, molar volume, oxygen molar volume and decreased oxygen packing density. where η is viscosity in dPa·s, A, B and T (K) are 0 The density increase was attributed to the higher rel- constants. These constants were obtained from Eq. ative molecular mass of the glass containing more (8) by resolving a couple equations, using tempera- La O . The decrease of the oxygen packing density tures and viscosity values obtained by HSM (Table 4). 2 3 indicated a less tightly packed the glass network. The These equations are used to calculate the viscosity increase in the Tg could be attributed to the greater

Figure 2. Log η versus reciprocal temperature curves for the glass samples: a) 1, b) 2, c) 3 and d) 4.

114 S.V. SMILJANIĆ et al.: EFFECT OF La2O3 ON THE STRUCTURE… Chem. Ind. Chem. Eng. Q. 22 (1) 111−115 (2016) bond strength of the La-O (244 kJ mol-1) bond in [4] S. Samal, S. Kim, H. Kim, J. Am. Ceram. Soc. 95 (2012) comparison to the Sr-O bond (134 kJ mol-1). Parallel 1595-1603 with lanthanum content increase, the GS decreased. [5] M.J. Pascual, A. Duran, M.O. Prado, Physis. Chem. Glasses 46 (2004) 512-520 Acknowledgement [6] C. Lara, M.J. Pascual, M.O. Prado, A. Duran, Solid State The authors are grateful to the Ministry of Edu- Ionics 170 (2004) 201-208 cation, Science and Technological Development of [7] C. Lara, M.J. Pascual, A. Duran, J. of Non-Crys. Solids the Republic of Serbia for the financial support (Pro- 348 (2004) 149-155 jects 172004 and 34001). [8] S. Bale, S. Rahman, A.M. Awasthi, V. Sathe, J Alloy Compd. 460 (2008) 699-703 REFERENCES [9] A. Goel, D.U. Tulyaganov, V.V. Kharton, A.A. Yarem- chenko, J.M.F. Ferreira, Acta Mater. 56 (2008) 3065- [1] M. Kaur, O. P. Pandey, S. P. Singh, J. Non-Crys. Solids –3076 358 (2012) 2589-2596 [10] N. Sasmal, M. Garai, A.R. Molla, A. trafder, S.P. Singh, [2] J.E. Shelby, Introduction to Glass Science and Techno- B. Karmakar, J. Non-Crys. Solids 387 (2014) 62-70 logy, The Royal Society of Chemistry, Cambridge, 2005 [11] S. Smiljanić, S. Grujić, M. Tošić, V. Živanović, J. Stoja- [3] S.V. Smiljanić, S.R. Grujić, M.B. Tošić, V.D. Živanović, nović, S. Matijašević, J. Nikolić, Ceram Int. 40 (2014) S.D. Matijašević, J.D. Nikolić, V. Topalović, Effect of 297-305

La2O3/SrO ratio on properties of La2O3–SrO–B2O3 glasses, [12] A. Kozmidis-Petrovic, J. Sestak, J. Therm. Anal. Calorim. in Proceedings of 12th International Conference on Fund- 110 (2012) 997–1004. amental and Applied Aspects of Physical Chemistry, Belgrade, Serbia (2014), Vol. II, pp. 667-670

1 SONJA V. SMILJANIĆ UTICAJ La2O3 NA STRUKTURU I SVOJSTVA SNEŽANA R. GRUJIĆ1 STRONCIJUM-BORATNIH STAKALA MIHAJLO B. TOŠIĆ2 2 VLADIMIR D. ŽIVANOVIĆ Izabrana lantan-stroncijum-boratna stakla su dobijena uobičajenom tehnikom topljenja i 2 SRĐAN D. MATIJAŠEVIĆ naglog hlađenja rastopa stakla. Sastavi ispitivanih stakala su: 5,7; 9,5; 14,3; 19,1 mol% 2 JELENA D. NIKOLIĆ La2O3; 22,9; 19,1; 14,3; 9,5 mol% SrO i 71,4 mol% B2O3. Ispitivan je uticaj izmene SrO sa 2 VLADIMIR S. TOPALOVIĆ La2O3 na: gustinu, molarnu zapreminu, molarnu zapreminu kiseonika, gustinu pakovanja kiseonika, odnose kiseonik/bor kao i strukturalne transformacije u mreži stakla. Gustina i 1Univerzitet u Beogradu, molarna zapremina se povećavaju sa porastom sadržaja lantan-oksida. Primećen je trend Tehnološko-metalurški fakultet, Karnegijeva 4, 11000 Beograd, rasta molarne zapremine kiseonika dok gustina pakovanja kiseonika opada. Za određi- Srbija vanje karakterističnih temperatura korišćene su diferencijalna termijska analiza (DTA) i 2 Institut za tehnologiju nuklearnih i termomikroskop (TM). Sa porastom sadržaja lantan-oksida temperature transformacije drugih mineralnih sirovina, Bulevar stakla su se menjale na isti način kao i molarna zapremina. Parametri stabilnosti stakla Franša d’Eparea 86, 11000 izračunati su na osnovu temperatura određenih TM i DTA. Na osnovu rezultata dobijenih Beograd, Srbija termomikroskopom postavljene se krive viskoznosti upotrebom Vogel-Fulcher-Tamman (VFT) jednačine. NAUČNI RAD Ključne reči: staklo, DTA, TM, viskoznost stakla.

115

Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 22 (1) 117−126 (2016) CI&CEQ

MARIJA ŠLJIVIĆ-IVANOVIĆ Ni(II) IMMOBILIZATION BY BIO-APATITE ALEKSANDRA MILENKOVIĆ MATERIALS: APPRAISAL OF CHEMICAL, MIHAJLO JOVIĆ THERMAL AND COMBINED TREATMENTS• SLAVKO DIMOVIĆ

ANA MRAKOVIĆ Article Highlights IVANA SMIČIKLAS • Apatite materials derived from bovine bones were studied as Ni(II) ions sorbents • Raw bones were compared with chemically, thermally and chemically/thermally University of Belgrade, Vinča treated samples Institute of Nuclear Sciences, • Different sorption mechanisms were identified by sorption data and FT-IR spectra Belgrade, Serbia analysis • Combined chemical/thermal treatment produced material with the highest sorption SCIENTIFIC PAPER capacity • Sorbed Ni(II) was very stable at low sorbent loads, while largely mobile at high UDC 637.5’62:621:66.081 loadings

DOI 10.2298/CICEQ150323024S Abstract Animal bones are a natural and rich source of calcium hydroxyapatite (HAP), which has been found to be a good sorbent material for heavy metals and radio- nuclides. Various treatments can reduce the content of bone organic phase and improve sorption properties. In this study, sorption capacities of raw bovine bones

(B) and samples obtained by chemical treatment with NaOH (BNaOH), by heating at 400 °C (B400) and by combined chemical and thermal treatment (BNaOH+400), were compared, using Ni(II) ions as sorbates. Maximum sorption capacities inc-

reased in the order B < BNaOH < B400 < BNaOH+400. Based on different sorption data and FT-IR analyses, the mechanism of Ni(II) sorption was found to be complex, with participation of both HAP and organic phase (when present). Sequential extraction analysis was applied for testing the stability of Ni(II) ions sorbed by

BNaOH+400. Majority of Ni(II) was found in residual phase (65%) at lower level of sorbent loading, while with the increase of sorbent saturation carbonate fraction

became dominant (39%). According to the results, BNaOH+400 can be utilized in water purification systems. As an apatite based material with low organic content and high efficiency for Ni(II) sorption, it is also a good candidate for in situ soil remediation, particularly at lower contamination levels. Keywords: bovine bones, treatments, apatite, Ni(II), sorption, sequential extraction.

Nickel is naturally occurring heavy metal, which Single-dose oral lethality studies indicate that soluble is in trace amounts essential for living organisms [1]. nickel compounds are more toxic than less-soluble On the other hand, exposure to high nickel concen- nickel compounds. Oral LD50 values of 46 or 39 mg trations may cause various health effects, even death. Ni per kg as nickel sulfate in male and female rats [2] and 116 and 136 mg Ni per kg as nickel acetate in female rats and male mice, respectively [3] have been Correspondence: M. Šljivić-Ivanović, University of Belgrade, reported for soluble nickel compounds. In contrast, Vinča Institute of Nuclear Sciences, P.O.Box 522, 11000 the oral LD50 values in rats for less-soluble nickel Belgrade, Serbia. E-mail: [email protected] oxide and subsulfide were >3,930 and >3,665 mg Ni Paper received: 23 March, 2015 per kg, respectively [2]. The concentration of nickel Paper revised: 20 June, 2015 Paper accepted: 8 July, 2015 and nickel compounds in the environment increases • Part of this paper was presented at the 12th International Con- due to anthropogenic activity. For example, this metal ference on Fundamental and Applied Aspects of Physical is frequently found in industrial products such as Chemistry, September 22-26, 2014, Belgrade, Serbia.

117 M. ŠLJIVIĆ-IVANOVIĆ et al.: Ni(II) IMMOBILIZATION BY BIO-APATITE… Chem. Ind. Chem. Eng. Q. 22 (1) 117−126 (2016)

stainless steel, metal alloys, catalysts, rechargeable of the chemical reagent (H2O2 or NaOH), concentra- batteries, and various products of common use, even tion of the reagent (0.1 or 2 mol L-1), reaction tem- jewelry [4]. Moreover, long lived radioactive isotopes perature (20 or 60 °C), contact time (1 or 3 h) and 59Ni and 63Ni are frequent constituents of liquid radio- sample annealing (without or at 400 °C), were con- active waste [5]. Thus, the decontamination of waste- sidered. By simultaneous variation of process vari- water streams containing Ni(II) ions is essential and it ables between lower and higher level, materials with can be conducted using different separation pro- different properties were obtained. Impact of treat- cesses, including sorption onto selective and high ment factors was compared by statistical analysis, capacity materials. In addition to conventional sorb- and it was concluded that annealing had predominant ents, different waste products were considered for influence on surface properties, as well as sorption Ni(II) removal like red mud [6], fly ash [7], tea factory capacity towards Cd2+ [20]. waste [8], animal bones [9], etc. It was shown that Considering the results achieved so far, this utilization of waste products as sorbent materials can study aims to compare performances of bone sorb- be particularly a cost-effective way of wastewater ents obtained under conditions of chemical and ther- purification. mal treatments that were found to provide the highest Animal bones are a natural source of hydroxy- sorption capacities [15,18-20]. In addition, synergistic apatite (HAP), which has been found to be a suitable effect of both treatments, applied one after the other, matrix for heavy metal immobilization [10]. The stu- was tested and the results were compared with the dies on Ni(II) immobilization using synthetic hydroxy- efficiency of raw, powdered bones. To evaluate sorp- apatite [11,12], fluoroapatite [13] and apatite derived tion kinetics and maximum capacities, experiments from fish and animal bones [9,14], have been were conducted in wide ranges of initial Ni(II) concen- reported so far. Thus, beside the synthetic apatites trations and contact times. Various sorption data, and phosphate rocks, usage of biogenic apatite forms spectroscopic and sequential extraction analyses represents one of the alternatives. were considered in order to get insight into the Ni(II) Raw animal bones contain 30-40% of organic sorption mechanisms. constituents, mostly fats and proteins (collagen). Since the nanoparticles of HAP are well “packed” in EXPERIMENTAL the organic matrix, specific surface area of bones is extremely low [15]. Consequently, crushed, raw bones Preparation and characterization of the sorbents were found to be poorer sorbents, compared to syn- At first, bovine femur bones, collected from the thetic apatite forms [16,17]. In order to reduce organic butchers shop, were cleaned from meat and cut using content, extraction of HAP has been carried out by a circular saw. Pieces of approximate size 2-3 cm different chemical or physical treatments. Heating in were boiled three times for about 3 h in distilled water, air atmosphere is one of the methods for decompo- for the removal of fats. After drying at 80 °C, one part sition of organic compounds. The influence of heating of the material was left for the preparation of referent temperature on bone physicochemical and sorption (untreated) sorbent, whereas the remaining quantity properties has been investigated, and the optimal was exposed to different treatments. Thermally treated ° temperature was found to be 400 C [15,18]. At lower sample (B400) was obtained by heating the obtained temperatures organic phase was removed incom- residues at 400 °C, in the electrical oven for 4 h. pletely, while higher temperatures resulted in sin- Chemically treated sample (BNaOH) was prepared by tering of HAP nanoparticles and deterioration of sorp- mixing 50 g of boiled bones with 1 L of 2 mol L-1 tion capacities. NaOH, for 3 h, at 60 °C. The obtained suspension Beside thermal decomposition processes, org- was filtered on the Buhner funnel. Solid residue was anic solvents such as ethanol and hexane have been thoroughly rinsed with 2 L of distilled water and then applied for fat tissue removal, while collagen degrad- dried at 80 °C. Finally, the sample denoted as ation was studied using NaOH or H2O2 solutions [19]. BNaOH+400 was produced by applying previously des-

The comparison of chemical agents efficiency has cribed thermal treatment on the sample BNaOH. The revealed that highest capacity sorbent was obtained referent sorbent (B) and treated sorbents were pow- using hot (60 °C) 0.1 mol L-1 NaOH solution [19]. dered in an electric mill, and after sieving, the fraction Recently, the effects of various treatment con- with particle size 45-200 μm was used for further ditions on bioapatite properties were compared [20]. experiments. Using experimental design methodology, the influ- Ca/P mole ratio of apatite sorbents was deter- ence of five process variables was investigated. Type mined as a measure of HAP stoichiometry. Sorbents

118 M. ŠLJIVIĆ-IVANOVIĆ et al.: Ni(II) IMMOBILIZATION BY BIO-APATITE… Chem. Ind. Chem. Eng. Q. 22 (1) 117−126 (2016) were dissolved by the process of microwave-assisted were performed using a Perkin Elmer 3100 atomic digestion in the mixture of HNO3 and H2O2, described absorption spectrometer. The amounts of Ni(II) rem- previously in detail [20]. Contents of Ca and P were oved from the solution were calculated as the differ- measured by ICP-OES (Thermo Scientific iCAP 6500 ences between the initial and the equilibrium concen- Duo ICP). trations. Specific surface areas (SSA) of samples B, B 400 FT-IR analysis of unloaded and Ni-loaded and B have been previously reported [15,20]. NaOH+400 bio-apatites Additionally, the SSA of BNaOH was determined via In order to determine interactions of bio-apatite sorption–desorption isotherm of N2, at –196 °C, by the McBain gravimetric method, where the sample was surfaces with Ni(II) ions, Fourier Transform Infrared firstly degassed at 100 °C and vacuumed for 24 h. (FT-IR) Spectroscopy was performed. Unloaded sorbents, and the solid residues obtained after equilib- Determination of BNaOH mineral composition was performed using X-ray diffraction (XRD). The Philips ration of sorbents with the most concentrated Ni-solution (6×10–3 mol L-1) were scanned. FTIR spectra of PW 1050 diffractometer with CuKα1,2 radiation was used, employing step/time scan mode of 0.05 °/s, and the samples were recorded at ambient conditions in -1 exposure time of 6 s, in the 2θ range 20–60°. the mid-IR region (400-4000 cm ) with a Nicolet IS 50 Obtained diffraction peaks were compared to Powder FT-IR spectrometer operating in the ATR mode and -1 Diffraction File database (PDF2). XRD patterns of using resolution of 4 cm with 32 scans. Major functional groups were identified in the samples B, B400 and BNaOH+400 have already been published [15,20]. FT-IR spectra, and the surfaces of starting materials and fully loaded samples were compared. Sorption experiments Sequential extraction of the sorbed Ni(II) For the evaluation of Ni(II) sorption, separate batches were prepared in 50 mL polypropylene centri- Stability of Ni(II) ions, sorbed onto the material fuge tubes. Each one contained 0.1000 g of sorbent with the highest sorption capacity (BNaOH+400), was analyzed by a sequential extraction protocol. For this and 20 mL of solution prepared from NiCl2 salt and distilled water. Initial pH values of metal solutions purpose, the sorbent was firstly equilibrated with -4 -1 -3 -1 were fixed at 6.0±0.1 in all experiments. Adjustments either 1.5×10 mol L or 6×10 mol L Ni(II) solution, of initial pH were performed by adding small aliquots in order to obtain samples with different degrees of of HCl or NaOH solutions. The suspensions were saturation. Batches containing 1.000 g of BNaOH+400 mixed on the rotary overhead shaker at 10 rpm. and 20 mL of each Ni(II) containing solution, were The kinetics of Ni(II) sorption was examined equilibrated for 24 h. After centrifugation, Ni(II) loaded using 6×10-3 mol L-1 Ni(II) solution, and varying con- bio-apatite samples were rinsed with 20 mL of distil- tact time between 15 min and 24 h. led water, centrifuged again, and dried at room tem- The effect of initial Ni(II) concentrations was inves- perature. The sequential extraction analysis was per- tigated by varying concentrations in the range 10-4– formed according to a modified Tessier procedure –6×10-3 mol L-1, while contact time was fixed at 24 h. [21]. Sorbed Ni(II) ions were portioned into 5 oper- After a given reaction time, liquid phases were ationally defined phases: exchangeable (F1), acid sol- separated from spent sorbents by centrifugation at uble (F2), reducible (F3), oxidizable (F4) and residual 7000 rpm for 10 min. Equilibrium pH values were (F5). The modification of the original Tessier protocol measured in clear supernatants. Determinations of refers to residual phase extraction, which was per- residual Ni(II) concentrations, as well as the concen- formed by digestion in 6 M HCl [22,23]. The sum- trations of Ca(II) ions released from bio-apatite phase, marized procedure is presented in Table 1.

Table 1. The modified sequential extraction procedure applied in this study (mass of dry sorbent sample 1.00 g)

Phase Fraction Experimental procedure

F1 Exchangeable Loaded samples were treated with 8 mL of 1 M MgCl2 (pH 7.0), 20 °C, 1 h.

F2 Acid soluble 8 mL of 1 M CH3COONa (pH 5, adjusted with CH3COOH), 20 °C, 5 h

F3 Reducible 20 mL of 0.04 M NH2OH·HCl in 25 vol.% CH3COOH, 96±3 °C, 6 h

F4 Oxidizable 3 mL of 0.02 M HNO3 and 5 mL of 30% H2O2 (pH 2, adjusted with HNO3), 85±2 °C, 2 h

3 mL aliquot of 30% H2O2 (pH 2, adjusted with HNO3), 85±2 °C, 3 h

5 mL of 3.2 M CH3COONH4 in 20 vol.% HNO3 was added and the sample was diluted to 20 mL, 20 °C, 30 min

F5 Residual 50 ml of 6 M HCl, 85±2 °C, 9 h

119 M. ŠLJIVIĆ-IVANOVIĆ et al.: Ni(II) IMMOBILIZATION BY BIO-APATITE… Chem. Ind. Chem. Eng. Q. 22 (1) 117−126 (2016)

RESULTS AND DISCUSSION decreases; therefore, the rate of process is stabilized.

Equilibrium was attained after 3h using BNaOH+400, Sorbents characteristics whereas approximately 24 h was required for other

XRD analysis of the sample BNaOH is presented samples. Sorbed amounts of Ni(II) ions at equilibrium -1 in Figure 1. Peaks characteristic for HAP crystalline increased in the order B (0.22 mmol g ) < BNaOH (0.28 -1 -1 phase (PDF2, card No. 09-0432) with the intensive mmol g ) < B400 (0.30 mmol g ) < BNaOH+400 (0.35 -1 background were identified. mmol g ).

Figure 1. XRD pattern of the sample BNaOH.

The reported XRD patterns of samples B, B400

[15] and BNaOH+400 [20] also showed the presence of HAP as a major crystalline phase. HAP peaks were generally of low intensities and fused, indicative of small grain size, low crystallization degree and high defectiveness of bone apatite crystals. Although the apatite samples might contain some other calcium- phosphate phases [24], their presence was not confirmed by XRD analysis. Furthermore, different intensities of the background in the XRD spectra can be ascribed to the different amounts of organic matter.

Ca/P mole ratio of the sample BNaOH+400 was found to be 1.23 [20], whereas molar ratios of 1.20, 1.17, and 1.16 were calculated for B, BNaOH and B400, respectively. Compared to the stoichiometric HAP Figure 2. a) Time-dependent sorption of Ni(II) ions on different Ca/P ratio = 1.67, considered bio-apatite samples bone sorbents. b) Data fitting using pseudo-second order kinetic -3 were Ca-deficient. Determined SSA of BNaOH was 35 model. Solid/liquid ratio 1/200, initial Ni(II) concentration 6×10 2 -1 -1 m g , which is much higher than SSA of raw bones mol L , initial pH 6.0. Symbols: (■) B, (●) BNaOH, (▲) B400 and 2 -1 (0.1 m g [15]), but lower than SSA of thermally (▼) BNaOH+400. 2 -1 2 -1 treated bones (85 m g for B400 [15] and 78 m g for

BNaOH+400 [20]). The experimental results were analyzed using pseudo-second order kinetic equation [25] which is Sorption kinetics widely used for sorption data modeling. The linear Sorption kinetics curves (Figure 2a) were of form of pseudo-second order model is given by the typical shape: a sharp increase at the beginning fol- following equation: lowed by slower metal uptake. With time, the active tt1 sites on the sorbent surface become increasingly =+ (1) q k q 2 q occupied and Ni(II) concentration in the liquid phase t 2e e

120 M. ŠLJIVIĆ-IVANOVIĆ et al.: Ni(II) IMMOBILIZATION BY BIO-APATITE… Chem. Ind. Chem. Eng. Q. 22 (1) 117−126 (2016)

-1 where qt and qe (mmol g ) are sorbed amounts at erally increased in the order B < BNaOH < B400 < -1 time t and at equilibrium, respectively, k2 (g mmol < BNaOH+400. Ni(II) removal from aqueous media using -1 min ) is the pseudo-second-order rate constant. The samples BNaOH and BNaOH+400 was especially enhanced denominator of the first member at right side is in the low concentration range (10-4–5×10-4 mol L-1). denoted as the initial sorption rate h (mmol g-1 min-1). Equilibrium pH values (Figure 3c) decreased Results of data fitting are presented in Figure 2b along with the increase of initial metal concentration, and Table 2. Equilibrium sorbed amounts calculated which may be related to the phenomenon known as by the model were 0.21, 0.29, 0.32 and 0.36 mmol g-1 specific cation sorption [29]. Specifically sorbed cat- for B, BNaOH, B400 and BNaOH+400, respectively, closely ions are attached strongly to the surface functional matching the values obtained experimentally. In addi- groups causing the release of H+ ions: tion, high R2 values indicated good agreement between S−− OH + Ni2+ = S ONi + + H + experimental results and mathematical model. −−+2+ + + SOH2 + Ni = SONi + 2H Table 2. Pseudo-second order model parameters for Ni(II) sorp- where S−OH and S−OH + respectively denote neutral tion by differently treated bones 2 or protonated surface functional groups of both HAP Sorbent (phosphate) and organic phase, if present. Parameter B BNaOH B400 BNaOH+400 Additionally, hydrolysis of Ni ions takes place -1 qe / mmol g 0.21 0.29 0.32 0.36 according to the reactions: -1 -1 k2 / g mmol min 0.043 0.40 0.071 0.18 + Ni2+ + H O = Ni() OH + H + , pK = 9.86 [33] h / mmol g-1 min-1 0.0020 0.0035 0.0072 0.0230 2 + R2 0.991 0.998 0.999 0.999 Ni() OH + H O = Ni () OH + H+ , pK = 9.14 [33] 2 2

Calculated values of k2 and h where in the range Calculation of Ni-species distribution in respect 2+ 0.043-0.40 g mmol-1 min-1 and 0.0020–0.023 mmol g-1 to solution pH showed that Ni are dominant up to pH min-1, respectively. Treatments have improved sorp- 8 [6]. Hydrolysis starts at pH > 8, reaching the max- tion of Ni(II) and led to an increase in the speed of imum amount of insoluble Ni(OH)2 at pH 10. Taking sorption. Rate constant k2 increased in the order this into account, large sorption of Ni(II) by BNaOH+400

B < B400 < BNaOH+400 < BNaOH, while h increased as observed in the low concentration range, can be follows: B < BNaOH < B400 < BNaOH+400. linked to the equilibrium pH values close to the Good correlation between pseudo-second-order Ni(OH)2 precipitation threshold. model and experimental data has already been rep- The sorption of Ni(II) from the solutions of differ- orted for the sorption of divalent cations onto various ent initial concentrations, was followed by almost linear sorbents: Pb2+, Cd2+, Zn2+ and Sr2+ by synthetic HAP increase of aqueous Ca(II) concentrations (Figure [26], Cr3+ and Sr2+ by bone char [17,27], Co2+ by ani- 3d). Direct linear proportionality between the amounts mal bones [15], Pb2+ and Cu2+ onto magnetic egg- of sorbed and released ions points toward ion- 2+ 2+ 2+ shell-Fe3O4 powder [28], Pb , Zn and Cd onto -exchange as one of the operating sorption mechan- Fe(III)-modified zeolite [29], etc. Theoretically, agree- isms. The ion exchange mechanism was already ment between the sorption kinetic data and mathe- recognized as the sorption mechanism characteristic matical models such as pseudo-second order implies for HAP phase in the case of divalent Cd and Zn that chemical reaction is the rate-controlling step [30]. sorption onto synthesized HAP [34]. Also, Cheung et 2+ 2+ However, sorption processes governed by different al. [35] investigated sorption of Cu and Zn onto mechanisms (surface-complexation, dissolution/preci- bone char, the heterogenous sorbent produced from pitation, ion-exchange, etc.) were equally well des- the destructive distillation of dried, crushed cattle cribed by this model [17,31,32]. This basically means bones. Knowing that the main sorbent components that the applicability of this model is not sufficient evi- are calcium hydroxyapatite, CaCO3 and carbon, the dence for mechanistic interpretations, but it is suitable authors concluded that the main sorption mechan- for mathematical description of the process, predict- isms are ion-exchange in HAP lattice and chemi- ion of qe values, and data comparison. sorption onto carbon surface. Moreover, Al-Asheh et al. [9] reported that the main sorption mechanism of Sorption equilibrium Ni(II) ions onto raw animal bones was ion-exchange. The sorption isotherms of Ni(II) ions onto raw Our study strongly supports the ion-exchange sce- and differently treated bovine bones are presented in nario. Mole ratios Ca(II):Ni(II) were less than 1:1, for Figure 3a and b. The sorbed amounts of Ni(II) gen- all investigated sorbents, which can be explained by

121 M. ŠLJIVIĆ-IVANOVIĆ et al.: Ni(II) IMMOBILIZATION BY BIO-APATITE… Chem. Ind. Chem. Eng. Q. 22 (1) 117−126 (2016)

Figure 3. Ni(II) sorption isotherms: a) values predicted by Freundlich equation; b) values predicted by Langmuir model (symbols - experimental points, lines - fitting by theoretical models, error bars - deviations between experimental and predicted values); c) relationship between equilibrium pH values and initial Ni (II) concentrations; d) relationship between amounts of released Ca(II) and

sorbed Ni(II) ions. Symbols: (■) B, (●) BNaOH, (▲) B400, and (▼) BNaOH+400.

Ca-deficient bio-apatite crystal lattice [20] and the Calculated parameters are summarized in Table participation of other sorption mechanisms. 3. Based on the correlation coefficients (R), a good Sorption isotherms were described using Lang- agreement exists between the models and experi- muir and Freundlich theoretical models, in the fol- mental data. Calculated maximum sorption capacities lowing linear forms: increased in the order B < B400 < BNaOH < BNaOH+400, which is somewhat different in respect to experiment- cc 1 ee=+ (2) ally obtained order, probably as a consequence of q emmLqqK linearization and fitting errors. The determined KL and 1 Kf values increased in the same order as experi- lnq =+ lnKc ln (3) efn e mentally determined maximum sorption capacities: B < BNaOH < B400 < BNaOH+400. -1 where ce (mmol L ) denotes the equilibrium con- From the Langmuir constant KL, the dimen- centrations of Ni(II) ions in the liquid phase, qm (mmol sionless separation factors RL can be calculated: -1 -1 g ) is the maximum sorption capacity, KL (L g ) is the Langmuir constant related to the energy of adsorp- = 1 RL (4) 1−(1/n) 1/n −1 1+CK tion, while Kf (mmol ·L ·g ) and n are the 0L Freundlich constants related to the capacity and R is related to the nature of sorbate/sorbent attract- intensity of the sorption process. L ion and isotherm type and it gives the information on

whether the process is: unfavorable (RL > 1), linear

122 M. ŠLJIVIĆ-IVANOVIĆ et al.: Ni(II) IMMOBILIZATION BY BIO-APATITE… Chem. Ind. Chem. Eng. Q. 22 (1) 117−126 (2016)

Table 3. Ni(II) sorption parameters calculated using Langmuir and Freundlich isotherms

Langmuir model Freundlich model Sorbent -1 -1 2 1−(1/n) 3/n −1 2 qm / mmol g KL / L mmol R Kf / mmol ·dm ·g n R B 0.274 1.06 0.974 0.118 1.78 0.993

BNaOH 0.332 1.20 0.952 0.162 1.65 0.865

B400 0.321 2.66 0.993 0.200 2.02 0.974

BNaOH+400 0.357 4.89 0.990 0.264 3.14 0.991

(RL = 1), favorable (0 < RL < 1), or irreversible (RL = FT-IR analysis = 0). The calculated RL values increased in the order: BNaOH+400 (RL = 0.033) < B400 (RL = 0.059) < B (RL = Since peaks in the XRD spectra of bone = 0.14) ≈ BNaOH (RL = 0.12) and indicated that all samples appeared to be wide and fused, and did not investigated processes were favorable. provide information about the content and the com- The qm values for commercially available syn- position of organic matter, FT-IR analysis was per- thetic HAP and HAP synthetized in the laboratory formed. The contribution of the bone organic and -1 were found to be 0.184 and 0.274 mmol g [12]. mineral phases to the overall FT-IR spectrum can be Furthermore, sorption capacities of 0.039 [36] and analyzed almost separately as their peaks occur in -1 0.617 mmol g [37] for different nano HAP were also different regions of the spectra. FT-IR analysis of all reported. Consequently, it can be concluded that unloaded samples showed peaks characteristic for treated bio-apatites applied in this study can be used HAP phase, whereas the content and qualitative com- as an alternative sorbent for synthesized hydroxy- position of organic matter was related to the applied apatite. treatment (Figure 4).

Figure 4. FT-IR analysis of investigated sorbents before (black lines) and after (gray lines) Ni(II) sorption.

123 M. ŠLJIVIĆ-IVANOVIĆ et al.: Ni(II) IMMOBILIZATION BY BIO-APATITE… Chem. Ind. Chem. Eng. Q. 22 (1) 117−126 (2016)

The HAP phase is characterized by the most the majority of sorbed metal was found in F5 and F2, intense peaks in the spectra at about 1020, 960, 600 about 65 and 17%, respectively. In F1 and F3 phase, and 560 cm−1, which correspond to various modes of Ni(II) ions were present with about 7.5% while almost 3- PO4 vibrations. In addition, peaks at 1410, 1450 and insignificant amounts were found in F4 phase. Ni(II) −1 2− near 870 cm can be attributed to the CO3 group, distribution extremely changed with the increase of demonstrating carbonate substitution in HAP crystal sorbent loading. After equilibration with 6×10-3 mol L-1 lattice [15,38]. The occurrence of −OH vibrations from Ni(II) solution, the percentages in F1, F2 and F3 different sources and Amide A and B vibrations was increased up to 24, 39 and 26%, respectively, evident from a broad peak at 3700-3000 cm−1 [38]. whereas content in F5 was reduced to 7%. The largest number of bands characteristic for organic phase functional groups was observed in spectra of untreated bones (Figure 4a). Peaks at approximately 1640, 1540 and 1240 cm−1 belong to amide I, II and III bands, respectively, peak at 1740 cm−1 is characteristic for carbonyl group, and doublet at near −1 2920 and 2850 cm originates from −CH2 vibrations [38]. Bone treatments caused the reduction of organic phase content (Figure 4b-d). The intensity of −OH and amide A and B stretching vibrations at high wave numbers (3700-3000 cm−1) was markedly reduced, and almost completely lost in the sample

BNaoH+400. After chemical treatment (Figure 4b) −CH2 vibrations were still visible, so as small intensity bands of Amide I and II. In the Figure 4c, only traces of amide I and II vibrations were visible, thus, the Figure 5. Effect of initial metal concentration on the distribution thermal treatment appeared to be more efficient than of Ni(II) ions sorbed by BNaOH+400. chemical, for the exclusion of bone organic phase.

Higher organic content of BNaoH in respect to B400 can As a result of strong bonds with Ni(II) ions, be associated with its lower SSA. The spectrum of BNaOH+400 can be recommended for utilization in water BNaOH+400 was organic phase free (Figure 4d), i.e., it purification systems. Based on the results of sequen- resembles the spectrum of synthetic carbonate con- tial extraction analysis, the regeneration of the sorb- taining HAP [38]. ent loaded with Ni(II) ions could hardly be feasible. The greatest changes in the appearance of Utilization of Ca2+ solutions might cause the removal FT-IR spectrum before and after Ni(II) sorption were of a part of exchangeable Ni(II) ions (maximum 7– observed for sample B. Reduced intensities of abs- –24%, depending on the previously sorbed amount). orption peaks coming from −OH, amide and carbonyl On the other hand, removal of Ni(II) ions bound to groups implies their participation in Ni(II) complex- phases F3-F5 requires more aggressive conditions ation mechanism. Similarly, removal of Ni(II) by pig- (Table 1). Thus, the removal of at least 40-60% eon pea pod biosorbent was attributed to the pre- (depending on the sorbed amount) of Ni(II) would sence of C=O, C–O, O–H bonds which were identified result in deterioration and dissolution of HAP phase. as responsible for coordination with Ni(II) [39]. The Consequently, proper disposal of spent sorbent FT-IR spectra of treated bone samples were almost needs to be considered. Furthermore, such high sta- unaffected by the presence of Ni(II) sorption. The bility of Ni(II) in sorbed form is preferential for soil fundamental apatite structure was preserved after remediation processes [40]. Knowing that the pol- sorption of Ni(II) ions, which is in agreement with ion- lutants bonded to F1 phase are considered as mobile exchange mechanism. and potentially bioavailable [40], it can be concluded Sequential extraction analysis that stabilization of Ni(II) ions is more efficient when lower amounts were sorbed. Thus, usage of BNaOH+400 Previous experiments have shown that BNaOH+400 as an amendment for in situ remediation of Ni(II) was the most efficient sorbent, thus it was selected to contaminated soil can be recommended especially for investigate the stability of sorbed Ni(II) ions (Figure lower contamination levels. 5). After metal sorption from 0.15 mmol L-1 solution,

124 M. ŠLJIVIĆ-IVANOVIĆ et al.: Ni(II) IMMOBILIZATION BY BIO-APATITE… Chem. Ind. Chem. Eng. Q. 22 (1) 117−126 (2016)

Ni(II) sorption mechanism showed that metal was preferentially found in residual Experimental results indicate high complexity of fraction. This implies that BNaOH+400 can be considered Ni(II) sorption mechanism by variously treated bone as an additive for soil remediation, at lower levels of sorbents. Relationships between sorbed metal quan- contamination. tities, equilibrium pH values, quantities of released Acknowledgement Ca(II) ions and organic phase composition, indicated This work was supported by the Ministry of Edu- that several sorption mechanisms were operating: cation, Science and Technological Development of specific cation sorption, ion-exchange with Ca2+ from the Republic of Serbia (Project III 43009). HAP surface, Ni(II) hydrolysis (with possible precipitation) and complexation with organic functional groups. REFERENCES Also, coprecipitation of new crystal phase cannot be excluded. However, the new Ni-containing HAP [1] M. Cempel, G. Nikel, Polish J. Environ. Stud. 15 (2006) phases would be hard for detection using XRD anal- 375-382 ysis due to the fact that the XRD patterns of inves- [2] E. Mastromatteo,. Am. Ind. Hyg. Assoc. J. 47 (1986) 589- tigated samples have intensive background and broad –601 peaks. Ion-exchange and specific surface sorption [3] R.T. Haro, A. Furst, H.L. Falk, Proc. W. Pharmacol. Soc. were common mechanism for all sorbents as they 11 (1968) 39-42. relate to HAP phase. On the other hand, chemical [4] J.R. Davis, ASM Specialty Handbook: Nickel, Cobalt, and bonding by organic functional groups was mainly Their Alloys, ASM International, Materials Park, OH, detected in case of untreated bones (sample B). The 2000, p. 8 presence of functional groups such as -OH, amide, [5] O. Fišera, F. Šebesta, J. Radioanal. Nucl. Chem. 286 carbonyl, etc., can explain relative high sorption cap- (2010) 713-717 acity of sample B considering very low specific sur- [6] S. Smiljanić, I. Smičiklas, A. Perić-Grujić, B. Lončar, M. Mitrić, Chem. Eng. J. 162 (2010) 75–83 face area of powdered bovine bones (0.1 m2 g-1 [15]). [7] E. Pehlivan, S. Cetin, Energ. Source. 30 (2008) 1153 Due to various treatments, organic phase removal ––1165 caused the increase of specific surface area [15,20], [8] E. Malkoc, Y. Nuhoglu, J. Hazard. Mater. B, 127 (2005) which resulted in improved Ni(II) sorption capacities. 120–128 In addition to above, Ni(II) removal from aqueous [9] S. Al-Asheh, F. Banat, F. Mohai, Chemosphere 39 (1999) media using sample BNaOH+400 was especially enhanced 2087–2096 in the low concentration range, due to Ni(II) hydrolysis [10] F. Monteil-Rivera, M. Fedoroff, in Encyclopedia of Sur- and probably its precipitation. face and Colloid Science, P. Somasundaran, Ed., Marcel Dekker, Inc, New York, 2002, p. 1 CONCLUSION [11] T. Suzuki, T. Hatsushida, Y. Hayakawa, J. Chem. Soc. Faraday Trans. 77 (1981) 1059-1062 Efficiency of bovine bones towards Ni(II) sorp- [12] O. Rosskopfová, M. Galamboš, L. Pivarčiová, M. Čaplo- tion was compared to the performance of materials vičová, P. Rajec, J. Radioanalytical Nucl. Chem. 295 produced following chemical, thermal and combined (2013) 459-465 treatments. Sorption equilibrium was satisfactorily [13] J. Perrone, B. Fourest, E.Giffaut, J. Colloid Interface Sci. described by Langmuir and Freundlich isotherm 239 (2001) 303–313 models, while kinetics obeyed pseudo-second order [14] J. Oliva, J. De Pablo, J.-L. Cortina, J.Cama, C. Ayora, J. kinetics. All tested treatments improved sorption cap- Hazard. Mater. 194 (2011) 312–323 acity toward Ni(II) ions in respect to raw bones and [15] S. Dimović, I. Smičiklas, I. Plećaš, D. Antonović, M. Mit- different synthetic HAP samples. For initial metal con- rić, J. Hazard. Mater. 164 (2009) 279–287 centrations higher than 10-4 mol L-1, sorption cap- [16] F. Banat, S. Al-Asheh, F. Mohai, Sep. Purif. Technol. 21 (2000) 155–164 acities increased in the order B < BNaOH < B400 < [17] K. Chojnacka, Chemosphere 59 (2005) 315–320 < BNaOH+400. The material with the highest sorption capacity towards Ni(II) was obtained by synergetic [18] I. Smičiklas, S. Dimović, M. Šljivić, I. Plećaš, B. Lončar, effect of chemical and thermal treatments indicating M. Mitrić, J. Nucl. Mater. 400 (2010) 15–24 great utilization possibilities in water purification sys- [19] B. Kizilkaya, A. Adem Tekinay, Y. Dilgin, Desalination 264 (2010) 37–47 tems. The mechanism of Ni(II) sorption was extremely complex, and involved inorganic (HAP) and organic [20] M.Š. Ljivić-Ivanović, I. Smičiklas, A. Milenković, B. Dojči- nović, B. Babić, M. Mitrić, J. Mater. Sci. 50 (2015) 354- phase, when present. Sequential extraction analysis –365 of BNaOH+400 loaded with lower amounts of Ni(II)

125 M. ŠLJIVIĆ-IVANOVIĆ et al.: Ni(II) IMMOBILIZATION BY BIO-APATITE… Chem. Ind. Chem. Eng. Q. 22 (1) 117−126 (2016)

[21] A. Tessier, P.G.C. Campbell, M. Bisson, Anal. Chem. 51 [31] N. Bektas, S. Kara, Sep. Purif. Technol. 39 (2004) 189– (1979) 844–860 –200 [22] P. Polić, P. Pfendt, J. Serb. Chem. Soc. 57 (1992) 697- [32] S. Saxena, M. Prasad, S.F. D’Souza, Ind. Eng. Chem. –703 Res. 45 (2006) 9122–9128 [23] D. Petrović, M. Todorović, D. Manojlović, V.D. Krsma- [33] M.M. Benjamin, Water Chemistry, McGraw-Hill, New nović, Pol. J. Environ. Stud. 18 (2009) 873–884 York, 2002 [24] J. Zhou, X. Zhang, J. Chen, S. Zeng, K. de Groot, J. [34] K. Viipsi, S. Sjöberg, K. Tõnsuaadu, A. Shchukarev, J. Mater. Sci. Mater. Med. 4 (1993) 83–85 Hazard. Mater. 252–253 (2013) 91–98 [25] Y.S. Ho, G. McKay, Process Biochem. 34 (1999) 451–465 [35] C.W. Cheung, J.F. Porter, G. McKay, Sep. Purif. Technol. [26] I. Smičiklas, A. Onjia, S. Raičević, D. Janćković, M. Mitrić, 19 (2000) 55–64 J. Hazard. Mater. 152 (2008) 876–884 [36] S. Zamani, E. Salahi, I. Mobasherpour, Can. Chem. [27] I. Smičiklas, S. Dimović, M. Sljivić, I. Plećaš, J. Environ. Trans. 1 (2013) 173-190 Sci. Health, A 48 (2008) 210–217 [37] I. Mobasherpour, E. Salahi, M. Pazouki, J. Saudi Chem. [28] J. Ren, M.F. Bopape,K.Setshedi, J.O. Kitinya, M.S. Soc. 15 (2011) 105-112 Onyango, Chem. Ind. Chem. Eng. Q. 18 (2012) 221−231 [38] M.M. Figueiredo, J.A.F. Gamelas, A.G. Martins, in: Infra- [29] M. Mihajlović, S. Lazarević, I. Janković-Častvan, B. Jokić, red Spectroscopy - Life and Biomedical Sciences, T. Đ. Janaćković, R. Petrović, Chem. Ind. Chem. Eng. Q. 20 Theophanides, Ed., InTech, Rijeka, 2012, p. 315 (2014) 283−293 [39] J. Aravind, C. Lenin, C. Nancyflavia, P. Rashika, S. Sara- [30] H. Chen, A. Wang, J. Colloid Interface Sci. 307 (2007) vanan, Int. J. Environ. Sci. Technol. 12 (2015) 105–114 309–316 [40] R.W. Peters, J. Hazard. Mater. 66 (1999) 151–210.

MARIJA ŠLJIVIĆ-IVANOVIĆ IMOBILIZACIJA Ni(II) BIO-APATITNIM ALEKSANDRA MILENKOVIĆ MATERIJALIMA: PROCENA EFIKASNOSTI MIHAJLO JOVIĆ HEMIJSKOG, TERMIČKOG I KOMBINOVANOG SLAVKO DIMOVIĆ ANA MRAKOVIĆ TRETMANA IVANA SMIČIKLAS Životinjske kosti su prirodni i bogat izvor kalcijum-hidroksiapatita (HAP), koji predstavlja Univerzitet u Beogradu, Institut za pogodan materijal za sorpciju teških metala i radionuklida. Sadržaj organske faze kostiju nuklearne nauke "Vinča", Beograd, se može redukovati različitim tretmanima i na taj način se poboljšavaju sorpciona svojstva. Srbija U ovoj studiji, upoređeni su sorpcioni kapaciteti sirovih goveđih kostiju (B) i uzoraka dobijenih hemijskim tretmanom pomoću NaOH (B ), žarenjem na 400 °C (B ) i kombino- NAUČNI RAD NaOH 400 vanim hemijskim i termičkim tretmanom (BNaOH + 400), korišćenjem Ni (II) jona kao sorbata.

Maksimalni kapacitet sorpcije povećavao se u nizu B

(II) jona sorbovanih uzorkom BNaOH + 400 ispitana je primenom sekvencijalne ekstrakcije. Pri nižem opterećenju sorbenta najveća količina Ni(II) je detektovana u rezidualnoj fazi (65%), dok karbonata frakcija postaje dominantna (39%) sa porastom zasićenja sorbenta. Rezul-

tati ukazuju na mogućnost primene BNaOH + 400 u prečišćavanju vode. Kao materijal na bazi apatita, sa niskim sadržajem organske materije i visokom efikasnošću sorpcije Ni(II), takođe je dobar kandidat za in situ remedijaciju zemljišta, posebno pri nižim koncentracijama metala u zemljištu.

Ključne reči: goveđe kosti, tretmani, apatit, Ni (II), sorpcija, sekvencijalna eks- trakcija.

126