Hydrothermal Upgrading of Algae Paste: Inorganics and Recycling Potential in the Aqueous Phase

Hydrothermal Upgrading of Algae Paste: Inorganics and Recycling Potential in the Aqueous Phase

Bhavish Patel, Miao Guo, Chinglih Chong, Syazwani Hj Mat Sarudin, Klaus Hellgardt*

Imperial College London, Department of Chemical Engineering, Exhibition Road, South Kensington, London SW7 2AZ, UK.

*Corresponding Author – Email: ; Tel: +44 (0)20 7594 5577

Abstract

Hydrothermal Liquefaction (HTL) for algal biomass conversion is a promising technology capable of producing high yields of biocrude as well as partitioning even higher quantity of nutrients in the aqueous phase. To assess the feasibility of utilising the aqueous phase, HTL of Nannochloropsis sp. was carried out in the temperature range of 275 to 350°C Residence Times (RT) ranging between 5 and 60 minutes The effect of reaction conditions on the ions as well as Chemical Oxygen Demand (COD) and pH was investigated with view of recycling the aqueous phase for either cultivation or energy generation via Anaerobic Digestion (AD), quantified via Lifecycle Assessment (LCA). It addition to substantial nutrient partitioning at short RT, an increase in alkalinity to almost pH 10 and decrease in COD at longer RT was observed. The LCA investigation found reaction conditions of 275°C/30 min and 350°C/10 min to be most suitable for nutrient and energy recovery but both processing routes offer environmental benefit at all reaction conditions, however recycling for cultivation has marginally better environmental credentials compared to AD.

Keywords: Hydrothermal Liquefaction (HTL), nutrient recycling, biocrude, LCA, Anaerobic Digestion (AD), microalgae

Graphical Abstract

1.0 Introduction

It is well established that the nutrients required for algal cultivation incur significant costs and results in significant environment burdens owing to the energy and resources employed during their synthesis/production (Clarens et al. 2010). Nitrogen and Phosphorus based salts in particular are required in large quantities to support algal growth (Clarens et al. 2010, Johnson et al. 2013 & Patel et al. 2012), and when implemented at scale, these inorganics may well become a limiting factor, especially considering the impending peak Phosphorus theory suggesting a shortage in coming decades with no real substitute possible (Beardsley, 2011). Even if sea water were used for commercial cultivation, the replenishment of nutrients, such as nitrogen-based ammonium salts, phosphorus and sulphates would still be necessary (Chebil and Yamasaki, 1998). Demirbas (2010) estimated that for commercial production of microalgae, approximately 8 to 15 tons of fertilizers are required per hectare per year. The production of fertilizers is energy intensive and requires substantial resource inputs, which is one of the bottlenecks for achieving a high economic return for mass cultivation of algae to date (Clarens et al. 2009). Consequently, the preparation of growth medium also contributes considerably to the high biofuel production cost (Molina Grima et al, 2003), particularly where a saline growth medium is needed. Most of these nutrients are absorbed by algal cells to synthesise and maintain their chemical composition as well as survive and therefore nutrient reclamation would involve direct removal/extraction from algal biomass, ideally without additional treatment or affecting the product pool.

Amongst various processing technologies for biomass transformation to fuel/chemicals, treatment of wet algae paste under elevated temperature and pressure via Hydrothermal Liquefaction (HTL) is considered to be a promising route, especially since HTL also makes the recovery and recycling of inorganic nutrients possible (Patel et al. 2015 & Peterson et al. 2008). Nutrient recycling is especially important because not all the input nutrients are consumed completely by algae and some end up in the aqueous phase after HTL. Furthermore, during hydrothermal treatment the inorganics bound to the biomass can be retrieved as it is not desirable to have these in the oil phase, thus alleviating special treatment. Therefore, recycling the nutrient-rich and carbon-containing aqueous phase after HTL could be a solution via nutrient reclamation for algal cultivation, which would help to reduce the input of fertilizers; another potential use is aqueous stream recycling for energy generation via Anaerobic Digestion (AD).

The concept of using HTL produced water for growth has already been demonstrated by Biller et al. (2012) who recycled HTL (at 300 and 350°C) aqueous phase for cultivation of four algae species, namely Chlorella vulgaris, Scenedesmus dimorphus, Spirulina platensis and Chlorogloeopsis fritschii. This study indicated the feasibility of recycling the HTL aqueous phase based on the high concentration of a range of nutrients required for algae growth, this has been confirmed by other studies which showed successfully cultivated algae using various concentrations of HTL aqueous phase (Alba et al. 2013, Biller et al. 2012 and Nelson et al. 2013). However, the effect of Residence Time (RT) on inorganics concentration in the aqueous phase is unknown. Particularly, recent studies (Faeth et al. 2014, Patel & Hellgardt, 2015 and Patel & Hellgardt, 2013) have suggested that HTL can be carried out at shorter RT and as a result, it is necessary to observe the partitioning of inorganics with respect to RT and to evaluate the feasibility to re-use the aqueous phase obtained at these reaction conditions. An alternative to nutrient reclamation is to digest the carbon-containing aqueous phase and recover energy via AD. However, deciding the most desirable method for HTL aqueous phase recovery should take into account economic and environmental variables. To elucidate the knowledge gap, this study focuses on the environmental aspects, where the holistic environmental impacts of two alternatives for aqueous phase treatment are compared using an LCA approach. Economic evaluation will be explored in future research.

This manuscript investigates the concentration of recoverable nutrients and the potential to re-use/recycle the aqueous phase. Experiments were carried out using a batch reactor to investigate the effects of processing temperatures (275, 300, 325 and 350°C) and residence time (5, 10, 15, 20, 30, 45 and 60 minutes) on the concentration of sodium, potassium, sulphates, phosphates, nitrates and trace metals in the aqueous phase. In addition, the experimental data is fed into an LCA model to investigate the environmental impacts for aqueous phase utilisation as cultivation medium and AD feedstock. To our best knowledge, the analysis of inorganics with respect to RT of HTL reaction and subsequent environmental assessment of recovered aqueous phase has not been addressed in any publically available literatures.

2.0 Experimental

Freeze dried Nannochloropsis sp. algae was obtained from the University of Almeria. All chemicals used for standards, sample preparation and extraction protocol were obtained from Sigma Aldrich Ltd. (unless stated otherwise) and used as purchased (purity >98%). The batch reactors were constructed in-house using Stainless Steel (316L) tubes and cap endings on either side, purchased from Swagelok®.

2.1 Reaction

The hydrothermal liquefaction of the microalgae was carried out in ½ inch outer diameter 316L stainless steel batch reactors of volume 6 cm3 in an isothermal recirculating oven. To expose the reactors to hydrothermal conditions and condition them, all fabricated reactors were filled with De-Ionised Water (DIW) and placed in the oven for 3 hours at 380°C. In a typical run, a fresh stock supply of 10 wt.-% algae paste was made and 3 gm of this stock solution was added to the reactor and the reactor sealed. It was then placed in the oven at required temperature (275, 300, 325, 350°C) and removed at designated Residence Time (RT) 5, 10, 15, 20, 30, 45, 60 min). The reactor was immediately quenched in ice and left overnight to equilibrate. The mass of reactor (+ contents) before and after the reaction was measured to ensure no leakage occurred during the reaction.

2.2 Product Extraction

The quenched batch reactor was slowly opened to release any produced gas and rinsed with Dichloromethane (DCM) to extract the products. A spatula was used to remove char stuck to the reactor walls. The product was left to settle for 2 hours to allow the organic rich DCM and aqueous phase to separate. A 2ml aliquot of the top aqueous phase was taken for analysis and filtered through a 0.22µm membrane filter. The remaining aqueous phase was stored in a fridge. After allowing the filtrate to separate in a separating funnel, the lower DCM solubilised biocrude phase was filtered, separated and run through anhydrous MgSO4 column to remove residual water. Moisture free biocrude was obtained by evaporating DCM over a steady stream of Nitrogen at room temperature for 5 hours. The pH of the extracted aqueous phase was also measured.

2.3 Feedstock Inorganic Composition

The initial inorganic (trace metals) composition was determined as follows. A sample of 0.15-0.30 gm freeze dried algae was added to a 60ml Teflon vessel (Savillex), with 2ml of concentrated Nitric Acid at 70°C for 4 hours followed by the addition of 5ml of concentrated Hydrochloric Acid and left to evaporate for 3 hours. The dried sample was then subjected to addition of 1ml 70% Perchloric Acid at 100°C and left for 2 hours in an open vessel. The vessel was then tightly closed and kept at 130°C for 48 hours. After cooling adding 0.2ml 40% Hydrofluoric Acid the vessel was placed in an oven at 130°C for 4 hours. The samples were then dried at 150°C prior to preparation for Inductively Coupled Plasma Atomic Emission Spectroscopy (ICP-AES) analysis using 7% Nitric Acid.

2.4 Analysis

2.41 Feedstock Inorganic Composition

ICP-AES analysis was conducted on a Thermo Scientific iCap Duo 6500. The calibration samples were prepared from single element stock solution in 5-10% Nitric Acid. Au and Rh at 5 ppm. was added inline as internal standard to account for matrix effects.

2.42 Ash Content

The ash content of the initial algal biomass was determined according to the National Renewable Energy Laboratories Laboratory Analytical Procedure (Sluiter et al. 2005). The algae was treated in a furnace at 105°C for 12 min after which the temperature was increased at 10°C/min to 250°C and held for 30 min. A final temperature of 550°C and hold time of 180 min was applied at a ramp pate of 20°C/min. Finally, after allowing the furnace to cool to 105°C, samples were removed and measured.

2.43 Trace Metal Composition

The concentrations of metal ions; sodium (Na+), potassium (K+), and nickel (Ni2+), were determined using a ICP-OES (Perkin-Elmer Optima 4300 DV). Due to the detection limit of ICP-OES (concentrations up to 30 ppm), the aqueous phase samples were diluted to give 0.5 % of its original concentration (dilution factor of 200). Commercially available standard solutions were analysed to deriving calibration curves.

2.44 Inorganic Salts Composition

The concentrations of anions nutrients, namely nitrate (NO3-), sulphate (SO42-) and phosphate (PO43-), contained in the aqueous phase were determined using a 882 Compact IC Plus Conductivity Detector Ion Exchange Chromatography (ICP - IEC) unit. A dilution factor of 40 was used to stay within the detection range.

2.45 Chemical Oxygen Demand (COD)

The COD test is used to determine the amount of organics present in the aqueous phase which involves two steps - oxidation of the samples using potassium dichromate (K2Cr2O7) digestion solution and COD determination using Ultraviolet – visible (UV-VIS) spectrophotometer. The digestion solution was prepared by adding 10.2 gm of dried (103°C for 2 hours) K2Cr2O7, to 167 ml of concentrated sulphuric acid (H2SO4) and 33.3 gm of mercury sulphate (HgSO4) to 500 ml of distilled water. This digestion solution was then diluted by adding distilled water to make up a 1000 ml solution. 0.6 ml of this solution was added to 1 ml of aqueous sample in a COD tube which was followed by the addition of 0.4 ml of concentrated H2SO4. The tube was capped and inverted several times to ensure a well-mixed solution. The tubes were then heated at 150°C for 2 hours and left to cool overnight. The COD concentration was determined by UV-VIS measurement where potassium hydrogen phthalate (KHP, C8H5KO4) was used for calibration.

3.0 LCA methodology

Generally, two types of LCA have been distinguished: Attributional LCA (ALCA) and Consequential LCA (CLCA) approaches. The former is focused on describing the environmentally relevant flows to and from a life cycle and its subsystems, while the latter aims to study how environmentally relevant flows change in response to possible decisions. In this study, an ALCA approach was applied to examine two potential options for utilizing aqueous phase – direct nutrient recycling for algae growth vs. energy recovery from AD treatment. As illustrated in Figure 1, nutrient recovery and AD units were included in this LCA comparison study along with its appropriate product. The functional unit was defined as ‘treating aqueous phase generated from production of per kg biocrude oil at biorefinery gate’. A ‘system expansion’ approach was applied to demonstrate the potential environmental benefits of nutrient or energy recovery. The surplus electrical and thermal energy generated from combustion of biogas at AD unit was modelled as net co-products export after accounting for the energy requirement for AD operation and assumed to substitute an equivalent amount of electricity and heat generation from UK national grid. The ‘functional equivalent’ quantity of commonly applied inorganic N or P fertilizers in the UK was allocated as an ‘avoided burden’ to nutrient recovery unit. Life cycle impact assessment (LCIA) methodologies can be categorised as midpoint and endpoint-oriented approaches which are also termed as ‘problem-oriented’ and ‘damage approach’ respectively. The former is chosen along with environmental mechanisms between the life cycle inventory (LCI) results and endpoints (ISO, 2000) and the latter is defined at the level of protection area (Finnveden et al., 2009). A midpoint approach developed by the Centre of Environmental Science (CML) of Leiden University - CML 2 baseline 2000 - was applied as characterisation method in this study at LCIA stage where the evaluation focused on six impact categories - abiotic depletion, global warming potential (GWP100), acidification, eutrophication, ozone depletion (ODP) and photochemical oxidation (POCP). The LCA model was implemented in Simapro 7.3 (PRe Consultants).

Figure 1 - LCA system boundary defined for AD energy recovery and nutrient recycling from HTL aqueous phase

It should be noted that the gas production was ignored in this instance as it does not form part of the study due to its minute volume and composition (mostly CO2).

4.0 Results

4.1 Initial Feedstock Composition

Table 1 shows the trace metal concentration present in the initial algal feedstock. Metals of interest (underlined) were chosen for analysis in the processed aqueous fraction after HTL. The ash content of the algae was found to be 11 wt.-%. The biocrude yields can be found in SI-3.

Table 1 –Concentration of metals present in untreated Nannochloropsis sp. biomass feedstock

4.2 Aqueous Phase:

4.21 Nitrates

Nitrates are derived from the nitrogen-containing substrates, mostly proteins present in the alga cell. Ideally, it would be beneficial to reclaim nitrates in the aqueous phase so that produced biocrude contains low nitrogen and also facilitates the reclamation of nitrogen rich aqueous phase. For the purpose of elucidation of nitrogen partitioning during HTL, it is assumed that all the nitrogen in the cells is converted to nitrates, which gives the initial (t0) concentration of nitrates to be approximately 37,000 ppm based on N content given in Table 1. The calculation can be found in SI-2.

Figure 2 shows that the concentration of nitrates decreases with increasing residence time and temperature. For reactions carried out at 325°C and 350°C, significant changes in nitrates concentration were observed for RT up to 15 minutes. For reactions carried out at 350°C, the most significant decreases in nitrate concentration, approximately 84% and 61%, occurred between 5 to 10 minutes and 10 to 15 minutes, respectively. This suggests that most of the protein degradation of the components in Nannochloropsis sp. that yields nitrate was completed in less than 15 minutes at high reaction temperatures. The short protein degradation time is consistent with that observed in the oil phase by Eboibi et al. (2015) and Patel and Hellgardt (2013). At reaction temperatures 275°C and 300°C, the concentrations of nitrate decreases gradually from a RT of 5 to 60 minutes. This indicates that the degradation reactions that produce nitrates were not completed in a short time period as is the case for higher temperature, elucidating that severe conditions result in less nitrate in the aqueous phase.

Figure 2 – Concentration of Nitrates in HTL processed aqueous phase

However, nitrogen could also be present as nitrite, ammonium and (dissolved) ammonia from degradation of proteins (Lourenço et al.2004). The inverse relationship between nitrate concentrations and RT as well as temperatures is the opposite to that of ammonia concentration. It was reported that in general the concentration of ammonia increases with temperature and residence time. This is due to the more favourable formation of ammonia from organic nitrogen at higher liquefaction temperatures (Valdez et al. 2012).

4.22 Phosphates

The concentration of phosphate at t0, which is 2040ppm, was calculated using the same method as nitrate (SI-2) and all the phosphorus was assumed to be converted to phosphates. There is a rapid decrease within the first 5 minutes of the reaction after which the concentration stabilises as seen from Figure 3. A substantial amount, about 39 to 51% (794 to 1040ppm) of the phosphorus in Nannochloropsis sp., has partitioned into the aqueous phase. The rest of phosphorus could have ended up in the bio-crude and solids (Valdez et al. 2012). The values reported in previous studies (Alba et al. 2013, Biller et al. 2012 and Valdez et al. 2012) vary significantly, which can be explained by the varying RT and different analysis method. Biller et al. (2012) found the phosphorus concentration in the aqueous phase decreases with increasing temperature given a fixed residence time and suggested that higher HTL temperature is not favourable for phosphorus retention. However, based on the temperature and RT investigated in this study, the data suggest this is not the case and the major change occurs within the first 5 minutes.

Figure 3 – Concentration of phosphates in HTL processed aqueous phase

4.23 Sulphate

The sulphate equivalent of the sulphur present in Nannochloropsis sp. is 2000ppm (SI-2). From Figure 4, the sulphate concentrations remained largely unchanged with temperature and residence time. Sulphate ions once formed associate with the metal ions present, such as sodium and potassium, and partition into the aqueous phase (UNIDO, 1980). The thermal stability of sodium sulphate is high and depending on the heating rate, the decomposition starts only at temperature much higher (around 850°C) than that used for HTL. Since magnesium ions are commonly present in seawater, the presence of magnesium sulphate in the aqueous phase is also possible. The onset decomposition temperature of magnesium sulphate was reported to be around 540°C (Ebert et al, 1997). Therefore, the high thermal stabilities of the sulphate ions could be the reason why the sulphate concentration in the aqueous phase does not change with temperature. Secondly, some sulphur based compounds also partition to the organic phase. Valdez et al. (2011) found significant production of dimethyl disulphide in the DCM extracted organic phase after 60 min of HTL at 350°C.