3.1 The conversion of cellulose into HMF in biphasic system consisting of H2O/THF
Because the aluminium salts such as AlCl3 and Al2(SO4)3 have been reported to show impressive activity on catalyzing the conversion of polysaccharides into platform chemicals(Chen et al., 2019; Huang et al., 2018; Li et al., 2014; Shen et al., 2020; Yang et al., 2017; Yang et al., 2012; Zhang et al., 2016), here we employed Al2(SO4)3 as the catalyst for cellulose-to-HMF process. In the beginning, we studied the conversion of cellulose with H2O/THF volume ratio as 2/20 because that the high Vorg/Vaque could efficiently protect the formed HMF from further degradation. As shown in Table 1, with the addition of Al2(SO4)3, the cellulose conversion sharply increased from only 14.9% without catalyst to 95.9%, accompanied with HMF yield increased from only 1.6% to 56.3 %, suggesting that the Al2(SO4)3 could efficiently convert cellulose into HMF. The excellent performance of the reaction system here was ascribed to the high Vorg/Vaque, high concentration of Al2(SO4)3 in aqueous phase and the special catalytic performance of Al2(SO4)3 on conversion of cellulose. As shown in Table 1, only around 0.4 mL amount of aqueous phase was formed, indicating that around 80% of the added water were dissolved into the organic phase, and the Vorg/Vaque was as high as 43.2/1. Besides, considerable amount of the added Al2(SO4)3 was precipitated out from the aqueous phase, suggesting that the aqueous phase was saturated by Al2(SO4)3. As has been reported previously(Choudhary et al., 2013; Huang et al., 2018; Tang et al., 2015; Yang et al., 2017), the Al2(SO4)3 in the aqueous phase could undergo hydrolysis to form [Al(OH)2]+ and H+ in aqueous phase, in which [Al(OH)2]+ is efficient on isomerization of glucose into fructose, while H+ is responsible for hydrolysis of cellulose into glucose and dehydration of fructose into HMF. Thus, the high concentration of Al2(SO4)3 in the aqueous phase could efficiently catalyze the conversion of cellulose into HMF, while the high Vorg/Vaque could guarantee the efficiency on extraction of the formed HMF into the organic phase to prevent its further degradation.
Table 1
The results of conversion of cellulose into HMF in H2O-THF mixture [a]
Entry
|
Solvent (mL)
|
Al2(SO4)3·18H2O dosage (g)
|
Vaque
|
H2O content in organic phase (wt.%)
|
Al2(SO4)3 distributed in organic phase (%)
|
Concen-Al2(SO4)3-aque (g/mL)
|
Vorg/Vaque
|
Cellulose conversion (%)
|
HMF yield (%)
|
H2O
|
THF
|
1
|
2.0
|
20.0
|
0.0
|
0
|
10.1
|
0
|
-
|
-
|
14.9
|
1.6
|
2
|
2.0
|
20.0
|
0.4
|
0.4
|
8.7
|
1.6
|
Saturated [b]
|
43
|
95.9
|
56.3
|
3
|
3.0
|
19.0
|
0.4
|
0.6
|
12.5
|
1.7
|
0.35
|
36
|
97.6.
|
59.3
|
4
|
4.0
|
18.0
|
0.4
|
0.8
|
16.7
|
4.1
|
0.26
|
27
|
98.2
|
61.1
|
5
|
6.0
|
16.0
|
0.4
|
0.9
|
26.4
|
6.7
|
0.23
|
23
|
71.1
|
43.8
|
6
|
8.0
|
14.0
|
0.4
|
1.2
|
35.4
|
13.4
|
0.17
|
17
|
64.9
|
32.5
|
7 [c]
|
2.0
|
20.0
|
0.4
|
1.1
|
4.8
|
0.8
|
0.19
|
19
|
58.2
|
32.2
|
8 [c]
|
4.0
|
18.0
|
0.4
|
3.3
|
4.2
|
0.7
|
0.06
|
6
|
32.7
|
25.5
|
[a] Reaction condition: 1 g cellulose, 0.4 g Al2(SO4)3·18H2O, 180℃, 1 h. [b] Considerable Al2(SO4)3 was precipitated out from the solvents. [c] 35% of NaCl was added (based on the mass of added water) |
We further studied the conversion of cellulose with higher H2O/THF volume ratio. With the increasing of H2O/THF volume ratio from 2/20 to 3/19 and 4/18, the HMF yield slightly increased from 56.3 % to 59.3% and 61.1%, respectively. The increase of HMF yield was ascribed to that the increasing water content could adjust the Vaque and the concentration of Al2(SO4)3 in aqueous phase to proper values for HMF production. It should note that the H2O content in the organic phase increased from 8.7–16.7% with the increase of H2O/THF volume ratio from 2/20 to 4/18 but the HMF yield didn’t decrease, indicating that the presence of 16.7% of H2O in the organic phase show very little impact on the stability of HMF in the organic phase. However, further increasing the H2O content by increasing the H2O/THF volume ratio to 6/16 and 8/14 resulted in a clear decrease in HMF yield to 43.8 % and 35.3%, respectively, accompanied with the decreasing of cellulose conversion to below 75%. The decreasing of efficiency on cellulose-to-HMF process with the increasing of H2O/THF volume ratio was ascribed to the decrease of Vorg/Vaque, the decrease of concentration of Al2(SO4)3 in aqueous phase and the increase of concentration of Al2(SO4)3 and content of H2O in organic phase(Yang et al., 2017). As shown in Table 1, with the increase of H2O/THF volume ratio from 4/18 to 8/14, the volume of aqueous phase increased from 0.8 mL to 1.2 mL, leading to the Vorg/Vaque decreased from 26.8/1 to 16.9/1. Besides, the H2O content and the solubility of Al2(SO4)3 in the organic phase all increased with the increase of H2O/THF ratio. With the increase of H2O/THF volume ratio from 4/18 to 8/14, the H2O content in the organic phase increased from 16.7–35.4%, while the amount of Al2(SO4)3 distributed in the organic phase increased from 4.1 % to 13.4%, all of which could decrease the stablity of HMF in the organic phase. On the other hand, the increase of Vaque and the transfering of Al2(SO4)3 into the organic phase all lead to the decreasing of concentration of Al2(SO4)3 in the aqueous phase from saturated to 0.17 g/mL, resulting the decreasing of the efficiency on hydrolysis of cellulose. In brief, the decreasing concentration of Al2(SO4)3 in the aqueous phase and the increasing content of H2O and Al2(SO4)3 in organic phase lead to the decreasing of efficiency on conversion of cellulose into HMF with high H2O/THF volume ratio.
To further confirm that the dissolving of H2O and Al2(SO4)3 into organic phase could decrease the stability of HMF in the reaction system, we studied the stability of HMF with various H2O/THF volume ratio. As shown in Fig. 1(a), after the same reaction time, the remained HMF with the H2O/THF volume ratio of 8/14 was always lower than that with the H2O/THF volume ratio of 4/18, clearly confirming that the stability of HMF decreased with the dissolving of H2O and Al2(SO4)3 into organic phase.
The HMF yield obtained here was obviously higher than that of 43.5% reported by Shen et al., who also employed Al2(SO4)3 to catalyze the conversion of cellulose pretreated by Al2(SO4)3-assisted ball milling in biphasic system consisted by H2O and γ-valerolactone (GVL) (Shen et al., 2020). In Shen’s work, 20% of NaCl (based on the mass of added water) was added into the reaction mixture to prevent the inter-miscible of H2O and GVL. Here we also studied the influence of the addition of NaCl on conversion of cellulose into HMF in H2O-THF biphasic system, and the results clearly showed that the addition of NaCl could sharply decrease the efficiency on conversion of cellulose into HMF (Table 1, entry 7–8). When 35% of NaCl (based on the usage of water) was added, the cellulose conversion decreased from over 95 % to below 60 %, accompanied with HMF yield decreased from over 56 % to below 35% (Table 1, entry 2–4 and 7–8), suggesting that the addition of NaCl is against the conversion of cellulose into HMF in the biphasic system. Further study found that the addition of NaCl could increase the Vaque of the reaction system because that the NaCl could suppress the dissolving of H2O into the organic phase. Thus, for the reaction systems with the addition of NaCl, the concentrations of the Al2(SO4)3 in the aqueous phase were much lower than those reaction systems without NaCl, and the Vorg/Vaque were also much lower than those conditions without NaCl, all of which were against for the conversion of cellulose into HMF. The result was not accord with Yang’s work(Yang et al., 2017), who found that the addition of NaCl could slightly improve the yield of furfural from xylan with Al2(SO4)3 as catalyst. We propose that the difference was caused by the fact that the hydrolysis of cellulose was much difficult than that of xylan, which presented as the rate-determining step for cellulose-to-HMF process (Shi et al., 2013) and need high concentration of Al2(SO4)3 in the aqueous phase to execute this reaction. Therefore, we proposed that NaCl should not be added into the H2O-THF biphasic system for preparation of HMF from cellulose.
3.2 Influence of CHX on the conversion of cellulose into HMF in biphasic system
As shown above, the dissolving of Al2(SO4)3 into the organic phase could decrease the catalytic efficiency on the conversion of cellulose into HMF. Thus, suppressing the dissolving of Al2(SO4)3 into the organic phase by addition of some non-polar solvent into organic phase may improve the HMF yield. CHX is one common and cheap non-polar solvent, and it is intersoluble with THF in any proportion but insoluble in water, so we studied the effect of addition of CHX into the reaction mixture on preparation of HMF from cellulose. Unexpectedly, the addition of CHX into the reaction mixture efficiently increased the HMF yield (Table 2). Change the reaction mixture from VH2O:VTHF =2:20 to VH2O:VTHF:VCHX =1:20:1, the HMF yield increased from 56.1–62.5%, and even higher HMF yield of 64.1% and 67.2 % could be achieved with VH2O:VTHF:VCHX = 2:18:2 and 3:16:3, respectively (Table 2, entry 1–3). Further study proved that the addition of CHX into reaction mixture could decrease the solubility of Al2(SO4)3 in the organic phase. Only less than 1% of Al2(SO4)3 was transferred into the organic phase with the addition of CHX, and over 99% of Al2(SO4)3 was remained in the organic phase. Obviously, the suppressing of solubility of Al2(SO4)3 in organic phase could increase the stability of HMF in organic phase (Fig. 1(b)), thus improved the efficiency on conversion of cellulose into HMF.
Table 2
The results of conversion of cellulose into HMF in H2O-THF-CHX mixture [a]
Entry
|
Solvent (mL)
|
Al2(SO4)3·18H2O dosage (g)
|
Vaque (mL)
|
Al2(SO4)3 distributed in organic phase (%)
|
Concen-Al2(SO4)3-aque (g/mL)
|
Vorg/Vaque
|
Cellulose conversion (%)
|
HMF yield (%)
|
H2O
|
THF
|
CHX
|
1
|
1.0
|
20.0
|
1.0
|
0.4
|
-
|
0.3
|
Saturated [b]
|
-
|
77.4
|
62.5
|
2
|
2.0
|
18.0
|
2.0
|
0.4
|
0.6
|
0.4
|
0.34
|
34
|
92.3
|
64.1
|
3
|
3.0
|
16.0
|
3.0
|
0.4
|
2.8
|
0.7
|
0.07
|
6
|
96.9
|
67.2
|
4
|
4.0
|
14.0
|
4.0
|
0.4
|
4.8
|
0.6
|
0.04
|
4
|
85.9
|
49.8
|
5
|
5.0
|
12.0
|
5.0
|
0.4
|
5.7
|
0.4
|
0.04
|
3
|
82.9
|
36.6
|
6
|
6.0
|
10.0
|
6.0
|
0.4
|
7.1
|
0.4
|
0.03
|
2
|
77.3
|
25.3
|
[a] Reaction condition: 1 g cellulose, 0.4 g Al2(SO4)3·18H2O, 180℃, 1 h. [b] Great amount of Al2(SO4)3 was precipitated out from the solvents. |
Some previous studies also reported that addition of organic phase modifier into the reaction could minimize HMF degradation by suppressing the humins formation (Atanda et al., 2015; Roman-Leshkov et al., 2006). Atanda et al. reported that addition of N-methyl-2-pyrrolidone (NMP) into H2O/THF biphasic system could increase the HMF yield from 60–80%. They proposed that NMP acts as an aqueous phase modifier that can suppress humin formation in water and simultaneously enhance partitioning of HMF into the organic layer(Atanda et al., 2015). Here we proposed that the effect of CHX on the efficiency of cellulose-to-HMF process was ascribed to that CHX could suppress the dissolution of acidic salts Al2(SO4)3 into organic phase, leading to higher stability of HMF in organic phase. We also studied the stability of HMF in the reaction mixture at 180℃, and the results were shown in Fig. 1(b). As shown, over 75% of the HMF were remained after 60 min, clearly confirmed that the addition of CHX into the reaction system improved the stability of HMF in the reaction mixture.
Further increase the H2O and CHX content (for the conditions of VH2O:VTHF:VCHX = 4:14:4, 5:12:5 and 6:10:6) lead to sharp decrease of HMF yield to below 50% (Table 2, entry 4–6). The sharp decreasing of HMF yield with the increase of H2O and CHX content was ascribed to that the increasing of H2O and CHX content could increase the Vaque of the reaction system, leading to relatively low Al2(SO4)3 concentration in the aqueous phase and low Vorg/Vaque. As shown in Table 2, for the conditions of VH2O:VTHF:VCHX = 4:14:4, 5:12:5 and 6:10:6, the Vorg/Vaque have decreased to below 5:1, and the Al2(SO4)3 concentration in the aqueous phase have decreased to below 0.04 g/mL, all of which are against for conversion of cellulose into HMF.
It was noted that for the condition of H2O/THF/CHX volume ratio as 3:16:3, the concentration of Al2(SO4)3 in aqueous phase was only 0.07 g/mL and the Vorg/Vaque was only 6.4:1, all of which were much lower than those without the addition of CHX, but the cellulose conversion (96.9 %) and HMF yield (67.2%) were all higher than those conditions without CHX. The high cellulose conversion suggested that the Al2(SO4)3 concentration in the aqueous phase at reaction temperature should be much higher than that measured at room temperature. Because the temperature could influence the liquid-liquid phase equilibrium of polycomponent system(García-Flores et al., 2015), the real Vorg/Vaque and the concentration of Al2(SO4)3 in aqueous phase at reaction temperature should be different with those measured at room temperature. Thus, we tried to measure the actual Vorg and Vaque of the reaction mixture, and the results were shown in the next section.
3.3 The real state of the biphasic system at reaction temperature
Table 3 showed the Vorg and Vaque of the reaction system under the room temperature and 180℃. Because H2O and THF are complete intersoluble, only one phase was formed at both room temperature and reaction temperature with the absence of CHX and Al2(SO4)3. Aqueous phase was formed when Al2(SO4)3 was added into mixture of H2O/THF, but the Vaque was always much lower than that of the added H2O, indicating that considerable added H2O was dissolved into organic phase. For the condition of H2O/THF/Al2(SO4)3 as 2/20/0.4 and 4/18/0.4, only 0.4 mL and 0.8 mL of aqueous phase were formed at room temperature, respectively, which sharply decreased to only 0.2 mL and 0.4 mL at 180 ℃, respectively. The decreasing of Vaque with the increase of temperature was ascribed to the increased solubility of H2O in THF with the increase of temperature. Further increasing of H2O/THF volume ratio to 6/16 and 8/14 slightly increased the Vaque to 0.9 mL and 1.2 mL at room temperature, respectively, and increase of temperature to 180 ℃ only slightly decrease the Vaque to 0.8 mL and 1.1 mL, respectively, suggesting that the affinity between H2O and THF become insensitive with temperature when the H2O content in organic phase reached over 26%.
Table 3
The real state of the biphasic system at room temperature and 180 ℃
Composition of the biphasic system
|
Biphasic system at room temperature
|
Biphasic system at 180 ℃
|
Solvent H2O (mL)
|
Salts (g)
|
Vorg (mL)
|
Vaque (mL)
|
Concen-Al2(SO4)3-Aque (g/mL)
|
Vorg/Vaque
|
Vorg (mL)
|
Vaque (mL)
|
Concen-Al2(SO4)3-Aque (g/mL)
|
Vorg/Vaque
|
H2O/THF = 2/20
|
0
|
22.0
|
0
|
0
|
-
|
27.7
|
0
|
0
|
-
|
H2O/THF = 4/18
|
0
|
22.0
|
0
|
0
|
-
|
27.4
|
0
|
0
|
-
|
H2O/THF = 6/16
|
0
|
22.0
|
0
|
0
|
-
|
27.1
|
0
|
0
|
-
|
H2O/THF = 8/14
|
0
|
22.0
|
0
|
0
|
-
|
26.8
|
0
|
0
|
-
|
H2O/THF = 2/20
|
0.4g Al2(SO4)3
|
21.6
|
0.4
|
Saturated
|
54
|
27.4
|
0.2
|
1.03
|
137
|
H2O/THF = 4/18
|
0.4g Al2(SO4)3
|
21.4
|
0.8
|
0.26
|
27
|
26.9
|
0.4
|
0.51
|
68
|
H2O/THF = 6/16
|
0.4g Al2(SO4)3
|
21.1
|
0.9
|
0.23
|
23
|
26.3
|
0.8
|
0.26
|
33
|
H2O/THF = 8/14
|
0.4g Al2(SO4)3
|
20.3
|
1.2
|
0.17
|
17
|
25.7
|
1.1
|
0.19
|
23
|
H2O/THF/CHX = 1/20/1
|
0
|
22.0
|
0
|
0
|
-
|
27.8
|
0
|
0
|
-
|
H2O/THF/CHX = 2/18/2
|
0
|
22.0
|
0
|
0
|
-
|
27.6
|
0
|
0.
|
-
|
H2O/THF/CHX = 3/16/3
|
0
|
19.2
|
2.8
|
0
|
7
|
27.4
|
0
|
0
|
-
|
H2O/THF/CHX = 4/14/4
|
0
|
17.4
|
4.6
|
0
|
4
|
25.8
|
1.5
|
0
|
17
|
H2O/THF/CHX = 1/20/1
|
0.4g Al2(SO4)3
|
21.6
|
0.4
|
Saturated
|
54
|
27.6
|
0.2
|
1.03
|
138
|
H2O/THF/CHX = 2/18/2
|
0.4g Al2(SO4)3
|
21.4
|
0.6
|
0.34
|
36
|
27.4
|
0.3
|
0.68
|
91
|
H2O/THF/CHX = 3/16/3
|
0.4g Al2(SO4)3
|
19.2
|
2.8
|
0.07
|
7
|
26.8
|
0.6
|
0.34
|
45
|
H2O/THF/CHX = 4/14/4
|
0.4g Al2(SO4)3
|
17.3
|
4.8
|
0.04
|
4
|
25.3
|
1.9
|
0.11
|
13
|
H2O/THF/CHX = 5/12/5
|
0.4g Al2(SO4)3
|
16.3
|
5.7
|
0.04
|
3
|
23.2
|
3.8
|
0.05
|
6
|
H2O/THF/CHX = 6/10/6
|
0.4g Al2(SO4)3
|
14.8
|
7.1
|
0.03
|
2
|
20.4
|
6.5
|
0.03
|
3
|
When CHX was added into H2O-THF mixture, only monophase was formed when the H2O/THF/CHX volume ratio was 1:20:1 and 2:18:2, while further increase the H2O and CHX content lead to the generation of biphase system. For the condition of H2O/THF/CHX volume ratio as 3:16:3, biphasic system was formed at room temperature but only one phase was observed at 180℃. However, for the condition of H2O/THF/CHX volume ratio as 4:14:4, biphase system was formed at both room temperature and at 180 ℃. The phase behavior could be explained by the H2O-THF-CHX phase diagram shown in Fig. 2. It is well-known that THF is complete miscibility with H2O and CHX, while the H2O and CHX are almost immiscible, so the H2O-THF-CHX phase diagram could be divided into two region of monophase region and biphase region(García-Flores et al., 2015). Besides, increasing of temperature could enlarge the monophase region of the phase diagram of H2O-THF-CHX, indicating that a certain biphase region at room temperature could become monophase state at high temperature. Thus, the H2O-THF-CHX phase diagram at different temperature could be divided into three region of monophase region, biphase region and transition region, as shown in Fig. 2(García-Flores et al., 2015). In the monophase region, the H2O/THF/CHX mixture presented as monophase at both room temperature and 180℃. In the transition region, the H2O/THF/CHX mixture presented as biphase at room temperature but become monophase at 180℃. In the biphase region, H2O/THF/CHX mixture presented as biphase at both room temperature and 180℃. Obviously, the plots of H2O/THF/CHX volume ratio as 1:20:1 and 2:18:2 are presented in monophase region of the H2O-THF-CHX phase diagram, the plot of H2O/THF/CHX volume ratio as 3:16:3 is presented in the transition region, while the plots of H2O/THF/CHX ratio = 4:14:4 is presented in the biphasic region.
For most of the biphasic systems, the Vaque at 180 ℃ were lower than the Vaque at room temperature. For example, the Vaque at room temperature was 0.8 mL for the condition of H2O/THF/Al2(SO4)3 ratio (mL/mL/g) as 4/18/0.4, which decreased to only 0.2 mL at 180 ℃. What is more exaggerated, for the condition of the H2O/THF/CHX/Al2(SO4)3 ratio (mL/mL/mL/g) as 3/16/3/0.4, the Vaque sharply decreased from 2.8 mL at room temperature to only 0.6 mL at 180 ℃. The variation of Vaque with the increasing of temperature could be described by the following equation:
![](data:image/png;base64,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)
Where the ∆maque denotes the change of the mass of aqueous phase, Vaque and Vorg denote the volume of aqueous phase and organic phase at room temperature, respectively, while ∆CTHF→aque and
denote the variable quantity of solubility of THF in aqueous phase and the variable quantity of solubility of H2O in organic phase with the increase of temperature, respectively. Because the Vorg was much higher than Vaque in the biphasic system in this study, so the above equation could be simplified as follow:
![](data:image/png;base64,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)
Thus, the ∆maque should decrease with the increase of temperature for most conditions in this study.
Be different from the aqueous phase, the Vorg measured at 180 ℃ were all higher than those measured at room temperature, which was caused by the volume expansion of the organic solvents and the transferring of H2O from aqueous phase to organic phase. The increasing of Vorg and decreasing of Vaque with the increase of temperature jointly lead to the Vorg/Vaque ratios at 180 ℃ be higher than those at room temperature. For the conditions of H2O/THF/Al2(SO4)3 as 2/20/0.4 and 4/18/0.4, the Vorg/Vaque ratios at 180 ℃ all reached as high as over 60/1, while for the conditions of H2O/THF/Al2(SO4)3 as 6/16/0.4 and 8/14/0.4, the Vorg/Vaque ratios were all below 33/1. The difference of the Vorg/Vaque and the concentration of Al2(SO4)3 in aqueous phase should be one main reason for the excellent performance of the former reaction systems and poor performance of the later reaction systems.
The decreasing of Vaque with the increase of temperature could increase the Al2(SO4)3 concentration in the aqueous phase, which could also influence the cellulose-to-HMF process. Assuming that the distribution of Al2(SO4)3 in the biphasic system at high temperature was the same as that at room temperature, we could calculate the Al2(SO4)3 concentration in the aqueous phase at 180 ℃, and the results were also shown in Table 3. Due to the decrease of Vaque with the increase of temperature, the concentrations of Al2(SO4)3 in aqueous phase at 180 ℃ were all much higher than those at room temperature. For cases of H2O/THF/Al2(SO4)3 as 4/18/0.4, the concentrations of Al2(SO4)3 in aqueous phase at 180 ℃ was 1.03 g/mL, much higher than that of 0.26 g/mL at room temperature. What’s more exaggerated, for the condition of H2O/THF/CHX/Al2(SO4)3 as 3/16/3/0.4, the concentrations of Al2(SO4)3 in aqueous phase increased from 0.07 g/mL to 0.34 g/mL. The high concentrations of Al2(SO4)3 in aqueous phase at 180 ℃ of the reaction system should also be responsible for the high efficiency on cellulose conversion.
The excellent performance of the reaction system of H2O/THF/CHX/Al2(SO4)3 as 3/16/3/0.4 could be explained by the relatively high Vorg/Vaque (44.6/1) and high concentration of Al2(SO4)3 in aqueous phase (0.34 g/mL) at reaction temperature. As shown in Table 3, with the increase of temperature from room temperature to 180 ℃, the Vaque sharply decreased from 2.8 mL to 0.6 mL. The sharp decrease of Vaque could be explained by the H2O-THF-CHX phase diagram shown in Fig. 2. The plot H2O/THF/CHX = 3/16/3 was in the transition region of the phase diagram, so the mixture presented as biphase at room temperature to generate 2.8 mL of aqueous phase. However, the mixture of H2O/THF/CHX = 3/16/3 present as monophase at 180℃, so no aqueous phase could be observed with the absence of Al2(SO4)3. When Al2(SO4)3 was added into the H2O/THF/CHX mixture at 180 ℃, Al2(SO4)3 could capture H2O from the organic phase to generate aqueous phase, leading to the formation of 0.6 mL aqueous phase at 180 ℃. The sharp decrease of Vaque leading to the Vorg/Vaque sharply increased from 6.9/1 at room temperature to 44.6/1 at 180 ℃, and the concentration of Al2(SO4)3 in aqueous phase increased from 0.07 g/mL to 0.34 g/mL, which are responsible for the excellent efficiency of the reaction system. The changeable Vaque for the conditions of H2O/THF/CHX/Al2(SO4)3 as 3/16/3/0.4 shows some advantages for cellulose-to-HMF process. The Vaque was low at reaction temperature, so the high Vorg/Vaque ratio and high concentration of Al2(SO4)3 in aqueous phase could guarantee the high efficiency on converting cellulose into HMF. However, the Vaque was high at room temperature, thus the catalyst could be extracted to the aqueous phase and recycled easily.
The above results showed that the reaction mixture of H2O/THF/CHX/Al2(SO4)3 as 3/16/3/0.4 was much more efficient than that of 4/14/4/0.4, which could be explained by the difference of Vorg/Vaque and the concentration of Al2(SO4)3 in aqueous phase at 180 ℃. For the condition of H2O/THF/CHX/Al2(SO4)3 as 3/16/3/0.4, the Vorg/Vaque and the concentration of Al2(SO4)3 in aqueous phase were 44.6/1 and 0.34 g/mL, but for the condition of H2O/THF/CHX/Al2(SO4)3 as 4/14/4/0.4, the Vorg/Vaque and the concentration of Al2(SO4)3 in aqueous phase were only 16/1 and 0.13 g/mL. Obviously, the relatively lower Vorg/Vaque and the concentration of Al2(SO4)3 in aqueous phase should be responsible for the poor performance of the condition of H2O/THF/CHX/Al2(SO4)3 as 4/14/4/0.4. Combining the catalytic performance and the real state of the reaction system, we could found that the concentration of Al2(SO4)3 in aqueous phase and the Vorg/Vaque ratio shows obvious influence on the conversion of cellulose into HMF. It seems that when the concentration of Al2(SO4)3 in aqueous phase was higher than 0.34 g/mL and the Vorg/Vaque was higher than 44.6/1, the reaction mixture was quite efficient on conversion of cellulose into HMF. However, the efficiency of cellulose-to-HMF process decreased with the decrease of concentration of Al2(SO4)3 in aqueous phase and Vorg/Vaque ratio when the concentration of Al2(SO4)3 in aqueous phase were below 0.34 g/mL and Vorg/Vaque ratio were below 44.6/1.
In brief, because of the increased solubility of H2O in the organic phase with the increase of temperature, the Vorg/Vaque and the concentration of Al2(SO4)3 in aqueous phase at reaction temperature were all much higher than that measured at room temperature. For the mixture of H2O/THF/CHX/Al2(SO4)3 as 3/16/3/0.4, the concentration of Al2(SO4)3 in aqueous phase increased from 0.07 g/mL at room temperature to 0.34 g/mL at 180 ℃, while the Vorg/Vaque increased from 6.9/1 at room temperature to 44.6/1 at reaction temperature, which were responsible for the excellent performance of the reaction system. Combining the real state of the biphasic system and the reaction performance on cellulose-to-HMF process, we found that the reaction systems with Al2(SO4)3 concentration in the aqueous phase over 0.34 g/mL, Vorg/Vaque ratio over 44/1 and H2O content in organic phase below 17% all showed excellent performance on conversion of cellulose into HMF.
3.4 The microemulsion-like systems and liquid film catalytic processes
It should be mentioned that because the Vaque at 180 ℃ was quite low (below 0.6 mL for all the conditions with excellent performance), the aqueous phase even could not wet the surface of cellulose. Thus, we proposed that the reaction system should be the so-called microemulsion-like system instead of the biphasic system (Fig. 3). Under stirring, the organic phase acts as the dispersion medium, while the cellulose particles and the aqueous phase are dispersed in the organic phase. Because of the hydrophilic property of the cellulose particles(Shi et al., 2013), the aqueous phase covers on the surface of the cellulose particles to generate one acidic liquid film, which could efficiently catalyze the hydrolysis of cellulose and isomerization-dehydration of hexose, while the formed HMF is efficiently extracted into the organic phase to suppress its further degradation. The CHX in the organic phase decreased the solubility of Al2(SO4)3, thus improved the selectivity of HMF from cellulose and suppressed the formation of humins.
3.5 Screening of acidic metal salts for cellulose conversion
In order to further improve the efficiency on cellulose-to-HMF process, a variety of other acidic metal salt catalysts were tested for the conversion of cellulose in the H2O/THF/CHX biphasic system, and the results were shown in Fig. 4. The cellulose conversions were lower than 65% with ZrOCl2, MgCl2, FeCl3, Fe2(SO4)3 and ErCl2 as catalyst, suggesting that the catalytic activity on hydrolysis of cellulose of these catalyst were relatively low. The low catalytic activities of these catalysts were ascribed to the relatively lower acid strength of the aqueous phase generated by hydrolysis of these metal salts(Yang et al., 2017). Although HfO(PO4)2.0 is reported to be excellent on conversion of cellulose into HMF as one solid acid(Cao et al., 2019), here we found that HfCl4 is also not efficient on conversion of cellulose in biphasic system, which is similar with previous report(Zhang et al., 2012). The ErCl2 also showed low efficient on conversion of cellulose, which could be due that the ErCl2 is more favor for conversion of cellulose into lactic acid(Lei et al., 2014). The AlCl3, CuSO4·5H2O and SnCl4·H2O all showed moderate activity on conversion of cellulose into HMF, and moderate HMF yield of over 40% was obtained from these three catalysts. For the same metal cations of Al3+ and Fe3+, the chlorine salts all showed lower efficiency than the sulfates on the cellulose-to-HMF process. Yang et al. reported similar phenomenon on xylan-to-furfural process (Yang et al., 2017). The lower efficiency of the chlorine salts was proposed to be caused by the higher solubility of the chlorine salts in the organic phase than that of the sulfates(Yang et al., 2017).
Obviously, the catalytic activity of these explored salts were all lower than Al2(SO4)3, clearly showing that the Al2(SO4)3 should be one promising candidate for large-scale producing HMF from cellulose.
3.6 Influence of catalyst usage, reaction temperature, reaction time and cellulose dosage on HMF production
The influence of catalyst usage, reaction temperature, reaction time and cellulose dosage was studied with H2O/THF/CHX volume ratio as 3/16/3. As shown in Fig. 5(a), increase the Al2(SO4)3 dosage to 0.05 g and 0.3 g, the HMF yield sharply increased to 34.3% and 62.3%, but further increase the dosage of Al2(SO4)3 from 0.3 g to 0.5 g only lead to slight increase of HMF yield from 62.3% to 71.2 %. The increased HMF yield with the increase of Al2(SO4)3 dosage was ascribed to that the increase of Al2(SO4)3 in reaction system could increase the concentration of Al2(SO4)3 in the aqueous phase, which could accelerate the hydrolysis of cellulose and dehydration of hexose. Further increasing the Al2(SO4)3 to 0.6 g lead to the decrease of HMF yield, which was ascribed to that the excessive Al2(SO4)3 loading accelerated the side reactions (e.g., rehydration or condensation of HMF) (Yang et al., 2017).
As shown in Fig. 5(b), the cellulose conversion was low at low reaction temperature of 160℃ and 170℃, suggesting that the temperature below 170℃ was not enough for Al2(SO4)3 to catalyze the conversion of cellulose. As the reaction temperature was increased, the conversion of cellulose and yield of HMF gradually increased and reached a plateau at 180℃ and 190 ℃, indicating that the reaction rates increased with the increasing of reaction temperature. However, at temperatures above 200 ℃, the HMF yields decreased to below 40%, indicating that the formed HMF became unstable with reaction temperature of 200 ℃. The temperature could influence the cellulose-to-HMF process through two main routes. On one hand, the increase of temperature could accelerate the hydrolysis of cellulose and dehydration of glucose, thus accelerate the cellulose-to-HMF process. On the other hand, the increase of temperature could further decrease the Vaque of the system, resulting higher Vorg/Vaque and higher concentration of Al2(SO4)3 in aqueous phase, which could also accelerate the reaction process.
The influence of reaction time on the yield of HMF showed similar trend (Fig. 5(c)). Low cellulose conversion of 75.3% was achieved within 30 min, accompanied with HMF yield of 52.7%. Prolonging the reaction time to 60 min leads to the increasing of HMF yield to 71.2%, but further increasing of reaction time to 90 min lead to decreasing of the HMF yield to only 54.1%, which was also ascribed to the degradation of the formed HMF into humin type polymer(Shi et al., 2020; Shi et al., 2013).
We further studied the influence of cellulose/solvent ratio on HMF yield from cellulose because that the high cellulose/solvent ratio is favored for HMF production in large scale. Not surprisingly, the cellulose conversion and HMF yield all decreased with the increase of cellulose/solvent ratio. With the increase of cellulose/solvent ratio (g/mL) from 1:22 to 2:22, 4:22 and 5:22, the HMF yield decreased from 71.2–58.7%, 49.8% and 40.3%, while the HMF concentration in organic phase increased from 28.5 g/L to 49.2 g/L, 64.6 g/L and 63.4 g/L, respectively (Fig. 5(d)). The decrease of HMF yield with the increase of cellulose/solvent ratio was ascribed to the accelerated condensation/polymerization of HMF with the increase of HMF concentration in the reaction system, which has been reported by previous literatures(Fang et al., 2018; Shi et al., 2013). Many literatures employed extremely low cellulose/solvent ratio (below 1:50) to obtain relatively high HMF yield (Li et al., 2018a), which could result in quite low HMF concentration in the reaction mixture. Cao et al. reported high HMF yield of 69.8% from cellulose with HfO(PO4)2 as catalyst, but the cellulose/solvent ratio was 1/50 (g/mL), leading to HMF concentration in the reaction mixture only 1.1 g/mL (Cao et al., 2019). Lai et al. achieved an excellent HMF yield of 67.5% from cellulose with ChH4CeW12O40 as catalyst, but the cellulose/solvent ratio was 1/70 (g/mL), also leading to HMF concentration below 0.75 g/mL(Lai et al., 2020). Some other literatures even conducted the cellulose-to-HMF process with cellulose/solvent below 1/80 (g/mL)(Li et al., 2018a; Li et al., 2018b), leading to extremely low concentration of HMF in the reaction mixture. In this study, acceptable HMF yield and high HMF concentration in organic phase could be obtained with high cellulose/solvent ratio, indicated that the reaction system has great prospects for industrialization.