Effect of Phosphine Doping and the Surface Metal State of Ni on the Catalytic Performance of Ni/Al₂O₃ Catalyst

Abstract

Ni-based catalysts as replacement for noble metal catalysts are of particular interest in the catalytic conversion of biomass due to their cheap and satisfactory catalytic activity. The Ni/SiO₂ catalyst has been studied for the hydrogenolysis of glycerol, and doping with phosphorus (P) found to improve the catalytic performance significantly because of the formation of Ni₂P alloys. However, in the present work we disclose a different catalytic phenomenon for the P-doped Ni/Al₂O₃ catalyst. We found that doping with P has a significant effect on the state of the active Ni species, and thus improves the selectivity to 1,2-propanediol (1,2-PDO) significantly in the hydrogenolysis of glycerol, although Ni-P alloys were not observed in our catalytic system. The structure and selectivity correlations were determined from the experimental data, combining the results of X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), hydrogen temperature-programmed reduction (H₂-TPR) and ammonia temperature-programmed desorption (NH₃-TPD). The presence of NiO species, formed from P-doped Ni/Al₂O₃ catalyst, was shown to benefit the formation of 1,2-PDO. This was supported by the results of the Ni/Al₂O₃ catalyst containing NiO species with incomplete reduction. Furthermore, the role the NiO species played in the reaction and the potential reaction mechanism over the P-doped Ni/Al₂O₃ catalyst is discussed. The new findings in the present work open a new vision for Ni catalysis and will benefit researchers in designing Ni-based catalysts.

1. Introduction

Over the last decade, biodiesel, a potential fuel, has attracted much attention. However, high cost in production makes it difficult for plants to take a step forward. In order to help overcome the problems, glycerol, a by-product in the manufacture of biodiesel fuel by transesterification (10% in weight), was used as a low-cost feedstock to produce valued fine chemicals [1,2]. The substance 1,2-propanediol (1,2-PDO), a hydrogenolysis product of glycerol, is widely applied in polyester resins, liquid detergents, pharmaceuticals, cosmetics, tobacco humectants, functional fluids (antifreeze, de-icing, and heat transfer) and its market share is expected to expand owing to the growing concern of the toxicity of ethylene glycol-based antifreeze to humans and animals on de-icing. Traditionally, 1,2-PDO is produced by the hydration of propylene oxide derived from propylene by either the chlorohydrin process or the hydroperoxide process, but these processes bring about many environmental problems [3]. As a green process, the hydrogenolysis of glycerol is an industrially relevant route for the production of 1,2-PDO.

Until now, great efforts have been made in the hydrogenolysis of glycerol [4,5,6,7,8,9]. Noble metals such as Ru, Rh, Pt have been much explored for the hydrogenolysis of glycerol. Although the noble metal catalysts are effective for the hydrogenolysis of glycerol at a relatively low temperature, they do not fit the requirements of industrial application due to their lower selectivity and high cost. By contrast, a non-noble metal has an advantages in price. Some catalysts such as Cu-based, Co-based, and Ni-based catalysts have been explored for the hydrogenolysis of glycerol. Ni-based catalysts have had a great deal of attention due to their moderate activity.

Perosa et al., reported their primary research on the hydrogenolysis of glycerol without solvent over Raney nickel catalyst, and they obtained 12% conversion of glycerol and 93% selectivity to 1,2-PDO at 150 °C, 8.0 MPa H₂. The selectivity to 1,2-PDO decreased on enhancing the reaction temperature, reducing to 70%–80% and 48% at 190 °C and 210 °C, respectively [10]. Ryneveld et al., investigated the transformation of glycerol to lower alcohols in a fixed-bed reactor over Ni/Al₂O₃ catalysts, and they obtained 1,2-PDO with a high selectivity of 80% at 230 °C. As the temperature increased to 320 °C, the 1,2-PDO selectivity decreased to 2%, while the selectivity for ethanol and propanol increased to 52% [11]. Yu et al., used KBH₄ treated carbon thermal reduced Ni/AC catalyst to catalyze the hydrogenolysis of glycerol; they obtained 76.1% selectivity to 1,2-PDO in water at 200 °C [12]. Huang et al., reported that the Ni/Al₂O₃ showed high selectivity to CH₄ and CO for vapor phase glycerol hydrogenolysis in a fixed-bed reactor [13]. The selectivity of the desired product of 1,2-PDO is not always satisfactory over the Ni-based catalysts.

Doping a second element to the metal catalysts is considered to be an efficient means as well as being widely used to adjust the catalytic performance. Morales et al., reported that selectivity to 1,2-PDO was enhanced significantly over Ni-Ce/SBA catalyst compared with that of the Ni/SBA due to the increase of surface acidity on the addition of Ce [14]. Huang et al., reported that the addition of phosphorus to Ni/SiO₂ catalyst could increase the selectivity to 1,2-PDO to 85.9% from 49.9% at 220 °C, they suggested that the formation of a Ni₂P alloy phase and its electronic and geometrical properties could affect the hydrogenolysis of C–O and C–C bonds [15]. Montassier et al., found that Ru/C catalyst doped with sulfur could increase the selectivity of 1,2-PDO from 12%–80% because doping of sulfur could prevent the carbonyl hydrogenation and C–C and C–O hydrogenolysis based on electron transfer from Ru to S [16]. Moreover, the effects of promoters like Al, Ba and Zn on the Cu-Cr catalysts in the hydrogenolysis of glycerol was investigated in 2-propanol at 220 °C, in the presence of Ba, the highest selectivity to 1,2-PDO reached 85% at 34% conversion, it was suggested that the increased acid sites and the BaCrO₄ phase might be responsible for the improved activity and selectivity [17].

In this paper, the catalytic performance of Ni/Al₂O₃ and phosphorus doping Ni/Al₂O₃ (Ni-P/Al₂O₃) catalysts were studied and compared for the hydrogenolysis of pure glycerol. The effect of phosphorus doping on the catalytic performance was determined based on the variation of the structure and electronic state of the Ni species. It was confirmed that the doping of phosphorus could induce the formation of NiO species in Ni-P/Al₂O₃ catalysts, which was found to be responsible for the improved selectivity to 1,2-PDO (91%).

2. Results and Discussion

2.1. Influence of P Doping on the Catalytic Performance

The catalytic activity of a series of 17.5% Ni-yP/Al₂O₃-500 catalysts was determined for the hydrogenolysis of glycerol in the absence of solvent. Usually, 1,2-PDO and ethylene glycol (EG) were produced as the main products over the Ni-based catalysts accompanied by minor amounts of byproducts of 1-propanol, ethanol, methanol, and acetol in the hydrogenolysis of glycerol; herein we mainly discuss the factors which affected the selectivity to 1,2-PDO and EG. The influence of P doping on the hydrogenolysis of glycerol is listed in Table 1. It is clear that the activity did not change with the P doping initially but it declined remarkably after the P content was increased up to 4.4%. However, the selectivity to 1,2-PDO increased linearly with the content of P, and it reached 91.1% over the 17.5% Ni-4.4% P/Al₂O₃-500, which was much higher than the value (66.5%) of 17.5% Ni/Al₂O₃-500. Figure 1 shows the changes of the conversion and product selectivity with reaction time over the 17.5% Ni-2.9% P/Al₂O₃-500. The conversion increased linearly with reaction time, indicating that the present hydrogenolysis of glycerol is zero order with respect to glycerol. The selectivity to 1,2-PDO remained around 85%, and changed very slightly with conversion of glycerol up to 95.7%.

Table 1. Effect of P-doping on the catalytic performance in hydrogenolysis of glycerol.

Entry	Catalyst	Conversion (%)	Selectivity (%)
			1,2-PDO a	EG a	Others b
1	17.5% Ni/Al2O3-500	42.4	66.5	21.6	11.9
2	17.5% Ni-1.7% P/Al2O3-500	42.6	77.5	11.9	10.6
3	17.5% Ni-2.9% P/Al2O3-500	42.7	85.7	6.3	8.0
4	17.5% Ni-4.4% P/Al2O3-500	25.8	91.1	3.2	5.7

Reaction conditions: glycerol 38.0 mmoL, catalyst 0.25 g, 230 °C, 5 MPa H₂, reaction time 6 h. ^a: 1,2-PDO: 1,2-propylene glycol, EG: ethylene glycol. ^b: The others include 1-propanol, ethanol, methanol and acetol.

Table 1. Effect of P-doping on the catalytic performance in hydrogenolysis of glycerol.

Entry	Catalyst	Conversion (%)	Selectivity (%)
			1,2-PDO a	EG a	Others b
1	17.5% Ni/Al2O3-500	42.4	66.5	21.6	11.9
2	17.5% Ni-1.7% P/Al2O3-500	42.6	77.5	11.9	10.6
3	17.5% Ni-2.9% P/Al2O3-500	42.7	85.7	6.3	8.0
4	17.5% Ni-4.4% P/Al2O3-500	25.8	91.1	3.2	5.7

Reaction conditions: glycerol 38.0 mmoL, catalyst 0.25 g, 230 °C, 5 MPa H₂, reaction time 6 h. ^a: 1,2-PDO: 1,2-propylene glycol, EG: ethylene glycol. ^b: The others include 1-propanol, ethanol, methanol and acetol.

Figure 1. Glycerol conversion and product selectivity as a function of reaction time over 17.5% Ni-2.9% P/Al₂O₃-500. Reaction conditions: glycerol 38.0 mmoL, catalyst 0.25 g, 230 °C, 5 MPa H₂. (■) conversion of glycerol, selectivity to (○)1,2-Propylene Glycol (1,2-PDO), (Δ)Ethylene Glycol (EG), and (▽) other byproducts include 1-propanol, ethanol, methanol and acetol.

Figure 1. Glycerol conversion and product selectivity as a function of reaction time over 17.5% Ni-2.9% P/Al₂O₃-500. Reaction conditions: glycerol 38.0 mmoL, catalyst 0.25 g, 230 °C, 5 MPa H₂. (■) conversion of glycerol, selectivity to (○)1,2-Propylene Glycol (1,2-PDO), (Δ)Ethylene Glycol (EG), and (▽) other byproducts include 1-propanol, ethanol, methanol and acetol.

It was reported that transition-metal phosphides such as Ni₁₂P₅ and Ni₂P can be formed on high temperature treatment under the reducing atmosphere of hydrogen [18,19]. In addition, Huang et al., reported that the formation of the Ni₂P alloy phase could improve the selectivity to 1,2-PDO in the glycerol hydrogenolysis over P-doped Ni/SiO₂ catalyst [15]. Thus, powder X-ray diffraction measurements of 17.5% Ni-yP/Al₂O₃-500 catalysts were carried out to check the formation of Ni-P alloys. As seen in Figure 2, the diffraction peaks appearing at 37.3° and 67.1° were attributed to γ-Al₂O₃ (JCPDS, 13-0373), and the peaks at 44.5°, 51.9°, and 76.4° were assigned to metallic Ni⁰ (JCPDS, 04-0850). The intensity of the Ni diffraction peaks changed slightly with doping of P, and all the sizes of the Ni particles were about 10.9 nm as calculated from the half-width of Ni (111) by the Scherrer equation, thus indicating that doping and increasing P content affect slightly the dispersion of crystalline Ni particles. However, the diffraction peaks of Ni₂P (at 40.8°, 44.6° and 47.3°, (JCPDS, 65-3544)), and Ni₁₂P₅ ( at 48.9°, 43.9° and 41.7°, (JCPDS, 65-1623)) were not found.

Figure 2. X-ray diffraction patterns of (a) 17.5% Ni/Al₂O₃-500, (b) 17.5% Ni-1.7% P/Al₂O₃-500, (c) 17.5% Ni-2.9% P/Al₂O₃-500, and (d) 17.5% Ni-4.4% P/Al₂O₃-500.

Figure 2. X-ray diffraction patterns of (a) 17.5% Ni/Al₂O₃-500, (b) 17.5% Ni-1.7% P/Al₂O₃-500, (c) 17.5% Ni-2.9% P/Al₂O₃-500, and (d) 17.5% Ni-4.4% P/Al₂O₃-500.

Moreover, the valence state and the chemical environment of Ni and P were examined by XPS, Figure 3I lists the XPS profiles of the Ni 2p3/2 level for the comparative samples of 17.5% Ni/Al₂O₃-500 and 17.5% Ni-4.4% P/Al₂O₃-500. The peaks appeared at binding energy of 850–870 eV, regions could be deconvoluted into three peaks: 852.8 eV corresponds to the metallic Ni [20], 856.3 eV corresponds to nickel oxide/hydroxide [21,22], besides, shake-up satellites can be seen at the higher binding energy of 6.1 eV. The electronic binding energy of Ni 2p_2/3 in 17.5% Ni-4.4% P/Al₂O₃-500 was the same as that in 17.5% Ni/Al₂O₃-500. The XPS spectra of P 2p presented at binding energy of 134.2 eV was assigned to the oxidized P species (P⁵⁺) [23,24]. However, both the reduced phosphides (element P) at 129.5 eV and AlPO₄ at binding energy of 134.7 eV were not detected [19,23,24]. The above XPS results indicated that, in 17.5% Ni-4.4% P/Al₂O₃-500, Ni species existed as Ni⁰ and NiO, and P species existed as PO₄³⁻. Huang and Sawhill reported that Ni-P alloys such as Ni₁₂P₅ and Ni₂P were formed with the addition of phosphorus to Ni/Al₂O₃ at a reduction temperature of 850 °C [15,19]. However, in the present work, Ni-P alloys were not formed due to a lower reduction temperature based on the results of XRD and XPS.

Figure 3. XPS spectra of (I) Ni 2p_3/2 levels for 17.5% Ni/Al₂O₃-500 (a) and 17.5% Ni-4.4% P/Al₂O₃-500 (b); and (II) P 2p levels for 17.5% Ni-4.4% P/Al₂O₃-500.

Figure 3. XPS spectra of (I) Ni 2p_3/2 levels for 17.5% Ni/Al₂O₃-500 (a) and 17.5% Ni-4.4% P/Al₂O₃-500 (b); and (II) P 2p levels for 17.5% Ni-4.4% P/Al₂O₃-500.

Therefore, in order to figure out the exact reasons for the improvement of the selectivity to 1,2-PDO over the 17.5% Ni-yP/Al₂O₃-500 catalysts, H₂-Temperature Programmed Reduction (H₂-TPR) and NH₃-Temperature Programmed Desorption (NH₃-TPD) measurements were studied. It is clear that the reduction peaks changed largely after doping of P ( H₂-TPR profiles in Figure 4I). For all the samples, there are three reduction peaks. With the doping of P, the reduction peak at the lower temperature (329 °C) moved to the higher temperature (390–420 °C), and the peak at 662 °C moved to the lower ones but the peak areas at the high reduction temperature increased greatly. It suggested that three kinds of NiO species were formed and that P doping could influence the interaction between these NiO species with Al₂O₃ support. At the reduction temperature of 580 °C, the NiO species could not be completely reduced to Ni⁰. But no diffraction peaks of NiO were detected in the reduced 17.5% Ni-yP/Al₂O₃-500 catalysts because of its low amount and poor crystallinity. In addition, the doping of P has an effect on the acidity of the catalyst, as seen in Figure 4II. The total acid sites increased with the increase in P content, and a new acid site presented at around 380–400 °C for the catalysts with higher P content.

Figure 4. The H₂-TPR (I) and NH₃-TPD (II) of 17.5% Ni/Al₂O₃-500 (a), 17.5% Ni-1.7% P/Al₂O₃-500 (b), 17.5% Ni-2.9% P/Al₂O₃-500 (c), and 17.5% Ni-4.4% P/Al₂O₃-500 (d).

Figure 4. The H₂-TPR (I) and NH₃-TPD (II) of 17.5% Ni/Al₂O₃-500 (a), 17.5% Ni-1.7% P/Al₂O₃-500 (b), 17.5% Ni-2.9% P/Al₂O₃-500 (c), and 17.5% Ni-4.4% P/Al₂O₃-500 (d).

Based on the above results of the H₂-TPR and NH₃-TPD, doping of P could influence the electronic environment of the Ni species and the surface acidity of Ni/Al₂O₃, which may have been responsible for the improved selectivity to 1,2-PDO over the Ni-P/Al₂O₃ catalysts. Therefore, we designed experiments to discuss the effects of the Ni species on the catalytic properties as follows.

2.2. Influence of Ni Metal State on the Catalytic Performance

As already known, the loading of metal and calcination temperature could change the redox properties and surface acidity of the catalysts. Thus, a series of xNi/Al₂O₃-450 catalysts with different Ni loadings were prepared first of all and their properties were examined, as shown in Figure 5I; the diffraction peaks of Ni became sharper with increasing Ni loading, indicating that the Ni particles increased with increase in Ni content. As calculated from the half-width of Ni (111) by the Scherrer equation, the average particle size of Ni increased from 7.1–16.6 nm when Ni loading increased from 5%–40%. It was found that when Ni loading was 5%, the main reduction peak appeared at 680 °C as shown in Figure 5II. While, on increasing Ni loading, the main reduction peak shifted to lower temperature, and moved to 335 °C at Ni loading of 40%. The change of reduction temperature with different loadings revealed that NiO has a strong interaction with Al₂O₃ [25,26]. The interaction decreased with increasing Ni loading because the NiO crystallites aggregated or grew to large ones, inducing a weak interaction with Al₂O₃ [27,28]. Therefore, the reduction temperature changed to the low one with increasing Ni loading. In addition, NH₃-TPD examination was performed to compare the acidity of the Ni/Al₂O₃ catalyst with different loadings (Figure 5III); there existed a main desorption peak at about 100 °C. The amount of surface acid sites decreased with increasing Ni loading because the acid sites of Al₂O₃ were overlaid by Ni particles. Furthermore, the influence of Ni loading on product selectivity was compared at a similar conversion around 50% as shown in Table 2. The selectivity to 1,2-PDO decreased and the selectivity to EG increased initially with increasing Ni loading, then it changed slightly at Ni loading above 17.5%, which may correlate with the surface properties such as the Ni state and the acidity of the catalysts.

Figure 5. The XRD (I); H₂-TPR(II); and NH₃-TPD (III) of Ni/Al₂O₃-450 with different loadings of Ni. (I) Al₂O₃ (a), and 5% (b), 9.5% (c), 17.5% (d), 29.6% (e) and 40% (f).

Figure 5. The XRD (I); H₂-TPR(II); and NH₃-TPD (III) of Ni/Al₂O₃-450 with different loadings of Ni. (I) Al₂O₃ (a), and 5% (b), 9.5% (c), 17.5% (d), 29.6% (e) and 40% (f).

Table 2. Influence of Ni loading on the hydrolysis of glycerol over Ni/Al₂O₃-450 catalysts.

Ni loading (%)	Time (h)	Conversion (%)	Selectivity (%)
			1,2-PDO	EG	Others
5	12	50.6	84.2	4.8	11.0
9.5	7.3	51.5	75.8	14.3	9.9
17.5	5	52.8	61.4	25.1	13.5
29.6	4	47.2	60.5	26.2	13.3
40	7	45.9	59.1	27.0	13.9

Reaction conditions: glycerol 38.0 mmoL, catalyst 0.25 g, 230 °C, 5 MPa H₂.

Table 2. Influence of Ni loading on the hydrolysis of glycerol over Ni/Al₂O₃-450 catalysts.

Ni loading (%)	Time (h)	Conversion (%)	Selectivity (%)
			1,2-PDO	EG	Others
5	12	50.6	84.2	4.8	11.0
9.5	7.3	51.5	75.8	14.3	9.9
17.5	5	52.8	61.4	25.1	13.5
29.6	4	47.2	60.5	26.2	13.3
40	7	45.9	59.1	27.0	13.9

Reaction conditions: glycerol 38.0 mmoL, catalyst 0.25 g, 230 °C, 5 MPa H₂.

Following this, the influence of the calcinating temperature was determined based on the results of TPR, XRD, and the NH₃-TPD results of the 17.5% Ni/Al₂O₃-T catalysts. From the H₂-TPR results, three reduction peaks were found for all the samples checked. With increasing calcination temperature from 450–600 °C, the reduction peaks moved to higher temperatures (Figure 6I), which was ascribed to the increasing interaction of the NiO crystal with the support during higher calcination [29,30]. Therefore, the complete reduction of NiO needs a much higher temperature. It is in agreement with the results of XRD patterns in Figure 6II, in which the NiO species was found in the 17.5% Ni/Al₂O₃-T catalysts calcined at 600 °C. Furthermore, when the Ni/Al₂O₃ was partially reduced via a shorter reduction time (17.5% Ni/Al₂O₃-450-PR), the diffraction peaks of NiO at 43.1° appeared clearly. The reduction degree decreased with increasing calcination temperature due to the stronger interaction between NiO and Al₂O₃ support. An increase in the amount of NiO species on the surface of the catalyst contributed to the high selectivity to 1,2-PDO in the hydrogenolysis of glycerol. As seen in Table 3, on increasing the calcination temperature from 450–600 °C, it is obvious that the conversion decreased gradually from 59.8%–27.1%, but the selectivity to 1,2-PDO increased from 61.4%–78.6% (Entries 1,3,4), and a similar result was also obtained for the 17.5% Ni/Al₂O₃-450-PR catalyst; with this the conversion decreased to 23.4% but the selectivity to 1,2-PDO increased to 78.0% (Entry 2). Thus, we can conclude that the existence of NiO could enhance the selectivity to 1,2-PDO, but decrease the reaction conversion. In addition, the acidity of the catalysts calcined at different temperatures was also examined. As shown in Figure 6III the amount of surface acid sites decreased with increasing calcination temperature, which may be attributed to the reduction of the surface hydroxyl density and specific surface. It is worth noting that the surface acid sites decreased but the selectivity to 1,2-PDO increased, which is different from the 17.5% Ni-yP/Al₂O₃-500 catalysts, in which the acidity increased with doping of P, and the selectivity to 1,2-PDO increased as well.

Table 3. Effect of calcination temperature on the catalytic performance of 17.5% Ni/Al₂O₃-T catalysts.

Entry	Calcination temperature (°C)	Time (h)	Conversion (%)	Selectivity (%)
				1,2-PDO	EG	Others
1	450	6	59.8	61.4	25.1	13.5
2 a	450	6	23.4	78.0	9.3	12.7
3	500	6	42.4	66.5	21.6	11.9
4	600	6	27.1	78.6	9.4	12.0

Reaction conditions: glycerol 38.0 mmoL; catalyst 0.25 g, 230 °C, 5 MPa H₂. ^a 17.5% Ni/Al₂O₃-450-PR.

Table 3. Effect of calcination temperature on the catalytic performance of 17.5% Ni/Al₂O₃-T catalysts.

Entry	Calcination temperature (°C)	Time (h)	Conversion (%)	Selectivity (%)
				1,2-PDO	EG	Others
1	450	6	59.8	61.4	25.1	13.5
2 a	450	6	23.4	78.0	9.3	12.7
3	500	6	42.4	66.5	21.6	11.9
4	600	6	27.1	78.6	9.4	12.0

Reaction conditions: glycerol 38.0 mmoL; catalyst 0.25 g, 230 °C, 5 MPa H₂. ^a 17.5% Ni/Al₂O₃-450-PR.

Figure 6. (I) H₂-TPR; (II) XRD and (III) NH₃-TPD of fresh reduced 17.5% Ni/Al₂O₃-T with different calcination temperatures: 450 °C (a), 500 °C (b), 600 °C (c), and calcinated at 450 °C but reduced from room temperature to 580 °C without maintaining at 580 °C (17.5% Ni/Al₂O₃-450-PR) (d).

Figure 6. (I) H₂-TPR; (II) XRD and (III) NH₃-TPD of fresh reduced 17.5% Ni/Al₂O₃-T with different calcination temperatures: 450 °C (a), 500 °C (b), 600 °C (c), and calcinated at 450 °C but reduced from room temperature to 580 °C without maintaining at 580 °C (17.5% Ni/Al₂O₃-450-PR) (d).

The effects of acidity have been widely discussed for the hydrogenolysis of glycerol. It was suggested that the surface acid site could benefit from the formation of an intermediate acetol, which is then converted to 1,2-PDO [5,17,31,32]. In other words, the selectivity to 1,2-PDO may depend on the amount of surface acid sites. Herein, with decreasing of Ni loading and the doping of P, the amount of surface acid sites increased, resulting in an enhancement of the selectivity to 1,2-PDO. However, with rising calcination temperature, the surface acid sites decreased but the selectivity to 1,2-PDO still increased. Based on the above results, we propose that both the acidity and NiO species have an influence on the production of 1,2-PDO; the amount of surface acid sites is not the decisive factor, but the NiO species is the main factor for the enhancement of selectivity to 1,2-PDO and the decrease in the reaction conversion in the hydrogenolysis of glycerol.

When the reaction was conducted in the absence of hydrogen over Ni/Al₂O₃-500 and Ni-1.7%P/Al₂O₃-500 catalysts even though acetol was detected, both the reactions gave very low conversion <1% under 230 °C for 10 h, thus direct dehydration of glycerol to acetol is very difficult with the present catalysis system. However, when acetol was used as the starting reactant, it was converted completely to 1,2-PDO within 20 min at same reaction conditions. So we deduced that the hydrogenolysis of glycerol in the present catalysis system follows the Langmuir-Hinshelwood-Hougen-Watson (LHHW) mechanism [33,34,35] in which the adjacent two hydroxyls of glycerol adsorbed on the NiO surface, react with atomic hydrogen on the Ni surface, and the C-O bond cleavage and hydrogen atom addition occur simultaneously to produce 1,2-PDO, as shown in Scheme 1. For the formation of EG, glycerol first of all dehydrogenates to glyceraldehydes, then decarbonylates through retro-aldolization to produce EG [16]. The NiO on the Ni-base catalysts reduced the active sites for dehydrogenation which was the first step in producing EG, so the selectivity to EG decreased.

Scheme 1. Illustration of the hydrogenolysis of glycerol to 1,2-propanediol (1,2-PDO) over Ni-P/γ-Al₂O₃.

Scheme 1. Illustration of the hydrogenolysis of glycerol to 1,2-propanediol (1,2-PDO) over Ni-P/γ-Al₂O₃.

The NiO benefits for the adsorption of glycerol, and activates the hydroxyl groups, eventually accelerating the hydrogenolysis and lowering the rate of dehydrogenation. The surface acid sites are also a benefit for the hydrogenolysis of glycerol to 1,2-PDO. Thus the selectivity to 1,2-PDO increased when the NiO species and acid sites were at higher levels on the surface of the catalysts. Therefore, 17.5% Ni-yP/Al₂O₃-500 catalysts presented a higher selectivity to 1,2-PDO compared with 17.5% Ni/Al₂O₃-500 because doping of P could induce an increase in the amount of surface NiO and acid sites.

For the activity and selectivity of 17.5% Ni-yP/Al₂O₃-500 catalysts in the glycerol hydrogenolysis, two factors need to be considered. The first one is the active site for the adsorption of hydroxyl of glycerol, which is related to the NiO and acid sites, and the second one is the active site of Ni⁰ species for cleavage of hydrogen. With doping of P and raising of the calcination temperature or shortening reduction time (partially reduction), the amount of the catalytic active sites of Ni⁰ decreased and NiO increased, resulting in a decrease in reaction rate and an increase in the selectivity to 1,2-PDO.

The stability of 17.5% Ni-2.9% P/Al₂O₃-500 was compared with 17.5% Ni/Al₂O₃-500, the results are shown in Table 4. Due to the catalyst being susceptible to air, the stability was measured by increasing the ratio of substrate to catalyst and extending the reaction time simultaneously. It is obvious that the reaction rate decreased slightly for both catalysts, it decreased to 0.79 from 0.99 h⁻¹ for 17.5% Ni/Al₂O₃-500, and decreased to 0.73 from 1.00 h⁻¹ for 17.5% Ni-2.9% P/Al₂O₃-500. However, the selectivity to the main product 1,2-PDO did not change, it was 66% for 17.5% Ni /Al₂O₃-500 , and >85% for 17.5% Ni-2.9% P/Al₂O₃-500. Ni leaching was not detected in the filtrate based on the Inductively Coupled Plasma Atomic Emission Spectrometry (ICP-AES) analysis with a detection limit of 0.1 ppm. The loss in activity may be caused by the surface coverage and the pore blockage with carbonaceous deposition on the surface of the catalyst.

Table 4. The stability of the catalysts.

Entry	Catalysts	Time (h)	Conv. (%)	Selectivity (%)			Rate (h−1) a
				1,2-PDO	EG	Others
1 b	17.5% Ni/ Al2O3-500	6	42.4	66.5	21.6	11.9	0.99
2 c	17.5% Ni/ Al2O3-500	18	33.9	66.1	20.8	13.1	0.79
3 b	17.5% Ni-2.9% P/Al2O3-500	6	42.7	85.7	6.3	8.0	1.00
4 c	17.5% Ni-2.9% P/Al2O3-500	18	31.2	89.7	5.7	4.6	0.73

Reaction conditions: 38.0 mmoL glycerol, 230 °C, 5 MPa H₂. ^a The rate was calculated with converted gram of glycerol divided by gram of catalyst used and reaction time. ^b 0.25g catalyst, ^c 0.083 g catalyst.

Table 4. The stability of the catalysts.

Entry	Catalysts	Time (h)	Conv. (%)	Selectivity (%)			Rate (h−1) a
				1,2-PDO	EG	Others
1 b	17.5% Ni/ Al2O3-500	6	42.4	66.5	21.6	11.9	0.99
2 c	17.5% Ni/ Al2O3-500	18	33.9	66.1	20.8	13.1	0.79
3 b	17.5% Ni-2.9% P/Al2O3-500	6	42.7	85.7	6.3	8.0	1.00
4 c	17.5% Ni-2.9% P/Al2O3-500	18	31.2	89.7	5.7	4.6	0.73

Reaction conditions: 38.0 mmoL glycerol, 230 °C, 5 MPa H₂. ^a The rate was calculated with converted gram of glycerol divided by gram of catalyst used and reaction time. ^b 0.25g catalyst, ^c 0.083 g catalyst.

The hydrogenolysis of glycerol with a supported Ni catalyst has been reported in several literature works and the results are compared in Table 5. Ni₂P/SiO₂ was reported to be the most efficient catalyst with 95.1% conversion and 85.9% selectivity to 1,2-PDO; the formation of Ni₂P alloy phase and its electronic and geometrical properties should have an effect on the hydrogenolysis of C–O and C–C bonds [15]. Herein, P doping Ni/Al₂O₃ exhibited a comparable conversion and selectivity to 1,2-PDO. However, Ni₂P alloy phase was not formed, but the existence of amorphous and/or crystal NiO was responsible for the high selectivity to 1,2-PDO at a comparable conversion.

Table 5. Comparison of the Ni-based catalysts in the hydrogenolysis of glycerol.

Catalyst	Reaction condition	Conversion (%)	S1,2-PDO (%)	Ref.
Raney Ni	190 °C, 1 MPa, 44 h	97	71	[10]
Ni/AC-CB	200 °C, 5 MPa, 24 h	63.2	77.4	[12]
7.5%Ce-Ni/SBA	200 °C, 2.4 MPa, 8 h	51	29	[14]
Ni2P/SiO2	220 °C, 3 MPa, 1.13 h	95.1	85.9	[15]
17.5% Ni-2.9% P/Al2O3	230 °C, 3 MPa, 15 h	95.7	84.7	This work

Table 5. Comparison of the Ni-based catalysts in the hydrogenolysis of glycerol.

Catalyst	Reaction condition	Conversion (%)	S1,2-PDO (%)	Ref.
Raney Ni	190 °C, 1 MPa, 44 h	97	71	[10]
Ni/AC-CB	200 °C, 5 MPa, 24 h	63.2	77.4	[12]
7.5%Ce-Ni/SBA	200 °C, 2.4 MPa, 8 h	51	29	[14]
Ni2P/SiO2	220 °C, 3 MPa, 1.13 h	95.1	85.9	[15]
17.5% Ni-2.9% P/Al2O3	230 °C, 3 MPa, 15 h	95.7	84.7	This work

3. Experimental Section

3.1. Materials

γ-Alumina (Sinopharm Chemical Reagent Co., Shanghai, China) was used as the support without special treatment. Ni(NO₃)₂·6H₂O (Beijing Yili Fine Chemicals Co., Beijing, China, 98%), NH₄·H₂PO₄ (Beijing Chemicals, Beijing, China, 98%), glycerol (Xilong Chemicals, Guangzhou, China, 99%) were used directly without any treatment. Deionized water and the gases of H₂ (99.999%), N₂ (99.999%) and NH₃ (99.999%) were used as delivered.

3.2. Catalysts Preparation

A series of Ni/Al₂O₃ catalysts with different Ni loadings of 5%, 9.5%, 17.5%, 29.6% and 40% were prepared by the incipient wetness impregnation method. A solution of Ni(NO₃)₂·6H₂O was added dropwise to a γ-Al₂O₃ slurry, and then the mixture was dried at 80 °C overnight, followed by calcination in air at different temperatures (450 °C, 500 °C and 600 °C ) for 5 h. The samples were labeled as xNi/Al₂O₃-T, where x denotes the loading of Ni and T denotes the calcined temperature.

In addition, the P-doped Ni/ Al₂O₃ catalysts with a Ni loading of 17.5% were prepared with the method as described above, except for dropwise addition of NH₄·H₂PO₄ and Ni(NO₃)₂·6H₂O together in the first impregnating step. After calcination (500 °C, 5 h), the samples were stored in a vessel and labeled as 17.5% Ni-yP/Al₂O₃-500, y denotes the loading of phospine. To eliminate the diffusion effect, the catalysts were ground and screen separated by 180 (<80 μm) and 200 mesh (<75 μm). The results show that the reaction rate kept at a constant value with a particle size smaller than 180 mesh, indicating that the inner transfer resistance was removed with a catalyst smaller than 180 mesh in size. As a result the catalysts used in the reactions were ground and separated by 180 mesh.

3.3. Characterization

XRD patterns of all involved catalysts were performed on Bruker D8 ADVANCE diffractometer, using Cu-Kα radiation (λ = 0.154 nm). Prior to the test, samples were pre-reduced at 580 °C for 2 h. Crystallite sizes of Ni particles were determined from the most intense peaks at 2θ = 44.5° (h k l = 1 1 1) using the Scherrer equation, i.e., Dc = K λ/β cosθ, where the constant K here is 0.9, λ is the wavelength of the X-ray radiation, β is the width of the peak at half-maximum, and θ is the Bragg angle.

H₂-TPR and NH₃-TPD were carried on TP-5080 (Tianjin Xianquan Industry and Trade Development Co. Ltd, Tianjin, China), with a thermal conductivity detector. Before a H₂-TPR run, 30 mg of the catalyst was loaded in the quartz tube and pretreated with nitrogen at 150 °C for 10 min. After cooling to room temperature, the sample cell was heated to 800 °C with a ramping rate of 10 °C/min in flow gases of H₂/N₂ with a flow rate of 30 mL/min. For the NH₃-TPD test, 100 mg of the sample was pre-reduced in pure H₂ at 580 °C for 1 h. After the catalyst had cooled to room temperature, the absorption was conducted by flushing the sample with NH₃ for 30 min, and then sweeping with N₂ for 1 h to remove physisorbed and/or weakly bound species. Desorption of NH₃ was carried out at 800 °C with a heating rate of 10 °C/min in a nitrogen flow.

For XPS measurements, a VG Microtech 3000 Multilab was used. Prior to measurement, samples were pre-reduced at 580 °C for 2 h. The electronic properties of Ni and P on the surface of the catalysts were determined. All XPS spectra were corrected to the C 1s peak at 284.6 eV.

3.4. Hydrogenolysis of Glycerol

The hydrogenolysis of glycerol was performed in a 50 mL stainless steel reactor with an inner Teflon tube in solvent-free conditions. The calcinated catalysts were pre-reduced under a H₂ flow at 580 °C for 2 h, and kept without exposing to air before use. The reduced catalysts were added to an autoclave containing 38.0 mmoL pure glycerol and pre-heated at 80 °C for about 5 min. The autoclave was then flushed with H₂ more than three times and pressured to 5.0 MPa of H₂. After the reactor had been preheated to 230 °C, the reaction time was counted by starting stirring with an agitation speed of 1200 rpm, under these conditions the mass and heat transfer limitations were verified to be negligible. After reaction, the reactor was cooled to room temperature, and the gas phase was collected by a gas-bag and analyzed using a gas chromatograph (Shimadzu, Kyoto, Japan, GC-14C) equipped with an active carbon column and a thermal conductivity detector (TCD). The liquid phase was diluted with water and centrifuged to remove the catalyst and analyzed using a gas chromatograph (Shimadzu, Kyoto, Japan, GC-14C) equipped with a capillary column (Restek Stabilwax, Bellefonte, PA, USA, 30 m × 0.53 mm × 1 μm, carrier: N₂) and a flame ionization detector (FID). The carbon balance of the present reaction was about 90% as calculated, in which the gas products were not included as it was difficult to detect the exact content at the high concentration of H₂. For the specific system, the conversion of glycerol and the selectivity of each liquid product were calculated based on the following Equations (1) [5].

$$\begin{array}{c}\text{Coversion }\left(\text{\%}\right)=\frac{\text{Sum of C}-\text{based moles of all liquid products}}{\text{Sum of C}-\text{based moles of reactant and all products}}\times 100\text{\%}\\ \text{Selectivity }\left(\text{\%}\right)=\frac{\text{ C}-\text{based moles of each liquid products}}{\text{Sum of C}-\text{based moles of all liquid products}}\times 100\text{\%}\end{array}$$

(1)

4. Conclusions

This work explored the hydrogenolysis of pure glycerol over P-doping Ni/Al₂O₃. It was found that doping of P to Ni/Al₂O₃ catalyst decreased the reaction rate slightly but increased the selectivity to 1,2-PDO greatly. It was confirmed that the doping of P lowered the reduction degree of Ni/Al₂O₃, and left some NiO species on the reduced Ni-P/Al₂O₃ which improved selectivity to 1,2-PDO. The role of NiO was proven by designed experiments including Ni/Al₂O₃ catalysts with NiO species, obtained by shortening reduction time and enhancing calcination temperature. Thus, a possible function of NiO was proposed, in that the hydroxyl of glycerol favored adsorption onto the NiO species and then dehydroxylation and hydrogenation occurred to produce 1,2-PDO. Moreover, doping of P induced an increase in the amount of surface acid sites, resulting in an enhancement of the selectivity to 1,2-PDO. The results described here provide an important reference on the rational design of Ni-based catalysts for glycerol hydrogenolysis.