The 29th Special CU-af Seminar 2021

Utilization of Bacteria for Self-healing Concrete Pitcha JONGVIVATSAKUL and Peem NUAKLONG

The 29th Special CU-af Seminar 2021 August 25, 2021 Utilization of Bacteria for Self-healing Concrete Pitcha JONGVIVATSAKUL1* and Peem NUAKLONG1* Abstract This study aimed to investigate the effectiveness of microbially induced carbonate precipitation (MICP) process for repairing plastic shrinkage cracks. Self-healing composites were prepared by mixing bacterial spores with nutrients. Bacillus sphaericus spores were added at 0%, 0.5% and 1% by weight of cement. The nutrients used were yeast extract, urea, and calcium nitrate tetrahydrate (Ca(NO3)2⸱4H2O). The results showed that the MICP process was successful in healing the plastic shrinkage cracking when 1% bacterial spore was incorporated. The XRD patterns showed that the CaCO3 crystals favored the formation of calcite phase. However, the 28-day compressive strength decreases with the inclusion of nutrients used to growth the bacteria. 1Department of Civil Engineering Faculty of Engineering, Chulalongkorn University Bangkok, Thailand 137

The 29th Special CU-af Seminar 2021 August 25, 2021 Introduction and Objectives At early ages, fresh concrete is prone to shrink when exposed to hot, dry, and windy weather conditions. The primary cause is the loss of water by evaporation when the concrete is still in the plastic state. When restrained, the shrinkage induces tensile stresses and generates cracks that are typically 0.1 to 3 mm width[1]. The cracking caused by this phenomenon is known as “plastic shrinkage cracking”, which normally occurs within the first 12 hours after placing[2]. The plastic shrinkage is probably the most common cause of cracking in concrete structures. It is one of the major concerns for designers and engineers since the presence of cracks in concrete increases its vulnerability to attack by external aggressive agents. As a result, this may lead to serviceability problems in the structure due to the reduction in concrete’s ability to prevent reinforcing steels from corrosion. Microorganisms, especially bacteria, are expected to play an important role in making concrete repairing materials. Recently, extensive research has been focused on using of microorganisms due to their potential to produce biominerals through metabolic activities in the presence of liquid medium (nutrients and precipitation precursors). Among them, Bacillus sphaericus, Bacillus pasteurii, and Bacillus megaterium are the most commonly used for bacteria for precipitation of carbonate crystals. Chemical reactions to form the crystal are discussed in detail in previous studies[3,4]. Application of microbially induced carbonate precipitation (MICP) process to repair cracking of concrete can be divided roughly into the following two groups. The first group focused on a way to apply bacteria on concrete surfaces. Repair of existing cracks in concrete can be accomplished by dropping healing agent (bacteria and liquid medium) on the crack surface[5] and by ponding or immersion. The second group focused on a method for application of bacteria inside concrete, commonly referred to as the “self-healing concrete”. During manufacture, the healing agent was incorporated directly into the concrete mix. The crack healing process may exist in the system after crack formation. The self-healing concrete acquires the healing property by means of MICP, which can close the load-induced cracks up to about 1 mm[6,7]. Approximately 15-60% of strength was recovered after the healing process[8-10]. Also, the MICP products are effective in reducing permeability of the cracked concrete. However, the performance of self-healing concrete subjected to plastic shrinkage is doubtful. The plastic shrinkage crack is initiated near the top surface of concrete. This is unlike the crack under stress which forms at random positions according to the local weak zones in concrete. The efficiency of MICP process has been challenged by sufficient healing agent nearby the crack zone, and a question which must be faced certainly is the crack healing capacity. Thus, the aim of this research is to study the remediation of cracks on self-healing concrete subjected to plastic shrinkage. Bacillus sphaericus spores were mixed with concrete during specimen casting. The plastic shrinkage was created using the method recommended by ASTM C157911). The performance of MICP for healing the shrinkage crack was observed through image processing software. Microstructure analyses, SEM/EDS and XRD, were used to observe the morphology and polymorph of precipitated CaCO3 crystals. In addition, the influence of the bacterial agent on strength and permeability of concrete was also investigated. 138

The 29th Special CU-af Seminar 2021 August 25, 2021 Methods Materials In this research, the mortar was used instead of concrete to study the healing of shrinkage crack, since the mortar is more prone to shrink because of the absence of coarse aggregates restrained shrinkage movements. The conventional mortar mixture involves Portland cement type I with specific gravity of 3.15 as a binder, well graded river sand with fineness modulus of 2.42 as fine aggregate and tap water. A spore-forming bacterium obtained from Bacillus sphaericus LMG 22257 (Belgian Coordinated Collection of Microorganisms, Ghent) was used to produce self-healing property. The YU medium used for promoting cell growth was comprised of 20 g/l yeast extract and 20 g/l urea, while the solid media contained 3 g/l nutrient broth, 0.85 g/l NaHCO3, 10 g/l urea, and 15 g/l agar. Encapsulation of bacteria In this study, the bacterial spores were encapsulated by using extrusion and freeze-drying techniques following procedures outlined in Pungrasmi et. al.[12]. The details of each technique are provided as follows: 1. Extrusion technique The spores were suspended to a homogeneous suspension in 200 ml of 2% (w/v) sodium alginate solution and conveyed in a silicone tube, which compresses the force of the peristaltic pump at 100 rpm, to a syringe needle (6 mm diameters) for extrusion as free-fall droplets into 2% (w/v) CaCl2 solution and left to harden at room temperature for 30 min. The solid capsules formed were washed with sterile distilled water, dried on clean paper until completely dry and then stored in a desiccator. 2. Freeze-drying technique Before sublimation, the spore suspension in 2% (w/v) sodium alginate solution was frozen at −46 °C in liquid ethanol and then the mixed solution was freeze dried for 24-48 h in a laboratory-scale freeze dryer (Christ Alpha 2–4/LD Plus, Germany) at a condenser temperature below 0 °C and a chamber pressure of 0.1 mbar. The dried products in sheet form were spun to small pieces and stored in a desiccator. Mix proportions and manufacturing process Details of mixtures are shown in Table 1. The ratio of mixing water to cement was kept constant at 0.6 by mass, and a constant ratio of sand to cement of 2.75 was used for all mixtures. It should be noted that the quantity of water used in the S-0, S-0.5 and S-1 mixtures is calculated based on the stoichiometry of the phases since the nutrient compounds consist of water. The spore-forming bacteria were used as additive material at levels of 0%, 0.5%, and 1% by weight of cement. Nutrients used to growth the bacteria spores in concrete were yeast extract, urea, and calcium nitrate tetrahydrate (Ca(NO3)2⸱4H2O), and their contents were 0.85%, 4%, and 8% of cement mass, respectively. 139

The 29th Special CU-af Seminar 2021 August 25, 2021 Table. 1: Mix proportions of mortars The mixing sequence was done according to ASTM C305[13] with a slight modification of incorporating nutrients and bacterial spores. It consisted of introducing cement to the water and nutrients, and mixing for 30 s. Then, added the river sand and bacterial spores. The slurry, sand, and spores were mixed together for 60 s, followed by 90 s of rest and remixing for an additional 60 s. After demolding, specimens were cured in water at 25±2 ºC until the age of 28 days. Testing procedure 1. Permeability and compressive strength After 28 days, the compressive strength was conducted on three cubes of 100 × 100 × 100 mm3, per ASTM C109[14]. Another set of three cubic specimens (100 mm) was tested in accordance with ASTM C642[15] to determine permeability of specimens, namely, volume of permeable voids (VPV) and water absorption. 2. Plastic shrinkage and environmental conditions In order to determine the plastic shrinkage cracking, specimens of size 280 × 177.5 × 50 mm3 with stress risers were cast (see Figure 1). According to ASTM C1579[11], the specimens were stored in a climate-controlled room (Figure 1) at a temperature of 36±3 ºC and 30±10% relative humidity (RH). The fans were used to achieve a wind speed of more than 4.7 m/s and an evaporation rate of at least 1.0 kg/m2⸱h. The testing is terminated when the final setting time of specimens was reached. During the test period of 28 days, the water (approximately 100 ml) was sprayed on the crack area every 12 hours. The average crack width was computed by processing of digital images in Adobe photoshop and MATLAB software. Figure 1: The plastic shrinkage mold and a climate-controlled room. 140

The 29th Special CU-af Seminar 2021 August 25, 2021 Results and Discussion Influence of the spore-loaded capsules and nutrients on behavior of the mortar specimens 1. Microstructures In this section, the microstructure of broken specimens for testing of compressive strength was observed using SEM/EDS technique. The morphology of the CON mixture after 28 days of curing is presented in Figure 2. It showed predominance of small fibrous crystals of calcium silicate hydrate (C-S-H) gel and portlandite with a plate shape. The needle-shaped crystals of ettringite were also detected. On the other hand, cementitious products were fluffy when urea-containing nutrients were added, which may be the result of a mixture of C-S-H gel and calcite. This is consistent with the test results reported by Wang et al.[16]. They also found that the hydration of the cement slurry was delayed by the presence of urea. The results showed that the use of bacterial spores did not affect the micro-morphology of the specimen compared with the S-0 mixture made with only nutrients. Although their photographs under the SEM were very similar, the carbon content in the reaction products of the S-1 specimen was higher than that of the S-0 mixture. The reasons for this will be discussed later. Figure 2: SEM images of specimens. As seen from Fig. 2, elemental compositions of specimens scanned by EDS indicated that the CON mixture contained calcium (Ca), silicon (SiO2), aluminum (Al2O3), and oxygen (O). Traces of carbon (C) were also detected. This implies that, in addition to cement hydration products, calcium carbonate (CaCO3) crystals caused by the reaction between cement and CO2 from air[17] were formed in the specimen. With regard to the effect of nutrients, it was noticed that the carbon content was increased as the nutrient was added. The results showed the carbon content of 16.3% for S-0 mixture compared with 13.5% for CON mixture. This indicates an increased amount of CaCO3 in the system. The urea in nutrient can react with water to form carbonic acid and ammonia. After that, the CaCO3 formed as a result of the reaction between portlandite and the acid remaining in the composite[18,19]. A significant increase in carbon content was also evident after the addition of bacterial spores. The carbon content of the reaction products in the S-1 mixture was 17.3%. It is possible to believe that the MICP process occurs since the EDS results are taken from those broken specimens after the compression test. When the crack appears, the embedded capsules in the crack zone will break, and the bacterial spores will be activated upon contact with oxygen, moisture and nutrients. However, no obvious formation of CaCO3 crystals could be observed under the SEM. 141

The 29th Special CU-af Seminar 2021 August 25, 2021 Permeability and strength Results of VPV of mortar specimens are shown in Table 2. The VPV values of S-0, S-0.5, and S-1 mixtures at 28 days were 10.3%, 10.5%, and 12.2%, which increased by 2.9%, 3.1%, and 4.8% respectively as compared to the CON mixture. This can be attributed to the effect of using yeast extract as a nutrient to produce calcium carbonate in concrete matrix. According to literatures[20,21], the attraction of water molecules was reduced by the presence of the amphiphilic compounds in the yeast extract leading to the formation of air bubbles in the system. Table 2: Results of permeability and compressive strength. The permeability of specimens was also measured in terms of the volume of absorbed water. The water absorption of tested specimens showed a similar trend to the VPV as shown in Table 2. The values of percentage water absorption for the S-0, S-0.5, and S-1 mixtures varied from 20.2 to 23.6%, whereas the value for the C mixture was 16.0%. This is due to the fact that the higher the VPV the higher the volume of absorbed water measured in the specimens[22]. The 28-day compressive strength of specimens is also presented in Table 2. It can be seen that the inclusion of nutrients has a detrimental effect on compressive strength. The compressive strength reduced from 45.9 MPa for the CON mixture to 20.7 MPa for the mixture made with nutrients only (S-0). This is consistent with the results that the VPV of S-0 mixture was higher than that of the CON mixture. The observation that addition of nutrients had a detrimental effect on the strength may be attributed to the formation of the fluffy microstructure[16]. The addition of bacterial capsule increased the permeability, and it decreased the strength of the mixes. The effect was more pronounced for the specimen made with 1% spore-loaded capsules. The reason for this lies in the fact that incorporating the capsules reduces the homogeneity of the composite. Wang et al.[6] also found the reduction in the compressive strength, even when microcapsules (with bacterial spores loaded) with the size of 5 µm were used. Mineral production capacity The specimen used in the investigation was subjected to plastic shrinkage cracking. When the crack passes through the bacterial capsules, the bio self-healing mechanism occurs. In addition to the content of bacterial capsule, the healing capacity has been challenged by the presence of moisture and oxygen. Also, depending on the genera of bacteria and the medium composition used in each study, the MICP process can produce various calcium carbonate polymorphs (calcite, vaterite and aragonite). Thus, the mineral production of the MICP process was investigated in this section. After 28 day of repairing, microstructures of the mortar at the crack zone were analyzed using XRD and SEM. 142

The 29th Special CU-af Seminar 2021 August 25, 2021 1. Phase compositions The specimens were collected from the crack zone at different depths (see Figure 3). Then, they were ground to ⁓ 200 mesh. The XRD method was used to determine their phase compositions. Beginning with the crystalline phases formed in the S-0 mixture, the diffraction peaks of quartz (SiO2) and anorthite (CaAl2Si2O8) were clearly found, as shown in Figure 4. These products forming in the specimen are mainly attributable to the presence of the fine aggregate used in the investigation. The results also showed the calcite crystal with an intensity peak at 29.4 of 2θ degrees. It should be noted that there is no significant variation in the peak intensity of calcite crystal for the S-0 mixture, except for the specimens positioned at the middle layer. Figure 3: Specimens used to study the microstructures. Figure 4: XRD patterns of specimens collected from the crack zone. 143

The 29th Special CU-af Seminar 2021 August 25, 2021 The XRD patterns of the S-1 mixture are shown in Figure 4. It can be seen that the peak intensity of calcite tends to increase when the bacterial spores are present in the system. The carbonate phase formed on the top surface can be observed with the intensity of 1,099 counts in its peak. This is the result of MICP process due to the presence of bacteria and nutrients. In both cases, the most noticeable change in the phase compositions is the peaks of quartz, anorthite, and albite (NaAlSi3O8). Their intensities are affected by randomly distributed aggregates in the specimens. 2. Morphology of CaCO3 crystals In this section, the specimens used to study the morphology were collected from the top zone of the cracking area. The SEM image (Figure 5) showed that the morphology of calcite crystal grown in the S-0 mixture. Although the calcite peak did not show an obvious change when the nutrients were added, the crystallization of calcite was found on the top zone. The triangular-like crystals were about 1-3 µm with rough surfaces. A large number of CaCO3 crystals were also found in the S-1 mixture, but they were rhombohedral. Figure 5: Morphology of CaCO3 crystals collected from the top zone of cracking area. Self-healing properties in mortar specimens The variations of crack width along the crack path are presented in Figure 6 and Figure 7. In the current study, the plastic shrinkage generated cracks ranging from 0.1 to 0.82 mm width and from 10 to 13 cm in length. The maximum crack width of the specimen which can undergo healing was also reported. After 28 days of healing, the maximum crack width healed for S-0 and S-0.5 were 0.28 mm and 0.74 mm, respectively. For the specimen made from 1% bacterial spores, the shrinkage crack width of up to 0.80 mm was sealed after only 3 days of healing (Figure 7). Relationship between the percentage of crack healing and healing time is shown in Table 3 and Figure 8. The percentage of crack healing was determined by computing the crack area that was sealed during healing period and dividing it by the initial crack area. A high rate of crack healing at early ages was found for all specimens. For example, in the case of S-0 mixture, approximately 42% of crack healing was reached at 7 days of repairing, and 50% at 14 days. The results also showed that the percentage of crack healing increased significantly as the 0.5% bacterial spores were added. However, there is little gain in the percentage of crack healing beyond 7 days. It should be noted that the shrinkage crack formed in the S-1 mixture was completely filled with MICP product after 3 days of healing. 144

The 29th Special CU-af Seminar 2021 August 25, 2021 Table 3: Percentage of crack healing. Figure 6: MICP crack healing for S-0 mixture. 145

The 29th Special CU-af Seminar 2021 August 25, 2021 Figure 7: MICP crack healing for S-1 mixture. Figure 8: Relationship between the percentage of crack healing and healing time. 146

The 29th Special CU-af Seminar 2021 August 25, 2021 Conclusion This work studied the influence of MICP on the remediation of plastic shrinkage cracking. The encapsulated Bacillus sphaericus spores and the nutrients were used to produce self-healing concrete. The results showed that the MICP method is an effective method to heal the crack caused by plastic shrinkage. The addition of 1% bacterial spore induces calcite precipitation that can heal the crack within 3 days. However, it should be noted that the compressive strength of mortar decreased because of the addition of the nutrient used to growth the bacteria. References 1. Mora-Ruacho, J., Gettu, R., & Aguado, A. (2009). Influence of shrinkage-reducing admixtures on the reduction of plastic shrinkage cracking in concrete. Cem. Concr. Compos. 39(3): 141-146. 2. Wu, L., Farzadnia, N., Shi, C., Zhang, Z., & Wang, H. (2017). Autogenous shrinkage of high performance concrete: A review. Constr. Build. Mater. 149: 62-75. 3. Seifan, M., Samani, A. K., & Berenjian, A. (2016). Bioconcrete: next generation of self-healing concrete. Appl. Microbiol. Biot. 100(6): 2591-2602. 4. Vijay, K., Murmu, M., & Deo, S. V. (2017). Bacteria based self healing concrete–A review. Constr. Build. Mater. 152: 1008-1014. 5. Jongvivatsakul, P., Janprasit, K., Nuaklong, P., Pungrasmi, W., & Likitlersuang, S. (2019). Investigation of the crack healing performance in mortar using microbially induced calcium carbonate precipitation (MICP) method. Constr. Build. Mater. 212: 737-744. 6. Wang, J. Y., Soens, H., Verstraete, W., & De Belie, N. (2014). Self-healing concrete by use of microencapsulated bacterial spores. Cem. Concre. Res. 56: 139-152. 7. Rao, M., Reddy, V. S., Hafsa, M., Veena, P., & Anusha, P. (2013). Bioengineered concrete-a sustainable self-healing construction material. Res. J. Eng. Sci. ISSN. 2278: 9472. 8. Xu, J., Wang, X., & Wang, B. (2018). Biochemical process of ureolysis-based microbial CaCO3 precipitation and its application in self-healing concrete.Appl. Microbiol. Biot. 102(7): 3121-3132. 9. Feng, J., Chen, B., Sun, W., & Wang, Y. (2021). Microbial induced calcium carbonate precipitation study using Bacillus subtilis with application to self-healing concrete preparation and characterization. Constr. Build. Mater. 280: 122460. 10. Xu, J., & Wang, X. (2018). Self-healing of concrete cracks by use of bacteria-containing low alkali cementitious material. Constr. Build. Mater. 167: 1-14. 11. ASTM C1579-13, Standard Test method for evaluating Plastic Shrinkage Cracking of Restrained Fiber Reinforced Concrete (Using a Steel Form Insert), ASTM International, West Conshohocken, United States, 2013. 12. Pungrasmi, W. Intarasoontron, J., Jongvivatsakul, P., & Likitlersuang S. (2019). Evaluation of microencapsulation techniques for MICP bacterial spores applied in self-healing concrete. Scientific Reports, 9(1):1-10. 13. ASTM C305-14, Standard Practice for Mechanical Mixing of Hydraulic Cement Pastes and Mortars of Plastic Consistency,ASTM International, West Conshohocken, United States, 2014. 14. ASTM C109/C109M-16, Standard Test Method for Compressive Strength of Hydraulic Cement Mortars (Using 2-in. or [50-mm] Cube Specimens), ASTM International, West Conshohocken, United States, 2016. 15. ASTM C642-13, Standard Test Method for Density, Absorption, and Voids in Hardened Concrete, ASTM International, West Conshohocken, United States, 2013. 147

The 29th Special CU-af Seminar 2021 August 25, 2021 16. Wang, L. G., Ju, S. Y., Chu, H. Y., Liu, Z. Y., Yang, Z. Q., Wang, F. J., & Jiang, J. Y. (2020). Hydration process and microstructure evolution of low exothermic concrete produced with urea. Constr. Build. Mater. 248: 118640. 17. He, P., Shi, C., Tu, Z., Poon, C. S., & Zhang, J. (2016). Effect of further water curing on compressive strength and microstructure of CO2-cured concrete. Cem. Concr. Compos. 72: 80-88. 18. Kim, H. Y. (2017). Urea additives for reduction of hydration heat in cement composites. Constr. Build. Mater. 156: 790-798. 19. Mwaluwinga, S., Ayano, T., & Sakata, K. (1997). Influence of urea in concrete. Cem. Concre. Res.27(5): 733-745. 20. Chen, X., Yuan, J., & Alazhari, M. (2019). Effect of microbiological growth components for bacteria-based self-healing on the properties of cement mortar. Materials. 12(8): 1303. 21. Şahin, Y., Akkaya, Y., Boylu, F., & Taşdemir, M. A. (2017). Characterization of air entraining admixtures in concrete using surface tension measurements. Cem. Concr. Compos. 82: 95-104. 22. Nuaklong, P., Boonchoo, N., Jongvivatsakul, P., Charinpanitkul, T., & Sukontasukkul, P. (2021). Hybrid effect of carbon nanotubes and polypropylene fibers on mechanical properties and fire resistance of cement mortar. Constr. Build. Mater. 275: 122189. 148



Appraisal of Corrosion Degree and Structural Properties on Reinforced Concrete Beam using Mill Cut Steel Fiber Concrete Withit PANSUK

The 29th Special CU-af Seminar 2021 August 25, 2021 Appraisal of Corrosion Degree and Structural Properties on Reinforced Concrete Beam using Mill Cut Steel Fiber Concrete Withit PANSUK1,2* Abstract Corrosion of steel that occurred in the reinforced concrete structures is one of the most important causes accounted for the deterioration and reduction of the loading capacity of concrete structures. This research aims to investigate the flexural behavior of the mild corroded reinforced concrete beam containing mill cut steel fiber under static loads. Twelve beams were cast with the water-to-cement ratio of 0.4 containing mill cut steel fibers with volume contents of 0%, 0.5%, 1.0%, and 1.5% respectively. After that, the corrosion acceleration was conducted to induce mild corrosion by applying the impressed current method. The electric current of 250µA/ cm2 was applied for 0, 19, 38 days induced the variety corrosion of 0, 2, 5%, respectively. The relationships between load-deflection and volume contents fibers of beams corroded were analyzed and the experiments noted that the presentation of mill cut steel fibers significantly affected the behavior flexural of the reinforcement concrete beams corroded. 1Innovative Construction Materials Research Unit 2Department of Civil Engineering Faculty of Engineering, Chulalongkorn University Bangkok, Thailand 151

The 29th Special CU-af Seminar 2021 August 25, 2021 Introduction and Objectives The reinforced concrete structure is one of the most widely used for various construction. However, reinforcement steel is well-known for its low corrosion resistance. Corrosion of steel rebar in reinforced concrete structures appears to be a critical worldwide issue. It shortens the service life of a structure and concerns the budget for repair, rehabilitation, or the reconstruction of the structure. The deterioration leads to a change in the performance of structure which affects the mechanical interaction as the behavior of material and the durable. The problem is set to fully understanding the behavior of eroded structures to deal with safety for usage. Concrete with high alkalinity is assumed to be a good material to protect steel reinforcement and prevent it from corrosion[1,2]. However, the invasion of aggressive ions from exposing environment incorporating with the reduction of alkalinity in concrete can induce the corrosion of steel reinforcement in reinforced concrete[1,3]. It was concluded that the tensile strength of corroded steel bars was lower than non-corroded and the ductility of bars varying levels of corrosion decreased as an increasing the degree of corrosion[4]. The formation of corrosion products, which have a volume of about 2.5-5 times higher than the original steel, leads to reduction of bonding strength between steel rebar and concrete, generates cracks in concrete, and in turn ease the penetration of aggressive ions into concrete. It is important to note that the corrosion of reinforcement leads to an increase in the brittleness of structure[5-17] which can lead to sudden collapse and cause danger for the user. The problems were proposed to minimize cost maintenance and repairing for the usage, moreover, the structure needs to be provided prolongs life. Recently, steel fiber concrete has been studied and applied for some decades. It has been proven that steel fiber is an important element to improve many features that commonly consider as the disadvantages of plain concrete[18, 19]. Using steel fiber in concrete with appropriate content can increase crack resistance, ductility, cycle loading capacity, or impact and fatigue resistance for concrete[18-20]. However, there is little research focusing on the durability of steel fiber concrete exposed to a chloride environment. It is noticed that steel fiber is usually used in high-strength concrete which contains a low water-to-cement ratio[20-23]. And research focusing on the degradation of steel fiber concrete would take a very long time while the timeframe of the project is limited. Therefore, the impressed current method was introduced to shorten the duration of the corrosion process. This method is based on the Faraday law and was widely applied to accelerate corrosion within reasonable times[24]. Thanks to these aforementioned problems, the objective of this research is to investigate the correlation between flexural loading capacity and displacement of the reinforced concrete beam using steel fiber concrete at a mild corrosion degree of steel reinforcement. Moreover, the crack patterns of reinforced concrete beams at various corrosion degrees can be measured to evaluate the model of failure. The experimental data attained in this study can be used to develop a simulation model that can predict the flexural behavior of RC beams using mill cut steel fiber concrete in further research. To induce and accelerate the corrosion progress on steel reinforcement at various degrees, the impressed current method is applied with the current densities of 250 μA/cm2 to generate corrosion degree at 2.0%, 5.0%. Sodium chloride solution with a concentration of 3% is used as an electrolyte. Besides, mill cut steel fiber is carried out in this study due to its advantages such as no fiber balling while mixing, non-subsidence, and non-outcrop. 152

The 29th Special CU-af Seminar 2021 August 25, 2021 Methods This experimental research focuses on the flexural behavior of fiber-reinforced concrete beams corroded with ranging volume fiber. The mechanical properties material, as well as the experimental production, will be shown in this section. Moreover, the outcome experimental will also be presented and discussed. Material properties The mix design having a water/cement ratio of 0.4 with a slump of 14±2 cm was used to cast both the plain and fiber concrete beams. The proportion of the concrete in this study was shown in Table 2 and the mill cut steel fiber with ranging volume 0%, 0.5%, 1.0%, 1.5% utilized in this research having the properties as illustrated in Table 1. To improve the workability of fresh concrete with the presentation of steel fibers, high-range water reduction admixtures were added. 24 cylinders 100 mm in diameter and 200 mm in height were fabricated and cured in the water tank before testing compressive strength at 7 days and 28 days followed ASTM C 39 for each of the variables volume fibers. Table 1: The properties of mill cut steel fiber. Table 2: Mix proportion of utilizing concrete. Figure 1: Mill cut steel fiber. Figure 2: Cross-sectional views of specimens. 153

The 29th Special CU-af Seminar 2021 August 25, 2021 Test Specimen Fabrication The details of the test specimens are illustrated in Figure 2. The twelve beams had 1400 mm span length with the cross-section 150 x 200 mm and were reinforced with four longitudinal deformed bars diameter 12 (two placed in the compression zone and two in the tensile zone) and transverse round bars with 6 mm diameter spaced 100 mm within a span. The yield strength of the round and deformed bar was 240 MPa and 300 MPa, respectively. The stirrup and the anchor bar were coated with epoxy to avoid local pitting corrosion causing the shear failure which is known as the brittle failure. Figure 3: Fabrication form Figure 4: Curing specimens To produce the concrete, cement and aggregates were blended first without fibers then the mixture of water and superplasticizer added, finally the hook-end steel fibers with no tendency balling were scattered to the mixture to produce the concrete with good workability. The extra time was asked to apply for the steel fiber in comparison with the plain concrete. Two series of reinforced concrete beams were cast in the steel mold as well as impacted by vibrator during the process and tested in four-point flexure according to ASTM C1609 - Figure 5 to investigate the flexural behavior of the corroded beams and non-corroded beam. All the fiber concrete beams were removed from the mold after first 24 hours and cured in humidity condition for 28 days. (a) Specimens’ geometry and test set up (b) Test picture Figure 5: Four-point flexural testing set up. 154

The 29th Special CU-af Seminar 2021 August 25, 2021 Accelerated Corrosion Technique In this study to accelerate corrosion of steel rebar embedded in the concrete beam within a short period, the impressed current technique was used. This method is based on the electrochemical process by application of constant direct current densities. To obtain significant corrosion with reasonable time, the higher of impressed current densities was utilized. However, utilizing higher levels of impressed current resulted in significantly increasing the width crack, lower current densities levels are required to reach corrosion products within a longer time in comparison with the same percentages of mass loss. Eight beams were immersed into the 3% sodium chloride environment with the solution depth was equivalent to concrete cover thickness to allow the entrance both humidity and oxygen. Before applying constant current density to induce desirable corrosion all specimens needed to immerse into sodium for 24 hours to ensure these specimens were in a saturated condition. Figure 6 illustrated the detail of the test setup, during the test the positive terminal of the DC power source was connected to the steel bars while another cooper wire connected the negative terminal with a stainless-steel plate placed along the length of the beam. A constant current density of 250µA/cm2 utilized for all specimens to induce 2 and 5% levels of corrosion. The degree of corrosion can be determined theoretically following the expression based on Faraday’s law. This theory measured exactly the mass of rust per unit surface area of the bar by applying current density within the predicted time. The rebars were weighed before inducing corrosion, after achieving the desired degree of corrosion the sample was broken to retrieve the rebar. After the elimination of concrete was the rebars were immersed in a solution of 3,5g hexamethylene tetramine diluted in 500 ml of hydrochloric acid and 500 ml reagent water according to the procedures of ASTM G1-03. After this process, the bars were removed from the solution and washed with clean water then measuring the mass to determine the level of corrosion following the equation. (1) (a) Specimens geometry and test set up (b) Test picture Figure 6: Accelerated corrosion set up. 155

The 29th Special CU-af Seminar 2021 August 25, 2021 Table 3: Testing matrix, experimental and theoretical of corrosion Results and Discussion Mass loss and Crack Corrosion Figure 7: Retrieving steel after corrosion Figure 8: Relationship between volume fiber and width crack. It is important to show that the targeted corrosion levels based on calculation were in fact achieved. The value from the comparison between the actual mass loss and the corresponding theoretical mass loss based on the theory Faraday’s Law for a different level of corrosion was 156

The 29th Special CU-af Seminar 2021 August 25, 2021 shown in Table 3. The average experimental errors for mild corrosion were found that 7% when comparing the actual mass loss and theory calculation. Although corrosion acceleration test methods are not entirely effective to simulate the real corrosion phenomenon which occurs in the environment, the impressed current method used in this research seems to be a reliable technique to simulate the corrosion of steel rebar in concrete. The maximum crack width for each specimen due to corrosion was listed in Table 3 while the crack pattern due to uniform corrosion product on the side of the non-fiber beams as illustrated in Figure 9. The corrosive products of iron are expansive, and their formation caused cracking and further deterioration in concrete. Cracking cover concrete due to initial corrosion allows chloride ions to penetrate faster certain reached critical concentrations widen the crack. The propagation of the crack having a tendency parallel to the corroded tensile steel rebar where observed formulation red and brownish-red rusts along the side surface of the beam. The final crack pattern of these beams was not affected regardless of the level of corrosion. These cracks were observed ranging from 0.05 mm to 0.30 mm at different locations. The crack pattern and width crack due to corrosion of fiber concrete beam was analyzed in Figure 10. The number of cracks and crack width increased according to a higher level of corrosion, the width crack in corroded beams specimen with 2% and 5% were 2 mm and 6 mm, respectively. A greater concentration of the corrosion products around the steel reinforcing bars generated higher expansion and internal stresses, this leads to the disrepair of the cover concrete. However, increasing the volume of fiber not only reduces the number of cracks but also against widening the crack due to corrosion. The relationship between volume fiber usage and the amount of width crack was shown in Figure 8. The result indicated utilizing 1.0% steel fiber effectively decreased the width crack, the reduction crack width up to 63.55% and 27.96% corresponding with the degree of corrosion 2%, 5%, respectively in comparison with corroded nonfiber beam. The distribution of steel fiber is an important bridge to transfer the internal stress, it helps limit the width crack. The various value of the crack width is the consequences of the different resistance electricity agents of fiber concrete beams. Figure 9: The crack pattern due to corrosion of non fiber concrete beam 157

The 29th Special CU-af Seminar 2021 August 25, 2021 Figure 10: The crack pattern due to corrosion of fiber concrete beam 158

The 29th Special CU-af Seminar 2021 August 25, 2021 Mechanical behavior due to corrosion A four-point loading test was conducted to assess the mechanical behavior of corroded and non-corroded concrete beams using the mill-cut steel fiber. The deflection at mid-span equivalent to load capacity was recorded during the test with a linear variable differential transducer (LVDT) and listed in Table 4. Investigation ultimate deflection corresponding to the ultimate load as well as the relation yield load and yield deflection were summarized in the table. The result indicated the ultimate load capacities of corroded concrete beams utilizing the same amount of steel fiber decreased linearly in comparison with non-corroded beams. On the other hand, the deflection of the corroded concrete beam at ultimate load increased gradually with the higher level of corrosion steel rebar. One of the main reasons reducing the load capacity of the corroded beam is the loss of mass this led to the loss cross-sectional of steel contributing to the flexural strength. Table 4: Relationship between load capacity and deflection under various degree of corrosion Effect content fiber on mechanical behavior of corroded and non-corroded concrete beam It was found that the bending strength of non-corroded beam or corrosive beams at the same degree of corrosion went up with an increasing volume of fiber content. The Figure 11 shows the effect of volume fiber on flexural behavior of non-corroded reinforced concrete beam. By observation, it was found that the higher volume steel fibers were used, the greater load capacity was obtained. With various volume steel fiber, the beam NC-1.0F has recoded the greatest load at yield point while loading capacity of NC-1.5F reaching the peak at the ultimate stage. The flexural strength of concrete beams was enhanced due to the presentation of fiber increasing the cross-section playing important role under the tensile stress. The Figure 12 illustrated the relationship deflection and load capacity of the group concrete beam is corrosive 2, 5% respectively. Those beams under mild corrosion conditions were observed to have a tendency similar to a non-corrosive beam, load capacity increased linearly 159

The 29th Special CU-af Seminar 2021 August 25, 2021 with an increased volume of steel fiber. However, the concrete beams corroded 5% was found that with utilizing 1.5% fiber the load capacity equivalent at the yield point of beam 5C-1.5F was the most increasing up to 67% while the non-corrosive beam NC-1.5F enhanced only 25%. The result also indicated using volume fiber was no significant changed at ultimate deflection of non-corrosive beams, by contrast, the group beams under 5% corrosion, the ultimate deflection of beam 5C-1.5F was investigated improved up to 57% in comparison with the 5C-0F. Accordingly, loss of reinforcing steel in the beam cross-section due to corrosion inclusion the effect corrosion on the volume fiber, this cause corroded with higher volume fiber content were more deflected. Figure 11: Relationship between load and deflection of fiber concrete beam. (a) Degree of corrosion 2% (b) Degree of corrosion 5% Figure 12: Relationship between load and deflection of corroded fiber concrete beam. Effect corrosion on mechanical performances of fiber-non fiber concrete beam The experiments program clearly showed a level of corrosion significantly affect the behavior of the non-fiber corroded concrete beams as well as the flexural strength of the beams. Obviously, an increasing degree of corrosion resulted in reducing the load capacities of the reinforced concrete beams in contrast improving the flexural strength of the reinforced concrete beams. Due to corrosion, the loss of tensile reinforcement in the beam cross-section, the beam was observed to more deflected. The Figure 13 showed there was a slight decrease in load capacity and deflection of 2C-0F in comparison with NC-0F whereas level corrosion of steel reached 5%, the ultimate load capacity was reduced by 21.3% and the deflection increased up to 40%. 160

The 29th Special CU-af Seminar 2021 August 25, 2021 (a) Volume fiber 0% (b) Volume fiber 0.5 % Figure 13: Effect of level corrosion on the flexural strength of the concrete beam. The mechanical behavior of fiber concrete beams utilizing 0.5% volume under various degrees of corrosion was investigated. The Figure 13 compared the relationship between load and deflection of these beams, a gradually went down in yield load of the beams was found as an increased level of corrosion. Considering the beam concrete with steel rebar is corrosive 2%, the flexural behavior as load capacity and deflection of beam 2C-0.5F remained the same as beam NC-0.5F, however the ultimate deflection and ultimate load of the beam under 5% corrosion reduced by 10%. Although a decrease was found in 5C-0.5F the reduction of bending strength at the same level of corrosion is less than the non-fiber corroded beam aforementioned above. Thanks to the distribution of steel fiber served as replacing the loss of reinforced steel rebar due to corrosion and rising the amount of reinforcing steel to the beam cross-section, then the flexural strength was enhanced better than the non-fiber corroded beams. (a) Volume fiber 1.0% (b) Volume fiber 1.5 % Figure 14: Effect of level corrosion on the flexural strength of the concrete beam. The flexural strength of corroded beams utilizing 1.0% volume fiber was concerned in the Figure 14. The behavior of fiber concrete beam under mild corrosion was dramatically enhanced. It was observed that with the presentation of steel fiber the bending strength was improved significantly as the increasing degree of corrosion 2% and 5%, the reduction ultimate 161

The 29th Special CU-af Seminar 2021 August 25, 2021 load of the corroded beam with higher level of corrosion reached up 1% and 6%, respectively. On the other hand, the deflection corresponding to the ultimate load of the corroded beam experienced a slight increase. There was a slightly increasing ultimate deflection of the group corroded beams with volume fiber 0.5% and 1.0% respectively, whereas the deflection at an ultimate load of corroded beams with 1.5% fiber went up two times in comparison with non-corrosive beams. Moreover, the bending strength of 2C-1.5 and 5C-1.5 was not affected by mild corrosion, the load capacity at the yield point was remained the same as NC1.5F. This improvement is due to the compensation of the volume steel fibers to the loss of tensile reinforcement under corrosion, which contributed to against the bending load. Flexural Static Test The formation of the crack under the applied load bending of non-corroded beams is illustrated in Figure 15. The type of crack mainly classifies as the flexural cracks in the beam, this crack appears at bottom mid-span of the beam and developed perpendicular to the beam longitudinal plane as a result of increased longitudinal tensile stress due to load and Flexural Shear Crack. The beams that are mixed with low-volume steel fibers (0.5%) or non-steel-fibers found diagonal cracks in the longitudinal plane of the beams. Cracks that occur are more likely web shear cracks due to tensile forces in the horizontal combined with the flexural shear crack. The load increase gradually until the beam failure due to flexural tension stress. it can be seen that the beam has a different load-bearing behavior, and the crack pattern is reduced with an increase in the content of steel fiber blending. Figure 15: Crack pattern of non-corroded fiber concrete beam under static load 162

The 29th Special CU-af Seminar 2021 August 25, 2021 Figure 16: Crack pattern of fiber concrete beam with 2% degree of corrosion under static load Figure 17: Crack pattern of fiber concrete beam with 5% degree of corrosion under static load 163

The 29th Special CU-af Seminar 2021 August 25, 2021 Another crack in the vertical perpendicular to the main pulling force (Flexural Crack) increases in volume when the beam under severe corrosion. As the degree of corrosion increases, the tensile stress generated in concrete beams is also much higher due to the loss of reinforced steel which makes the concrete beam bear more load of tensile loads, resulted in cracks perpendicularly. The flexural strength in the beam remained unchanged with increased the amount of steel fiber. These fibers bridge transfer effectively the internal stress compared with concrete beams that are not mixed with steel fibers. Conclusions This study investigated the mechanical behavior of non-corroded and corroded concrete beams were enhanced with various volume steel fiber. These specimens were exposed to the chloride environment and were accelerated desirable corrosion by impressed current technique. Loss of mass, crack due to corrosion, load-bearing capacity of fiber, and non-fiber concrete beams under corrosive conditions were presented. 1. The steel fiber reduced effectively the number of cracks due to corrosion. The reduction width crack trend to linearly an increased fiber consumption. By using the steel fibers at 1.0%, the width of cracks due to corrosion was most reduceable such as when the steel rebar is corrosive 2%, there was a significant reduction up to 63.55% in the number of cracks. Moreover, with a higher degree of corrosion the steel fiber still bridges well the crack 5%, it experienced a decrease by 27.96%. 2. Mill cut steel fiber increased significantly both yield and ultimate strength of the non-corrosive concrete beams. The improvement yield load of fiber concrete beams utilizing 1.0% fiber reached up to 42.5%. Moreover, under the same corrosive condition, steel fiber also rose the bending strength as well as the deflection of the corroded beams. Especially with the higher degree of corrosion as the steel rebar is corrosive 5%, the yield load of corroded beams using 1.0% volume steel fiber increased by 78.7%. 3. The reduction of ultimate load capacity non-fiber beams is associated with the degree of corrosion and was the best proportional to the mass loss. The ultimate or yielding capacity of corroded reinforced beams decreased dramatically linearly with the level of corrosion. As the steel bar is corrosive 5% resulted reducing 24.6% yield load capacity comparing to the non-corroded beams. 4. The similar correlation between the degree of corrosion and the ultimate capacity load of the fiber concrete beams was found. However, adding more volume of steel fiber maintained the ultimate or yield load capacity of the concrete beam under corrosive conditions. Especially, the concrete beam is severed corrosion as 5% the reduction of yield load with 1.0% steel fiber was only 5.2% and enhanced to be more deflected. 5. With the reasonable amount of steel fiber perfectly enhanced some limits of the mechanical properties of the corroded concrete beam. It is not only reduced the corrosion crack but also taking account for compensate the flexural strength of the structure due to loss cross-section by corrosion effect and limits the propagating crack under bending. 164

The 29th Special CU-af Seminar 2021 August 25, 2021 References 1. J. Broomfield, Corrosion of Steel in Concrete: Understanding, Investigation and Repair, E & FN Spon, London, 1997. 2. S. T. Shill, Chloride penetration into concrete structures exposed to the marine atmosphere, Mater Thesis, Florida Atlantic University, 2014. 3. U. Angst, Chloride induced reinforcement corrosion in concrete Concept of critical chloride content – methods and mechanisms, Ph.D. Thesis, Norwegian University of Science and Technology, 2011. 4. A. A. Almusallam, Effect of degree of corrosion on the properties of reinforcing steel bars construction and Building Materials 15, 361 -368, 2001. 5. D. H. Vu, R. Francois, Prediction of ductility factor of corroded reinforced concrete beams exposed to long term aging in chloride environment, Cement and Concrete composites 53, 136-147, 2014. 6. L. Yu, R. Francois, D. H. Vu, V. L’Hostis, R. Gagne, Development of chloride-induced corrosion in pre-cracked RC beams under sustained loading: Effect of load-induced cracks, concrete cover, and exposure conditions, Cement and Concrete Research 67, 246-258, 2015. 7. M. G. Stewart, Mechanical behaviour of pitting corrosion of flexural and shear reinforcement and its effect on structural reliability of corroding RC beams, Structural Safety 31, 19-30, 2009. 8. M. G. Stewart, Q. Suo, Extent of spatially variable corrosion damage as an indicator of strength and time-dependent reliability of RC beams, Engineering Structures 31, 198-207, 2009. 9. D. Hajializadeh, E. J. OBrien, M. G. Stewart, The sensitivity of bridge safety to spatial correlation of load and resistance, Structures 5, 23-34, 2016. 10. L. Yu, R. Francois, H. D. Vu, V. L’Hostis, R. Gagne, Structural performance of RC beams damaged by natural corrosion under sustained loading in a chloride environment, Engineering Structures 96, 30-40, 2015. 11. W. Zhu, R. Francois, Experimental investigation of the relationships between residual cross-section shapes and the ductility of corroded bars, Construction and Building Materials 69, 335-345, 2014. 12. L. Yu, R. Francois, H. D. Vu , V. L’Hostis, R. Gagne, Distribution of corrosion and pitting factor of steel in corroded RC beams, Construction and Building Materials 95, 384-392, 2015. 13. N. A. Taha, M. Morsy, Study of the behavior of corroded steel bar and convenient method of repairing, HBRC Journal 12, 107-113, 2016. 14. H. S. Lee, T. Noguchi, F. Tomosawac, Evaluation of the bond properties between concrete and reinforcement as a function of the degree of reinforcement corrosion, Cement and Concrete Research 32, 1313-1318, 2002. 15. C. Fang, K. Gylltoft, K. Lundgren, M. Plos, Effect of corrosion on bond in reinforced concrete under cyclic loading, Cement and Concrete Research 36, 548-555, 2006. 16. G. Malumbela, M. Alexander, P. Moyo, Variation of steel loss and its effect on the ultimate flexural capacity of RC beams corroded and repaired under load, Construction and Building Materials 24(6), 1051-1059,2010. 17. W. Zhu, R.Francois, D.Cleland, D.Coronelli, Failure mode transitions of corroded deep beams exposed to marine environment for long period, Engineering Structures 96, 66-77, 2015. 18. P.S. Song, S. Hwang, Mechanical properties of high-strength steel fiber-reinforced concrete, Construction and Building Materials 18(9), 669-673, 2004. 19. K. Holschemacher, T. Mueller, Y. Ribakov, Effect of steel fibres on mechanical properties of high-strength concrete, Materials and Design 31(5), 2604-2615, 2010. 165

The 29th Special CU-af Seminar 2021 August 25, 2021 20. D. Y. Yoo, Y. S. Yoon, N. Banthin, Flexural response of steel-fiber-reinforced concrete beams: Effects of strength, fiber content, and strain-rate, Cement and Concrete Composites 64, 84-92, 2015. 21. H. C. Mertol, E. Baran, H. J. Bello, Flexural behavior of lightly and heavily reinforced steel fiber concrete Beams, Construction and Building Materials 98, 189-193, 2015. 22. Y. Sahin, F. Koksal, The influences of matrix and steel fibre tensile strengths on the fracture energy of high-strength concrete Construction and Building Materials 25, 1801-1806, 2011. 23. G. Campione, M. L. Mangiavillano, Fibrous reinforced concrete beams in flexure: Experimental investigation, analytical modelling and design considerations, Engineering Structures 30, 2970-2980, 2008. 24. E. Maaddawy, T.A. and K.A. Soudki, Effectiveness of impressed current technique to simulate corrosion of steel reinforcement in concrete. Journal of materials in civil engineering 15, 41-47, 2003. 166



Noncovalent Functionalization of Graphene Oxide for Photocatalytic Applications Pannee LEELADEE

The 29th Special CU-af Seminar 2021 August 25, 2021 Noncovalent Functionalization of Graphene Oxide for Photocatalytic Applications Pannee LEELADEE1* Abstract Modification of graphene oxide (GO) with porphyrins; TPP and ZnTPP (TPP = 5,10,15,20-tetraphenyl-H21,H23-porphine) was achieved by facile synthesis through a self-assembly process. Compared to unmodified GO and pristine porphyrins, the GO-porphyrins composites exhibited remarkable photocatalytic enhancement under irradiation using a light emitting diode. Moreover, the GO-TPP composite was shown to be a competent and reusable photocatalyst for selective alcohol oxidation under mild conditions, affording desired products in moderate to excellent yields. (Phuangburee, T., Solonenko, D., Plainpan, N., Thamyongkit, P., Zahn, D. R. T., Unarunotai, S., Tuntulania, T., Leeladee, P. New J. Chem. 2020(44): 8264–8272.) In another sub-project, copper nanoparticles fabricated onto reduced graphene oxide (Cu NPs/rGO) were successfully synthesized via a one-pot dimethylformamide (DMF) reduction approach with an addition of the nominal water. Small particle sizes and high dispersion of Cu NPs on rGO were confirmed by scanning transmission electron microscopy (STEM). In addition, our Cu NPs/rGO was competent to catalyze the Ullmann- coupling reaction (i.e., arylation of 3,5-dimethylphenol with 86% yield and turnover number of 2,642). (Suktanarak, P., Tanaka, T., Nagata, T., Kondo, R., Suzuki, T., Tuntulani, T., Leeladee, P., Obora, Y. Bull. Chem. Soc. Jpn. 2020(93): 1164–1170.) 1Department of Chemistry Faculty of Science, Chulalongkorn University Bangkok, Thailand 169

The 29th Special CU-af Seminar 2021 August 25, 2021 Introduction and Objectives Graphene oxide (GO) has emerged as a promising material for catalytic applications due to its unique electronic and chemical properties, high surface area, low cost, and environmental friendliness. In addition, GO as a heterogeneous catalyst offers easy separation and reusability of materials. It also allows reactions to be carried out in a nontoxic solvent like water. These beneficial properties can lead to sustainable catalytic processes. In particular, its potential use as a photocatalyst is of interest and receive a great attention. However, GO by itself usually exhibits inefficient photocatalytic performance due to its low visible absorption. Hence, GO is usually modified with visible light-absorbing molecules or photosensitizers to enhance the photocatalytic activity. Heterogeneous photocatalysis under visible-light irradiation has been considered as an alternative route for green synthesis, leading to environmentally friendly and energy sustainable processes.[1] It was also reported that photocatalytic processes can result in higher selectivity when compared to conventional thermal methods.[2] In particular, GO was used as a semiconductor functionalized with various photosensitizers such as metal nanoparticles, metal complexes as well as organic dyes.[3-6] These materials combine individual properties of each component to produce a synergistic effect, which can result in higher photocatalytic activity, selectivity, and stability of the hybrid catalysts.Among those photosensitizers, porphyrin derivatives and their corresponding metal complexes offer several advantages, including their strong absorptivity in a visible region, thermal stability, photostability, and versatile modification with GO through covalent and noncovalent interactions.[7-10] However, covalent functionalization of GO can limit its extensive applications due to the complicated synthesis and low yield. On the other hand, modification through noncovalent interactions, such as electrostatic interaction, π-π stacking, and hydrogen bonding, is usually a simpler and more facile process. Additionally, noncovalent functionalization does not alter structure and electronic properties of GO while introducing new chemical groups on its surface. For example, a nanocomposite between GO and 5,10,15,20-tetraphenyl-21H,23H-porphine cobalt (CoTPP) was easily synthesized via a self-assembly process assisted by ultrasonic oscillation.[11] Notably, the application of modified GO as photocatalysts for organic transformation has been less explored, compared to those for energy conversion, carbon dioxide reduction, proton reduction, and other applications.[12-13] In addition, identifying the actual active sites on these catalysts, getting insight into the reaction mechanism, and control of the catalyst activity through surface modification are still great challenges because of their structural complexity that originates from different types of functionalities, sites, edges, and defects. Throughout the course of this study, we also learned that GOs can easily interact with metals, presumably due to the presence of oxygen functionalities including hydroxyl and epoxide groups, and may allow for reasonable fabrication with metal NPs (M NPs). Given this hypothesis, we extended our study and carried out another project to prepare hybrid materials between Cu NPs and graphene-based materials for catalytic application. Hybrid nanocatalysts are likely to offer combined properties of both homogeneous and heterogeneous catalytic systems—exhibiting high activity and selectivity of homogeneous catalysts, whilst being reusable like heterogeneous catalysts.[14] Our materials were further examined for their catalytic activity toward the Ullmann-coupling reaction 170

The 29th Special CU-af Seminar 2021 August 25, 2021 Objectives 1. To prepare photocatalysts based on GO and visible-light absorbing molecules by a simple sonication method 2. To investigate and optimize the catalytic activity of the synthesized GO composites towards alcohol oxidation Methods Synthesis of Graphene Oxide (GO) GO was prepared according to a modified published method.[15] To a flask containing KNO3 (0.40 g, 4.7 mmol) and conc. H2SO4 (25.0 mL), graphite powder (0.80 g) was added and the mixture was cooled in an ice bath. Then, KMnO4 (2.64 g, 16.7 mmol) was slowly added into the reaction and the mixture was stirred at room temperature for 6 h. Another portion of KMnO4 (2.44 g, 15.4 mmol) was added into the mixture and the reaction was stirred for 18 h. The crude product was poured into a beaker containing 30% H2O2 (5.0 mL, 44 mmol) and ice. The mixture was stirred for additional 1 h prior to a washing process. For the washing step, the mixture was transferred to centrifugal tubes and ultrapure water was added. The mixture was stirred for 10 min before centrifugation at 4000 rpm for 10 min. The washing step was then repeated using the following washing solution respectively: 6 M HCl, ultrapure water, KH2PO4/K2HPO4 buffer solution (pH 7.0), and ultrapure water (twice). It should be noted that the washing with KH2PO4/K2HPO4 buffer solution required overnight stirring to ensure a complete removal of the trapped acid on graphene oxide sheets. The product was collected and vacuum dried to yield dark sheets of graphene oxide (1.26 g). UV-Vis (H2O): λmax = 230 nm, FTIR (ATR): ν˜ = 2950-3650 (br) cm-1 (O-H), 1743 (m) cm-1 (C=O), 1612 (s) cm-1 (C=C), 1364 (m) cm-1 (C-OH), 1250 (m) cm-1 (C-O) and 1050 (s) cm-1 (C-O-C) cm-1, Raman: ν˜ = 1357 cm-1 (D-band), 1600 cm-1 (G-band), XRD: 2theta = 9.8, TGA: 76.7 °C (steam) 146.4 °C (oxygen containing functional into CO2 and/or CO) 300-700 °C (decomposition of the stable sp2 framework) Preparation of GO-TPP and GO-ZnTPP Composite In a typical procedure, graphene oxide (11.0 mg) in ultrapure water (4.0 mL) was sonicated at room temperature for 1 h. In the meantime, a solution of TPP or ZnTPP (11.0 - 66.0 mg) in DMF (2.5 mL) was also sonicated for 1 h. Then, the porphyrin solution was mixed with the GO suspension and the mixture was further sonicated for 2 h to form the composite. The mixture was filtered through cotton wool to remove large particles. The crude product was obtained through centrifugation (10,000 rpm for at least 50 min) and re-dispersed in ultrapure water (10 mL). The suspension was centrifuged at 4000 rpm for 10 min. The solid was dried under vacuum to obtain a dark-purple product (45.2 mg). To determine the amount of TPP or ZnTPP on GO, the composite was extracted with CH2Cl2 (3 times). The combined organic extract was dried and the amount of TPP or ZnTPP was analyzed by UV-Vis spectroscopy. UV-Vis and Fluorescence Titration A solution of TPP or ZnTPP (2.50 µM for UV-Vis and 1.25 µM for fluorescence in 2.00 mL DMSO) was added successive aliquots of GO dispersed in water (2 mg/mL). After each addition, the solution was stirred for 2 min and immediately monitored by UV-Vis and fluorescence spectroscopy (λex = 420 nm). 171

The 29th Special CU-af Seminar 2021 August 25, 2021 Photocatalytic Oxidation of Alcohols Unless otherwise noted, all reactions were carried out under the ambient conditions. In a typical experiment, a mixture of the GO-TPP composite (5 mg, GO:TPP = 1:4 weight ratio) and an alcohol substrate (5 mg of benzyl alcohol, (1-phenyl)ethanol, (1-thionyl)methanol or cyclohexanol) was prepared in ultrapure water (2.00 mL). The reaction was stirred and irradiated with the LED white cold light (14 W) for 24 h. Then, the reaction was filtered through 0.45 µm and 0.22 µm Nylon syringe filters to remove the composite then NaCl (30 mg) was added. The filtrate (1.00 mL) was extracted with CH2Cl2 and an internal standard was added for the GC analysis. All product yields were based on the average of at least two data points. All reactions were carried out for at least in triplicates. For the oxidation of benzhydrol, the reaction was scaled up: GO-TPP40 (100 mg) and benzhydrol (100 mg, 0.54 mmol) to allow the product isolation by the column chromatography for the determination of the isolated yield. In the recycling experiment, GO-TPP40 (100 mg) was dispersed in ultrapure water (40.0 mL) and the dispersion was sonicated for 1 h. Benzaldehyde (100 mg, 0.92 mmol) was added into the GO-TPP dispersion. Then, the reaction was irradiated and stirred for 24 h. The catalyst was separated by centrifugation (10,000 rpm, 30 min). The supernatant was collected, extracted with CH2Cl2 and dried over MgSO4 before the GC analysis. The separated catalyst was washed with generous amount of water and stirred overnight to ensure a complete removal of the trapped compounds in the composite. After that, the catalyst was dried under vacuum prior to use in the next cycle. Preparation of Cu NPs DMF (50 mL) was pre-heated at 140 °C for 5 min before the addition of CuCl2·2H2O (0.1 M, 0.5 mL). Thereafter, the solution was vigorously stirred at 140 °C, 1,500 rpm for 24 h. The blue CuCl2 solution turned yellow, which is indicative of Cu NP formation. Synthesis of Cu NPs/rGO In a typical procedure, GO (40 mg) was dispersed in DMF (25 mL) by sonication for 3 h. Prior addition of a metal precursor (CuCl2 or Cu NPs), and additional DMF (25 mL) were added to the mixture solution and pre-heated at 140 °C for 5 min. The solution was vigorously stirred at 1,500 rpm for 24 h. Thereafter, the solvent was removed in vacuo and the product was washed with MeOH (×3). To enhance Cu NP deposition onto rGO, the solvent mixture comprising DMF and a varying nominal amount of water was employed to study the effect of water in the synthetic process. First, GO (40 mg) was dispersed in DMF (25 mL) by sonication for 1 h prior to the addition of a nominal amount of water (0.5–8 mL). The solution was then sonicated for an additional 2 h. CuCl2 (0.1 M, 0.5 mL) was then added to the pre-heated GO solution (5 min for pre-heating) and stirred at 1,500 rpm for 24 h. The solvent was removed in vacuo and the product was washed with MeOH (×3). The copper loading on rGO was determined by ICP-AES. The Cu NPs/rGO samples were dispersed in a 3 M HNO3 solution to extract the copper from the support material prior to ICP-AES analysis. The ICP-AES data for each sample were compared with a standard copper solution to obtain correct and reliable results. 172

The 29th Special CU-af Seminar 2021 August 25, 2021 Catalytic activity studies toward the Ullmann-coupling reaction In a typical reaction, Cu NPs/rGO (1 mg), Cs2CO3 (2 mmol), iodobenzene (1.5 mmol) and 3,5-dimethylphenol (1 mmol) were added into a pressure tube and dispersed in DMF (1 mL) by sonication for 10 min. The mixture solution was deoxygenated under Ar for a 2-3 minutes and then stirred at 110, 120 or 130 °C for 24 h. For quantitative analysis, the products were extracted with ethyl acetate, hexane and n-tridecane as an internal standard prior to GC analysis. TON values were calculated from the number of moles of the obtained product per mole of copper active sites on the catalyst. Results and Discussion Interaction between GO and porphyrins in solution In addition to the π-π stacking and hydrophobic interaction, it was reported that the oxygen-functional groups on GO can serve as ligands binding to metal centers of metalloporphyrinoid complexes (Chart 1).[16] Due to this additional interaction, we hypothesize that ZnTPP might interact with GO differently from TPP to some extent, which would lead to a distinct photocatalytic activity of their composites. Hence, UV-Visible (UV-Vis) and fluorescence spectroscopy were carried out to get insight into the interaction between porphyrins (TPP and ZnTPP) and GO in solution. Figure 1a demonstrated UV-Vis spectra of TPP upon sequential addition of GO. Soret bands located at 415 nm are gradually shifted to 434 nm. This bathochromic shift indicates interactions between TPP and GO via the π-π stacking, which could originate from electron transfer and molecular flattening mechanism.[11,16,17] This suggests that GO can be modified with TPP and ZnTPP via the noncovalent functionalization. Chart 1: (a) Proposed structures of graphene oxide (GO), chemical structure of (b) TPP and (c) ZnTPP. Figure 1: Spectral change upon addition of GO (dispersion in H2O) into a TPP solution in DMSO monitored by (a) UV-Vis and (b) fluorescence spectroscopy; (c) Modified Stern-Volmer plots of fluorescence quenching: (black) TPP and (red) ZnTPP solution; (d) UV-Vis spectra of GO, TPP, and GO-TPP. 173

The 29th Special CU-af Seminar 2021 August 25, 2021 Fluorescence data also helped to confirm that TPP and ZnTPP could be assembled onto GO. Fluorescence intensity of TPP and ZnTPP gradually decreased as GO was added. This fluorescence quenching effect was further examined by a modified Stern-Volmer equation (eq.1).[11] F0/F = 1 + (KD+KS)[Q] + KDKS[Q]2 (eq.1) Where F0 = initial fluorescence intensity, F = fluorescence intensity after GO addition, KD = dynamic quenching constant, KS = static quenching constant, and Q = GO concentration The result after solving a quadratic equation for TPP indicated coexisting static/dynamic quenching mechanisms with a static quenching constant (KS = 2.60 L g-1) being larger than a dynamic quenching constant (KD = 0.80 L g-1). Experiments on ZnTPP with GO also gave similar results with KS = 3.49 L g-1 and KD = 0.34 L g-1. Moreover, KS/KD of TPP and ZnTPP in the presence of GO are 3.24 and 10.23, respectively. The much higher KS/KD of GO-ZnTPP compared to that of GO-TPP implies that ZnTPP tends to form the complex with GO more efficiently than TPP, possibly due to coordination between oxygen on GO and a Zn center in ZnTPP. Our results are consistent with those of similar systems reported previously.[11,16,18] Synthesis and characterization of porphyrin noncovalently functionalized graphene oxide composites Since our UV-Vis and fluorescence data indicated that TPP and ZnTPP could be assembled onto GO and suggested that ZnTPP could interact with GO more strongly than TPP, we decided to prepare the GO-TPP and GO-ZnTPP composites based on the self-assembly process (Chart 2). This sonication method was modified from a published procedure which reported that ultrasonic oscillation could accelerate the assembly between the porphyrin macrocycle and GO.[11] Chart 2: Preparation of graphene oxide composites with porphyrin derivatives. Successful preparation of GO-TPP and GO-ZnTPP was confirmed by various techniques including UV-Vis, FTIR, Raman and photoluminescence (PL) spectroscopy, TGA and SEM. The change in UV-Vis absorption features of both composites in comparison to GO and pristine TPP was clearly observed as shown in Figure 1d. In particular, peak broadening and red shift, compared with the absorption of the pristine porphyrins, were indicative of TPP and ZnTPP being assembled onto/into GO.[16] In addition, FTIR spectra of the composites revealed typical signals for both GO and the corresponding porphyrins. Slight bathochromic shift of the signals of the phenyl groups in the porphyrin rings was observed, suggesting the π-π interaction between these groups with sp2 carbons on GO sheets.[19] 174

The 29th Special CU-af Seminar 2021 August 25, 2021 In order to understand how the molecules are oriented on the surface of GO, we performed Raman measurements for the samples with two weight loads of TPP molecules deposited onto the GO flakes, i.e., 1:2 and 1:4 GO:TPP weight ratio. As shown in Figure 2, the Raman spectrum of GO exhibited well-known broad phonon modes around 1357 cm-1 and 1600 cm-1, originating from D and G bands of graphite, respectively.[42] Moreover, the spectra of the composites differed depending on the weight load. The spectrum of the sample with the GO:TPP weight ratio of 1:2 (GO-TPP20) showed numerous spectral bands, attributed to the TPP molecules (Figure 2), while the one of the sample with the GO:TPP weight ratio of 1:4 (GO-TPP40) exhibited the same modes but with lower intensity. The spectral signature of GO was visible only in the latter case. For a given system, the Raman intensity can be considered to be directly proportional to the amount of materials probed. Thus, the sample with the high amount of TPP should exhibit Raman bands of higher intensity. However, the situation was opposite in this case. Considering a large absorption coefficient of the porphyrin molecules in the visible optical range, such outcome could be explained by the molecule segregation at the GO surface. In the case of GO-TPP20, the porphyrin molecules formed thin, well-covered and relatively homogeneous layers on the GO surface. Therefore, no GO signal was observed since the scattered light originated from GO could not effectively escape the molecular film. On the other hand, the presence of GO signals in the spectrum of GO-TPP40 could indicated uncovered surface areas due to molecular aggregation of the porphyrin macrocycles during deposition and thus the incident light could easily reach there. This reasoning was in a perfect agreement with the SEM results performed for these samples (Figure 2). A similar situation was observed for the composite samples with the ZnTPP molecules, where the aggregation was also observed despite the differences in the intermolecular interactions of the TPP and ZnTPP molecules. Figure 2: (left) Raman spectra of GO flakes, GO-TPP20 and GO-TPP40. The spectra are stacked for clarity; (right) SEM Images of (a) GO-TPP20, (b) GO-ZnTPP20, (c) GO-TPP40 and d) GO-ZnTPP40. These images were taken at 2500x magnification, 20kV. 175

The 29th Special CU-af Seminar 2021 August 25, 2021 Photocatalytic activities toward selective alcohol oxidation As we speculated that different interaction between TPP and ZnTPP on the GO surface, as well as the amount of porphyrins in the composites might affect their catalytic activity, the GO-TPP and GO-ZnTPP composites with 1:1 to 1:6 GO:porphyrin weight ratios were examined for their catalytic activities toward the oxidation of benzyl alcohol in water under LED irradiation and ambient condition. Then, the products were analyzed by gas chromatography. Results for both GO-TPP and GO-ZnTPP composites showed that as the amount of porphyrins increased, the yield of the product was improved and started to plateau at the GO:porphyrin weight ratio of 1:4 (Figure 3a). Moreover, the results revealed that the reactions catalyzed by GO-TPP gave up to four times higher yield than those by GO-ZnTPP. This observation led to the question on the catalytic mechanism of these composites. Previous studies on the catalytic properties of GO using the DFT calculations revealed that the reaction proceeded by hydrogen atom transfer from the alcohol substrate to the epoxy groups on the GO surface, resulting in the formation of diols and followed by dehydration to release water to give aldehyde or ketone as the products.[20] In other words, the epoxy groups on GO served as the active site for the alcohol oxidation. Therefore, in our cases, it is likely that the coordination between the oxygen functional groups, including the epoxides, of GO and the Zn center of ZnTPP as mentioned in the previous sections hindered the active sites for the GO-catalyzed oxidation of alcohols. On the other hand, TPP aggregation on the GO surface left parts of the GO surface uncovered and, therefore, exposed the accessible active sites. In fact, our catalytic studies are in good agreement with the proposed mechanism by the DFT[20] andsurface characterization of our GO-TPP and GO-ZnTPP. Figure 3: Photocatalytic oxidation of benzyl alcohol using GO-TPP, GO-ZnTPP, TPP and ZnTPP as photocatalysts with (a) different GO:porphyrin weight ratio and (b) various LED light source To demonstrate synergistic roles of irradiated GO and TPP on the photocatalytic oxidation of benzyl alcohol, the reaction was carried out in water under ambient condition and irradiated with LED white cold light for 24 h. Then, the conversion of benzyl alcohol to benzaldehyde was determined by proton nuclear magnetic resonance (1H NMR) spectroscopy. The result clearly showed that in the presence of only TPP or GO, no or little product was observed. In contrast, when GO-TPP40 was present, high and selective conversion (approximately 76% yield) to benzaldehyde was achieved without any over-oxidized products (Figure S9). Also, it should be noted that a lower amount of catalyst (100 wt% GO-TPP) was used in our reaction compared to that of the previous study (200 wt% GO) using high temperature for activation.[21] 176

The 29th Special CU-af Seminar 2021 August 25, 2021 Furthermore, a variation of the LED light source demonstrated that irradiating the reaction using the white cold LED gave better photocatalytic activities over other light sources (Figure 3b). Thus, our optimal condition for the photocatalytic oxidation of alcohols included GO-TPP40 (100 wt% GO-TPP) as the catalyst, the white cold LED as the light source, water as the solvent and reaction time of 24 h. To expand the scope of the substrates, a variety of alcohols were tested using our GO-TPP40 under the optimal condition. Table 1 shows that our system also exhibited high compatibility with various substrates and our catalytic performance was on par with the previous report using GO as the catalyst, which was carried out in higher catalyst loading under high temperature.[21] It should be noted that our reactions could be performed in a large scale as shown in the oxidation of benzhydrol (entry 3) and afforded the high yield after isolation by column chromatography. Table 1: Oxidation of various alcohols.a (a) Conditions: 1:1 weight ratio of GO-TPP40:substrate, white cold LED, 24 h, ambient condition. (b) Product yield was determined by GC. (c) Isolated yield after column chromatography After the catalytic reactions, GO-TPP was successfully recovered and reused up to three cycles as shown in Figure 4, demonstrating the stability and recyclability of the catalyst. A dramatic decrease in the catalytic performance was observed in the 4th cycle. Investigation of GO by FTIR after each cycle revealed that GO was partially reduced and converted to rGO. Especially, the C-O-C vibrational signal at 1050 cm-1 was significantly decreased upon each catalytic run. This result is consistent with the proposed mechanism that the epoxy groups serve as the active site of alcohol oxidation. 177

The 29th Special CU-af Seminar 2021 August 25, 2021 Figure 4: (a) Reusability of GO-TPP in photocatalytic oxidation of benzyl alcohol. (b) FTIR spectra of extracted GO after each cycle. In comparison to the catalytic systems for benzyl alcohol oxidation reported previously, our GO-TPP shows several advantages.[21] In terms of catalytic performance compared to the pristine GO, our GO-TPP photocatalysts exhibited a comparable, high catalytic activity with significantly lower catalyst loading (100%w vs 200%w) and milder condition (room temperature vs Teflon-lined vessel, 100 °C). Regarding the catalytic condition, our system is considered to be a greener alternative. Apart from ambient condition used, our catalyst contains no heavy or precious metals, hence being less toxicity and more cost-effective when compared to other efficient photocatalysts. Moreover, the reactions can be carried out in water as a sole solvent. Therefore, our catalyst is a promising candidate to be further developed for an environmentally-friendly and sustainable catalysis. Fabrication of Cu NPs onto rGO Initially, Cu NPs deposited on rGO were chemically synthesized using a one-pot reduction approach with DMF as a pure solvent. DMF is demonstrated to play three important functions including serving as solvent, a reducing agent and a stabilizer.[22-24] To prepare the efficient fabrication of Cu NPs onto rGO, the influence of copper precursor types (CuCl2 and Cu NPs) on the Cu deposition onto rGO was examined. From the Cu concentration, analyzed by X-ray fluorescence (XRF), the use of CuCl2 as a Cu precursor provided a significant enhancement of Cu on the rGO, when compared with the Cu NPs precursor. However, from a naked-eye observation, the solution color after washing the Cu NPs/rGO product still retained an intense yellow color indicating a low Cu content fabricated onto the rGO leading to a yellow solution of free Cu (Cu NPs). Particularly in the case of the Cu NPs precursor, a more intense yellow color was observed after washing the product with MeOH, which is in agreement with XRF and ICP (vide infra) results. Thus, CuCl2 was appropriate as a Cu precursor for the preparation of Cu NPs/rGO. Next, the influence of the solvent mixture ratio, between DMF and H2O, on the synthesis of such hybrid materials was investigated. The rationale was that the addition of water may assist facile exfoliation of the graphite oxide to a monolayer of GO, and in addition, influence particle dispersion by improving the GO dispersion in the solvent. Markedly, it was demonstrated that oxygen functionality on GO can be directly reduced by DMF, especially, under heating.[25-26] This could lead to an inefficient fabricating Cu NPs onto rGO since oxygen-containing functional groups on GO play an important role to absorb the transition metal via non-covalent interaction.[27] Therefore, we hypothesized that using a mixture of 178

The 29th Special CU-af Seminar 2021 August 25, 2021 solvent (i.e. DMF/H2O) may also help to decrease the degree of GO reduction by DMF, resulting in a higher copper content deposited on rGO. From our experimental data, the presence of a nominal amount of water can improve the Cu deposition on the support materials. Both XRF and inductively coupled plasma-atomic emission spectroscopy (ICP-AES) results revealed that the Cu concentration on the graphene sheets was obviously enhanced in the presence of a nominal amount of water in DMF. Notably, 28 ppm of Cu on rGO in pure DMF could be increased to 100 ppm in 4:50 (v/v) H2O/DMF (Figure 5). Figure 5: (Left) Synthesis of copper nanoparticles fabricated onto reduced graphene oxide (Cu NPs/rGO) using a CuCl2 precursor in N,N-Dimethylformamide (DMF)/H2O (50/4 v/v); (right) Quantitative analysis of Cu fabricated on rGO by inductively coupled plasma-atomic emission spectroscopy. The Cu NPs/rGO was prepared under various conditions (use of Cu NPs precursor in DMF, CuCl2 precursor in DMF and CuCl2 precursor in 4:50 (v/v) H2O/DMF. Characterization of the hybrid Cu NPs/rGO materials Next, Cu NPs/rGO was characterized by various techniques including IR, XPS and scanning transmission electron microscopy (STEM). Figure 6 shows the IR spectra of the starting GO and Cu NPs/rGO synthesized under 4:50 (v/v) H2O/DMF. The GO revealed IR vibrations related to oxygen functionalities including C-O (1106 cm−1), O-H (1231 cm−1), C=O (1699 cm−1) as well as sp2-hybridized C=C at ~1500–1600 cm−1, which are consistent with previous works. [28-30] Conversely, the intense oxygen vibrational bands at ~1,000 ‒ 1,800 cm−1 are weakened in Cu NPs/rGO indicating the incomplete reduction of GO (reduced graphene oxide form, rGO). Some oxygen functionalities retained their structures because of the inhibition for GO to be fabricated with Cu NPs. Additionally, STEM observations revealed a narrow and precise nano-sized distribution of Cu on rGO (~3–5 nm), Figure 6 (right). The small particle diameter implied a high specific surface area, which is a prerequisite for improved catalytic activity. The results also suggest that the Cu NPs are highly dispersed on rGO and are not aggregated to larger sizes, resulting in stabilizing of the Cu NPs with the support material (rGO) and DMF. 179

The 29th Special CU-af Seminar 2021 August 25, 2021 Figure 6: (left) Comparison of the FT-IR spectra of GO (red line) and Cu NPs/rGO (black line); (right) Scanning transmission electron microscopy image of Cu NPs/rGO exhibiting a high degree of Cu NP dispersion on rGO. Inset: size distribution histogram indicating an average particle size of ~3–5 nm. Catalytic activity of copper nanoparticles fabricated onto reduced graphene oxide (Cu NPs/rGO) toward the Ullmann-coupling reaction. Table 2: Cu NPs/rGO-catalyzed arylation of phenol1 1Reaction conditions: 3,5-dimethylphenol (1 mmol), iodobenzene (1.5 mmol), Cs2CO3 (2 mmol), catalysts: Cu NPs/rGO or GO (1 mg/1 mL DMF) or Cu NPs (0.1 mM, 1mL), reactions were performed under heating for 24 h. 2Reaction conditions: 3,5-dimethylphenol (5 mmol), iodobenzene (7.5 mmol), Cs2CO3 (10 mmol), catalysts: Cu NPs/rGO (1 mg/1 mL DMF), reactions were performed under heating for 24 h. 3Quantitative product analysis by GC. 4TON = turnover number. aCu NPs were fabricated onto rGO under 4:50 (v/v) H2O/DMF utilizing the CuCl2 precursor. bCu NPs were fabricated onto rGO under DMF utilizing the Cu NPs precursor. cCu NPs were fabricated onto rGO under DMF utilizing the CuCl2 precursor. 180

The 29th Special CU-af Seminar 2021 August 25, 2021 In addition to the observed high dispersion, precise narrow size distribution and high stability, the number of Cu active sites (Cu NPs) fabricated onto the support material was also important to improve catalytic activity. All of the Cu NPs/rGO materials prepared in 4:50 (v/v) H2O/DMF exhibited a higher Cu loading on rGO when compared with the other reaction conditions, and in addition, provided higher activity toward the Ullmann-coupling reaction. First, the optimal catalytic conditions were studied as a function of temperature (110–130 °C), Entries 6–8. The results show that product yield increased at higher reaction temperatures with the optimum temperature at 130 °C. As shown in Table 2, the catalytic activity data are in good agreement with the hypothesis that Cu NPs/rGO prepared in 4:50 (v/v) H2O/DMF exhibited significantly higher efficiency for the catalytic arylation of phenols than for Cu NPs/rGO prepared in pure DMF in the presence of the same precursor or the Cu NPs precursor. Additionally, the data show that the desired product was not obtained in the absence of any catalyst or in the presence of GO only, implying an insignificant number of any active sites (Cu), Entries 1–3. To further increase the turn over number (TON) and test the robustness of the Cu NPs/ rGO catalyst, the substrate (3,5-dimethylphenol and iodobenzene) was increased up to 5-fold under the same amount of the Cu NPs/rGO catalysts prepared in 4:50 (v/v) H2O/DMF. The findings show that 86% of the product was still obtained, with TON increased to 2,642 implying the degree of robustness of the Cu NPs/rGO catalysts prepared in 4:50 (v/v) H2O/DMF toward the Ullmann-coupling reaction. Conclusion GO-based photocatalysts, GO-TPP and GO-ZnTPP, were successfully prepared through the noncovalent functionalization by a simple method. The spectroscopic data and SEM revealed the additional interaction between ZnTPP and GO through the coordination of the oxygen-functional groups to the Zn center, apart from the π-π stacking through the porphyrin rings. This led to the higher photocatalytic performance of GO-TPP toward the oxidation of benzyl alcohol as compared to that of GO-ZnTPP. In addition, GO-TPP was shown to be the competent photocatalyst for the selective oxidation of alcohols in water under ambient conditions. Using the photoactivation also helped to lower catalyst loading when compared to the case of the conventional thermal process. Reusability of the photocatalysts was also feasible, and our system could be applied to various alcohol substrates. Furthermore, our findings suggested that although the surface modification of GO via the noncovalent interaction might not change the structure and properties of GO, it could alter the accessibility of the substrate to the GO active sites which, in turn, had a strong effect on the catalytic efficiency. Insight into the interactions between GO and the photosensitizer molecules will ensure the rational design and development of efficient graphene-based photocatalysts for sustainable catalysis. In addition, a new approach was presented to improve the fabrication of Cu NPs/rGO upon addition of a nominal amount of water in DMF via a one-pot DMF reduction approach. Water decreases the reducing properties of DMF and further protects the GO oxygen functional groups, which are important for the adhesion of Cu on rGO. The influence of water content on the fabrication of Cu NPs on rGO was examined by XRF and ICP-AES analyses of the Cu concentrations. XPS elucidates the structural properties of the functional groups of the support and hybrid materials and provides an explanation for the synthetic mechanism. Furthermore, Cu NPs/rGO exhibited a narrow and precise nano-sized distribution (3–5 nm) 181

The 29th Special CU-af Seminar 2021 August 25, 2021 with highly dispersed Cu on rGO, as observed by STEM. The analytical data reported herein demonstrates the successful synthetic approach for Cu NPs/rGO. Finally, to show an example for practical application toward Ullmann-coupling reaction, Cu NPs/rGO was demonstrated to catalyze arylation of 3,5-dimethylphenol with high yield (86% yield and turnover number = 2,642). References 1. Lang, X., Chen, X., Zhao, J., Chem. Soc. Rev., 2014(43): 473-486. 2. Palmisano, G., Garcia-Lopez, E., Marci, G., Loddo, V., Yurdakal, S., Augugliaro, V., Palmisano, L., Chem. Commun. (Camb.), 2010(46): 7074-7089. 3. Fernando, K. A., Watson, V. G., Wang, X., McNamara, N. D. , JoChum, M. C., Bair, D. W., Miller, B. A., Bunker, C. E., Langmuir, 2014(30): 11776-11784. 4. Al-Nafiey, A., Kumar, A., Kumar, M., Addad, A., Sieber, B., Szunerits, S., Boukherroub, R., Jain, S. L., J. Photochem. Photobiol. A. Chem., 2017(336): 198-207. 5. Kumar, A., Aathira, M. S., Pal, U., Jain, S. L., ChemCatChem, 2018(10): 1844-1852. 6. Pan, Y., Wang, S., Kee, C. W., Dubuisson, E., Yang, Y., Loh, K. P., Tan, C.-H., Green Chemistry, 2011(13): 3341. 7. Georgakilas, V., Tiwari, J. N., Kemp, K. C., Perman, J. A., Bourlinos, A. B., Kim, K. S., Zboril, R., Chem. Rev., 2016(116): 5464-5519. 8. Zhu, M., Cao, C., Chen, J., Sun, Y., Ye, R., Xu, J., Han, Y.-F., ACS Appl. Energy Mater., 2019(2): 2435-2440. 9. Jurow, M., Manichev, V., Pabon, C., Hageman, B., Matolina, Y., Drain, C. M., Inorg. Chem., 2013(52): 10576-10582. 10. Di Mauro, A., Randazzo, R., Spano, S. F., Compagnini, G., Gaeta, M., D’Urso, L., Paolesse, R., Pomarico, G., Di Natale, C., Villari, V., Micali, N., Fragala, M. E., D’Urso, A., Purrello, R., Chem. Commun. (Camb.), 2016(52): 13094-13096. 11. Shu, J., Qiu, Z., Wei, Z., Zhuang, J., Tang, D., Sci. Rep., 2015(5): 15113. 12. Machado, B. F., Serp, P., Catal. Sci. Technol., 2012(2): 54-75. 13. Su, C., Loh, K. P., Acc. Chem. Res., 2013(46): 2275-2285. 14. Mohammadi, O., Golestanzadeh, M., Abdouss, M., New J. Chem., 2017(41): 11471. 15. Hirata, M., Gotou, T., Horiuchi, S., Fujiwara, M., Ohba, M., Carbon, 2004(42): 2929-2937. 16. Yang, J.-H., Gao, Y., Zhang, W., Tang, P., Tan, J., Lu, A.-H., Ma, D., J. Phys. Chem. C, 2013(117): 3785-3788. 17. Xu, Y., Zhao, L., Bai, H., Hong, W., Li, C., Shi, G., J. Am. Chem. Soc., 2009(131): 13490-13497. 18. Ge, R., Wang, X., Zhang, C., Kang, S.-Z., Qin, L., Li, G., Li, X., Colloids Surf. A Physicochem. Eng. Asp., 483, 45-52. 19. Mondal, B., Bera, R., Nayak, S. K., Patra, A., J. Mater. Chem. C, 2016(4): 6027-6036 20. Boukhvalov, D. W., Dreyer, D. R., Bielawski, C. W., Young-Woo, S., ChemCatChem, 2012(4): 1844-1849. 21. Dreyer, D. R., Jia, H. P., Bielawski, C. W., Angew. Chem. Int. Ed. Engl., 2010(49): 6813-6816. 22. Isomura, Y., Narushima, T., Kawasaki, H., Yonezawa, T., Obora, Y., Chem. Commun. 2012(48): 3784. 23. Oka, H., Kitai, K., Suzuki, T., Obora, Y., RSC Adv., 2017(7): 22869. 24. Zhang, Y., Sun, H., Zhang, W., Gao, Z., Yang, P., Gu, J., Appl. Catal. A: General 2015(496): 9. 25. Zhou, D., Cheng, Q.-Y., Han, B.-H., Carbon 2011(49): 3920. 182

The 29th Special CU-af Seminar 2021 August 25, 2021 26. Ai, K., Liu, Y., Lu, L., Cheng, X., Huo, L., J. Mater. Chem. 2011(21): 3365. 27. Basiuk, V. A., Alzate-Carvajal, N., Henao-Holguín, L. V., Rybak-Akimova, E. V., Basiuk, E. V., Appl. Surf. Sci, 2016(371): 16. 28. Zhang, P., Song, T., Wang, T., Zeng, H., Appl. Catal., B: Environmental 2018(225): 172. 29. Guo, M., Zhao, Y., Zhang, F., Xu, L., Yang, H., Song, X., Bu, Y., RSC Adv. 2016(6): 50587. 30. Zhang, H., Hines, D., Akins, D. L., Dalton Trans. 2014(43): 2670. 183

Synthesis and electrocatalytic activities of MgNiCoCuZn high entropy oxides for hydrogen and oxygen evolution reactions Padtaraporn CHANHOM and Numpon INSIN

The 29th Special CU-af Seminar 2021 August 25, 2021 Synthesis and electrocatalytic activities of MgNiCoCuZn high entropy oxides for hydrogen and oxygen evolution reactions Padtaraporn CHANHOM1* and Numpon INSIN1* Abstract High entropy metal oxides (HEOs) are materials that exhibit high stability and perform good catalytic performance. In corporation of earth-abundant metals especially alkali-earth metals into HEO could lead to materials with high catalytic performance in a lower production cost. In this article, MgNiCoCuZn HEO was selected to demonstrate this hypothesis. MgNiCoCuZn HEO was prepared using two different approaches, high temperature solid state reaction and hydrothermal method. Features of HEO, including pure single crystalline phase and uniformly distribution of the metal compositions, were obtained from both methods as confirmed using X-ray diffraction and Scanning Electron Microscope equipped with X-ray Energy Dispersive Spectroscope. Hydrothermal method yielded the HEO of smaller size and spherical shape with lower preparation temperature. The HEOs performed as electrochemical catalyst when used in carbon electrode as the resulted electrodes exhibited substantially lower in overpotential in oxygen evolution reaction. Earth-abundant HEOs have potential to used as catalysts for further studies. 1Division of Inorganic Chemistry, Department of Chemistry Faculty of Science, Chulalongkorn University Bangkok, Thailand 185