A Novel Integrated Hydrothermal Liquefaction and Solar Catalytic Reforming Method for Enhanced Hydrogen Generation from Biomass

Veiw figure View Figure

Additionally, the rapid recombination of the photogenerated electrons with the holes limits H₂ production. Using co-catalyst such as platinum on TiO₂ surface traps photogenerated electrons and partially prevents electron-hole recombination ^[20]. The PR of biomass model compounds, such as the primary, secondary and tertiary alcohols, glycerol, sugars and a few carboxylic acids have been studied for H₂generation ^[21]. It is expected that H₂ and CO₂ production from biocrude components will follow the general reforming scheme:

(1)

Green solar photoreforming of the biocrude to H₂ will not only remediate the biocrude but also permit significantly higher H₂ abstraction from the biomass. In this study, HTL of cellulosic biomass was carried out at 250°C in the presence of NiSO₄ catalyst for the production of H₂ and biocrude containing C₁-C₃ carboxylic acids. The biocrude and its carboxylic acids were individually photoreformed to demonstrate additional yields of H₂. A schematic describing the combined HTL and PR process is shown in Figure 1. The results show that integrated HTL and PR process has a potential to develop into technologically feasible route for enhanced H2 production.

2. Materials and Methods

2.1. Materials

Commercially available cotton balls (98% cellulose) were selected as biomass feedstock. Nickel sulphate hexahydrate and dihydrogen hexachloroplatinate (IV) hydrate, Premion (99.999% metal basis) were purchased from Alfa Aesar, MA. Titanium (IV) oxide (P25 nanoparticles), 99.7% anatase (particle size ~ 25 nm, surface area 45-55 m²/gm) was obtained from Sigma Aldrich. Filtrasorb 600 activated charcoal (Iodine number 850 mg/g) was purchased from Calgon Carbon.

2.1.1. Elemental Composition of Biomass

Biomass, cotton (1 gram) was dried at 45°C overnight in the oven and weighed the next day to determine moisture content. The moisture free cotton was supplied to the Midcontinent Testing Labs, Inc., Rapid City, SD, USA for the elemental analysis.

2.1.2. Preparation of Platinum Coated TiO₂ Nanoparticles

Platinum coating on TiO₂ nanoparticles was achieved by the wet incipient method. TiO₂ nanoparticles (0.5 g) were placed in water (12 ml) and sonicated for 30 min. To this dispersion, hexachloroplatinic acid (1 wt% Pt equivalent) was added and sonication was continued for additional 3 hrs. The dispersion was dried at 100°C and calcined at 600°C for 1 hrs. The platinum loading was determined by EDX analysis using Zeiss Supra 40VP scanning electron microscope/energy dispersive X-ray spectrometer.

2.1.3. Hydrothermal Liquefaction (HTL) of Biomass

A 300 ml high-temperature, high-pressure (HTHP) stainless steel (SS316) reactor (PARR Instrument Co, Illinois) fitted with gas inlet, outlet valves, a liquid sampling valve, pressure gauge, rupture disc, magnetic drive and a thermocouple was used for HTL of biomass. Slurry of 1.0 gram cotton balls cut into 0.5 cm² pieces was made in 150 ml of distilled water containing 5 wt% NiSO₄ catalyst. The slurry was loaded in the reactor and stirred by an impeller at a constant speed of 1300 rpm. The reactor was pressurized to 40 psi with N₂ at room temperature and heated to 250°C to perform HTL for 120 min. During the HTL processing, gas and liquid samples were withdrawn periodically and analyzed by GC, HPLC and GC-MS. The processed slurry was centrifuged at 13000 rpm for 20 min to recover supernatant biocrude.

2.1.4. Characterization of Product Gas from HTHP Reactor Headspace

While HTL of biomass was in progress, product gas samples were withdrawn from the HTHP reactor headspace and analyzed on the Agilent7890A gas chromatograph (GC) equipped with thermal conductivity detector (TCD) and Porapak-Q stainless steel packed column (6 ft x 0.25 in x 2.1 mm) from Supelco. The GC oven was operated at 40°C and N₂ at 1 ml/min was used as a carrier gas. He (99.99% purity) was used to characterize gaseous carbon species. Calibration curves were prepared using specific gas standards purchased from the Matheson Gases, SD, USA and the product gas volumes were determined at the NTP (20°C and 1 atm) conditions.

2.1.5. Activated Carbon (AC) Treatment of Biocrude and Characterization of AC-treated Biocrude

To the biocrude obtained from HTL, activated carbon was added at two different concentrations of 10 mg/ml and 50 mg/ml. The mixture was left on a stirrer for 12 hours after which it was centrifuged. The clear supernatant was kept at 4°C until ready to use. The biocrude was analyzed on a HPLC (Shimadzu) equipped with Rezex ROA 300 x 7.8 mm column and UV-vis detector. A mobile phase of 0.005M H₂SO₄ was used. Ethyl acetate extract of the biocrude was analyzed on an Agilent GC-MS (7890 GC/5975C MSD) equipped with Agilent HP- 5ms capillary column. Helium (99.999%, UHP grade) with column flow rate of 1.197 ml/min was used as carrier gas. The oven was programmed as follows: 40°C (8 min hold), ramped up (2°C/min) to 120°C (8 min hold), ramped up (5°C/min) to 220°C (no hold) and (20°C/min) to 250°C (no hold). The temperature of injector and detector was maintained at 300°C. The compounds were identified using National Institute of Standards and Technology (NIST) mass spectral library (2008, version 2.0f). The total organic carbon (TOC) content of the AC-treated HTL derived cotton biocrude was determined using a Sievers InnovOx Laboratory TOC analyzer (GE Instruments). Similar analyses were carried out with the samples of untreated.

2.1.6. Photocatalytic Reforming (PR) of Biocrude and Carboxylic Acids to H₂

Pt/TiO₂ photocatalyst (4.0 mg) was dispersed in 10 ml de-ionized water for 30 minute using ultrasonic bath and the dispersion was purged with N₂ for 30 min to displace oxygen. Lactic (LA), formic (FA), acetic (AA) and propionic (PA) acids were individually and in a mixed form dissolved in water to a final concentration of 50 mM. In another experiment LA, AA, FA PA were mixed in ratio of 3.65:1.80:1.27:1 to form biocrude acids stimulant. The reaction mixture (7 ml) was added to a quartz reactor and sealed with a butyl septum. The reactor headspace (4 ml) was purged with N₂. The headspace gas was sampled to confirm absence of O₂.The mixture was stirred by a magnetic needle and exposed to simulated solar irradiation obtained with 240 W Xe lamp and AM1.5G filter. The power density measured at the center of the reactor window was 100 mW/cm². During PR, headspace gas was withdrawn periodically and analyzed for H₂ on the Agilent 7890GC using the same column and method used for the analysis of product gas from HTL. Photreforming of AC-treated and untreated biocrude for H₂ generation was carried out using the same procedure as described earlier for individual acid and acid mixture. To achieve maximal reforming of the biocrude to H₂, the headspace was evacuated after 3 hour and purged with N₂ prior performing the next photoreforming cycle. The photoreforming cycles were continued until H₂ levels were below detection limit in GC analysis.

3. Results and Discussion

3.1. HTL of Cotton in the Presence of NiSO₄ Catalyst

Elemental analysis of cotton revealed that C, H, N and O were 45 wt%, 7.11 wt%, 0.2 wt% and 47.69 wt%, respectively. Figure 2a shows the product gas analysis from HTL of 1 gram cotton in the presence of 5 wt% NiSO₄ at 250°C and 650 psi. At “zero” reaction time, practically no H2 was observed, however, after 60 min of reaction 4.3 ml H₂ volume was detected. A total H₂ volume of 11.54 ml was observed after 120 min reaction time suggesting an almost linear increase in H₂ volume generation. The concentration of CH₄ in product gas was almost negligible at zero reaction time and did not change over the period of 2 hrs of reaction time. However, the volume of CO and CO₂ observed at zero reaction was about 25 ml. As the reaction progressed further, the volume of CO increased to 28 ml and remained constant.

Starting from zero reaction time to 2 hrs, the volumetric ratio of H₂: CO gaseous products increased from 0.0438 to 0.43 indicating 10 fold increase in H₂ over CO. Thus it appears that NiSO₄ has increased H₂ selectivity in the product gas phase. It is to be noted that in the absence of catalyst negligible quantities of H₂ were observed (data not shown). The CO₂ volume observed at zero reaction time (25 ml) was increased to 33 ml after 120 min reaction time. These results indicated 9.2%, 23% and 90% (volume basis) increase in CO, CO₂ and H₂, respectively over the period of 2 hrs reaction time at 250°C. The total product gas was estimated to be 10 wt% of the 1 gram cotton that was processed at 250°C for 2 hrs. After the HTL, the reactor contents were cooled to room temperature and centrifuged to recover the biocrude and residue. Based on the TOC analysis of the biocrude (1720 mg/L) and carbon content of cotton (estimated TOC 2811 mg/L), 61.2 wt% liquefaction of cotton was estimated (Figure 2b). The observed char residue was 28.8 wt%.

Figure 2. a) Product gas components observed during HTL of cotton at 250°C in the presence of NiSO₄ catalyst and b) mass balance showing liquefied carbon, total product gas, and char

Download as

Veiw figure View Figure

3.2. Composition of the Cotton Biocrude

The liquefied cotton biocrude was analyzed by HPLC and GC-MS and the results obtained are presented in Table 1 and Table 2, respectively. HPLC analysis revealed presence of acetic, formic, lactic and propionic acids. Among these acids, lactic acid is the main component with the concentration of 1675 mg/L, whereas propionic acid was observed in minimum concentration of 459 mg/L. The total acids were estimated to be 79.88 wt% of cotton biocrude as per the following equation-

Table 1. HPLC Analysis of C₁-C₃ acids in the Cotton Biocrude Obtained after HTL at 250°C in presence of 5 wt% NiSO₄

Download as

View current table in a new window

Tables index

Veiw figure View Table

View next table

HMF and furfural were observed to be 1.69 wt% of liquefied biocrude. The remaining 18.4 wt% of biocrude was assigned to C₆ sugars (mainly glucose) and oxygenated hydrocarbons, which were thoroughly analyzed by GC-MS using HP-5ms column and inbuilt NIST library. The complete mass balance of different compounds observed in liquefied cotton biocrude is shown in Figure 2b. The GC-MS analysis of the biocrude showing only major compounds (peak area >1% and confidence level >95%) is presented in Table 2. It indicates cyclic ketones and substituted cyclic ketones, quinone derivatives, phenols and substituted phenols, alcohols and acids. However, based on the mass balance analysis, it appears that the pathways leading to the formation of C₁-C₃ carboxylic acids dominated with the use of NiSO₄ catalyst and under the experimental conditions employed.

Table 2. Major compounds observed in the biocrude and AC-treated biocrude as analyzed by GCMS; >95% confidence level

Download as

View current table in a new window

Tables index

Veiw figure View Table

View previous table

3.3. Activated Charcoal (AC) Treatment of Biocrude

The cotton biocrude was treated with 10 mg/ml and 50 mg/ml activated charcoal and the visible difference in biocrude coloration is shown in inset of Figure 3. Activated charcoal at 10 mg/ml did not completely remove biocude color while colorless biocrude was obtained with 50 mg/ml activated charcoal treatment for 12 hours. Equal volumes (1 ml) of untreated biocrude and AC-treated samples were tested for photocatalytic H₂ generation using 0.4 mg/ml Pt/TiO₂ catalyst concentration. The gas phase H₂ volume (ml at NTP) was estimated for 150 ml biocrude. As seen from Figure 3, PR of untreated biocrude did not generate any H₂ whereas 10 mg/ml and 50 mg/ml AC-treated biocrude generated 0.6 ml and 2.14 ml of H₂, respectively after 180 min.

Figure 3. H₂volume generated during PR of untreated cotton biocrude and biocrude treated with activated carbon at different concentrations

Download as

Veiw figure View Figure

Although the colorless biocrude provided a six fold increase in H₂ volume generation, increasing AC concentration to 50 mg/ml contributed to loss of the C and H content of the biocrude. HPLC analysis of AC-treated cotton biocrude (Table 1) indicates significant removal of C₁-C₃ carboxylic acids as compared to the untreated biocrude. Alternate methods to selectively remove only the interfering compounds without affecting the carboxylic acids and sugars are currently under investigation. It can also be noticed from Table 2 that several oxygenated hydrocarbon compounds including the major ones such as 1H-pyrazole, 1,5-dimethyl-, 1,23-benzenetriol, 1,2,4 benzenetriol and 1,2 cyclopentanedione, 3-methyl have been removed by the AC treatment. It is to be noted that the AC has significant Ni ions adsorption capacity (97.8%) ^[18]; therefore, we inferred that negligible Ni ions were present during PR of AC-treated biocrude.

3.4. Photocatalysis of Biocrude Acid Component and AC-treated Biocrude

To further understand the role of the four major acid components of the biocrude towards the H₂ production, photoreforming of 50 mM LA, AA, FA and PA was carried out using procedure outlined in the Experimental section. Clearly, as seen from Figure 4 the acids differ in relative proficiencies as electron donors. Formic acid which is structurally simplest C₁ acid with one α hydrogen appears to be best electron donor producing maximum H₂ volume during 180 min reaction time. Under identical PR conditions, formic acid yielded 9.7 ml H₂ at NTP, which was 133 fold, 12.77 fold, and 2.95 fold higher than AA, PA and LA, respectively. LA, which is major acid in the biocrude, is also a better electron donor compared with the other two C₁-C₃ acids. On the other hand, acetic acid is relatively recalcitrant to photoreforming to H₂. LA, a C₃ acid differs from PA by the presence of a single hydroxyl group. The hydroxyl group in addition to the carboxyl group probably increases adsorption affinity of LA on the TiO₂ surface. This could facilitate better transfer of the photogenerated electrons to LA generating higher H₂yields. A mixture containing 50 mM of FA, AA, PA and LA was also investigated for H₂ generation under identical PR conditions. Although the total H₂ content of this mixture was higher, the H₂ volume generated (6.65 mL) was about 1.5 fold lower than produced by the FA during PR, however, it was higher than LA, PA, and AA. This suggests that recalcitrant acids such as AA and PA in the acid biocrude probably reduce H₂ generation. This warrants further investigation of Pt/TiO₂ surface reactions of acid mixture during PR. In a separate experiment LA, PA, AA and FA were mixed in the concentration ratio found in the cotton biocrude. This acids mixture is referred to as ‘biocrude acid simulant’. During 180 min of PR reaction, 1.16 mL H₂ was generated from the biocrude simulant. The biocrude stimulants were prepared with 0.1 and 10 fold acids concentrations to study the effect of concentration on H₂ evolution. The 0.1 fold biocrude stimulant generated significantly higher H₂ volume (4.13 mL), whereas 10-fold biocrude stimulant generated only 0.8 ml at NTP during PR. As the catalyst loading was held constant (0.4 mg/ml), higher number of active sites were available for the acid moieties present in 0.1 fold biocrude stimulant, which reflected into higher H₂ volume generation.

Figure 4. Comparative H₂ volume generated during PR of carboxylic acids and their mixture

Download as

Veiw figure View Figure

It can be concluded that under the experimental conditions employed in this study, NiSO₄ catalyzed HTL process generates sufficient quantities of acids (~79 wt%) in the downstream biocrude. The biocrude can be used without dewatering or concentrating the acids for PR to enhance H₂ generation.

3.5. Combined Yields of H₂ from Integrated HTL and PR Processing of Cotton

To investigate total H₂ volume generation capacity of the biocrude simulant and actual biocrude during PR, multiple PR cycles were performed in the time interval of 180 min. After 180 min of PR reaction, the reactor headspace was evacuated by purging with N₂ until H₂ was undetectable in the GC analysis. The reactor was again exposed to simulated solar irradiation and next PR cycle was performed for 180 min. The H₂ volume generated from the biocrude acid simulant during four PR cycles is shown in Figure 5a. During PR cycle-1, cycle-2, cycle-3 and cycle-4, the maximum H₂ volume of 5.35 ml, 3.75ml, 1.98 ml, and 0.75 ml, respectively was observed after 180 min reaction time. The H₂ volume generated from the actual biocrude during four PR cycles is shown in Figure 5b. The maximum H₂ volume observed during PR cycle-1, cycle-2, cycle-3 and cycle-4 was 2.17 ml, 1.6 ml, 0.87 ml and 0.56 ml, respectively. These results indicate lower H₂ volume generation from the actual biocrude as compared with the biocrude acid simulant, which infers that the additional oxygenated compounds (Table 2) present in actual biocrude probably inhibited some of the H₂ generating reactions. The volumes of H₂ generated from each PR cycle were combined to obtain the total H₂ volume.

Figure 5. a: H2 volume generated during PR of C₁-C₃ acids mixture, and b) AC-treated cotton biocrude

Download as

Veiw figure View Figure

The H₂ yield (vol%) was determined as per the following equation-

The total H₂ volume of acids mixture was estimated based on the theoretical H₂ content of all acids and their concentrations used to prepare the acids mixture. For biocrude, the H₂ yield was determined as the ratio of total H₂ volume generated in all PR cycles and H₂ volume content of 150 ml AC-treated biocrude. The total H₂ volume for the AC-treated biocrude was estimated based on the acids concentration reported in Table 1. These acids were present in the biocrude after AC treatment. The H₂ yields in terms of vol% are presented in Figure 6. It shows a total H₂ vol% produced from actual biocrude during four PR cycles was 17.82%. When H₂ vol% from PR and HTL is combined, a total of 20.93 vol% H₂ was extracted from cotton biocrude. This shows that by integrating HTL and PR processes, higher H₂ volume generation could be achieved. A maximum of 57.5 vol% H₂ was observed from biocrude acid simulant that contained only carboxylic acids during four PR cycles. This leads to the notion that the acids can be separated from the biocrude and treated with PR for higher H₂ volume generation.

Figure 6. Total H₂ (vol%) extracted by integrating HTL of biomass and PR of HTL derived biocrude

Download as

Veiw figure View Figure