Hydrodeoxygenation–Isomerization of Methyl Palmitate over SAPO-11-Supported Ni-Phosphide Catalysts

: Ni-phosphide catalysts on SAPO-11 were studied in the hydrodeoxygenation–isomerization of methyl palmitate (C 15 H 31 COOCH 3 —MP). The catalysts were synthesized using temperature-programmed reduction (TPR) of a phosphate precursor ((NH 4 ) 2 HPO 4 and Ni(CH 3 CH 2 COO) 2 ), TPR of a phosphite precursor (H 3 PO 3 and Ni(OH) 2 ), and using phosphidation of Ni/SAPO-11 by PPh 3 in the liquid phase. The samples were characterized by ICP-AES chemical analysis, N 2 physisorption, NH 3 -TPD, XRD, and TEM. First, the screening of the catalysts prepared by the TPR method was carried out in a semi-batch autoclave to determine the inﬂuence of the preparation method and conditions on one-pot HDO–isomerization (290–380 ◦ C, 2–3 MPa). The precursor’s nature and the amount of phosphorus strongly inﬂuenced the activity of the catalysts and their surface area and acidity. Isomerization occurred only at a low P content (Ni/P = 2/1) and blocking of the SAPO-11 channels by unreduced phosphates at higher P contents did not allow us to obtain iso-alkanes. Experiments with liquid phosphidation samples in a continuous-ﬂow reactor also showed the strong dependence of activity on phosphidation duration as well as on Ni content. The highest yield of isomerized products (66% iso-C 15–16 hydrocarbons, at complete conversion of O-containing compounds, 340 ◦ C, 2 MPa, and LHSV = 5.3 h − 1 ) was obtained over 7% Ni 2 P/SAPO-11 prepared by the liquid phosphidation method.


Introduction
Renewable sources have attracted much attention in the production of fuels and chemical products. Hydrodeoxygenation (HDO) of fatty-acid-based feedstocks (vegetable oils, animal fats, tall oils) leads to the formation of normal alkanes [1]. HDO is carried out over Ni(Co)Mo/Al 2 O 3 catalysts, resulting in products with a high cetane index but poor low-temperature properties (high cloud and pour points) [2][3][4]. Isomerization of these products is needed to meet the requirements of commercial fuels. Noble metal catalysts on silicoaluminophosphates (SAPOs) or zeolites are used in these processes [5]. Overall, the process is two-step [6], and there is an attractive approach to combine HDO and isomerization into one step over bifunctional catalysts [7].
The aim of this work is to study the influence of the preparation method on the catalytic activity of SAPO-11-supported Ni-phosphide catalysts in HDO-isomerization of methyl palmitate (C15H31COOCH3) (MP). Our study involved the screening of the catalysts using an autoclave reactor to verify the activity of TPR catalysts and determine the conditions for optimal HDO-isomerization, the preparation and testing of liquid phosphidation catalysts in a continuous-flow reactor and a comparison with TPR catalysts, and physicochemical analysis of the most promising catalysts.

TPR Catalyst Screening and Search for Optimal Conditions in an Autoclave
The catalysts were prepared by incipient wetness impregnation of the support with two different precursors (phosphate and phosphite) with subsequent TPR. The samples were labeled NiP_A (phosphate) and NiP_I (phosphite).
To avoid blocking the SAPO-11 pores, we tested 3 wt.% NiP_I/SAPO-11 (TPR at 450 °C). To verify whether SAPO-11 is a reliable support, we tested 3 wt.% Ni/SAPO-11. Conversion values of XMP = 75% and XO = 58% were reached over Ni/SAPO-11 after 5 h of SAPO-11 is a microporous material (Table S1) and micropores as well as mesopores can be blocked by excess P. For example, it was shown that Ni/zeolite catalysts have a larger surface area than Ni(H 2 PO 4 ) 2 /zeolite catalysts [51]. Indeed, S BET significantly decreased after phosphide preparation (29 m 2 /g, Table S1) and after reaction (8 m 2 /g, Table  S1). The initial Ni/P ratio was 1/2. During TPR at high temperatures, P can form volatile P-containing compounds (PH 3 and products of its interaction with H 2 O) [52,53]. However, excess P can also remain on the support surface after TPR in the form of unreduced phosphates (PO x ) (H n PO 4 (3 − n)− , P 2 O 7 4− , and (PO 3− ) n ) [36,54,55]. To avoid blocking the SAPO-11 pores, we tested 3 wt.% NiP_I/SAPO-11 (TPR at 450 • C). To verify whether SAPO-11 is a reliable support, we tested 3 wt.% Ni/SAPO-11. Conversion values of X MP = 75% and X O = 58% were reached over Ni/SAPO-11 after 5 h of reaction ( Figure 2). The NiP_I/SAPO-11 remained inactive. XRD analysis showed the presence of metallic Ni in Ni/SAPO-11 and, due to the small particle sizes, NiP_I/SAPO-11 did not have any reflexes of Ni-phosphide phases ( Figure S1). This experiment showed that excess P blocks the active component in the support. As it is impossible to decrease the P content in NiP_I samples without additives (due to the poor solubility of phosphites), we used phosphate NiP_A precursors to prepare samples with different Ni/P ratios. reaction ( Figure 2). The NiP_I/SAPO-11 remained inactive. XRD analysis showed the presence of metallic Ni in Ni/SAPO-11 and, due to the small particle sizes, NiP_I/SAPO-11 did not have any reflexes of Ni-phosphide phases ( Figure S1). This experiment showed that excess P blocks the active component in the support. As it is impossible to decrease the P content in NiP_I samples without additives (due to the poor solubility of phosphites), we used phosphate NiP_A precursors to prepare samples with different Ni/P ratios.  Depending on the Ni/P ratio, the conversion values of MP were in the range of 45-49% and the conversion values of O-containing compounds were 33-45% over the NiP_A/SAPO-11 catalyst after 5 h of reaction ( Figure 3). not have any reflexes of Ni-phosphide phases ( Figure S1). This experiment showed that excess P blocks the active component in the support. As it is impossible to decrease the P content in NiP_I samples without additives (due to the poor solubility of phosphites), we used phosphate NiP_A precursors to prepare samples with different Ni/P ratios.  Depending on the Ni/P ratio, the conversion values of MP were in the range of 45-49% and the conversion values of O-containing compounds were 33-45% over the NiP_A/SAPO-11 catalyst after 5 h of reaction ( Figure 3). The Ni/P ratio also influenced the selectivity ( Table 1). The main products of the reaction were n-alkanes. The C 16 /C 15 molar ratio for NiP_A 1/2 was 0.365, for NiP_A 1/1 it was 0.168, and for NiP_A 2/1 it was 0.137. These numbers are close to the results in the literature for Ni 2 P/SAPO-11 [40][41][42]. Figure 4 shows product distributions vs. time. MP, n-pentadecane (n-C 15 ), and n-hexadecane (n-C 16  changed as follows: for Ni/P = 1/2, S C15 = 66% and S C16 = 24%; for Ni/P = 1/1, S C15 = 76% and S C16 = 13%; and for Ni/P = 2/1, S C15 = 60% and S C16 = 8% ( Table 1). The overall selectivity and yield to alkanes decreased as the P content decreased, confirming the similar trend and activity dependence for SiO 2 -supported Ni-phosphides [53,56]. As the P content decreased, the selectivity to O-containing intermediates and side products increased from 10% (for Ni/P = 1/2) to 31% (for Ni/P = 2/1).  The next step was to determine the conditions under which isomerization takes place. Screening of the NiP_A samples with different Ni contents and Ni/P ratios was conducted at 340 and 380 • C ( Figure 5). At these temperatures, X MP was 100%, so the selectivities are equal to the yields of the products. Catalytic experiments showed that, for the NiP_A catalyst with Ni/P = 1/2, only n-alkanes were detected both at 340 and 380 • C. Decreasing the Ni content resulted in the formation of cracked products, and no isomerized alkanes were detected (3 wt.% NiP_A, Ni/P = 1/2). Decreasing the P content in the 7 wt.% samples also contributed to the cracking; however, in the case of Ni/P = 1/1, iso-C 15 were detected with a selectivity of 13%. For the series of 3 wt.% NiP_A catalysts with different Ni/P ratios, decreasing the P content yielded a larger number of cracked alkanes and more iso-alkanes (because of the increase in the catalyst's acidity, see Figure S2). Thus, the 3 wt.% NiP_A catalyst with Ni/P = 2/1 has the highest selectivity to iso-alkenes (35%) (33% of iso-C 15 and 2% of iso-C 16 ).
equal to the yields of the products. Catalytic experiments showed that, for the NiP_A cat-alyst with Ni/P = 1/2, only n-alkanes were detected both at 340 and 380 °C. Decreasing the Ni content resulted in the formation of cracked products, and no isomerized alkanes were detected (3 wt.% NiP_A, Ni/P = 1/2). Decreasing the P content in the 7 wt.% samples also contributed to the cracking; however, in the case of Ni/P = 1/1, iso-C15 were detected with a selectivity of 13%. For the series of 3 wt.% NiP_A catalysts with different Ni/P ratios, decreasing the P content yielded a larger number of cracked alkanes and more iso-alkanes (because of the increase in the catalyst's acidity, see Figure S2). Thus, the 3 wt.% NiP_A catalyst with Ni/P = 2/1 has the highest selectivity to iso-alkenes (35%) (33% of iso-C15 and 2% of iso-C16). At this point, 3 wt.% NiP_A/SAPO-11 (Ni/P = 2/1) showed the highest isomerization activity, but the number of cracked products was the highest among all tested catalysts. We carried out HDO-isomerization over the 3 wt.% NiP_A series at a lower temperature (340 °C) and pressure (2 MPa) with a larger catalyst mass (1.7 g). Iso-alkanes were only obtained over the sample with Ni/P = 2/1 (35%, Figure 6). However, for this catalyst, the products were not balanced with the initial amount of MP, and no cracked products were detected. Thus, we think that 15% of the MP transformed into carbon deposits. At this point, 3 wt.% NiP_A/SAPO-11 (Ni/P = 2/1) showed the highest isomerization activity, but the number of cracked products was the highest among all tested catalysts. We carried out HDO-isomerization over the 3 wt.% NiP_A series at a lower temperature (340 • C) and pressure (2 MPa) with a larger catalyst mass (1.7 g). Iso-alkanes were only obtained over the sample with Ni/P = 2/1 (35%, Figure 6). However, for this catalyst, the products were not balanced with the initial amount of MP, and no cracked products were detected. Thus, we think that 15% of the MP transformed into carbon deposits. From the experimental results, we can conclude that, to obtain active catalysts simultaneously for HDO and isomerization, one needs to avoid large amounts of P in order not to block SAPO-11 pores and the active component. However, at the same time, one needs enough P to form the Ni-phosphide and prevent cracking and cocking. The liquid phosphidation method of Ni/SAPO-11 can meet both requirements [49]. The phosphidation degree can be regulated by the temperature and duration of this procedure, and the formation of POx phosphate residues can be avoided [57]. This method requires a contin- From the experimental results, we can conclude that, to obtain active catalysts simultaneously for HDO and isomerization, one needs to avoid large amounts of P in order not to block SAPO-11 pores and the active component. However, at the same time, one needs enough P to form the Ni-phosphide and prevent cracking and cocking. The liquid phosphidation method of Ni/SAPO-11 can meet both requirements [49]. The phosphidation degree can be regulated by the temperature and duration of this procedure, and the formation of PO x phosphate residues can be avoided [57]. This method requires a continuous-flow reactor for both phosphidation and catalytic tests.
The NiP_P/SAPO-11 TPP380 catalyst did not show any isomerization activity (Figure 7), and 2% of the MP transformed into carbon deposits. Under the same conditions, 8% of the MP transformed into carbon deposits over the NiP_A/SAPO-11 catalyst, but the selectivity to iso-alkanes was 27% (25% of iso-C 15 , 2% of iso-C 16 ). Thus, we decided to lower the phosphidation degree by decreasing the phosphidation temperature (250 • C, TPP250) and duration (2 h). This allowed us to obtain 40% of the iso-alkanes (33% of iso-C 15 and 7% of iso-C 16 over 3 wt.% NiP_P, Figure 7), but the selectivity to carbon deposits was 6%. Increasing the Ni content proved to be successful in avoiding the formation of carbon deposits and cracked products (7 wt.% NiP_P and 12 wt.% NiP_P, Figure 7) such that the mass balance was 100%. The optimal content of Ni to obtain the highest number of iso-alkanes (54% of iso-C 15 and 12% of iso-C 16 ) was 7 wt.%. The MP conversion and selectivities of the products vs. time on stream (TOS) are shown in Figure S3. Ni/SAPO-11 was also tested in continuous-flow mode in order to compare it with NiP_P/SAPO-11 (Figure 7). Despite the high selectivity to iso-alkanes (59% total), Ni/SAPO-11 had very high cracking and cocking activity (37% selectivity). Unfortunately, it was not possible to determine exactly the number of cracked products because the solvent (n-dodecane) also cracked and contributed to these products ( Figure S4).
Due to the presence of metallic Ni in Ni/SAPO-11, it has high activity in methanation. There is no CO in gas phase products of Ni/SAPO-11, and high concentrations of CH4, C2H6, and C3H8 are observed compared with Ni2P/SAPO-11 catalysts (Figure 8). C2H6, C3H8, and additional amounts of CH4 are produced by cracking reactions of MP, its intermediate and final products, and the n-dodecane solvent. Ni/SAPO-11 was also tested in continuous-flow mode in order to compare it with NiP_P/SAPO-11 (Figure 7). Despite the high selectivity to iso-alkanes (59% total), Ni/SAPO-11 had very high cracking and cocking activity (37% selectivity). Unfortunately, it was not possible to determine exactly the number of cracked products because the solvent (n-dodecane) also cracked and contributed to these products ( Figure S4).
Due to the presence of metallic Ni in Ni/SAPO-11, it has high activity in methanation. There is no CO in gas phase products of Ni/SAPO-11, and high concentrations of CH 4 , C 2 H 6 , and C 3 H 8 are observed compared with Ni 2 P/SAPO-11 catalysts (Figure 8). C 2 H 6 , C 3 H 8 , and additional amounts of CH 4 are produced by cracking reactions of MP, its intermediate and final products, and the n-dodecane solvent. vent (n-dodecane) also cracked and contributed to these products ( Figure S4).
Due to the presence of metallic Ni in Ni/SAPO-11, it has high activity in methanation. There is no CO in gas phase products of Ni/SAPO-11, and high concentrations of CH4, C2H6, and C3H8 are observed compared with Ni2P/SAPO-11 catalysts (Figure 8). C2H6, C3H8, and additional amounts of CH4 are produced by cracking reactions of MP, its intermediate and final products, and the n-dodecane solvent. The NiP_A and NiP_P catalysts had quite similar amounts of CO and CH4, and the most active 7 wt.% NiP_P catalyst had the lowest concentrations ( Figure 8). Interestingly, there was no C2H6, C3H8, CH3OCH3, and CH3OH over the NiP_P TPP380 sample. Isomerization of n-alkanes (C15H32 and C16H34) occurs through dehydration to alkenes with the subsequent formation of n-alkylcarbenium ions (C15H31 + and C16H33 + ), which form in protonation over acid sites of SAPO-11. An alternative route of n-alkylcarbenium ion formation is protonation of n-alkanes to n-alkylcarbonium ions (C15H33 + and C16H35 + ), even though this route is significantly slower [58]. The subsequent transformation of n-alkylcarbenium ions leads to monobranched alkanes, dibranched alkanes, and cracked products. According to GC-MS analysis, only trace amounts of dibranched alkanes were detected over Ni2P/SAPO-11 catalysts. Almost all iso-alkanes are monobranched ones.

Scheme 1. Methyl palmitate HDO-isomerization.
As phosphidized catalysts (TPP) are more promising and active than TPR samples, we characterized NiP_P catalysts with different Ni contents by physicochemical methods.  Table 2). Increasing the relative Ni content results in a decrease in the relative P content because P is lighter than Ni. Phosphidation increases the amount of P, but its effect was negligible compared with the overall P content in SAPO-11 and changed due to the Ni loading. HDO intermediates as well as palmityl palmitate are detectable at low reaction temperatures (290 • C, Figure 4). At higher temperatures (340 and 380 • C), the HDO reactions are quite fast and the conversion of O-containing compounds is complete (Figures 5-7).

Physicochemical Analysis of the TPP Catalysts
Isomerization of n-alkanes (C 15 H 32 and C 16 H 34 ) occurs through dehydration to alkenes with the subsequent formation of n-alkylcarbenium ions (C 15 H 31 + and C 16 H 33 + ), which form in protonation over acid sites of SAPO-11. An alternative route of n-alkylcarbenium ion formation is protonation of n-alkanes to n-alkylcarbonium ions (C 15 H 33 + and C 16 H 35 + ), even though this route is significantly slower [58]. The subsequent transformation of n-alkylcarbenium ions leads to monobranched alkanes, dibranched alkanes, and cracked products. According to GC-MS analysis, only trace amounts of dibranched alkanes were detected over Ni 2 P/SAPO-11 catalysts. Almost all iso-alkanes are monobranched ones.
As phosphidized catalysts (TPP) are more promising and active than TPR samples, we characterized NiP_P catalysts with different Ni contents by physicochemical methods.

Chemical Analysis, Textural Properties, NH 3 -TPD, and XRD
The initial SAPO-11 contained 24.1 wt.% P ( Table 2). Increasing the relative Ni content results in a decrease in the relative P content because P is lighter than Ni. Phosphidation increases the amount of P, but its effect was negligible compared with the overall P content in SAPO-11 and changed due to the Ni loading. The S BET of the NiP_P samples was significantly lower than the S BET of the original SAPO-11 (Table 2). This may have been due to blocking of the SAPO-11 pores by the active component and excess P. It resulted in the absence of micropores, as V micro = 0 cm 3 /g and S micro = 0 m 2 /g. The pore size distributions of the support and the catalysts are shown in Figure S5.
The acidity of the catalysts was lower than for SAPO-11 (Table 2, Figure 9). In the NH 3 -TPD curve of SAPO-11, there are two main signals with maximums at~180 • C and 300 • C. The first signal corresponds to weak acid sites, and the second signal corresponds to medium-strength acid sites [26,43]. These signals remain in the NiP_P samples, but an additional broad signal appears after phosphidation at~500 • C. This may correspond to the formation of volatile P-containing compounds at high temperatures of NH 3 -TPD. The acidity of the 12% sample is the highest among the NiP_P catalysts (145 µmol/g), and the acidity of the 3 and 7% samples is similar (122 and 110 µmol/g, respectively). Figure 10 shows XRD patterns of NiP_P samples with different Ni contents. All catalysts contain the Ni 2 P phase. Table 1 lists the D XRD of Ni 2 P particles, and it is quite large (30-59 nm). It should be noted that small particles (<3 nm) could not be detected by XRD; therefore, the D XRD is larger than the D TEM .
additional broad signal appears after phosphidation at ~500 °C. This may corr the formation of volatile P-containing compounds at high temperatures of NH3 acidity of the 12% sample is the highest among the NiP_P catalysts (145 μmol/g acidity of the 3 and 7% samples is similar (122 and 110 μmol/g, respectively).  Figure 10 shows XRD patterns of NiP_P samples with different Ni conten alysts contain the Ni2P phase. Table 1 lists the DXRD of Ni2P particles, and it is q (30-59 nm). It should be noted that small particles (<3 nm) could not be detecte therefore, the DXRD is larger than the DTEM.

TEM and SEM Analysis
The initial SAPO-11 support consisted of agglomerates of elongated particles with sizes of ~1 μm ( Figure S6). TEM images of the NiP_P catalysts show that the SAPO-11 particles are evenly covered with Ni2P nanoparticles (Figure 11).

TEM and SEM Analysis
The initial SAPO-11 support consisted of agglomerates of elongated particles with sizes of~1 µm ( Figure S6). TEM images of the NiP_P catalysts show that the SAPO-11 particles are evenly covered with Ni 2 P nanoparticles ( Figure 11).  Figure 11a-c show TEM images and particle size distributions of NiP_P samples with different Ni contents. All the distributions have a maximum at 4.5 nm. As the Ni content increased, the distributions became broader and bimodal for 7% NiP_P and 12% NiP_P. The mean values (D TEM ) of particle sizes are listed in Table 1. The D TEM increased from 6.5 nm to 9.9 nm as the Ni content increased. EDX mapping showed the presence of Niphosphide particles on SAPO-11 (Figure 11d-f). All particles were covered with an oxidized layer, which mainly consisted of Ni 3 (PO) 4 according to the EDX analysis (Figure 11g-i). Interplanar distances were identified for several particles (Figure 11g

SAPO-11 Synthesis
SAPO-11 was synthesized via conventional hydrothermal synthesis (HTS) in an autoclave with a volume of 10 L. Aluminum and phosphorus precursors were mixed together with the subsequent addition of DPA and a silica source. All precursors were mixed until the reaction mixture became homogeneous. The final reaction mixture underwent hydrothermal synthesis at 200 • C for 24 h under stirring during HTS.
The initial molar composition contained Al 2 O 3 /P 2 O 5 /SiO 2 = 1/1/0.1. The amount of molecular template used was in the range of 0.9-1.5 moles to 1 mol of Al 2 O 3 .
After synthesis, the product was collected, washed with distilled water from the mother liquor, dried, and calcined at 650 • C with a heating rate of 3 • C/min to remove the organic template.
NiP_A (phosphate samples). To prepare phosphate precursors, (NH 4 ) 2 HPO 4 was dissolved in distilled water. Then, Ni(CH 3 CH 2 COO) 2 ·4H 2 O was added. A green-yellow precipitate of Ni-phosphate was formed, which was then dissolved by the dropwise addition of concentrated HNO 3 . The supports were impregnated by this solution. The samples were dried overnight at room temperature, then at 110 • C for 3 h. Afterwards, the samples were calcined at 500 • C for 3 h. To form phosphides, the precursors were reduced in a hydrogen flow (250 mL/min) using the following temperature program: heating to 380 • C on a 3 • C/min ramp, then heating to 600 • C on a 1 • C/min ramp. The content of Ni in the reduced catalysts was~3 or~7 wt.%. The content of P was varied using different initial Ni/P molar ratios (2/1, 1/1, and 1/2) in the precursors.
NiP_I (phosphite samples). To prepare phosphite samples, the precursor Ni(OH) 2 was dissolved in a H 3 PO 3 water solution. The initial Ni/P molar ratio was 1/2. The supports were impregnated by this solution. The samples were dried overnight at room temperature, then at 80 • C for 24 h. Then, the precursors were reduced in a hydrogen flow (250 mL/min) using the following temperature program: heating to 600 • C on a 1 • C/min ramp. The content of Ni in the reduced catalysts was~3 or~7 wt.%.
NiP_P (liquid phosphidation samples). To prepare metallic samples for liquid phosphidation, Ni(CH 3 CH 2 COO) 2 ·4H 2 O was dissolved in distilled water. The green solution was used to impregnate the supports. The samples were dried overnight at room temperature, then at 110 • C for 3 h. Then, the precursors were calcined at 400 • C for 3 h. The content of Ni after calcination was~3,~7, or~12 wt.%. The oxide precursors were reduced in situ in a continuous-flow fixed-bed reactor at 400 • C for 2 h (on a 1 • C/min ramp). After cooling, phosphidation was carried out using 2 wt.% PPh 3 in dodecane (LHSV = 5.3 h −1 , P H2 = 0.5 MPa, H 2 /feed = 600 Ncm 3 /cm 3 ) with heating to 250 • C at 1 • C/min and holding at 250 • C for 2 h.
Textural properties were determined by N 2 physisorption on an ASAP 2400 («Micromeritics», Norcross, GA, USA). NH 3 -TPD analysis was carried out using a Chemosorb («Neosib» LTd., Novosibirsk, Russia). A total of 200 mg of sample was preheated for 1 h in He flow (60 mL/min) at 500 • C. The adsorption of NH 3 was carried out at 100 • C for 1 h. Then, to remove physically adsorbed ammonia, the sample was kept at 100 • C for 1 h in He flow (60 mL/min). Desorption was carried out at a heating rate of 10 • C/min in He flow (60 mL/min). Desorbed NH 3 was determined by a thermal conductivity detector.
TEM microphotographs were obtained on a JEM-2010 («JEOL», Tokyo, Japan) transmission electron microscope with an accelerating voltage of 200 kV and a resolution of 0.14 nm.

Catalytic Experiments
Screening of the HDO-isomerization activity of the NiP_A and NiP_I catalysts was carried out in a 300 mL «Autoclave Engineers» (Erie, PA, USA) Bolted Pressure Vessel Closure (Autoclave reactor) in semi-batch mode (H 2 continuous flow). For a typical experiment, the catalyst was reduced ex situ, then transferred under Ar to the reactor and heated to 180 • C in H 2 flow (100 mL/min). Subsequently, 100 mL of 10% MP in n-dodecane was added to the reactor, the pressure was set to 3 MPa, and the temperature was set to 290-380 • C. When the temperature reached the reaction temperature, mixing was started (500 RPM).
NiP_P catalysts were prepared and tested in a continuous-flow fixed-bed reactor (i.d. = 12 mm) with a coaxial thermocouple with a diameter of 3 mm. The precursors were reduced and phosphidized; then, after the catalyst was cooled and washed with n-dodecane, HDO-isomerization of MP (10 wt.% in n-dodecane) was carried out at 250-340 • C, 2 MPa, LHSV = 5.3 h −1 , and H 2 /feed = 600 Ncm 3 /cm 3 .
The conversion of MP was calculated as follows: where C 0 MP is the initial concentration of MP and C MP is the MP concentration in the products. The conversion of O-containing compounds was calculated as follows: where C 0 O is the initial concentration of oxygen and C O is the concentration of oxygen in the products.
The selectivity was calculated as follows: where C i is the concentration of the i-th compound in the products. For cracked products and carbon deposits, selectivity was estimated based on mass balance. The yield was calculated as follows:

Conclusions
The activity in HDO-isomerization of MP over SAPO-11-supported Ni-phosphide catalysts proved to depend on the preparation method and the phosphidation degree of the catalysts. NiP_I/SAPO-11 (the phosphite precursor) did not show activity in both HDO and isomerization. This was attributed to unreduced PO x residues that blocked the surface of the catalyst and the active component. NiP_A/SAPO-11 (the phosphate precursor) activity correlated with the Ni content and Ni/P molar ratio. A low Ni loading (3 wt.%) and low P content (Ni/P = 2/1) resulted in the formation of iso-C 15 and iso-C 16 hydrocarbons. The selectivity to iso-alkanes at 340 • C, 2 MPa, and at full conversion in a semi-batch autoclave was 35%. At the same time, the selectivity to carbon deposits was 15%. In the continuous-flow reactor, the same catalyst produced 27% of the iso-alkanes and 8% of the carbon deposits (at T = 340 • C, 2 MPa, LHSV = 5.3 h −1 , and H 2 /feed = 600 Ncm 3 /cm 3 ). This catalyst has a metallic character that is too strong, which results in a high degree of cracking activity. In situ liquid-phase phosphidation by PPh 3 allowed us to finely tune the phosphidation degree of the Ni 2 P/SAPO-11 catalysts (NiP_P) in order to make them suitable for simultaneous HDO-isomerization. The highest selectivity to iso-alkanes was 66% (54% of iso-C 15 and 12% of iso-C 16 ) for 7 wt.% NiP_P/SAPO-11 phosphidized at 250 • C for 2 h. No carbon deposits formed over this catalyst.
The evidence from this study suggests that liquid-phase phosphidation is an effective way to prepare Ni 2 P/SAPO-11 catalysts for one-pot HDO-isomerization of fatty esters.