Development of Pd-supported Catalysts for the Conversion of Palm Oil to Biohydrogenated Diesel in a Microscale-based Reactor

Y. Sa-ngasaeng,a R. Nernrimnong,a N. Sirimungkalakul,b G. Jovanovic,c and S. Jongpatiwuta,d,* aThe Petroleum and Petrochemical College, Chulalongkorn University, Bangkok 10330, Thailand bInnovation Institute, PTT Public Company Limited, Ayutthaya 13170, Thailand cSchool of Chemical, Biological, and Environmental Engineering, Oregon State University, Corvallis, OR 97331, USA dCenter of Excellence on Petrochemical and Materials Technology, Chulalongkorn University, Bangkok 10330, Thailand


Introduction
Microstructured reactors have emerged as an innovative platform technology for process intensification, which involves a substantial improvement in equipment size, energy consumption, costs, and safety. The microscale-based reactors give high surface to volume ratio 10 4  ). 1,2 These reactors allow reactants to flow in a thin film between reactor walls (100 -1000 µm apart), resulting in the superior heat and mass transfer rate and reduction of hot-spot during operation. 1,3 Microscale-based reactors are suitable for operation with highly endothermic or exothermic, and even explosive reaction processes. 4,5 Other advantages are short residence times, narrow residence time distributions, higher space-time yield, lower energy consumption, and lower capital cost. 1,[6][7][8] Furthermore, microscale-based reactors have an effective way of increasing the ca-pacity of production by numbering-up the units with the identical design. 9,10 There are several applications for catalytic reaction systems in microscale-based reactor, such as hydrogenation, oxidation process, Fischer-Tropsch synthesis, methanol synthesis, steam reforming, dehydration of alcohols, epoxide synthesis, and photocatalytic process. 7,[11][12][13][14][15] These reaction processes are often supported with thin (~1 µm) solid-catalyst layers coated on the microreactor walls. 13 The transesterification process in microreactor has been widely studied. [16][17][18][19] However, there are only few studies 3,[20][21][22] about hydrodeoxygenation of triglycerides to produce liquid fuel in the microreactor. One noted challenge in the process of hydrodeoxygenation of vegetable oils in packed bed reactors is a limited mass transfer between gas, liquid, and solid catalyst phase. Limited mass transfer is primarily caused by a very low diffusion coefficient for bulky molecules of triglycerides. 3 In one study, Sinha and co-worker 3 studied the hydroprocessing of jatropha oil to kerosene and diesel over NiMo/Al 2 O 3 catalyst using microchannel, and monolithic reactors. The results showed that reaction selectivity had improved when compared to the conventional trickle bed reactor. Another study by Zhou et al. 20 investigated the hydrodeoxygenation of microalgae oil to green diesel using pre-sulfided NiMo/Al 2 O 3 catalyst in micro fixed-bed reactor. They found that the space-time yield (STY) of product had enhanced in a microreactor (internal diameter < 1 mm) when compared to macroscopic reactor (internal diameter > 1 mm) because of the microreactor's superior mass transfer characteristics. The extraordinary mass transport performance of microscale-based reactors is best viewed by rivalling the characteristic diffusion times in a typical microreactor (t dif = 10-100 ms) and classic reactors (t dif = 10,000-100,000 ms).
One of the challenges in heterogeneous catalytic reaction processes in a microscale-based reactor, is the coating of catalyst on the reactor walls. 23 To achieve a long-term stability of the catalyst coating, a strong adhesion between catalyst and the substrate is desired. 13 To improve the coating adhesion, a thermal pretreatment was used to create an oxide layer before catalyst coating was applied. 24 While there are several methods to coat the catalyst on the surface of reactors (suspension, sol-gel deposition, hybrid method between suspension and sol-gel, electrophoretic deposition (EPD), chemical vapor deposition (CVD), and physical vapor deposition (PVD) 14,[24][25][26], the suspension and the sol-gel methods are the most widely used. 14,24 Suspension method is a simple technique that uses a slurry of the catalyst support (Al 2 O 3 , TiO 2 ) or finished catalyst, binder, acid, and solvent (water) as main ingridients. 14,24,27 Sol-gel method is a low-temperature process, in which a small amount of precursor is used. 28,29 Sol-gel coating method provides a thinner layer than the suspension method. 24 The suspension and sol-gel solutions are typically coated on reactor walls by washcoating, dip-coating, or spray coating procedures.
In the hydrodeoxygenation process, noble-metal catalysts are considered more active and more environmentally favorable than sulfided transition metal catalysts. 30,31 Palladium (Pd) is one of the noble metals with high catalytic performance for hydrogenation reaction. 31,32 A Pd-supported titania (TiO 2 ) catalyst showed high catalytic activity for biohydrogenated diesel production. 33 An alternative catalyst support, alumina (Al 2 O 3 ), is also widely used as a catalyst support in the hydrotreating process of vegetable oils. 20,34,35 In this study, comparison of catalytic activity of Pd supported by alumina or titania, prepared by two different coating slurries (suspension and sol-gel), is a point of interest in our investigation. The physical and chemical properties of the coated catalysts were analyzed by 3D-optical profiler measurement, SEM-EDX, XRD, BET, and adhesion test. The catalytic activity in the microscale-based reactor was tested for converting vegetable oil to biohydrogenated diesel at 325 ºC, 3.4 MPa, and H 2 /feed molar ratio of 96.

Microscale-based reactor
Experiments were performed in a microscale-based reactor with dimensions of 30 cm x 20 cm x 5 cm (PCT patent application No. PCT/ US2016/061814, United States: 15/774,756, Japan: JP2018-525363, Thailand: 1801002705). Fig. 1 shows the design of the experimental microscale-based rector, a) 3D 'exploded' view of the reactor, b) top view of the catalyst plate, c) cross-section view of catalyst holder with the coated catalyst on the substrate plate, and d) top view of bottom clamp plate. The catalyst plate (14.5 cm x 18 cm) was made of stainless steel 316 with a multitude of 300 μm high microposts. The diameter of microposts was 300 µm, and microposts were arranged in an equilateral (1 mm) triangular pitch. The assembly of two catalyst plates generated 600 µm size of microscale-base reactor. The complete experimental assembly consisted of two catalyst plates, two mixer stainless steel plates, and two cover plates. All plates were laser welded to provide excellent sealing and plate alignments (Fig. 1c). The catalyst holder was put into the bottom clamp plate. Afterward, the bottom clamp plate was assembled with the top clamp plate using a graphite seal and tightened with 18 screws (Fig. 1d).

Catalyst plate pretreatment
Before catalyst coating, the catalyst plates were pretreated by rinsing with DI water, followed by ethanol rinse to remove impurities from the plate. The plate was then sonicated in 20 wt% citric acid (C 6 H 8 O 7 · H 2 O, RCI Labscan Limited) for 30 min, dried, and calcined in a furnace at 800 ºC for 2 h in the air environment. This thermal treatment of the SS catalyst plates creates a thin oxide layer, which substantially improves the catalyst adhesion.

Washcoating of alumina catalyst support
Alumina support was coated on catalyst plate by washcoating of the alumina slurry. The slurry was prepared by suspension (SUS) or sol-gel (SG) method. For suspension, polyvinyl alcohol (PVA, 98-99 % hydrolyzed, Polysciences) was slowly dissolved in hot water at 85 ºC and continuously stirred for 3 h. Acetic acid (CH 3 COOH, 100 %, QReC) was then F i g .

Washcoating of titania catalyst support
Titania support was also coated on pretreated catalyst plate by washcoating method using titania slurries prepared by suspension or sol solution from the sol-gel method. For the suspension, the slurry was prepared by the same procedure as alumina, except using nanopowder titania (P25, 21 nm, Sigma-Aldrich). After the washcoating, the sample was dried in an oven at 110 ºC overnight, followed by calcination at 500 ºC in air for 4 h with a heating ramp rate of 10 ºC min -1 .
For the sol-gel method preparation, acetylacetone (ACA, Merck) was homogeneously mixed with tetra isopropyl titanate (TIPT, Aldrich) with a 1:1 molar ratio. Afterward, a laurylamine hydrochloride (LAHC, 98 %, Merck) aqueous solution with a pH of 4.2 was prepared by mixing LAHC with ethanol and distilled water (3:2 volume ratio) and hydrochloric acid (HCl, 37 %, RCI Labscan). The prepared LAHC aqueous solution was added to the ACA-modified TIPT solution, in which the molar ratio of TIPT to LAHC was 4 to control the porosity of TiO 2 . The mixture was continuously stirred at room temperature overnight to obtain transparent yellow sol as a homogeneous sol solution. 33 The solution was coated onto a pretreated plate, then dried at 80 ºC in a closed system for 3 h to complete gel formation. The coated plates with gel were dried in the open system at the same temperature (80 o C) overnight to evaporate water, followed by calcination at 500 ºC in air for 4 h with a heating ramp rate of 10 ºC min -1 to remove the LAHC template.

Impregnation of Pd
The inclusion of Pd on Al 2 O 3 and TiO 2 coated films on the catalyst plate was done by the im pregnation method. Palladium nitrate dihydrate (Pd(NO 3 ) 2 · 2H 2 O, Aldrich) solution was added to the coated Al 2 O 3 or TiO 2 support to acquire the 1 wt% of Pd loading. After impregnation, the coated plates were dried at 110 ºC overnight, followed by calcination at 500 ºC in air for 4 h with a heating ramp rate of 10 ºC min -1 .

Catalyst preparation for fixed bed reactor
The TiO 2 support was prepared by the sol-gel method. After obtaining transparent yellow sol, the sol was dried at 80 ºC in a closed system for a week to complete the gel formation. Subsequently, the gel was dried in air at 80 ºC overnight, and calcined at 500 ºC for 4 h to obtain TiO 2 support powder. The incorporation of 1 wt% Pd on TiO 2 powder was prepared by the incipient wetness impregnation method. The Pd/TiO 2 catalyst was then pelletized and sieved to obtain 20-40 mesh of particle size before placing it in a fixed bed reactor.

Catalyst characterization methods
The prepared catalysts were tested for their physical and chemical properties. The coated catalyst plates were analyzed for their morphology, composition, and adhesion by scanning electron microscope (SEM), energy-dispersive X-ray analysis (EDX), 3D optical profiler measurement, and adhesion test. SEM-EDX (JEOL JSM-6610LV) provides the catalyst layer coating images and elemental composition results. The 3D optical profiler and surface measurement using ZeGage machine provide measurements of the surface texture, roughness, and thickness of catalyst layer. The adhesion between the coated catalyst and catalyst plate was evaluated by weight remaining after sonication in dodecane using an ultrasonic bath (Elma, 25 kHz and 750 W) for 30 min, as defined in the following equation.
For Brunauer-Emmett-Teller (BET) surface determination and X-ray diffraction (XRD) analysis, the catalyst sample was detached from the plate and provided in powder form. The BET surface area analyzer (BEL, Belsorp Mini-II) was used to analyze the impregnated Pd catalyst's surface area. The samples were outgassed under vacuum condition at 350 ºC for 4 h before analysis. The XRD analysis method was used to identify the crystalline material by Rigaku X-ray diffractometer system (RINT-weight of catalyst after sonication Adhesion (%) = 100 weight of catalyst before sonication ⋅ (1) 2200) with CuK α radiation (1.5406 Å) and a nickel filter. The catalysts were measured in the 2θ range of 20-80º with a scanning speed of 5º min -1 .

Catalyst activity testing
The catalyst was activated before testing. Two catalyst plates were installed and sealed with two mixing plates and two cover plates, as shown in Fig. 1c. The assembled reactor was placed in a furnace at 200 ºC and exposed to a flow of 99.99 % of hydrogen for three hours. The stream of 50 wt% of palm oil in dodecane was then fed to the reactor using a high-pressure pump (Waters 515). The flow of hydrogen gas and the reaction pressure were controlled by a mass flow controller (Brooks, SLA5800 Series) and a backpressure regulator (Brooks, SLA5800 Series). The reaction was performed at 325 °C, 3.4 MPa, H 2 /feed molar ratio of 96, and WHSV of 4.6, 9.7, 15.8, and 27.9 h -1 . The liquid product was separated by a gas-liquid separator, and analyzed hourly by an Agilent 7890A gas chromatograph equipped with an FID detector, as illustrated in Fig. 2.
For the conventional fixed bed reactor operation, the catalyst pellets were reduced by flowing 99.99 % hydrogen at 200 °C for three hours. The stream of pure palm oil was fed to the reactor (3/4inch OD, stainless steel) with a flow of hydrogen at an H 2 /feed molar ratio of 30. The reaction was operated at 325 °C, 3.4 MPa, and WHSV of 0.7 h -1 . The liquid product was collected in a condenser, and analyzed by a gas chromatograph equipped with an FID detector (Agilent 7890A). The conversion and space-time yield were calculated using Eq. 2 and Eq. 3:

Conversion:
weight of feed converted Conversion (%) = 100 weight of feed input ⋅ Space-time yield: weight of BHD produced/time STY = weight of catalyst (3) Results and discussion

Surface morphology of catalyst plates
To improve the adhesion between the catalyst and catalyst plate, the catalyst plates were thermally pretreated to increase surface roughness and generate an oxide layer. 36,37 The catalyst plates before and after pretreatment with citric acid followed by thermal pretreatment were analyzed for their surface morphology using SEM-EDX and 3D-optical profiler measurement. Table 1 summarizes surface roughness and elemental composition of catalyst plate before and after pretreatment. The results showed that the surface roughness of the catalyst plate before pretreatment and catalyst plate cleaned with citric acid was not significantly different. In contrast, the roughness increased from 3.055 μm to 3.283 μm after thermal pretreatment. Fig. 3 shows SEM images of the catalyst plate before and after thermal pretreatment at a magnification of 500 and 3,000. The surface morphology of catalyst plate before pretreatment was smooth compared to the one after thermal pretreatment. The contents of oxygen increased after thermal pretreatment from 7.71 wt% to 36.05 wt%, while the composition of other elements relatively decreased. The SEM images (Fig.  3c) and elemental composition suggests the presence of iron oxide layer with higher roughness after thermal pretreatment. Similar effects were reported by Schmidt et al. 27 The thermal pretreatment en-hances adhesion between the catalyst layer and catalyst plate due to increasing anchoring sites on the stainless-steel surface caused by iron oxide layer formation. 38,39 Coated catalyst characterization The morphology and thickness of the catalyst layer on the catalyst plate was investigated by 3D optical profiler measurement. Fig. 4 Table 2. Fig. 4a shows the 3D optical profile images of a blank catalyst plate whose microposts height are 300 μm. Fig. 4b shows that the coated Pd/Al 2 O 3 (SUS) catalyst was thin and uniform. In contrast, the coated Pd/Al 2 O 3 (SG) catalyst (Fig. 4c) was thick and uneven on the catalyst plate. The thickness of Pd/Al 2 O 3 (SUS) was 17-20 μm, which was lower than Pd/Al 2 O 3 (SG) (25-28 μm). On the other hand, the quantity of Pd/Al 2 O 3 (SUS) on the catalyst plate was higher than Pd/Al 2 O 3 (SG), which could be due to the high density of Al 2 O 3 commercial powder used in suspension method. As shown in Fig. 4d, the Pd/TiO 2 (SUS) provided a thick, uniform and smooth layer on the catalyst plate. It was observed from Fig. 4e that the Pd/TiO 2 (SG) presented a non-uniform and cracked layer on the surface. In agreement with 3D optical profile results, the thickness of the Pd/TiO 2 (SUS) catalyst layer was 44-47 μm, which was higher than that of Pd/ TiO 2 (SG) (34-38 μm). In line with the thickness, the Pd/TiO 2 (SUS) weight was higher than Pd/TiO 2 (SG). SEM-EDX confirmed the morphology and elemental distribution of catalysts. Fig. 5 exhibits SEM images and SEM-EDX mapping results of different catalysts on the catalyst plate, which are Pd/Al 2 O 3 (SUS), Pd/Al 2 O 3 (SG), Pd/TiO 2 (SUS), and Pd/TiO 2 (SG), respectively. In agreement with 3D optical profile results, the SEM images in Figs. 5a and 5c indicate that the Pd/Al 2 O 3 (SUS) and Pd/TiO 2 (SUS) catalysts gave a uniform and homogeneous layer. The corresponding EDX results also show a high distribution of support composition (Al or Ti and O) on the catalyst plate. This could be due to the narrow particle size distribution of Al 2 O 3 and TiO 2 powder (<50 nm) used in the suspension slurry. As shown in Figs. 5b and 5d, the SEM images of Pd/ Al 2 O 3 (SG) and Pd/TiO 2 (SG) gave a cracked non-continuous layer, and some agglomeration as also observed by 3D optical profile images. The EDX mapping of Al or Ti and O also confirmed the presence of a cracked layer. This could be because the sol-gel synthesis generates rapid condensation in sol, resulting in a shrinkage layer of gel film. 40 However, the EDX mapping of all catalysts exhibited that Pd was well dispersed on the catalyst supports layer as well as on the catalyst plate. It was noticed that Cr and Fe were detected due to their composition in the stainless-steel catalyst plate.

-3D optical images of different catalysts on catalyst plate: (a) blank catalyst plate, (b) Pd/Al 2 O 3 (SUS), (c) Pd/Al 2 O 3 (SG), (d) Pd/TiO 2 (SUS), and (e) Pd/TiO 2 (SG)
It was noted that the homogeneous catalyst layer was more favorable than cracking surface, which could be because the detachment of catalyst and substrate can occur during catalytic testing. However, some cracking might provide the benefit to improve the contacting area of catalyst and reactants if that cracking can attach to the substrate in catalytic testing environment.
XRD identified the crystallinity of catalyst.  (SUS) showed the same nanopowder TiO 2 structure pattern, while the XRD pattern of the Pd/TiO 2 (SG) showed the crystalline structure of pure anatase phase. [41][42][43] The Pd showed diffraction peak at 33.87º, which referred to (100) plane of PdO phase. The BET surface areas of Pd/Al 2 O 3 (SUS), Pd/Al 2 O 3 (SG), Pd/TiO 2 (SUS), and Pd/TiO 2 (SG) catalysts are reported in Table 2. For both Al 2 O 3 -and TiO 2 -supported catalysts, the one prepared by solgel gave a significantly higher surface area than the one prepared by the suspension. These results are in line with previous studies 41, 44 showing a high surface area of support synthesized by sol-gel compared to support prepared by conventional synthesis. The adhesion test results of all prepared catalysts are shown in Table 2. For Al 2 O 3 -supported catalysts, Pd/Al 2 O 3 (SUS) and Pd/Al 2 O 3 (SG) exhibited 83.05 and 82.74 % adhesion. For TiO 2 -supported catalysts, Pd/TiO 2 (SUS) showed higher %adhesion, as compared to Pd/TiO 2 (SG). This could be because, during the coating steps, suspension slurries and Al 2 O 3 sol solution were added with PVA binder, which could improve the adhesion between catalyst and catalyst plate. Hydroxyl group in PVA could enhance the interaction between the alumina particles by having hydrogen bond formation (Al-O-H · · · · OH or Al-O-· · · · OH). This hydrogen bond bridge could improve the adhesion for alumina coating. 39 On the other hand, Pd/TiO 2 (SG) exhibited much lower %adhesion as compared to the others. This could be due to the absence of binder in the coating solution, which resulted in an unstable cracked layer observed by SEM and 3D-optical profile image.

Catalyst activity testing for BHD production
The catalytic activity of the prepared Pd-supported catalysts was investigated in the deoxygenation of the palm oil process. The deoxygenation was performed in a microscale-based reactor for eight hours at 325 ºC and 3.4 MPa. These operating conditions were recommended by our industrial partner, and suggested as optimal by many published studies [45][46][47] for deoxygenation of triglycerides. WHSV (4.6-27.9 h -1 ) and H 2 /oil molar ratio of 96 were selected to obtain comparable conversions for the catalytic activity over the differently prepared catalysts as well as to support the slug flow microfluidic behavior in the microscale-based reactor.
Since the amounts of catalyst coated on the catalyst plate were not constant to keep the flow pattern the same for all experiments, the WHSV was varied, as shown in Table 2. Product yield per unit weight of catalyst or space-time yield (STY) was considered to evaluate catalyst performance among the catalysts tested. Fig. 8 exhibits the catalytic activity, assessed through space-time yield, of different Pd-supported catalysts. For comparison, the catalytic activity was also illustrated in a spider chart together with the catalyst properties, as shown in Fig. 9. The results indicated that the conversion of palm oil over Pd/ Al 2 O 3 (SUS), Pd/Al 2 O 3 (SG), and Pd/TiO 2 (SG) catalysts were overall very close to 100 % for eight hours on-stream. Triglycerides conversion slightly decreased after six hours for the Pd/TiO 2 (SUS) catalyst. The Pd/TiO 2 (SG) gave the highest space-time yield of BHD followed by Pd/TiO 2 (SUS), Pd/Al 2 O 3 (SG), and Pd/Al 2 O 3 (SUS), as shown in Fig. 8. The Pd/TiO 2 catalysts exhibited higher STY when compared to Pd/Al 2 O 3 catalysts. This result illuminates the synergetic effect among Pd particles and titania support, which is described by the spillover mechanism. In the spillover mechanism, absorbed hydro- gen in the Pd nanoparticle diffuses into the titania support such that Ti 4+ is reduced to Ti 3+ Lewis acid sites. After the reduction, the oxygenated groups strongly interact with Ti 3+ , leading to the weakening of the C=O bond and deoxygenation occurs. 48,49 The Pd/TiO 2 (SG) gave higher STY than the Pd/TiO 2 (SUS), which could be explained by a strong metal-support interaction (SMSI) effect. The Pd/TiO 2 (SG) contains a pure anatase phase, which generally shows the SMSI effect. In contrast, the Pd/TiO 2 (SUS) had a mixed phase of anatase and rutile that did not significantly express the SMSI effect at low reduction temperature. 33,48,50 The Pd/TiO 2 (SG) has a high surface area compared to Pd/TiO 2 (SUS).
Higher surface area provides the opportunity for better metal dispersion on the catalyst surface. Also, some cracks of Pd/TiO 2 (SG) might increase diffusion in the catalyst layer, which helps catalyst performance before the catalyst is detached from the substrate. 38 For Pd/Al 2 O 3 catalysts, Pd/Al 2 O 3 (SG) gave higher STY as compared to Pd/Al 2 O 3 (SUS), which could be owing to the high surface area of the Pd/ Al 2 O 3 (SG) catalyst.
A brown precipitate was observed in the exit stream after a certain time on-stream. The precipitate was filtered and analyzed by SEM-EDX to investigate the catalyst deactivation. The result is shown in   34,35 and our experimental results prove that due to its excellent mass transfer, the contact time for the deoxygenation reaction could be substantially reduced. Thus, microscale-based reactors provide higher productivity for the BHD process as compared to conventional fixed bed reactors. It is interesting to compare the performance of our arguably best catalyst Pd/TiO 2 (SG) with catalysts used in similar reaction processes with Jatropha 3 and Microalgae 20 oil. Fig. 10 shows the catalyst properties and catalytic performance of the Pd/TiO 2 (SG) compared to two literature studies. The first study, Sinha et al., investigated the deoxygenation of jatropha oil over presulfided NiMo/Al 2 O 3 in microreactor to produce BHD and jet fuel at 380 °C, 3.8 MPa, and H 2 /feed ratio of 2,500 NL L -1 , WHSV of 46 h -1 . 2 The other study, Zhou et al., investigated the deoxygenation of microalgae oil over presulfided NiMo/Al 2 O 3 in microreactor for BHD production at 360 °C, 3.4 MPa, H 2 /oil ratio of 1,000 mL mL -1 , WHSV of 0.65 h -1 . 9 All reaction processes were performed in microscale-based reactors and yielded alike BHD products.
As shown in the diagram, all studies exhibited complete conversion, while this study presented the highest STY of BHD product. The Pd/TiO 2 (SG) exhibited the lowest surface area and degree of adhesion, and the highest thickness of catalyst layer as compared to the others. High space-time yield obtained with our Pd/TiO 2 (SG) catalyst implies promising overall process performance results. The relatively lower surface area of TiO 2 support as compared to Al 2 O 3 still provides enough active sites for the reaction. The low degree of adhesion was presented in Pd/TiO 2 (SG), which could be due to the high thickness of catalyst, surface cracking, and no additive in coating process.
Obviously, specific surface area (m 2 g -1 catalyst) and adhesion properties of Pd/TiO 2 (SG) could be improved to create a truly superior catalyst for BHD production from plant-based triglycerides.

Conclusions
Pd/Al 2 O 3 and Pd/TiO 2 catalysts with suspension and sol-gel coating methods were investigated for biohydrogenated diesel production in a microscale-based reactor. The SEM and 3D optical profile images indicated that the coating by suspension slurry provided a homogeneous catalyst layer, while the coating by sol-gel solution provided a non-homogeneous layer. In the adhesion test, the Pd/TiO 2 (SG) showed the lowest %adhesion due to the absence of PVA binder and the presence of a cracked non-homogeneous layer. The Pd/Al 2 O 3 (SUS), Pd/Al 2 O 3 (SG), and Pd/TiO 2 (SG) catalysts exhibited conversions closed to 100 %. For the Pd/ TiO 2 (SUS) catalyst, the conversion slightly decreased because of its low surface area. The highest space-time yield of BHD was obtained by Pd/TiO 2 (SG), which was due to the presence of pure anatase phase that strongly interacted with Pd, thus promoting the hydrogen spillover effect for the deoxygenation reaction. Despite low adhesion and coating non-uniformity, Pd/TiO 2 (SG) provided the highest catalytic performance for deoxygenation reaction. The reaction process in the microscale-based reactor exhibited a much higher space-time yield, when compared to classic fixed bed operation, due to its superior mass transfer. However, improved adhesion of the coated catalyst should be further developed. It could be concluded that Pd/TiO 2 (SG) coating was successful in deoxygenation of triglyceride application in microscale-based reactor.