High Yield to 1-Propanol from Crude Glycerol Using Two Reaction Steps with Ni Catalysts

The objective of the present work is to achieve high yield to 1-propanol (1-POH) by crude glycerol hydrogenolysis in liquid phase and find an alternative to the use of noble metals by employing Ni catalysts. Two Ni catalysts with different supports, alumina (γ-Al2O3), and a phosphorous-impregnated carbon composite (CS-P) were studied and characterized in order to determine their acid properties and metallic phases. With the Ni/γ-Al2O3 catalyst, which presented small particles of metallic Ni interacting with the acid sites of the support, it was possible to obtain a complete conversion of crude glycerol with high selectivity towards 1,2-propylene glycol (1,2 PG) (87%) at 220 ◦C whereas with the Ni/CS-P catalyst, the presence of AlPOx species and the Ni2P metallic phase supplied acidity to the catalyst, which promoted the C-O bond cleavage reaction of the secondary carbon of 1,2 PG to obtain 1-POH with very high selectivity (71%) at 260 ◦C. It was found that the employment of two consecutive reaction stages (first with Ni/ γ-Al2O3 at 220 ◦C and then with Ni/CS-P at 260 ◦C) allows reaching levels of selectivity and a yield to 1-POH (79%) comparable to noble metal-based catalysts.

With respect to 1,2-PG, it has been widely used as a raw material in cosmetic, pharmaceutical, food, and chemical industries [14]. No less important is 1-POH, a chemical utilized as an additive in the manufacture of printing inks, antifreezes, brake fluids, and cosmetic lotions that can also be employed as a solvent in the manufacture of rubber, lacquer and essential oils [15].
As previously reported by several authors, 1,2-PG is obtained by a first dehydration stage of C-O bond cleavage reaction from glycerol molecule and a subsequent hydrogenation stage, while 1-POH can be obtained through the subsequent hydrogenolysis of 1,2-PG [16][17][18]. However, in the literature, there are more scientific contributions focused on glycerol hydrogenolysis to produce 1,2-PG than to produce 1-POH [19,20], though the latter is also a high value-added product [21].
To obtain glycols or 1-POH, a bifunctional catalyst is required. In this reaction, the surface acid-base properties of the support play an important role. On the other hand, the choice of the metallic phase is also relevant. It has been reported that metallic phases based on noble metals such as Ru [22,23], NaCl [33] or Na 2 SO 4 [34] due to the neutralization of NaOH with inorganic acids, such as HCl or H 2 SO 4 respectively. Matter organic non-glycerol (MONG) such as mono, di-and triglycerides may also be present in crude glycerol [33]. In the case of NaOH, it was reported that the concentration of NaOH can lead to lactic acid formation, which is a degradation product [35]. With respect to the NaCl salt, the presence of Cl-can lead to a catalyst deactivation due to the incorporation of Cl-over the metallic particles [33]. As regards the MONG content, reports indicate that glycerides (mono, di, or triglycerides) make the catalyst surface dirty, block active sites, and even act as coke precursors [33]. In this way, the use of crude glycerol would be an important variable to study a catalytic material for the glycerol hydrogenolysis reaction.
The objectives of the present work are to achieve high yield to 1-POH by crude glycerol hydrogenolysis in liquid phase and to find an alternative to the use of noble metals by employing Ni catalysts. Two Ni catalysts with different supports, alumina, and a carbon composite were studied and characterized in order to determine their acid properties and metallic phases. In this sense, the acidic properties of alumina promote the dehydration of glycerol to acetol (AcCH 2 OH), which is the main intermediate in the hydrogenolysis reaction. Furthermore, if there is any transformation from alumina to boehmite, due to the hydrothermal conditions of this reaction, this gives it more acidity without negatively affecting the reaction. Carbon-based supports have been widely reported in liquid phase reactions at low temperatures, with advantages associated with their hydrophobicity, stability, textural properties, as well as easy recovery of the metal phase. They can also be acidified by different techniques such as functionalization with inorganic acids [36]. In fact, in previous studies, we found that a carbon composite treated with HF generates the presence of phenolic and lactonic groups on its surface, favoring the formation of 1-POH [37].
For each catalyst, the operative reaction conditions were optimized to reach total conversion and the higher selectivity to the main product.
Taking into consideration that 1,2 PG is the intermediate compound to produce 1-POH, the possibility of employing both catalysts was analyzed, using two consecutive reaction stages. Since it was not possible to observe phosphorus phases on the CS-P support, this composite was analyzed by NMR spectroscopy. Figure 2 shows the 31 P-NMR spectra for the CS-P support that presents a band at −27 ppm, which is associated with the presence of several orthophosphate species, generally denoted as AlPOx [38,39]. In the CS-P composite, a peak at 2θ = 43.7 • can be observed that is assigned to the hexagonal phase of graphite carbon ( ) (JCPDS . The presence of graphitic carbon is due to the pyrolysis in reducing atmosphere of the phenol-formaldehyde liquid resin employed for the synthesis of the CS support (more details in Supplementary Materials). An amorphous plateau between 2θ = 15 • and 30 • that is characteristic of amorphous silica is also observed, and a peak at 2θ = 21.8 • that is assigned to the tridymite phase of silica ( ) (JCPDS . The presence of silica is due to the use of TEOS in the synthesis of the CS support. Besides, the peaks at 2θ = 35.7 • , 41.4 • and 60.0 • can be assigned to silicon carbide ( ) (JCPDS . The formation of silicon carbide is explained due to a reaction between carbon and silica during the calcination process at 1580 • C.

Characterization of Supports and Catalysts
Since it was not possible to observe phosphorus phases on the CS-P support, this composite was analyzed by NMR spectroscopy. Figure 2 shows the 31 P-NMR spectra for the CS-P support that presents a band at −27 ppm, which is associated with the presence of several orthophosphate species, generally denoted as AlPOx [38,39]. Since it was not possible to observe phosphorus phases on the CS-P support, this composite was analyzed by NMR spectroscopy. Figure 2 shows the 31 P-NMR spectra for the CS-P support that presents a band at −27 ppm, which is associated with the presence of several orthophosphate species, generally denoted as AlPOx [38,39].  Table 1 shows the textural characterization of the supports by the adsorption-desorption of N2 and the surface acid properties determined by the decomposition reaction of isopropanol (IPA) and potentiometric titration.  Table 1 shows the textural characterization of the supports by the adsorption-desorption of N 2 and the surface acid properties determined by the decomposition reaction of isopropanol (IPA) and potentiometric titration. a Specific surface area (m 2 g −1 ) b Total pore volume (cm 3 g −1 ) c Initial potential (mV) d Total number of acid sites (mmol n-butylamine g −1 ) e Temperature ( • C) f Selectivity to propylene (%) g Selectivity to acetone (%), h Selectivity to di-isopropyl ether (%).
The adsorption-desorption isotherms of N 2 for γ-Al 2 O 3 and CS-P were of type IV, characteristic of mesoporous materials with hysteresis loops H1 and H3, respectively. Isotherms and pore size distributions are shown in Supplementary Figures S1 and S2.
Although the supports have a similar porosity around 200 m 2 g −1 , they present greater differences in their acidic surface properties. Although carbonaceous supports generally have specific surface area Catalysts 2020, 10, 615 5 of 15 (S BET ) values greater than 600 m 2 g −1 , the lower S BET in the support impregnated with Al(H 2 PO 4 ) 3 (CS-P) can be ascribed to a partial blocking of pores. Araujo et al. also reported that the impregnation with phosphoric acid noticeably reduced the mesoporous volume of the support [40].
The potentiometric titration technique with n-butylamine allowed determining the strength of the acid sites and the total number of acid sites present in a given solid. While the initial potential of the titration curve (Ei) indicates the maximum strength of the acid sites, the consumption of n-butylamine is related to the total number of acid sites (NS). The γ-Al 2 O 3 and CS-P supports present Ei values of 60 and 340 mV respectively. This indicates that the superficial acid sites are strong and very strong for γ-Al 2 O 3 and CS-P respectively [37]. In Table 1, it can be seen that the CS-P support has a higher total number of acid sites (NS) than the γ-Al 2 O 3 support.
The decomposition reaction of isopropanol (IPA) is known as an indirect method that allows characterizing the acid strength and type of sites of the surface of a solid. These are classified according to their ability to dehydrate and form propylene, di-isopropyl ether (DIPE) and/or acetone, or to dehydrogenize and form acetone and hydrogen.
As observed in Table 1, to achieve an IPA conversion level of 15%, the CS-P support requires a temperature of 170 • C, while γ-Al 2 O 3 requires 200 • C. The lower the temperature to reach a certain level of conversion, the greater the number of surface active sites [41].
On the other hand, the selectivity analysis towards the different reaction products allows determining the nature of the surface active sites. As observed in Table 1, the γ-Al 2 O 3 support has a selectivity of 73% towards propylene and 27% towards DIPE, which would indicate that this material has strong Lewis acid sites. The CS-P support exhibits 100% selectivity towards propylene, indicating the presence of strong Lewis and/or Brønsted acid sites [41]. Because NMR spectroscopy shows AlPOx species, the Lewis acidity could be assigned to cations Al + 3 and the Brønsted acidity to the P-OH groups, as reported by Moffat et al. [42].
As regards the Ni catalysts, Table 2 shows the results of the textural and physicochemical characterization. --a Ni content (wt.%) b Specific surface area (m 2 g −1 ) c Total pore volume (cm 3 g −1 ) d Initial potential (mV) e Total number of acid sites (mmol n-butylamine g −1 ) f Crystallite size (nm) g Average particle diameter (nm) h Temperature of 1st, 2nd and 3rd peak in TPR analysis ( • C).
By the AAS technique, it was determined that in both catalysts, the Ni content was very close to the nominal value (5 wt.%).
The results of the adsorption-desorption of N 2 show a drop in the value of S BET for the reduced Ni/γ-Al 2 O 3 catalyst, which is due to the blocking of the pores due to the presence of Ni. Conversely, for the Ni/CS-P catalyst, there is an increase in S BET, which could be assigned to the loss of some AlPO x species of the support, during the thermal treatment in H 2 flow, that were blocking the support pores. It has been reported in the literature that the thermal treatment of phosphoric acid species in carbon is employed to increase the specific area [43]. Isotherms and pore size distributions of the catalysts are shown in Supplementary Figures S1 and S2.
The results of the potentiometric titration (Table 2) indicate the preservation in the catalysts of the strong acid sites of the supports. Thus, the Ni/CS-P catalyst has a greater number of acid sites than Ni/γ-Al 2 O 3 and a higher acid strength of the sites. Figure 3 shows the X-ray diffractograms of the reduced catalysts. For the Ni/γ-Al 2 O 3 catalyst, a peak at 51.7 • is observed which corresponds to the Ni metal phase ( ) (JCPDS 4-850). Due to the low intensity of this peak, it was not possible to calculate an average crystallite size (using the Scherrer equation) but this would indicate that these particles are widely dispersed.
In the X-ray diffractogram of the Ni/CS-P catalyst, the peaks at 2θ° = 40.6°, 44.5°, 47.1° and 54.1° that correspond to the Ni2P phase (•) can be observed (JCPDS 74-1385). Several publications have reported that in Ni and P catalysts, the Ni/P molar ratio would determine the type of alloy formed. For Ni/P ratios greater than 1.4, the formation of the Ni12P5 and Ni3P phases has been indicated, while for Ni/P ratios less than 1, the formation of the Ni2P phase has been reported [44,45]. Since the Ni/CS-P catalyst has a Ni/P molar ratio equal to 0.36, it would be in agreement with the formation of the Ni2P phase determined by XRD. The average crystallite size of Ni2P is about 11 nm, obtained by the Scherrer equation, using the peak at 40.6 ° corresponding to plane (111) of Ni2P. Since the moles of P with respect to Ni are greater, in the Ni/CS-P catalyst all Ni is alloyed as Ni2P and P unalloyed as AlPOx species, even though it was not possible to detect these latter species by XRD. While AlPOx species were detected in the support by NMR, it was not possible to apply this technique in the catalyst due to the magnetic properties of Ni. Figure 4 shows the TEM images and the particle size distribution for the reduced catalysts. For the Ni/γ-Al2O3 catalyst, the histogram shows a narrow distribution with particle sizes between 3 and 5 nm, while the Ni/CS-P catalyst has a wider distribution, with particles from 3 to 21 nm. Table 2 shows the average particle diameter values, where Ni/γ-Al2O3 catalyst has an average diameter of 4 In the X-ray diffractogram of the Ni/CS-P catalyst, the peaks at 2θ • = 40.6 • , 44.5 • , 47.1 • and 54.1 • that correspond to the Ni 2 P phase (•) can be observed (JCPDS 74-1385). Several publications have reported that in Ni and P catalysts, the Ni/P molar ratio would determine the type of alloy formed. For Ni/P ratios greater than 1.4, the formation of the Ni 12 P 5 and Ni 3 P phases has been indicated, while for Ni/P ratios less than 1, the formation of the Ni 2 P phase has been reported [44,45]. Since the Ni/CS-P catalyst has a Ni/P molar ratio equal to 0.36, it would be in agreement with the formation of the Ni 2 P phase determined by XRD. The average crystallite size of Ni 2 P is about 11 nm, obtained by the Scherrer equation, using the peak at 40.6 • corresponding to plane (111) of Ni 2 P.
Since the moles of P with respect to Ni are greater, in the Ni/CS-P catalyst all Ni is alloyed as Ni 2 P and P unalloyed as AlPOx species, even though it was not possible to detect these latter species by XRD. While AlPOx species were detected in the support by NMR, it was not possible to apply this technique in the catalyst due to the magnetic properties of Ni. Figure 4 shows the TEM images and the particle size distribution for the reduced catalysts. For the Ni/γ-Al 2 O 3 catalyst, the histogram shows a narrow distribution with particle sizes between 3 and 5 nm, while the Ni/CS-P catalyst has a wider distribution, with particles from 3 to 21 nm. Table 2 shows the average particle diameter values, where Ni/γ-Al 2 O 3 catalyst has an average diameter of 4 nm. In contrast, the Ni/CS-P catalyst has an average diameter of 12 nm, which is a value very similar to the crystallite size determined by XRD.
The TPR results are summarized in Table 2. As observed, the Ni/γ-Al 2 O 3 catalyst has three characteristic peaks. The first of them at 322 • C indicates the presence of low interaction NiO with the support. The second peak, at 532 • C, would indicate the reduction of Ni +2 ions incorporated in the octahedral vacancies of γ-Al 2 O 3 . Finally, the third peak at 625 • C indicates the reduction of Ni +2 with high interaction with the support [46,47]. TPR peaks assigned to nickel oxides that are not completely integrated in the spinel structure, but have a certain degree of interaction with the support, were observed at temperatures between 500 and 600 • C [48].
As we have indicated, Ni/CS-P was directly activated in H 2 flow (50 cm 3 min −1 ) at 400 • C during 90 min, so the TPR analysis would allow to verifying the reduction of the metal phase. Indeed, no peak was observed, indicating that the activation method was effective to reduce all the Ni precursor.
Catalysts 2020, 10, x FOR PEER REVIEW 7 of 16 nm. In contrast, the Ni/CS-P catalyst has an average diameter of 12 nm, which is a value very similar to the crystallite size determined by XRD. The TPR results are summarized in Table 2. As observed, the Ni/γ-Al2O3 catalyst has three characteristic peaks. The first of them at 322 °C indicates the presence of low interaction NiO with the support. The second peak, at 532 °C, would indicate the reduction of Ni +2 ions incorporated in the octahedral vacancies of γ-Al2O3. Finally, the third peak at 625°C indicates the reduction of Ni +2 with high interaction with the support [46,47]. TPR peaks assigned to nickel oxides that are not completely integrated in the spinel structure, but have a certain degree of interaction with the support, were observed at temperatures between 500 and 600 °C [48].
As we have indicated, Ni/CS-P was directly activated in H2 flow (50 cm 3 min −1 ) at 400 °C during 90 min, so the TPR analysis would allow to verifying the reduction of the metal phase. Indeed, no peak was observed, indicating that the activation method was effective to reduce all the Ni precursor.

Catalytic Activity
The catalytic activity tests were carried out using both technical grade glycerol (99 % Cicarelli) and crude glycerol (Oleomud). The composition of this crude glycerol (~79 wt.%) is shown in Table 3. This raw material contains MeOH and MONG from the biodiesel manufacturing process. The MONG content includes MeOH, mono-, di-, triglycerides, etc. The ash content is mainly due to the presence of Na + salts produced during the neutralization of the NaOH catalyst with acid. Catalytic activity results in the glycerol hydrogenolysis reaction are shown in Table 4.  The reaction products in liquid phase that were identified and quantified were the following: from C1: methanol (MeOH), from C2: ethanol (EtOH), ethylene glycol (EG), from C3: acetone (AcO), 1-propanol (1-POH), acetol (AcCH 2 OH), and 1,2-propylene glycol (1,2-PG). The main gaseous products identified were CO 2 and CH 4 .
The results of Tests 1-6 were obtained by employing the Ni/γ-Al 2 O 3 catalyst. It can be seen that with this catalyst it was not possible to obtain 1-POH as the main product. However, the results in Table 4 indicate that, when the temperature is below 260 • C, the main product of glycerol hydrogenolysis is 1,2-PG (Tests 1,2, and 4-6).
At 260 • C (Test 3), it can be observed that there is a strong increase in the products of C1 and C2 (Gases (CH 4 and CO 2 ), MeOH, EtOH and EG), which shows the greater contribution of the C-C bond cleavage reactions. The Ni/γ-Al 2 O 3 catalyst has widely dispersed Ni 0 particles with average sizes of 4 nm, and this would promote C-C bond cleavage reactions.
The selectivity towards 1,2-PG reaches its maximum value (77.6%) at the temperature of 220 • C (Test 2). Similarly, Marinoiu et al. employed a Ni/SiO 2 -Al 2 O 3 catalyst in the range 170-250 • C at 1.5 MPa of H 2 and found that the yield to 1,2-PG reaches a maximum at 200 • C, due to a good balance between conversion and selectivity [49]. In this sense, Menchavez et al. employed a Ni/CeO 2 -MgO catalyst and reported that 1,2-PG selectivity was optimal in the range 200-215 • C at 6 MPa of H 2 , due to a balance between dehydration and hydrogenation reactions [50].
Furthermore, Long et al. used a Ni catalyst supported on a sepiolite modified with WO 3 and reported that at temperatures higher than 200 • C the selectivity to EG increases due to C-C bond cleavage reactions [51].
From Table 4, the results at 220 • C can be compared with different concentrations of glycerol (Tests 2 and 4), where it is observed that the level of conversion and selectivity improves with the lower content of water in the reaction mixture. Dasari et al. reported that with a Cu-Cr 2 O 4 catalyst, the glycerol conversion increases from 33% to 69%, if the water content decreases from 80 wt.% to 0 wt.% (at 200 • C, 1.4 MPa of H 2 and 24 h of reaction) [16]. Menchavez et al. employed a Ni/CeO 2 -MgO catalyst and reported an increase in glycerol conversion from 43% to 80% when the water content declined from 40 wt.% to 0 wt.% (at 215 • C, 6.9 MPa of H 2 and 24 h of reaction) [50].
In addition, in Table 4 it can be seen that using a glycerol solution (99% Cicarelli) at 80 wt.% and 220 • C (Test 5), it is possible to obtain total glycerol conversion with a 1,2-PG selectivity of 83%, after 5 h of reaction, which implies a yield towards 1,2-PG of 83%. This yield is superior to the ones reported in the literature for Ni-Raney (68.8%) and Ni/NaX (69.6%) catalysts [52,53].
The results of the hydrogenolysis of crude glycerol (79 wt.% Oleomud) employing the Ni/γ-Al 2 O 3 catalyst (Test 6) show that the total conversion of glycerol is also achieved with a selectivity towards 1,2-PG of 87%. The quality of this raw material, which contains impurities such as MeOH and Na + salts, slightly affects the results and this presence could have a positive effect. Feng et al. [54] determined that the addition of Li, Na, and K bases to the reaction medium improves the conversion of glycerol following the order: Li + > Na + > K + . With respect to MeOH, its presence in the reaction medium has been reported to be beneficial due to its H 2 donor capacity [55].
The second part of Table 4 (Tests 7-11) shows the results of the glycerol hydrogenolysis reaction using the Ni/CS-P catalyst. Results obtained at 220 • C with a 10 wt.% glycerol aqueous solution (Test 7) show that Ni/CS-P catalyst is much less active than Ni/γ-Al 2 O 3 catalyst (5% vs. 22% conversion), which could be due to the greater metallic dispersion of Ni/γ-Al 2 O 3 . However, in the tests carried out at higher temperatures and higher glycerol concentrations, it is possible to observe that the Ni/CS-P catalyst does not favor C-C cleavage reactions, and the selectivity to C3 products of glycerol hydrogenolysis is very high. The results of Tests 8-10 show that, with this catalyst, it is possible to obtain 1-POH as the main product of glycerol hydrogenolysis and the main byproduct is AcO. The presence of AcO is explained precisely by the high capacity of C-O bond cleavage of this catalyst.
As reported [16][17][18], 1,2-PG is the intermediate to form 1-POH and both are the main products of glycerol hydrogenolysis. This is corroborated with the results of Test 11 (Table 4) carried out with an 1,2-PG aqueous solution at 80 wt.%, in which a selectivity towards 1-POH greater than 90% is obtained and there is a very small amount of AcO if compared to the glycerol-fed reaction.
The presence of strong acid sites in the Ni/CS-P catalyst would favor the C-O bond cleavage reactions of the secondary carbon, promoting the formation of 1-POH [37]. On the other hand, the results obtained by XRD evidenced the presence of Ni 2 P in the Ni/CS-P catalyst. In the literature, it has been reported that the formation of a NiP alloy provokes an increment in the electrophilicity of Ni sites due to the charge transfer from nickel to phosphorous [56,57]. This property facilitates the adsorption of the O atom of the C-O bond, subsequently promoting the C-O bond cleavage reactions. This tendency increases in the sequence Ni 3 P < Ni 12 P 5 < Ni 2 P, indicating that for Ni 2 P the activity to C-O bond cleavage reactions is the highest among the different nickel phosphides [58]. With respect to the low contents of C1 and C2 reaction products, it was also found that Ni 2 P has a much lower activity for the C-C bond cleavage than Ni 0 [58,59]. Furthermore, the Ni/CS-P catalyst has particles with an average size of 12 nm, which is also a disadvantage for C-C bond cleavage reactions.
To combine the greater catalytic activity of both catalysts, Ni/γ-Al 2 O 3 and Ni/CS-P, to obtain 1,2-PG and 1-POH respectively, we proposed the reaction in two consecutive steps. In this way, we sought to increase the 1-POH yield above 71% which was the maximum that could be obtained with the Ni/CS-P catalyst.
In the first reaction step, the crude glycerol hydrogenolysis reaction was carried out in order to maximize the yield to 1,2-PG, using the Ni/γ-Al 2 O 3 catalyst (Test 6). The second stage was carried out with the reaction products obtained from the first stage this time using the Ni/CS-P catalyst (Test 12). The gaseous products (such as CH 4 , CO 2 ) produced in the first stage are the only reaction products that are lost because it is necessary to open the reactor to proceed with the replacement of the catalyst.
To calculate the selectivity to gaseous products reported in Test 12, the moles of gaseous products from the first reaction step were considered (Test 6).

Synthesis of Supports and Catalysts
Commercial γ-Al 2 O 3 (99.99%, Dytech, Corporation Ltd., Sheffield, UK) and a carbonaceous-based composite were used as supports.
The carbonaceous-based composite was synthetized using the gelling property of TEOS (SILBOND 40-AKZO Chemicals (Buenos Aires, Argentina) in ethanol to include a phenol-formaldehyde liquid resin (RL 43003, ATANOR, Santa Fe, Argentina) in its structure. The Supplementary Materials describe the preparation method used. This material was then modified by impregnation with an Al(H 2 PO 4 ) 3 aqueous solution Sigma-Aldrich (St. Louis, MI, USA). The concentration of P in the solution was calculated so as to obtain 7 wt.% P in the final solid. Finally, the solid was dried at 120 • C for 12 h and calcined at 400 • C for 30 min (heating rate of 10 • C min −1 ). Thus, this support was denoted as CS-P.
The Ni/γ-Al 2 O 3 catalyst was prepared by incipient wetness impregnation employing Ni(NO 3 ) 3 ·6H 2 O (Sigma-Aldrich), while the Ni/CS-P catalyst was prepared using Ni(CH 3 COO) 2 ·4H 2 O in order to decompose this nickel precursor at a lower temperature so as to avoid the gassing of the carbonaceous support.
The concentration of Ni in the solution was calculated so as to obtain 5 wt.% Ni in the final solid. The catalysts were dried at 120 • C during 12 h. Ni/γ-Al 2 O 3 was calcined in stagnant air at 550 • C for 90 min (heating rate of 10 • C min −1 ) and activated in H 2 flow (50 cm 3 min −1 ) at 550 • C for 90 min (heating rate of 10 • C min −1 ). Ni/CS-P was directly activated in H 2 flow (50 cm 3 min −1 ) at 400 • C for 90 min (heating rate of 10 • C min −1 ) so as to prevent support gasification during the calcination process.

Characterization of Supports and Catalysts
The adsorption-desorption measurements were carried out employing N 2 at -196 • C for the textural characterization. A Micromeritics ASAP 2020 equipment (Micromeritics Instrument Corporation, Norcross, GA, USA) was employed for the specific surface measurements and micro and mesoporous characterization. The samples were pretreated under vacuum in two 1-h stages at 100 • C and 300 • C, respectively.
The X-ray diffractograms were recorded on a Philips 1729 powder diffractometer, using CuKα radiation (λ = 1.5418 Å, intensity = 20 mA, and voltage = 40 kV). Spectra were collected in the range 2θ = 10-70 • . The crystallite sizes (d XRD ) in the reduced samples were calculated using the Scherrer equation: where K was taken as 0.90 and B was the full width of the diffraction line at half maximum intensity. The solid-state NMR experiments were performed at room temperature in a 7 T Bruker Avance II-300 spectrometer (Billerica, MA, USA) equipped with a 4-mm MAS probe. The operating frequency for 31 P was 121.5. AlPO 4 was used as an external reference. 31 P 1D spectra were recorded using a π/2 pulse (4.8 µs) and a 30 s delay between two pulses. The spinning rate for the samples was 10 kHz.
The potentiometric titrations were performed employing 0.05 g of sample suspended in acetonitrile (Merck KGaA, Darmstadt, Germany) previously stirred for 3 h. The titrations were suspended with solutions of n-butylamine 0.05 M in acetonitrile, employing a Metrohm 794 Basic potentiometric titrator with a double junction electrode.
The acid-base properties of the supports were determined by the isopropanol (IPA) decomposition test reaction and the potentiometric titration technique. The IPA decomposition test reaction was carried out in a continuous flow reactor at atmospheric pressure between 120 and 350 • C, employing an IPA flow (4.5%) in helium (40 cm 3 min −1 ).
The Ni content in the samples was determined by atomic absorption spectrometry (Spectrophotometer AA-6650 Shimadzu, Tokyo, Japan) employing an IL Model 457 spectrometer with single channel and single beam.
Temperature-programmed reduction experiments (TPR) were carried out using conventional equipment. Samples were heated from room temperature up to 800 • C with a heating rate of 10 • C min −1 in a mixture of H 2 /N 2 (1/9 ratio) with a flow of 50 cm 3 min −1 . The H 2 uptake was calculated using a thermal conductivity detector (TCD) previously calibrated.
Images were obtained by electronic transmission microscopy with a TEM JEOL 100 C instrument (JEOL Ltd., Tokyo, Japan) operating at 200 kV. The samples were suspended in 2-propanol and sonicated in an ultrasonic bath during 10 min prior to analysis. In order to estimate average diameter volume/area (d va ), the particles were considered spherical and the following expression was used for the calculation: where n i is the number of particles with diameter d i . The particle size distribution histograms were obtained from micrographs employing the bright field technique.

Crude Glycerol Characterization
Crude glycerol was supplied by Olemud S.A., a chemical plant situated in Buenos Aires, Argentina. The density, pH, and the contents of glycerol, methanol, water, ashes, and MONG (Matter Organic Non-Glycerol) of the crude glycerol were measured.
The pH of the sample was determined using a 1 g of crude glycerol dissolved in 50 mL of distilled water and was measured with a digital pH-meter (Oakton pH 11 series, Vernon Hills, IL, USA) at room temperature.

Catalytic Actitivy
The activity tests were carried out with a BR-100 (Berghof, Eningen, Germany) high-pressure stainless-steel batch reactor of 100 cm 3 . The magnetic stirring was fixed at 1000 rpm in order to ensure the kinetic control conditions. The Ni/γ-Al 2 O 3 and Ni/CS-P catalysts were reduced ex situ in H 2 flow and then cooled down to room temperature under hydrogen flow and immediately transferred to the reactor containing the glycerol. Finally, the reactor was closed, purged, and pressurized with H 2 (Air Liquide, 99.99%) up to the desired pressure. Afterwards, heating was started (6 • C min −1 ) and, when the reactor was at the desired temperature, the stirring began.
After reaction, the reactor was cooled down to ambient temperature. Gas samples were passed through a pipe with silica-gel so as to dry them and injected into a Shimadzu GC-8A gas chromatograph (Shimadzu Corporation, Tokyo, Japan) equipped with a thermal conductivity detector (TCD) and a Hayesep D 100-120 column operating isothermally at 40 • C. For the identification and quantification of gas samples, standards of CO, CO 2 , and CH 4 were used. Response factors for gas products were obtained using a calibration curve of the standards.
Once the gas samples were analyzed, the reactor was opened under hood, the catalyst was filtered and the liquid obtained was centrifuged and diluted in water (1:40). The liquid samples were injected into a Shimadzu GCMS-QP505A gas chromatograph equipped (Shimadzu Corporation, Tokyo, Japan) with a 50-m 19091S-001 HP PONA capillary column and FID detector. The identification of liquid products was performed by injecting aqueous solutions of the standards at 4 wt.%. The following heating program was used: from 50 • C to 150 • C with a ramp of 10 • C min −1 , then from 150 • C to 220 • C with a ramp of 40 • C min −1 and then isothermally at 220 • C up to the end of the analysis. The quantification was performed employing the integration of the areas of each compound multiplied by a response factor calculated according to the effective carbon number (ECAN) of each compound, as indicated by Scanlon and Willis [60].
The glycerol conversion (X) was determined as follows: Conversion (%) = moles of consumed glycerol moles of initial glycerol ×100% The selectivity to liquid products was defined as: Selectivity to specific product (%) = moles of carbon in specific product 3 × moles of consumed glycerol × 100% The yield was defined as: Yield to specific product (%) = Selectivity to specific product × Conversion 100 (6) The carbon balance for all the tests was in the range between 93 and 97 %.

Conclusions
The aim of this study was to obtain 1-POH by the hydrogenolysis of crude glycerol using two consecutive reaction stages with Ni catalysts, so as to selectively obtain 1,2 PG in the first stage, which is the intermediate compound necessary to obtain 1-POH in the second stage.
With the Ni/γ-Al 2 O 3 catalyst, which presented small particles of metallic Ni interacting with the acid sites of the support, it was possible to obtain a complete conversion of crude glycerol with high selectivity to 1,2 PG (87%) at 220 • C, whereas with the Ni/CS-P catalyst, the presence of AlPOx species and the Ni 2 P metallic phase supplied acidity to the catalyst, which promoted the C-O bond cleavage reaction of the secondary carbon of 1,2-PG to obtain 1-POH with very high selectivity (71%) at 260 • C.
It was found that the employment of two consecutive reaction stages (first with Ni/ γ-Al 2 O 3 at 220 • C and then with Ni/CS-P at 260 • C) allows reaching levels of selectivity and a yield to 1-POH (79%) comparable to noble metal-based catalysts.
Author Contributions: M.N.G. and J.L.C. carried out all the experimental work which was conceived and designed with G.F.S. and F.P.; N.N.N. wrote the paper. All authors have read and approved the final manuscript.

Conflicts of Interest:
The authors declare no conflict of interest.