Composite of Pt/AlSBA-15+zeolite catalyst for the hydroisomerization of n-hexadecane: The effect of platinum precursor

Abstract In this study, the Pt (0.5 wt%) catalysts supported on the bimodal composite materials consisting of AlSBA-15 and BEA zeolite, were prepared using three different precursors of Pt, i.e. H2PtCl6, Pt(NH3)4(NO3)2 and Pt(NH3)4Cl2. The obtained catalysts were characterized by means of N2 sorption, XRD, Py–FTIR, FTIR, H2 chemisorption and TEM to determine the effect of the Pt precursor on their physicochemical properties. Catalytic performance of obtained catalysts was investigated in hydroisomerization of n-hexadecane. It was found, that the Pt precursor had significant impact on catalytic activity and selectivity. The results of n-hexadecane hydroconversion showed that the catalyst obtained with Pt(NH3)4(NO3)2 provided the highest yield of the most desired high-cetane number products. In addition, it had the lowest selectivity to cracking products, which are undesirable in hydroconversion of long-chain alkanes.


Introduction
Hydroisomerization of the n-alkanes is an important process in the production of high-quality fuels [1][2][3]. The branching of normal long chain alkanes enables the production of diesel fuel with improved cold flow properties [4] -the branched alkanes are characterized by lower temperature of cloud point and cold filter plugging in comparison with their normal analogues [5]. Thus, in recent years the process became more common. Additionally, the researchers aim to investigate new sources of long chain n-alkanes, thus its production from alternative energy carriers, especially from biomass resources is growing attention. Such fuels include the fractions obtained from bio-syngas in Fischer-Tropsch (FT) synthesis as well as those produced by hydroconversion of vegetable oils (HVO -Hydrotreatment of Vegetable Oils) [6]. The FT synthesis from bio-syngas technology is currently applied by Linde Engineering Dresden and HVO technologies lived to see the realization of industrial processes inter alia by Neste Oil (NExBTL license) [7] and Honeywell/UOP (Ecofining license) [8]. Typically hydroisomerization of n-alkanes takes place over the bifunctional catalysts containing both metal, which provides active sites for hydrocarbons dehydrogenation/hydrogenation and the Brønsted acid sites on which skeletal isomerization of carbocations occurs. According to classical mechanism of isomerization and cracking of alkanes, hydroconversion of alkanes involves few steps (Scheme 1) [9][10][11]. In the first step, n-alkanes are dehydrogenated on metal sites to olefins. Next, olefins diffuse into the Brønsted acid sites where they are subsequently protonated to form of the corresponding alkylcarbenium ions. The resulting carbenium ions undergo: (i) skeletal rearrangement via protonated cyclopropane (PCP) intermediates, (ii) alkyl shift and (iii) hydride shift or β-scission of C-C bond. Subsequently, these ions are deprotonated and hydrogenated over metal sites to form i-alkanes (mono-and multibranched isomers) and cracking products. In order to improve the performance of isomerization catalysts of long-chain alkanes, much attention is paid to ensure an adequate metal/acid balance, which strongly affects the properties of a catalyst, especially its selectivity towards isomerization [12,13] and consequently the yield of isomers. It can be found in many reports, that the adequate metal/acid balance can be controlled by several factors: (i) the preparation method of support and catalyst [14][15][16], (ii) the textural and chemical properties of support (concentration of acid sites and diffusion rate in the pores) [17,18], (iii) the nature of the metal precursor (dehydrogenation/hydrogenation activity and number of metal sites) [19][20][21] and (iv) the distance between the metal and Brønsted acid sites [22,23].
Dehydrogenation/hydrogenation properties of bifunctional catalysts are generally provided by the noble metals in mono-or bimetallic systems, usually Pt and Pd. The concentration of metal sites must be sufficient to: (i) supply a maximum amount of alkene intermediates to the acid sites, (ii) hydrogenate rapidly the primary branched carbenium ions and (iii) limit the scission of C-C bond. It was proved, that even small amounts of Pt (about 0.5 wt%) may provide sufficient activity of the catalyst (dehydrogenation/hydrogenation) to balance the acid sites [24]. Thus, a single metallic site with high activity (Pt or Pd) can balance several acid sites, depending on their strength [25]. It is also known that with the increase of the amount of metal content, the dispersion decreases and consequently the size of the metal particles increases as a result of agglomeration and sintering [19]. The uneven distribution of the metal particles and the presence of its agglomerates on the outer surface of the support, deteriorates the hydrogenation/dehydration properties of the metallic sites. As a result, a decrease in selectivity to isomers can be observed. Furthermore, carbon deposition on the surface of the catalyst occurs. In order to ensure high metal dispersion, high surface area of active phase and finally the activity of the catalyst, several methods of metal incorporation can be used.  [32]. The type of metal precursor used for impregnation of the support, affects the location of the metal particles (i.e. on the outer or inner surface of the support) and the distance between the metallic and Brønsted acidic sites [14]. The distribution of metal sites also depends on the type of the support i.e. texture, distribution and strength of the acid sites [33]. Therefore, there are some discrepancies in the literature regarding the influence of platinum precursor differing with the valence state of Pt, its form (in cationic or anionic group), and the presence of chlorine or ammonium ions on size of metal particles and dispersion. The results of Belopukhov et al. [31] and M€ aki-Arvela et al. [16,34] [19] showed, that the catalysts prepared with use of an anionic precursors i.e. H 2 PtCl 6 or (NH 4 ) 2 PtCl 4 , were characterized with better activity and higher yield of isomers, than those made using a cationic Pt precursors. The authors attributed better catalytic properties of the catalysts prepared with use of anionic precursors of Pt, to smaller particle sizes and higher Pt dispersion. Chandler et al. [35] revealed, that the presence of chlorine in the Pt precursor affected the size of the metal particles. For catalysts containing chlorine ions, a reduction of the crystal size of metal particles was observed and as a consequence the Pt dispersion increased. A similar effect was not observed for catalysts containing chlorine ions and amine groups in precursors of Pt, due to autoreduction of Pt species by NH 3 during calcination stage. Wang et al. [19] and Antoniassi et al. [36] proved, that the valence state of Pt in precursor (in the form of cation Pt (II) or anion containing Pt(IV)) affected crystallographic orientation of the particle facets on support surface. The catalysts prepared with use of H 2 PtCl 6 , exhibited smaller Pt particles in crystal orientation (1 1 1) which are more active in hydrogenation/dehydrogenation of hydrocarbons. Usage of the cationic precursor caused creation of larger Pt particles with the crystallographic orientation of the particle facets Pt (1 0 0). Despite the development that was made in the synthesis of catalysts for n-alkanes branching, the precise design of the catalyst composition to achieve the optimal activity, practicability and economic productivity still remains valid. To the best of the authors knowledge, no papers considering the impact of Pt precursor in the case of micro-mesoporous carriers were published. Most of the works regarding the impact of Pt precursor were referred to the single component supports i.e. Al 2 O 3 [37], SiO 2 [38], SAPO-11 [39], SBA-15 [40], zeolites [19,39] and related materials (physical mixture of zeolite and ordered mesoporous material or zeolite and inert materials [41,42]) in powder form. Our investigation concerned the influence of platinum precursor on physicochemical properties and catalytic activity of Pt catalysts supported on Scheme 1. Hydroisomerization/hydrocracking process of n-hexadecane over bifunctional catalysts.

Catalyst preparation
The synthesis of composite material AlSBA-15 þ BEA zeolite (weight ratio zeolite: AlSBA-15 ¼ 1:4) was carried out following the procedure presented in Scheme 2 and described in detail in our recent work [18].
The powder of AlSBA-15 þ BEA zeolite were blended with binder -20 wt% of AlO(OH), peptized with 3% solution of HNO 3 and then shaped into the cylindrical extrudates. The resulting extrudates were dried (12 h at 110 � C) and calcined (6 h at 450 � C). The platinum catalysts were prepared by the dry impregnation method using formed and calcined extrudates with a particle size of 0.40-0.63 mm and three different platinum precursors: H 2 PtCl 6 , Pt(NH 3 ) 4 (NO 3 ) 2 and Pt (NH 3 ) 4 Cl 2 . For all catalysts the Pt loading was 0.5 wt%. The impregnated materials were dried overnight at RT, next 12 h at 110 � C and then calcined for 6 h at 450 � C. Following the preparation procedure described above and using three Pt precursors, i.e. H 2 PtCl 6 , Pt (NH 3 ) 4 (NO 3 ) 2 and Pt(NH 3 ) 4 Cl 2 , the catalysts were designated as: Pt/ SBA_BEA, Pt/SBA_BEAn and Pt/SBA_BEAc, respectively.

Texture
The pore structure and Brunauer-Emmett-Teller specific surface area (S BET ) of Pt/AlSBA-15 þ BEA were determined by nitrogen adsorptiondesorption at À 196 � C using Autosorb-1C Quantachrome analyzer. Before the adsorption measurements, samples were degassed for 6 h at 150 � C. Further, the samples were filled with nitrogen and analyzed for 13 h. Distribution of pore sizes was calculated according to Barrett-Joyner-Halenda method (BJH).

Characterization of the metallic function
Hydrogen chemisorption was performed using Micromeritics ASAP 2020. Prior to the measurement, the 25 mg of sample was reduced in situ using H 2 flow for 1 h at 450 � C. Next, the sample was cooled to 35 � C under He flow. The measurement was carried out at the pressure range from 10 to 450 Torr. The average metal particle size (d Pt ), metal dispersion (D) and Pt surface area (S Pt ) were calculated at the chemisorption stoichiometry of H:Pt ¼ 1 based on the procedure described by Hunt et al. [43].

Characterization of the acidic function
Acidity of catalysts was determined using pyridine infrared spectroscopy (Py-FTIR). The IR spectra were measured on a Bruker Vector 22 spectrophotometer. Prior to the measurement the catalysts samples Scheme 2. Synthesis of bimodal AlSBA-15þ BEA zeolite supports.
were pressed into self-supporting wafers. Next, the tablets were placed inside the FTIR quartz chamber and degassed under vacuum for 1 h at 400 � C. Then, the samples were subsequently exposed to pyridine vapor at 150 � C. The IR spectra were then recorded at 150, 200, 250, 300 and 350 � C in vacuum for 30 min in the range from 1400 to 1700 cm À 1 . The quantitative calculation of Brønsted and Lewis acid sites was made with respect of IR vibrational bands observed at about 1545 and 1454 cm À 1 , respectively.

FTIR analysis
Prior to the analysis, samples of the Pt catalysts were dried for 24 h at 80 � C. For FTIR analysis, KBr tablets were prepared with each containing 1.5 mg of the catalysts samples and 200 mg of KBr. The spectra was recorded on a Bruker spectrophotometer (FTIR IFS 66/s) in the mid IR range (400-4000 cm À 1 ).

TEM
The micro-mesoporous structure of supports and distribution of Pt particles of the samples was investigated with a Hitachi H-800 microscope, operating at 150 kV. Prior to imaging, the samples were dispersed in methanol and placed on the microscope copper grid covered with a carbon film.

Catalytic experiments
The hydroisomerization of n-hexadecane (n-C 16 ) was carried out in a high-pressure stainless-steel flow reactor with a fixed catalyst bed of approximately 80 mm long. Prior to the catalytic activity test, the catalysts (1.0 g, 0.40-0.63 mm) were activated at 250 � C (1 h), 350 � C (1 h) and 450 � C (3 h) under H 2 pressure of 5 MPa. The activity test was carried out under H 2 pressure of 5 MPa, the H 2 :CH molar ratio of 4.6 mol/mol, the WHSV of 3.5 h À 1 and at the temperature range from 260 to 320 � C. The liquid products of reaction were collected at 3.5 h intervals and analyzed using the gas chromatography (PerkinElmer Clarus 580) with Elite 1 column (60 m � 0.53 mm � 1.5 μm). The total hydroisomerization reaction time on steam was 14 h. Additionally, in Supporting information a comparison results of catalytic experiments for 3 different catalysts were compared: Pt/AlSBA-15 (Pt over pure AlSBA-15 Si/Al ¼ 7), Pt/zeolite BEA (Pt over pure BEA zeolite Si/Al ¼ 75), and Pt/ SBA_BEA(F) (composite material prepared by mechanical mixing of AlSBA-15 and BEA zeolite, AlSBA-15:zeolite ¼ 4:1). The Pt loading was 0.5 wt% and the Pt-precursor was H 2 PtCl 6 .
In the case of catalysts prepared with use of different Pt precursor, hydrocarbons with 14 and 15 carbon atoms (produced in hydrogenolysis of n-C 16 ) were not observed among the reaction products.
In order to verify the reproducibility of the catalytic test, the experiments were repeated 2 times for all Pt catalysts. The results of reproducibility were calculated as a standard deviation and this standard deviation based on n-C 16 conversion was below 1%.

Texture of the AlSBA-15þzeolit material and the Pt/AlSBA-15þzeolit catalyst
The textural properties of Pt catalysts prepared with use of various platinum precursors are collected in Fig. 1 and in Table 1. The occurrence of the highly ordered mesoporous structure of AlSBA-15 in bimodal support was confirmed by a low-angle XRD (Fig. 1A). The presence of intense three diffraction peaks at 2θ around 0.9, 1.6 and 1.8 � corresponded to (1 0 0), (1 1 0) and (2 0 0) planes in the hexagonal structure of SBA-15, respectively. It is also worth noting that in the bimodal supports, the crystalline structure of zeolite was preserved. The wide-range XRD patterns (Fig. 1B) of bimodal catalysts displayed diffraction peaks in the range of 12-14 � and 22.5 � (2θ) which were attributed to crystalline phase of BEA zeolite. For all catalysts the widerange XRD patterns did not show any diffraction peak characteristic for platinum, which may be due to the low Pt loading -below than the detection limit of the XRD technique. Additionally in tested catalysts, the presence of extra crystalline lattices were not observed. It suggests, that BEA zeolite had relatively high crystallinity and the Pt species were highly dispersed on bimodal supports, as expected. Nevertheless, the intensity of reflections typical for the BEA zeolite were much lower in the bimodal materials in comparison with pure zeolite due to: (i) the presence of only 20 wt% of zeolite in the support and/or (ii) the creation of mesopores in zeolite crystals as a result of zeolite digestion by HCl at the stage of synthesis of the bimodal support. Fig. 1C shows the low-temperature N 2 adsorption-desorption isotherms and pore size distributions of Pt catalysts. Regardless of the used Pt precursors, for all catalysts, the isotherms type IV with capillary condensation step around p/p 0 of 0.75 was observed, which is characteristic for mesoporous materials SBA-15 type. Furthermore, for all catalysts supported on AlSBA-15þ BEA zeolite, H1-type hysteresis loop was observed, which presence confirmed highly ordered mesoporous structure of the SBA-15 with double-opened cylindrical-shape pores. As was expected, all of the composite catalysts, showed lower S BET and total pore volume determined at p/p 0 > 0.99 (V T ) in comparison with the powder of AlSBA-15 þ BEA (sample designated as SBA_BEA P , Table 1). It might due to the addition of binder during support preparation (sample designated as SBA_BEA B , Table 1) as well as Pt deposition. Studied catalysts show the narrow pore size distribution with the maximum pore sizes were within the range of 7.4-7.7 nm (Fig. 1D). Among the catalysts, the largest S BET (560 m 2 g -1 ) and V T (0.94 cm 3 g -1 ) was found for Pt/SBA_BEAn prepared by impregnation with Pt(NH 3 ) 4 (NO 3 ) 2 . In the case of the Pt/SBA_BEAc prepared by impregnation with Pt(NH 3 ) 4 Cl 2 , a significant decrease of S BET and V T was observed in comparison with Pt/ SBA_BEA and Pt/SBA_BEAn. This phenomenon was probably caused by the partial blockage of the Pt/SBA_BEAc catalyst pores by large Pt crystals (d Pt about 5.1 nm). Table 1 summarises the results of the H 2 chemisorption measurments (Pt dispersion, D; Pt particle diameter, d Pt ; Pt surface, SPt). Obtained data indicated, that the Pt precursor significantly affect both the dispersion and preferred adsorption of Pt particles on composite support components i.e. AlSBA-15, BEA zeolite or Al 2 O 3, during the impregnation step. The Pt/SBA_BEA catalyst exhibited the highest Pt dispersion (58%) and the smallest particle size (2.0 nm). The Pt dispersion on the surface of catalysts obtained using Pt(NH 3 ) 4 (NO 3 ) 2 and Pt(NH 3 ) 4 Cl 2 were lower and equaled 32 and 23%, respectively. Average particle size located on Pt for Pt/SBA_BEAn and Pt/SBA_BEAc samples were much larger than for Pt/SBA_BEA catalyst and equaled 3.6 and 5.1 nm, respectively.

Characterization of the metal function
The reason for the differences in Pt particle size and their dispersion may be: (i) different values of PZC of support surface and pH of aqueous precursor solutions used for impregnation, i.e. 2.2; 6.0 and 6.5 for H 2 PtCl 6 , Pt(NH 3 ) 4 (NO 3 ) 2 , Pt(NH 3 ) 4 Cl 2 , respectively ( Fig. 2) and (ii) the presence of ammonium and chlorine ions in the calcination step during catalysts preparation. The pH value at which the carrier surface is inert is termed the point of zero charge (PZC). When choosing a metal precursor and a method of metal incorporation to a supports, it is important to know the PZC value and the pH of the metal precursor solution.
During the impregnation of the carrier, the hydroxyl groups located on its surface, become protonated (positively charged), or deprotonated (negatively charged) [45]. In case of impregnation with a solution with a pH value above the PZC value of the carrier, its surface will be polarized negatively and will preferentially adsorb the cations such as [Pt (NH 3 ) 4 ] 2þ . Meanwhile, using the impregnating solution with a pH below the PZC of the carrier, the surface of carrier is positively charged and will adsorb anions such as [PtCl 6 ] 2-. In addition, the important parameter that influences the dispersion and size of metal particles is the difference between the pH value of the impregnating solution and the PZC of the carrier. Higher difference between the pH and PZC values led to the stronger precursor-carrier interactions, and as a result to a higher concentration of the active phase introduced and its better dispersion [46]. The results of Hao et al. [45], Spieker et al. [47] and Samad et al. [25] indicated, that knowing the PZC value of the carrier and using solutions of metal precursors with different pH, preferential metal deposition on a support component can be predicted. In consequence, not only the size of metal particles can be controlled, but also the ratio and proximity of the metal and acid sites.
In the case of catalysts based on a bimodal material consisting of AlSBA-15 (PZCffi 5) [48], BEA zeolite (PZCffi 6) [49] and Al 2 O 3 (PZCffi 8) [46] it was found, that the high difference between the pH value of the H 2 PtCl 6 solution (pH ¼ 2.2) and the pH PZC value of all support components, promoted adsorption of [PtCl 6 ] 2ions on its surface. Meanwhile, the higher pH values of Pt(NH 3 ) 4 (NO 3 ) 2 (pH ¼ 6) and Pt (NH 3 ) 4 Cl 2 (pH ¼ 6.5) solutions and lack or small differences in their pH values, with respect to PZC value of the support components, were the cause of weaker interactions between the [Pt(NH 3 ) 4 ] 2þ ions and the carrier surface. Reduction of the strength of metal -support interactions in contrast led to decrease of metal dispersion and the increase in Pt crystals size (Table 1). Hence, Pt dispersion on the Pt/SBA_BEA catalyst was about 1.8 and 2.5 fold higher, than on the Pt/SBA_BEAn and Pt/SBA_BEAc, respectively. On the other hand, the use of cationic Pt precursors allowed to incorporate platinum into the zeolite and reduction of the proximity of acid and metallic sites. Thus, it appears, that in cases of impregnation of carriers containing zeolite, the usage of precursors such as Pt(NH 3 ) 4 (NO 3 ) 2 and Pt(NH 3 ) 4 Cl 2 was reasonable, because the incorporation of Pt around the acid sites was enabled. However, the incorporation of Pt into the zeolite crystals through ion exchange, disabled determination of Pt dispersion using chemisorption method. This phenomenon was confirmed by Wang et al. [19] and Geng et al. [39], who observed that not all Pt particles located on zeolite crystals are occupied by adsorbed CO or H 2 molecules and may negatively affect the chemisorption resultsdecrease Pt dispersion and increase the size of Pt crystals in comparison with the real values. As was earlier mentioned, the presence of chlorine ions in the Pt precursor  a V T -total pore volume determined at p/p 0 >0.99, b S MES -surface of mesopore from t-plot, c V MES -volume of mesopore from t-plot, d d BJH -pore diameter (BJH method), e D -dispersion, f SPt -Pt surface area, g dPt -average Pt particle diameter, P powdered supports without binder, B supports with binder.
impacts the metal dispersion. As it was described by Kanda et al. [50] and Jaroszewska et al. [51], the interactions between the chlorine residues and Al phase were stronger than the ones between the chlorine residues and Si species. Therefore, the chlorine containing precursors should be preferably adsorbed on the Al-rich surfaces; in the case of investigated here catalysts on Al 2 O 3 binder. Nevertheless our results proved, that in the case of the catalyst impregnated with H 2 PtCl 6 solution, Pt was located mainly over the AlSBA-15 surface (Fig. 4A). We believe, that in can be linked with some structural effect i.e. high surface area of AlSBA-15 and the presence of 80 wt% AlSBA-15 in the catalyst composition. In consequence, it can be expected, that the dispersion of Pt on AlSBA-15 will be high and a good Pt dispersion may provide high catalytic activity. However, the presence of Pt species located on Al 2 O 3 cannot be neglected because these Pt species can be involved in a sufficient delivery of spillover hydrogen onto the acid sites of AlSBA-15 and zeolite components. It is also known the role of chlorine ions in preserving Pt dispersion in reforming catalysts [52]. The chlorine ions react with oxidized Pt species distributed over the Al 2 O 3 support to form mobile platinum oxychloride [Pt IV O x Cl y ] s , which redistributes the Pt over the catalysts support. Chlorine ions are required for re-dispersion of the Pt atoms in spent catalysts but also to help maintain a high Pt dispersion during the process. The decomposition of ammonium ions during catalyst calcination leads to the release of ammonia and subsequently to remaining a proton on the surface (acid form of AlSBA-15 and/or zeolite). Calcination of the chlorine-containing precursor does not lead to its removal from the surface -Cl remains on it, e.g. bound to alumina or in the acid center of zeolite. Each time it causes changes in concentration, strength and distribution of acid centers on the surface (to vary degrees for individual carrier components).

FTIR and TEM measurements
In Fig. 3, the FTIR absorption spectra of pure zeolite and bimodal catalysts are presented. For all samples the FTIR spectrum contains a group of absorption bands with a high intensity in the range from 1750 to 400 cm À 1 . The peaks at around 570 and 445 cm À 1 indicate the presence of five-membered double rings, typical for BEA zeolite and were assigned to the internal flexions of T-O-T (where T symbolizes the atom of Si or Al) siloxane bonds in the rings [53,54]. The peak at 800 cm À 1 , was attributed to symmetric stretching vibrations of the siloxane groups, from internal bonds of the tetrahedral SiO 4 structural unit. The signals at around 1090 and 1200 cm À 1 were assigned to internal asymmetric stretching vibrations of the T-O-T bond and external asymmetric stretching vibrations of T-O-T, respectively [55]. For all samples, the signals at approximately 1500 and 1700 cm À 1 may correspond to the deformational vibrations of water trapped in the pores of materials.
It is also worth noting that in FTIR absorption spectra for both Pt/ SBA_BEAc and Pt/SBA_BEAn, a significant decline in the intensity of the main peaks corresponding to T-O-T vibrations in the zeolite skeleton was observed. It might be a result of the deposition of Pt on zeolite crystals caused by partial ion exchange. Jin et al. [56] and Gülec et al. [54] observed, that during the impregnation with aqueous metal precursor solutions, the partial ion exchange in zeolites occurs.
In order to further investigate the impact of Pt precursor on the platinum particle size and their location on the support, Pt catalysts were examined by means of TEM. The TEM micrographs of the catalysts are given in Fig. 4. On all TEM micrographs of bimodal catalysts, it can be observed, that the size of BEA zeolite crystals varied from 50 to 200 nm. However, it should be noted, that for some crystals their edges were not regular, which indicates the partial change of their structure. This phenomenon might be caused by interaction with the HCl solution during the synthesis of bimodal material -formation of secondary porosity within. Hence, in some of BEA zeolite crystals, the presence of randomly distributed mesopores can be observed. Additionally, TEM micrographs of Pt/SBA_BEAn and Pt/SBA_BEAc (Fig. 4C and F) demonstrated uniform and hexagonally ordered SBA-15-type mesostructure. In Fig. 4 A,C,E, also some areas of amorphous materials derived from the binder (Pural 400) were observed. Also in the case of catalysts obtained by impregnation with an aqueous solutions of Pt (NH 3 ) 4 Cl 2 and Pt(NH 3 ) 4 (NO 3 ) 2 , the Pt particles were mainly located on zeolite crystals (Fig. 4C-F). In contrast to the catalysts obtained by impregnation with solution containing [Pt(NH 3 ) 4 ] 2þ ions, the usage of solution containing [PtCl 6 ] 2ions allowed Pt to be deposited mainly on AlSBA-15 (80% of the mass of the supports with ratio AlSBA-15:zeolite ¼ 4:1), thus better dispersion of Pt particles and smaller particle size was obtained (Fig. 4A). The TEM micrograph of Pt/SBA_BEAn and Pt/ SBA_BEAc exhibited, that the average diameter of Pt particles on these catalysts was larger than that for the Pt/SBA_BEA catalyst. Both, TEM micrographs and H 2 chemisorption, implied that the diameter of Pt particles on catalysts obtained by impregnation with cationic Pt precursor, was larger, what might be a result of autoreduction of Pt by NH 3 ,

Acidity of the catalysts
The Py-FTIR spectra of pyridine desorption at 150 � C of the calcined and reduced Pt catalysts are shown in Fig. 5A. The quantitatively calculated values from the Py-FTIR spectra of pyridine desorption at 150, 250 and 350 � C in the range of 1700-1400 cm À 1 were compared in Table 2 and Fig. 5B and C. For all samples, pyridine absorption bands attributed to both Lewis and Brønsted acid sites were observed. Peaks at wavelength of 1623 and 1454 cm À 1 were assigned to pyridine coordinated to the Lewis acid sites and the bands at 1635 and 1545 cm À 1 were attributed to pyridine bounded to the Brønsted acid sites. The peak at 1490 cm À 1 can be assigned to pyridine associated with both Brønsted and Lewis acid sites [57]. Py-FTIR measurements revealed that for all investigated Pt catalysts, the concentration of Brønsted acid sites was higher than the concentration of the Lewis acid sites (Fig. 5. B and C). On the other hand, for all catalysts the strength of Brønsted acid sites was much lower than the strength of Lewis acid sites. In the case of composite Pt catalysts consisting of AlSBA-15, BEA zeolite and γ-Al 2 O 3 (a binder), the Lewis acid sites might be generated on the surface of each ingredients -Al in extra framework positions (AlO 6 structural unit). The presence of Brønsted acid sites was associated with the acidity of zeolite and AlSBA-15 material -in which Al was covalently bounded with four Si atoms via oxygen bridges (AlO 4 structural unit). The concentration of Brønsted acid sites depends also on the presence of chlorine ions from either decomposition of Pt precursor e.g. H 2 PtCl 6 and/or HCl used in the synthesis AlSBA-15. The results of Wang et al. [19] also showed that the deposition of platinum on the surface of the support may modify its acid function as a result of coating some of the acid sites arranged on its surface with platinum particles -Pt particles covered both Brønsted and Lewis acid sites. On the other hand, the results of Fang et al. [58] revealed, that Pt-supported atoms can form additional Lewis acid sites. Thus, the effect of "reducing the concentration of Lewis acid sites", caused by covering of Lewis acid sites by Pt was reduced.
The total concentration of acidic sites on the surface of catalysts with different Pt precursor decreases in the following order: Pt/SBA_BEAc > Pt/SBA_BEA > Pt/SBA_BEAn. In contrast, the ratio of the Brønsted acid sites concentration to the total Brønsted and Lewis acid sites concentration increased in the order Pt/SBA_BEAc � Pt/SBA_BEAn < Pt/ SBA_BEA ( Table 2, ratio PyHþ PyHþþPyL ). This phenomenon remained in correlation with the increase in Pt dispersion on these catalysts (Table 1). Thus, the concentration of Brønsted acid sites on the surface of the catalysts also depended on the dispersion of the metallic phase. Additionally, the nPt nΣPyÀ IR ratio was calculated to evaluate the balance between metal and acid sites in the catalysts for hydroisomerization of nalkanes. The obtained results were presented in Table 2. We observed significant differences in the values of nPt nΣPyÀ IR ratio among the three samples. The ratio decreased in order: Pt/SBA_BEA ( nPt nΣPyÀ IR ¼ 0.10) > Pt/SBA_BEAn ( nPt nΣPyÀ IR ¼ 0.06) > Pt/SBA_BEAc ( nPt nΣPyÀ IR ¼ 0.03). Pt/ SBA_BEA and Pt/SBA_BEAn were characterized by comparable concentration of total acid sites, thus the differences in ratio value was a result of various Pt dispersion (58% for Pt/SBA_BEA and 35% for Pt/SBA_-BEAn). On the other hand, for Pt/SBA_BEAc low value of ratio was a result of: high concentration of acid sites i.e. 206 μmolPyg À 1 , and low amount of accessible Pt atoms (dispersion of 23%). However, there are many works [10,59] on hydroisomerization of n-alkanes on bifunctional catalysts, where it was proved, that one metal site could balance even few acid sites. This phenomenon allows to explain how at such great differences in nPt nΣPyÀ IR ratio values for examined catalysts it is possible to achieve balance between the concentration of metal and acid sites.

Catalytic test
Hydroisomerization of n-C 16 was selected to evaluate the catalytic performances of bifunctional catalysts studied in this research. The conversions of n-C 16 over Pt/SBA_BEA, Pt/SBA_BEAn and Pt/SBA_BEAc are illustrated in Fig. 6 as a function of temperature. The distribution of the reaction products as a function of temperature and conversion is plotted in Figs. 6 and 7. It was observed, that activity and selectivity to C 16 isomers of studied catalysts depended not only on the reaction temperature but also on the Pt precursor used for catalyst preparation. For all catalysts the activities increased with rise of the temperature from 260 to 320 � C, providing n-C 16 conversion in the range of 8-97%. Irrespective on reaction temperature, the activity of the catalysts decreases in following Pt/SBA_BEA � Pt/SBA_BEAn > Pt/SBA_BEAc, what suggests that conversion of n-C 16 was not dependent only on the Pt dispersion (Table 1), but also on the concentration of acid sites and nPt nΣPyÀ IR ratio ( Table 2 and Fig. 5B and C). In the case of Pt/SBA_BEA catalyst the higher dispersion of Pt particles resulted in the creation of a large number of active metallic sites ( nPt nΣPyÀ IR ¼ 0.10), which consequently led to the formation of a large amount of intermediates as a result of dehydrogenation of the n-alkane. On the other hand, the useage of a Pt/SBA_BEAn catalyst, where Pt particles were deposited on zeolite crystals (in spite of lower dispersion of 32%), ensured smaller distance between acid and metal sites. This phenomenon, might increase the number of intermediates and improve the transfer of intermediates that took place between the two active sites. Among the examined catalysts, Pt/SBA_BEAc ( nPt nΣPyÀ IR ¼ 0.03) was less active, when compared to both others and required higher reaction temperature to ensure comparable conversion level (Fig. 6a). It appeared, that in the case of Pt/SBA_BEAc, the deposition of Pt particles on BEA zeolite crystals (Fig. 4E) did not increase its activity, which may be due too low Pt dispersion (D ¼ 23%), large Pt particle size (d Pt ¼ 5.6 nm) and low nPt nΣPyÀ IR ratio value. For all investigated catalysts, the n-C 16 conversion led to the  formation of hydroisomerization products and at higher reaction temperatures also hydrocracking products (Figs. 6 and 7). It was observed, that with the increase of reaction temperature, the yield of isomers increased and passed through the maximum at conversion of n-C 16 around 80% and then decrease. For Pt/SBA_BEA and Pt/SBA_BEAn maximum yield of i-C 16 equaled 59 and 68 wt%, respectively (Figs. 6c and 7b). While, the maximum yield of isomers on Pt/SBA_BEAc was 51 wt% with 82% of total n-C 16 conversion at 20 � C higher temperature compared to the other two catalysts (Figs. 6c and 7b). In Fig. 7a, the relationship between the selectivity of i-C 16 and conversion of n-hexadecane was presented. The selectivity of i-C 16 over the three catalysts decreased with the increase of conversion. Among of three Pt catalysts, at n-C 16 conversion <50%, Pt/SBA_BEA and Pt \SBA_BEAc showed the highest selectivity (S HI > 90%). This phenomenon was caused by the highest Pt dispersion i.e. 58% (Table 1) and nPt nΣPyÀ IR ratio ( Table 2) for Pt/SBA_BEA. The higher the number of  hydrogenating sites, the greater the chance for hydrogenation of isomerization products and limitation of the further cracking of the carbenium ions on the acid sites. We suppose that the high selectivity of Pt/ SBA_BEAc results from not only the favourable distribution of the acidic sites on the catalyst surface, but also from the decreased distance between metallic and acid sites. However, it was observed, that in the case of both Pt/SBA_BEA and Pt/SBA_BEAc the selectivity decreased significantly with the increase of n-C 16 conversion. Whereas for Pt/SBA_BEAn the isomerization selectivity was lower and remained constant i.e. 85% (Rys. 6b and 7a) in wider conversion range. It might be concluded, that for Pt/SBA_BEAn the better balance between metal and acid functions was obtained ( nPt nPyHþ ¼0.12). For all catalytic systems, the main i-C 16 products were MoBC 16 and DiC 16 , which are important components of diesel fuel, due to the high cetane number and good low temperature properties (Fig. 6d and e) (Fig. 6a-d), which may be a result of short distance between the metallic and acidic sites and consequently fast hydrogenation of isomerization products.
For all tested catalysts, at the conversion higher than 80%, the yield of multibranched isomers of hexadecane (MuBC 16 ) increased. Among multibranched isomers, the main fraction were those with two methyl groups (DiBC 16 ). The molar ratio of monobranched to multibranched hexadecane isomers (MoB/MuB) as a function of conversion of n-C 16 is shown in Fig. 7c. In all cases, the MoB/MuB ratio decreased with the increase of n-C 16 conversion, which corresponds to the n-alkanes monomolecular isomerization mechanism via the protonated cyclopropane (PCP) [60]. At the same conversion (ca. 80%), the MoB/MuB ratio decreased in the sequence Pt/SBA_BEAn > Pt/SBA_BEA � Pt/SBA_BEAc, which might be a result of different distance between metallic and acidic sites (Fig. 7c).
In the case of the tested catalysts, the isomerization was proceeded through branching from n-C 16  It is worth to mention, that in the whole temperature range of the reaction, on the all catalysts, the yield of isomers was much higher than yield of cracking products (Figs. 6 and 7). The yields of cracked products as a function of conversion are given in Fig. 7b. The values of yields C 3 -C 13 fraction on Pt/SBA_BEAn (Fig. 7e) and Pt/SBA_BEAc (Fig. 7f) catalysts were lower than on the Pt/SBA_BEA (Fig. 7d) and increased from 0.3 to 38.2 wt%. The yield of C 3 -C 13 hydrocarbons on the Pt/ SBA_BEA, varied between 0.8 and 48.8 wt%, what may be due to the greater distance between the metal and acid sites on the catalyst surface. The C 3 -C 13 fraction produced on the studied catalysts consisted mainly of i-alkanes (see i/n ratio in Fig. 7d-e). The values of i/n ratio for fraction C 3 -C 13 varied from 2.2 to 3.9 which corresponded to β-scissions of alkylcarbenium ions in agreement with degradation reaction type B and/or type C [18].
The distribution of the cracking products, according to the carbon numbers at conversion of ca. 80% is given in Fig. 7d-f. On the catalysts prepared via impregnation with solutions containing [Pt(NH 3 ) 4 ] 2þ ions, the distribution of cracked products was almost symmetrical with the maximum positioned at C 8 (in Fig. 7e-f). The symmetrical distribution of C 3 -C 13 fraction was observed for both catalysts and indicated the lack or very small extent of secondary cracking. However, it should be noted, that on the Pt/SBA_BEAc catalyst, the yield of cracking products was about 2-fold higher than on the Pt/SBA_BEAn, what may be due to lower activity of this catalyst ( nPt nΣPyÀ IR ¼ 0.03). Pt/SBA_BEAc catalyst achieved a comparable level of conversion as Pt/SBA_BEAn did at higher reaction temperature, thus the probability of hydrocarbon cracking increased. Meanwhile, the distribution pattern obtained on the Pt/SBA_BEA catalyst (in Fig. 7d) was slightly shifted towards hydrocarbons containing C 10 -C 12 of carbon numbers. In the case of [Pt(NH 3 ) 4 ] 2þ ions derived catalysts, lack of hydrogenolysis products, symmetrical distribution of C 3 -C 13 hydrocarbons and calculated HK ind values close to 2 (ratio of the moles of cracking products to the moles of cracked n-C 16 ) excluded the formation of hydrocarbons during secondary cracking.
To summarize, the activity of the studied catalysts is a function of many factors (type of precursor, size and location of Pt crystallites, texture of the support, type, concentration and strength of acid sites on its surface and distance between metallic and acidic sites, pH of the impregnation solution and surface PZC). The application of the Pt/ SBA_BEAn catalyst in n-C 16 hydroisomerization, provided the highest yield of isomers and simultaneously, the lowest yield of cracking products. The best catalytic properties of Pt/SBA_BEAn was a result of moderate acidity, sufficient Pt distribution and privileged location of Pt species on zeolite crystals in AlSBA-15 þ BEA material. Thus, the distance between metal and acid sites was reduced.

Conclusions
The Pt catalysts prepared by impregnation of bimodal supports using H 2 PtCl 6 , Pt(NH 3 ) 4 Cl 2 and Pt(NH 3 ) 4 (NO 3 ) 2 . The TEM, H 2 chemisorption, FTIR and Py-FTIR results indicated, that the Pt distribution (dispersion, localization of Pt particles and their size) were affected by the used metal precursor. Usage of H 2 PtCl 6 enabled the deposition of Pt particles mainly on the AlSBA-15 and Al 2 O 3 surface. For catalysts obtained from cationic Pt precursors i.e. Pt(NH 3 ) 4 Cl 2 and Pt(NH 3 ) 4 (NO 3 ) 2 , metal particles were distributed mainly over the zeolite crystals.
The activity and selectivity to hydroisomerization products of studied catalysts was affected by the distribution of Pt sites, which directly depended on: (i) PZC value of the support, (ii) location of Pt in cationic or anionic group, (iii) presence of chlorine or ammonia in precursor of Pt.
In the case of the composite support, in which each components shows different PZC, by choosing the Pt precursor (Pt in cationic or anionic group) and knowing its solution pH, it is possible to selectively incorporate the Pt particles over the surface of one of the components of the support.
The catalyst prepared with anionic precursor (H 2 PtCl 6 ) was characterized by high activity in hydroisomerization of n-C 16 , due to smaller Pt particle size and higher dispersion in comparison with the catalysts prepared with cationic Pt precursors. It appeared, that the presence of chlorine in platinum precursor, prevented Pt autoreduction during the calcination. While the usage of the precursors containing NH 3 or NH 3 and Cl simultaneously caused agglomeration of Pt sites on the catalysts surface. Despite the high catalytic activity of Pt/SBA_BEA, the maximum i-C 16 yield (including desirable MoBC 16 ) obtained using this catalyst was lower than yield of i-C 16 achieved on the Pt/SBA_BEAn catalyst. In addition, the amount of C 3 -C 13 hydrocarbons formed using Pt/SBA_BEA catalyst was significantly higher than for the catalysts obtained by impregnation with solutions containing [Pt(NH 3 ) 4 ] 2þ ions. This phenomenon resulted from the greater distance between metallic and acidic sites. It appeared, that the monobranched alkenes were formed on acidic sites, while diffusion into metallic sites, encountered several acid sites on their way. As a result, they were transformed into multibranched alkenes and hydrocarbons containing from 3 to 13 carbon atoms.
It was observed, that the catalytic activity, yield and distribution of hydroisomerization products, strongly depended on the Pt precursor used for catalyst preparation. The catalyst prepared with use Pt (NH 3 ) 4 (NO 3 ) 2 , enabled Pt deposition ovn the BEA zeolite, reducing the distance between the metallic and acidic sites. As a result, n-hexadecane hydroconversion activity and isomerization selectivity were improved.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.