Вы находитесь на странице: 1из 41

Journal Pre-proofs

Bio-Crude Oil from Hydrothermal Liquefaction of Wastewater Microalgae in a


Pilot-Scale Continuous Flow Reactor

Feng Cheng, Jacqueline M. Jarvis, Jiuling Yu, Umakanta Jena, Nagamany


Nirmalakhandan, Tanner M. Schaub, Catherine E. Brewer

PII: S0960-8524(19)31414-2
DOI: https://doi.org/10.1016/j.biortech.2019.122184
Reference: BITE 122184

To appear in: Bioresource Technology

Received Date: 2 August 2019


Revised Date: 18 September 2019
Accepted Date: 20 September 2019

Please cite this article as: Cheng, F., Jarvis, J.M., Yu, J., Jena, U., Nirmalakhandan, N., Schaub, T.M., Brewer, C.E.,
Bio-Crude Oil from Hydrothermal Liquefaction of Wastewater Microalgae in a Pilot-Scale Continuous Flow
Reactor, Bioresource Technology (2019), doi: https://doi.org/10.1016/j.biortech.2019.122184

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will
undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing
this version to give early visibility of the article. Please note that, during the production process, errors may be
discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2019 Elsevier Ltd. All rights reserved.


Bio-Crude Oil from Hydrothermal Liquefaction of Wastewater Microalgae in a Pilot-

Scale Continuous Flow Reactor

Feng Chenga, Jacqueline M. Jarvisb, Jiuling Yua, Umakanta Jenaa, Nagamany

Nirmalakhandanc, Tanner M. Schaubb, Catherine E. Brewera,*

a
Department of Chemical and Materials Engineering, New Mexico State University, P.O.

Box 30001 MSC 3805, Las Cruces, NM 88003, USA.


b
Department of Plant and Environmental Sciences, New Mexico State University, Las

Cruces, NM 88003, USA.


c
Department of Civil Engineering, New Mexico State University, P.O. Box 30001 MSC

3CE, Las Cruces, NM 88003, USA.

* Corresponding author. Tel.: +1 575 646 8637; fax: +1 575 646 7706.

E-mail: cbrewer@nmsu.edu

Declarations of interest: none.

Abstract

To explore the feasibility of scaling up hydrothermal liquefaction (HTL) of algal biomass, a

pilot-scale continuous flow reactor (CFR) was operated to produce bio-crude oil from algal

biomass cultivated in urban wastewater. The CFR system ran algal slurry (5 wt.% solids

loading) at 350 °C and 17 MPa for 4 hours without any clogging issues. Bio-crude oil

chemistry was characterized by high-resolution Fourier transform mass spectroscopy (FT-

MS), proton nuclear magnetic resonance spectroscopy (1H NMR), bomb calorimetry, and

Page 1 of 40
elemental analysis. Bio-crude oil yield of 28.1 wt.% was obtained with high heating values

of 38-39 MJ/kg. The quality of light bio-crude oil produced from the CFR system was

comparable in terms of molecular structures to bio-crude oil produced in a batch reactor.

Keywords: wastewater microalgae; hydrothermal liquefaction; continuous flow reactor;

bio-crude oil

1. Introduction

Many first- and second-generation biomass feedstocks are not suitable for energy

recovery due to high pretreatment costs and/or competition with food production (Demirbas

& Demirbas, 2011). To mitigate these issues, algae-based fuels have emerged rapidly

because of higher fuel product yields, higher fuel energy density, better fit with existing

fuel and industrial infrastructure, higher photosynthesis efficiency of algal biomass (Schenk

et al., 2008), and greater diversity of suitable growing conditions. Use of algae cultivated in

wastewater as a biofuel feedstock can synergistically enhance energy efficiency for both

wastewater treatment and for bio-energy production (Chen et al., 2014b; Selvaratnam et al.,

2016).

Technologies for algal biomass valorization include hydrothermal liquefaction (HTL)

(Peterson et al., 2008), catalytic hydrothermal gasification (Elliott, 2008), fast pyrolysis

(Bridgwater, 2012), and anaerobic digestion (Van Doren et al., 2017). HTL of wet biomass

has the advantages of minimized dewatering requirements (Peterson et al., 2008) and higher

conversion rates over other conversion techniques (Vardon et al., 2012). Bio-crude oil

yields of up to 66 wt.% (Faeth et al., 2013) were previously reported with higher heating

Page 2 of 40
values (HHV) of 36-40 MJ/kg (Cheng et al., 2018), which are comparable to those of

petroleum-derived oils. Even low-lipid wastewater algae can be converted into bio-crude

oils with high yields and comparable energy content as those produced from high-lipid

microalgae (Chen et al., 2014b). An acidophilic and thermophilic microalgae, Galdieria

sulphuraria, has been shown to be a promising algal species for wastewater treatment due

to its metabolic flexibility in utilizing dissolved organic carbon and nutrients in wastewater

as well as CO2 mixotrophically (Selvaratnam et al., 2016). Our previous study on batch

HTL of wastewater-treatment Galdieria sulphuraria (WWGS) algal biomass showed that

WWGS has potential for producing algae-derived biofuel as long as higher temperatures

(330-350 °C) and intermediate reaction times (~30 min.) are used to account for the higher

protein and carbohydrate contents (Cheng et al., 2019).

HTL reactions in the reported studies were usually conducted in batch reactors (Faeth

et al., 2013; Vardon et al., 2012). Scale up and implementation of HTL at an industrial

scale will require the development of continuous flow reactor systems. Even though work

on continuous reactors has been ongoing since the 1980s (Lindemuth, 1981), there are only

a few lab-scale continuous reactors, most of which were built recently (Ocfemia et al.,

2006). With advances in pumping, filtration and oil-aqueous phase separation systems, the

newest continuous reactors can handle biomass slurries with higher solids loadings (Billing

et al., 2016), and can effectively remove HTL solids, enabling continuous bio-crude oil

production for longer time periods between shut downs (Mørup et al., 2015).

Page 3 of 40
In this study, continuous HTL of both a monoculture and polyculture of G. sulphuraria,

grown in a pilot-scale testbed deployed at a local municipal wastewater treatment plant,

was carried out in a modified pilot-scale CFR system. The purpose of the study was to

evaluate continuous HTL performance against batch HTL under the same conditions in

terms of bio-crude oil yield and chemistry, and bioenergy recovery potential.

2. Materials and methods

2.1 Algae production

Two types of wastewater-grown microalgae, G. sulphuraria monoculture 5587.1

(WWGS), and G. sulphuraria polyculture (WWPC), were grown on primary settled

municipal wastewater in 700 L pilot photobioreactors at the Jacob A. Hands Wastewater

Treatment Facility (Las Cruces, NM). WWGS was grown during the warmer months of the

year under acidified conditions as described in (Cheng et al., 2019). WWPC was produced

as a “natural” polyculture from January to March 2018, believed to contain G. sulphuraria,

other unidentified algae species, and some bacteria. No growth medium components or

enrichment of the bioreactor headspace with carbon dioxide were used. Bioreactor

temperature, pH (without modification), ammonium nitrogen, phosphate, and biochemical

oxygen demand were recorded to monitor wastewater treatment performance. The settled

biomass was harvested by gravity settling followed by centrifugation (Avanti J-26 XP

centrifuge, Beckman Coulter, Brea, CA) in 6 L batches at 10,000 rpm for 5-10 min. The

centrifuged samples were frozen and stored at -20 °C prior to use. Hexane used was

analytical grade (>99%) (analytical grade, Pharmco-Aaper, Shelbyville, KY).

Page 4 of 40
2.2 Modified CFR configuration

The pilot-scale CFR system was modified from a supercritical fluid bioreactor

(Supercritical Fluid Technologies, Inc, Newark, DE, USA) and included five sub-systems

(Figure 1): 1) supply (stirred feed tank and feed pump), 2) reaction (preheater and plug flow

reactor), 3) separation (cylinder filters and blowdown pots), 4) letdown (cooler and

backpressure regulator), and 5) auxiliary (temperature/pressure gages and controls, pressure

relief, gas booster, etc.) Algae slurry is loaded into the feed tank equipped with an agitator.

During operation, the slurry passes through a cylinder filter to remove large inorganic

particles, then is pumped by a high-pressure pump into a preheater. The slurry is then

heated to the desired reaction temperature as it flows upward through a vertical plug-flow

reactor, where the tubing diameter increases from 0.21 cm to 0.51 cm. Exiting from the

reactor top, HTL products flow into one of two custom-made high-pressure filter vessels

(9.53 cm I.D.) where HTL char is removed from the liquid/gas products by filter elements

with surface areas of 405 cm2. During switches between parallel filter systems, HTL char is

collected in customized blowdown pots underneath the filter vessels (Figure 1). Char-free

liquid/gas products flow out of the filter vessel, through a water-cooled heat exchanger, and

then through a back-pressure regulator (BPR) capable of maintaining a reactor pressure of

20 MPa. The now-ambient temperature and pressure HTL product mixture is collected in a

product drum, and gases are vented through an activated carbon air filter. Bio-crude oil and

aqueous phases separate within the product drum. Additional details about the CFR system

and the modifications made are provided in the MethodsX article.

Page 5 of 40
2.3 Hydrothermal liquefaction of wastewater-grown algae

HTL of the algae slurries, with solids loading of 3 and 5 wt.%, was carried out at 325-

350 °C, at flow rates of 152-155 mL/min with theoretical reaction residence times (reactor

volume divided by slurry flow rate) of 3-9 min. Total reactor operation time was 7 hours.

HTL liquid products, 800 mL samples of bio-crude oil/aqueous phase mixture, from the

product drum and HTL char from the blowdown pots were collected. Hexane-soluble

“light” bio-crude oil (LBO) was recovered from the liquid samples using a hexane

extraction procedure (Cheng et al., 2017). LBO was stored in sealed glass containers at 4 ºC

prior to characterization.

2.4 Evaluation of mass and energy balances

Yield of LBO on a dry basis (d.b.) was calculated from:

LBO content in liquid sample × HTL liquid mass flow rate


𝑌LBO(d.b.) = × 100% (1)
Dry algal solids loading × algal mass flow rate

Yield of LBO on a dry ash-free basis (daf.) was calculated from:

Y
𝐿𝐵𝑂 (d.b.)
𝑌LBO(daf.) = (1−ash (2)
content)

Energy recovery (ER), defined as the ratio of the energy available in LBO to the energy

available in the feedstock, was calculated according to:

HHVLBO
ER = HHV × 𝑌LBO(d.b.) (3)
Feedstock

where HHV is higher heating value and LBO is hexane-soluble bio-crude oil.

Energy consumption ratio (ECR), defined as the ratio of the energy consumed for

feedstock conversion into LBO to the energy produced from LBO combustion, was

estimated by:

Page 6 of 40
𝑄
[𝑅𝑤 ×𝐶𝑝𝑤 +(1−𝑅𝑤 )×𝐶𝑝𝑠 ]×(𝑇−20)×(1−𝑟2 )+ 𝐿𝑜𝑠𝑠⁄𝐹+(𝑃𝑠𝑦𝑠𝑡𝑒𝑚 −𝑃𝑎𝑡𝑚 )/𝜌
ECR = (1−𝑅𝑤 )×𝑌LBO(d.b.) ×HHVLBO ×𝑟1
× 100% (4)

where RW is the moisture content in wt.%, Cpw and Cps are the average specific heats of

water (4.18 kJ kg-1 K-1) and dry algal biomass, (1.25 kJ kg-1 K-1), respectively (Minowa et

al., 1995), T is the reaction temperature in °C, YLBO(d.b.) is the oil yield on a dry basis in

wt.%, HHVLBO is the higher heating value of LBO in MJ/kg, r1 is the assumed efficiency of

the combustion process (0.6), r2 is the assumed efficiency of heat recovery (0.5) (Minowa

et al., 1998), QLoss is the estimated heat loss from the reactor to the environment (>580 W),

which can be found in the MethodsX, F is the mass flow rate in g/s., Psystem and Patm are the

operating pressure and atmospheric pressure in N/m2, respectively, and ρ is the density of

feedstock slurry in kg/m3. Energy consumption for the continuous HTL process includes

energy for pumping the slurry, heating the feedstock, and losses from the heaters. When

ECR is less than 1, net-positive energy can be gained due to LBO production.

2.5 Characterization of algal feedstock and HTL products

Moisture content of algal biomass was measured using a FreeZone 12 L Console

Freeze Dry System with stoppering tray dryer (Labconco, Kansas City, MO, USA). Ash

content was measured by combustion of samples at 585°C for 240 min using a

programmable box furnace (Cole–Parmer, Vernon Hills, IL, USA) (Van Wychen &

Laurens, 2013). Lipid and carbohydrate contents in the algal feedstock were measured

using methods developed by the National Renewable Energy Laboratory (Wychen &

Laurens, 2013; Wychen et al., 2013).

Proteins were extracted following (Slocombe et al., 2013) with several modifications.

Protein quantification was performed with PierceTM 660 nm Protein Assay Reagent

Page 7 of 40
(Thermo, US) following the user guide microplate procedure instructions

(Pierce_Biotechnology, 2013). 20 ± 5 mg lyophilized cells were homogenized in 1 mL ice-

cold 10% (w/v) TCA in acetone, supplemented with 1 mM PMSF, 25 mM DTT and 20

mM EDTA, with 0.5 mm glass beads for (3) 60 s cycles at 6500 rpm with a Precellys 24

homogenizer (Bertin, US). Cell homogenates were incubated at -20 °C for 90 min,

centrifuged (12000 x g, 20 min, 4 °C), and pellets were then washed twice with ice-cold

choloroform:methanol (2:1 v/v) to remove residual TCA. Pellets were air-dried for 5 min

to remove organic solvents, then resolubilized in lysis buffer (4% CHAPS, 8 M urea, 2 M

thiourea, 25 mM DTT, 25 mM tris pH = 7.5). Serial dilutions (1:1) of BSA (1 mg/mL) in

lysis buffer were quantified and used to construct a standard curve to quantify protein

samples.

Surface morphology of algae and HTL char particles was observed using an S-3400N

Type II scanning electron microscope (SEM) (Hitachi High-Technologies Corp.,

Pleasanton, CA, USA) and analyzed using FEI Nova NanoSEM 630 software. Viscosity of

algae slurry was measured using a CANNON® Model 2020 viscometer (Cannon

Instrument Company®, State College, PA, USA) with a LV-1 spindle. HHV of LBO

samples (20-30 mg) was determined using a Model 6725 semi-micro bomb calorimeter

(Parr Instrument Company, Moline, IL, USA). HHV of algal feedstock was estimated using

CHNS data (Eboibi et al., 2014).

Elemental CHNS content in HTL products was analyzed using a Series II 2400

elemental analyzer (Perkin Elmer, Waltham, MA, USA). The analyzer was calibrated using

cystine and acetanilide. Total organic carbon and total nitrogen of aqueous phase were

measured using a model TOC-VCPH analyzer (Shimadzu Corp., Kyoto, Japan) and a model

Page 8 of 40
TNM-1 analyzer (Shimadzu Corp., Kyoto, Japan), respectively. Inorganic contents of algal

biomass, HTL aqueous phase, and char were quantified using an Optima 4300 DV

inductively coupled plasma optical emission spectrophotometer (ICP-OES) (PerkinElmer,

Waltham, MA, USA) after acid digestion. The pH of the aqueous phase was measured with

a Surpass Electrokinetic Analyzer (Anton Paar USA Inc., Ashland, VA, USA). All

measurements were done in triplicate.

Thermogravimetric analysis (TGA) of dry algal feedstock and LBO were conducted

using a thermogravimetric analyzer (TGA-Q500, TA Instruments, New Castle, DE, USA)

to find the boiling temperature distribution, as described in (Cheng et al., 2018). Functional

groups of algal biomass and bio-crude oils were characterized by Fourier transfer infrared

spectroscopy (FT-IR) using a Spectrum Two spectrophotometer (PerkinElmer, Waltham,

MA, USA) and by nuclear magnetic resonance spectroscopy (NMR) using an Oxford 300

MHz NMR (Varian, Palo Alto, CA, USA) as described in (Cheng et al., 2017).

Molecular composition of LBO samples were analyzed by Fourier transform mass

spectrometry (FT-MS). The sample preparation procedure is described in (Sudasinghe et

al., 2014) with a few modifications. Bio-crude oil samples were diluted in 1:1

chloroform:methanol (HPLC grade) to a concentration of 1 mg/mL, then sequentially

diluted to a final concentration of 50 µg/mL in 1:3 chloroform:methanol with 1% formic

acid for positive-ion electron spray ionization (ESI), and to a concentration of 20 µg/mL in

1:3 chloroform:methanol with 1% ammonium hydroxide solution (28% in water) for

negative-ion ESI. Positive- and negative-ion ESI mass spectra were collected on the

Orbitrap Fusion mass spectrometer (Thermo Scientific, San Jose, CA) equipped with an

Advion TriVersa NanoMate system (Advion, Ithaca, NY). 300 microscans (eFT and full

Page 9 of 40
profile) at mass resolving power > 500,000 at m/z 200 were averaged to give the final mass

spectrum. The easy-IC ion source with fluoranthene (m/z 202) was used for internal

calibration of the mass spectra. Data was analyzed, and peak lists were generated with

PetroOrg® software.

3. Results and discussion

3.1 Characterization of algae feedstock

Algae feedstock characteristics are shown in Table 1. WWGS and WWPC had energy

contents (12-15.4 MJ/kg) comparable to other biomass (12-18 MJ/kg). The high N content

(3.4–5.6%) in the feedstock is expected to pose difficulty in upgrading the bio-crude oil

from the production of thermally stable N-containing aromatic compounds during HTL

(Cheng et al., 2017). The high ash content (34.7–42.0 wt.%) has the potential to cause

fouling in the CFR system and to inhibit the formation of bio-crude oil (Chen et al., 2014a).

WWPC had higher contents of C, H, and N, and lower O content than that of WWGS

(34.5% vs. 29.0%, 5.3% vs. 4.2%, 5.6% vs. 3.4%, and 18.6% vs. 20.3%, respectively),

resulting in a higher estimated HHV (15.7 MJ/kg vs. 12.1 MJ/kg) (Table 1). The higher N

content in WWPC indicates higher feedstock protein content than WWGS. Scanning

electron microscopy showed particle sizes of approximately 10-20 µm for the lyophilized

WWPC algae particles with a multilayer leaflet structure. The viscosity of algae slurry

increased with solids loading, from 6 to 22 cP as the solids loading increased from 8.5 to

10.5 wt.%, then rapidly to over 1000 cP, and to nearly 10,000 cP for a solids loading of 15

wt.%. High viscosity negatively impacts mass and heat transfer within the algae slurry,

creating potential for material build-up and hot spots in the CFR system.

Page 10 of 40
3.2 Characterization of light bio-crude oil

The HHV (38-39 MJ/kg) and LBO yield (28.1 wt.%, daf.) for continuous HTL of

WWGS (Table 2) was comparable to that of batch HTL of WWGS (38.4 MJ/kg and 25.7

wt.% daf., respectively) under similar operating conditions (5 wt.% solids loading, 350 °C,

and 5 min reaction time) (Cheng et al., 2019). Compared to LBO from the batch reactor,

LBO from the CFR had slightly lower C content (78 wt.% vs. 81 wt.%), similar H and N

contents (11.3 wt.% vs. 11.2 wt.%, and 4.5 wt.% vs. 4.4 wt.%), but higher O content (5.0

wt.% vs. 1.6 wt.%). This is attributed to the longer heating period in the batch reactor,

enhancing aromatization and deoxygenation. The elemental composition of 3% WWGS-

derived LBO is comparable to that of 5% WWGS-derived LBO (77.7% vs. 78.2% C,

11.2% vs. 11.4% H, and 5.2% vs. 4.2% O), suggesting a higher solids loading does not

negatively affect LBO composition relative to energy content. The energy recovery from

continuous HTL was higher (41-51%), but the energy consumption ratios of continuous

HTL were also higher (3.7-7.2), compared to those from batch HTL of WWGS (energy

recovery: 33.5% and energy consumption ratio: 2.55, respectively) (Cheng et al., 2019).

CFR has lower energy efficiency than a batch reactor mainly due to higher heat losses to

the environment and additional energy requirements for pumping. Processing higher solids

loading algal slurry, and enhancing both insulation and heat integration, are needed to

improve CFR energy consumption ratio.

WWPC-derived LBO had lower C and H contents (76.6% and 10.7%, respectively),

and less O (1.5%) than that of the WWGS-derived LBO, resulting in an overall slightly

higher HHV (38.9 MJ/kg vs. 38.2–38.8 MJ/kg). The extremely low yield of WWPC-

derived LBO is attributed to the substantial differences in feedstock composition, especially

Page 11 of 40
macromolecular structures, between WWPC and WWGS. There were also relatively high

concentrations of polar compounds within the unfractionated HTL liquid products that are

not soluble in hexane. The purpose of using hexane for LBO extraction here was to

maintain consistency with the batch process post-treatment methods. In the future,

extraction by volatile organic solvents is unlikely to be used at scale due to the additional

steps and costs, lower oil yields, and potential negative environmental impacts. Instead, (in-

situ) conversion of the heavy, polar organic compounds in the liquid product into light,

non-polar organic molecules to form a homogeneous oil phase will be a crucial area of

ongoing research.

Proximate analysis of the feedstocks and the boiling temperature distributions of

WWGS- and WWPC-derived LBO are shown in Figure 2 and Table 3. Compared to LBO,

WWGS feedstock contained much more fixed carbon and ash (Figure 2). LBO from HTL

of 5% WWGS contained more low-volatilization-temperature molecules (100-240 °C) than

the LBO from HTL of 3% WWGS due to more severe conditions for 5% WWGS (350 °C

for 6 min and 325 °C for 3 min, respectively), which contribute to greater decomposition of

large molecules. Like the WWGS feedstock, WWPC feedstock had high fixed carbon and

ash, as seen in the supplemental material. From the even more severe conditions (350 °C

for 9 min.), the WWPC-derived LBO contained more low-volatilization-temperature

molecules than the WWGS-derived LBO in this study and in (Cheng et al., 2018). WWPC-

derived LBO had 36.8 wt.%, the most abundant organic molecules, that volatilized in the

100-180 °C range, while the 5 wt.% and 3wt.% WWGS-derived LBO contained 39.4 wt.%

at 180-240 °C and 33.7 wt.% at 240-360 °C, respectively (Table 3).

Page 12 of 40
The primary functional groups of algal feedstock and LBO by FT-IR are shown in the

supplemental material. The C=O stretching (amide I band) at 1626 cm−1, C=H bending at

1456.22 cm−1, and the C-O-C stretching peak at 1187-910 cm−1 indicate the presence of

proteins, lipids, and carbohydrates, respectively. In the WWGS algae spectrum, a sharp

peak at 875 cm-1 was noticeable and was attributed to the interference of bicarbonate ion

present in wastewater sources (Mecozzi et al., 2011).

For all LBO samples, the FT-IR spectra were similar. C-H stretching (2956, 2924, and

2855 cm−1), C=O stretching (1709 cm−1), C-H bending (1457 and 1378 cm−1), and =C-H

out-of-plane bending (964 cm−1) peaks suggest the presence of unsaturated carboxylic acids

derived from lipids, and cyclic amides and/or carbonyl compounds from Maillard,

amidation, and rearrangement reactions of proteins with lipids and carbohydrates released

from the algae (Zhang et al., 2016). The broad C-O stretching peaks (1320-1120 cm−1)

indicate carbohydrates-derived molecules. The C (aromatic)-H out-of-plane bending peaks

(740-700 cm−1) indicate the presence of aromatic compounds produced from the cyclization

and condensation of proteins and carbohydrates (Yang et al., 2015). There was a broad N-H

bending peak (1613-1486 cm-1) for WWGS-derived LBO, which is almost not present for

WWPC LBO; this implies more protein-derived amide and amine compounds (Grierson et

al., 2011) which disappeared as operating conditions became more severe. WWPC LBO

possessed stronger peaks of the C (aromatic)-H out-of-plane bending at 740-700 cm−1,

indicating more aromatic compounds (in the supplemental material).

The 1H NMR spectra and the integrated peak area distributions for LBO from WWGS

and WWPC are shown in Table 4. In all LBO samples, aliphatic protons, mainly derived

from lipids, were prevalent (37-45%), followed by protons proximal to heteroatoms (17-

Page 13 of 40
20%), aromatics (14-17%), carbohydrates/ unconjugated olefins (9-12%), and

alcohol/methylene-dibenzene (7-9%), derived from proteins and carbohydrates. For

WWGS, increasing the reaction condition severity led to an increase in aliphatic protons,

and disappearance of the other types of protons, suggesting that bio-crude oil quality can be

improved by using more severe operating conditions. At 350 °C, WWGS LBO contained

more aliphatic protons (45.4% vs. 38.7%), and less of other proton types compared to

WWPC LBO. Peak heights at 0.9 ppm and 1.3 ppm were approximately equal for WWGS

LBO, whereas for WWPC LBO, the peak height at 1.3 ppm was much higher than at 0.9

ppm, indicating that methylene groups are more prevalent than methyl groups in WWPC

LBO. This suggests that branched aliphatic compounds are relatively predominant in

WWPC LBO, but not in WWGS LBO.

ESI FT-MS provides a more comprehensive and quantitative picture of bio-crude oil

chemistry (> 90% of detectable organic ions from the whole bio-crude oil) beyond the

general characteristics from FTIR and 1H NMR, and the volatile organics subset (30-40%

of the whole bio-crude oil) from gas chromatography–mass spectrometry (Herod et al.,

2007). Bio-crude oil composition differences between the CFR and previous batch

reactions (Cheng et al., 2019) are expected to be mostly attributed to reaction severity.

Based on the reaction ordinate, an integration equation used to quantify the severity of

reaction conditions (Cheng et al., 2017), continuous HTL of WWGS is less severe than the

batch process due to higher heating rates and shorter total retention times in the CFR (150-

180 °C/min vs. 2.8-3.2 °C/min, < 2 min vs. 110 min, and reaction ordinate: 7.22-8.19 vs.

8.55 under 350 °C and 5 min).

Page 14 of 40
130-180 different heteroatom classes and 18,000-24,000 unique chemical formulae

were recognized by FT-MS within the bio-crude oils analyzed (> 90% of the relative

abundance of detectable ions were assigned elemental formulae with known numbers of C,

H, N, O, and S atoms, and double bond equivalence (DBE). The most abundant heteroatom

classes were N1, N1O1, N2, N2O1 (Figure 3a), and O2 (Figure 3b) for both reactor

configurations. In the N1-2O0-1 classes, the relative abundances of species from batch HTL

are higher than that from CFR; the relative abundance decreased as the number of O atoms

per molecule increased. This indicates that the more severe conditions in the batch reaction

may facilitate bio-crude oil deoxygenation.

The heteroatom classes of N1, N1O1, N2, N2O1, and N1O1Na from (+)ESI, and O2 from

(-)ESI, accounting for ~70% of the relative abundance of detectable ions, were chosen for

analyzing molecular structural characteristics of bio-crude oil from CFR and batch

processing (Figure 4). The number of peaks decreased from batch to CFR. The most

abundant peaks are at higher DBE values in the N1 and N1O1 classes for batch. For the N2

class, the batch reactor generated abundant species that covered a broader compositional

space, whereas the CFR preferentially generated species with DBE 9. For the N2O1 class,

the batch reactor generated species with DBE values <6, whereas the CFR generated a

higher relative abundance of species with DBE values >6. These results imply that more

severe conditions contribute to more diverse hetero-compounds with higher aromaticity but

lower oxygen content. Also, the diversity of hetero-compounds (larger C number and DBE

space) increased as the number of N and O atoms per molecule increased in the classes of

N1-2O0-1, with the number of detected chemical species ranging from ~400 (N1) to 640

(N2O1). These results showed a good agreement with the previous study, which indicate

Page 15 of 40
that the number of species per class increases as heteroatom content increases (Cheng et al.,

2017).

Within the N1 class, batch HTL produced bio-crude oil with higher relative abundances

of detectable ions (expanded compositional space) than that of continuous HTL. Higher

relative abundances of detectable C8-21 amines with DBE of 0-1 were produced in CFR,

whereas higher relative abundances of detectable pyridine, indole, indoline and quinoline

derivatives (C7-15 with DBE of 4-8) were obtained from batch, attributed to amidation,

imine formation, Beckmann rearrangement, and dehydration among protein-derived amino

acids and monosaccharides from carbohydrates (Madsen et al., 2017).

Within the N1O1 class, the compositional space shifted towards lower carbon numbers

and high DBE with increased reaction severity. In continuous HTL, this was observed as

generation of C14-22 amides with DBE of 0-2, most of which are fatty acid degradation

products (Cheng et al., 2017). These results show that batch HTL under more severe

condition gives higher relative abundances of heterocyclic aromatic species, such as O-

containing pyridines and indoles, and indolinones. Thus, N-containing long-carbon-chain

molecules are preserved under milder condition, demonstrating that CFR with higher

heating rates favors releasing of C/H-rich compounds with lower aromaticity.

Within the N2 and N2O1 classes, the compositional space moved towards higher carbon

numbers and lower DBE as severity increased, indicating more light hetero-compounds.

Abundant O-containing indole derivatives (C11-19 with DBE of 7-9) were detected in

continuous HTL, which can be explained by decomposition of β-carboline alkaloids that

exist in microalgae (Volk, 2008). In batch HTL, indole derivatives degraded further, or

aromatized and polymerized into coke, and more pyrazine/pyridine derivatives and phenyl

Page 16 of 40
compounds (C8-15 with DBE of 4-7) were formed because of aldol condensation and Diels-

Alder reaction of monosaccharides (Wang et al., 2016). Indole and quinazoline derivatives

exhibit weak structural stability under severe conditions, leading to degradation or

formation of coke precursors. To improve bio-crude oil quality, efforts inhibit cyclization

and aromatization of carbohydrate-derived sugars and protein-derived amino acids need to

be considered.

In the classes of N1O1Na1 from (+)ESI and O2 from (-)ESI, more intact lipid-derived

fatty acids (C16-20 with 1-2 DBE), as well as resin acids (C20 with 5-6 DBE), were found in

continuous HTL. In batch HTL, palmitic, palmitoleic, and elaidic acids were

disintegrated/dehydrogenated into shorter carbon chains (C9-14) with higher DBE (3-4),

reducing the abundance of the O2 class, while increasing the abundance of the N1O1Na1

class due to the stability of amide structures. Comparing the N1O1Na1 and O2 classes with

the N1-2O0-1 classes, lipid-derived fatty acids are not cyclized/aromatized as easily as

carbohydrate- and protein-derived compounds, but they are still cracked into smaller

molecules through carbon-carbon bond cleavage under more severe batch conditions.

Even though continuous HTL preserves C- and H-rich components (78% and 11%,

respectively) with fair HHV (~39 MJ/kg), the higher abundances of O- and N-containing

cyclic compounds (>55%) observed in the bio-crude oil are still a serious problem when

compared to the higher C and H contents (83-85 wt.% and 10-14 wt.%, respectively), lower

N and O contents (0.1-2 wt.% and 0.05-1.5 wt.%, respectively), higher concentrations of

straight-chain and branched hydrocarbons (~60 wt.%), and higher HHV (42-48 MJ/kg) in

petroleum oils (Norman, 2001; Speight, 1999). Appropriate deoxygenation techniques,

Page 17 of 40
such as alkali catalysts and/or hydrogen-donor solvents, should be considered for removing

heteroatoms during continuous HTL.

3.3 Characterization of HTL aqueous phase

Total organic carbon and total nitrogen results showed that substantial amounts of C

and N were present in the HTL aqueous phase (2.8-5.3 g/L and 550-830 mg/L,

respectively). As expected, total organic carbon and total nitrogen increased with solids

loading, and also with reaction temperature and residence time. The CFR reached steady

state at approximately 70 min of runtime for WWGS (MethodsX). The N/C elemental ratio

decreased in the collected liquid products over the course of the reaction, from 0.24 to 0.15

at 325 °C and shorter residence time, and from 0.18 to 0.13 at 350 °C and longer residence

time. This indicates that more N-rich compounds were converted from the aqueous phase

under more severe conditions (> 325 °C) (Huang et al., 2016), which is consistent a

previous study, in which C and N elemental yields (on a dry feedstock basis) decreased

slightly with increased reaction severity as algae-derived organic compounds transferred

from the aqueous phase into the water-insoluble organic and gaseous phases (Cheng et al.,

2017).

3.4 Characterization of HTL solid phase

The multilayer leaflet structure of the WWPC algae feedstock had degraded into finer

particles with coarser surfaces after HTL, indicating the decomposition of algal

components. Inorganic elements in the algae feedstock mostly concentrated into the solid

HTL product. As expected from their solubility in water, Ca and Mg were more

concentrated in the solid phase, while K and Na were more concentrated in the aqueous

phase, leading to a slightly alkaline pH (7.4-7.7). Initially, there was concern about the

Page 18 of 40
potential of corrosion of the stainless-steel reactor, however, no increased concentrations of

elements from the type 316 steel (e.g. Cr, Fe, Mn, Mo, and Ni) were observed in the liquid

products. Notably, phosphorus was enriched in the solid product from 5% WWPC, which

could be considered as the potential nutrient resource for plant cultivation.

3.5 Process improvement

The minimum fuel-selling price for algal fuel is $3.53/gallon, slightly higher as

consider the average price of gasoline and diesel in the last decade ($2.17-$3.97/gallon

(USEIA, 2018)). As a promising HTL reactor system, CFR systems should be modified and

optimized via combining with upstream and downstream units to improve the economic

feasibility of algae-to-biofuel processes. Feedstock cost and biofuel yield are usually the

two primary contributors to the minimum fuel-selling price and greenhouse gases emission

with respect to sensitivity analysis (Jones et al., 2014; Ou et al., 2015; Zhu et al., 2014). To

lower algae feedstock costs, the costs of CO2, nutrients, cultivation time, dewatering, and

feedstock delivery should be reduced. CO2-rich off-gases from the algal HTL process or a

steam plant could be reused to feed algae. N fertilizer could be supplemented by N nutrients

recovered from HTL aqueous using catalytic hydrothermal gasification (Van Doren et al.,

2017). Extension of cultivation time in winter can be offset through growth of different

algae species (Albrecht et al., 2016), and incorporation of other organic wastes, such as

manure (Hoffmann et al., 2013), food waste (Posmanik et al., 2017), and sludge (Vardon et

al., 2011). Dewatering cost needs to be balanced, since higher solids loading of algal slurry

improves bio-crude oil productivity, lowers feedstock delivery cost, and requires higher

energy consumption. Feedstock delivery cost can be decreased by shortening transportation

distance by operating smaller-scale HTL reactors at the wastewater treatment plant (Fortier

Page 19 of 40
et al., 2014), followed by transferring bio-crude oil to a centralized upgrading unit in an

existing petroleum refinery (Zhu et al., 2014).

Current biofuel productivity of 26.9-65.2 million gal/year is much lower than the

supply of fossil-based transportation fuels (2.54×105 million gal/year (USEIA, 2017)).

Biofuel productivity can be enhanced through increasing bio-crude oil yield and solids algal

content in the CFR system. To obtain higher bio-crude oil yield, ash and carbohydrate

contents need to be decreased prior to HTL, as ash/carbohydrates inhibit bio-crude oil

formation and deteriorate bio-crude oil quality (Cheng et al., 2017). Short reaction times

and low reaction rates lead to under-reaction of biomass, resulting in less bio-crude oil,

which directly affects the oil-aqueous separation and oil productivity in the CFR system.

These two issues can be addressed by 1) modifying the plug flow reactor configuration to

increase residence time under the desired operating conditions or lowering the flow rate if

particle settling can be prevented; 2) adding hydrogen-donor solvents (e.g. ethanol or

glycerol) to increase bio-crude oil yield by transferring more organics from the aqueous

phase into the oil phase and inhibiting re-polymerization of smaller molecules; 3)

integrating a deoxygenation/denitrogenation/cracking catalytic packed bed column with the

filter vessels to upgrade bio-crude oil into low-polarity, low-heteroatom biofuel, leading to

formation of a thicker and more hydrophobic oil layer, and enabling easier oil separation

from aqueous phase.

Assuming a comparable bio-crude oil yield, a solids loading of 20 wt.% has been

considered as a fair trade-off between the cost of water removal from algae feedstock and

the capital costs of running HTL (Pearce et al., 2016). To improve the capability of

processing higher solids loads in a CFR system, one can: 1) use a double piston pump or

Page 20 of 40
hose-diaphragm pump instead of a metering pump (Billing et al., 2016) and adjust the

tubing/valve sizes based on slurry terminal velocity; 2) add a biodegradable water-soluble

thickener such as guar gum to improve particle suspension; or 3) co-process other

industrial/agricultural/municipal wet wastes containing lower-density particles.

A considerable amount of energy needed for the algae-to-biofuels process can be

recycled internally as surplus heat, electricity, and by-products (e.g. biogas, aqueous phase,

and char). To recover most of heat from product flow, the preheater and product chiller in

the CFR system can be replaced by a suitable heat exchanger between the feed stream after

the pump and the product stream before the back pressure regulator (Van Doren et al.,

2017), and hot surfaces better insulated. To prevent damage to the BPR diaphragm (product

stream too hot) or excessive fluid viscosity (product stream too cold), an intermediate

temperature (45-65 °C) is recommended. Concentrated solar power could also be used to

provide additional heat for HTL (Pearce et al., 2016).

For a byproduct recovery unit, Van Doren et al. compared the influences of anaerobic

digestion (AD) and catalytic hydrothermal gasification (CHG) on recycling of water-

soluble organics and heat from HTL aqueous (Van Doren et al., 2017). Their model

suggested AD requires less heat for operation and has lower investment cost, while CHG

generates more heat and recycles more nutrients from the aqueous phase (Beal et al., 2015).

Also, the HTL aqueous phase was treated less effectively in AD compared to CHG,

because the N/C ratio of the HTL aqueous phase was not suitable to the AD process (Frank

et al., 2013). CHG could be integrated with an HTL CFR system to minimize energy

demand by removing the need to cool/reheat and de-pressurize/re-pressurize (Jones et al.,

2014). The P-rich char could be converted into a value-added product.

Page 21 of 40
Figure 5 presents a wastewater-grown algae-to-biofuel process that incorporates the

proposed improvements. Wastewater-grown algae are harvested, and concentrated into a

slurry with a solids loading of 20 wt.% by concentration (e.g. coagulation and flocculation)

or by addition of dry co-feedstock. The feed stream is pretreated to remove ash and separate

carbohydrates, before being pumped into an on-site CFR system for bio-crude oil

production. Bio-crude oil is sent from the wastewater plant to a petroleum refinery for

upgrading. The HTL aqueous phase is sent to CHG to produce syngas and nutrient-rich

liquid; the liquid would be treated and returned to the algae growth pond. HTL char is sent

to the solid disposal plant for phosphorous recycle. Off-gases/syngas from different units is

sent to the combined heat and power plant to produce heat for all reaction units. An existing

hydrogen generation unit provides hydrogen for the upgrading unit. Most of natural gas,

heat, and electricity supplied for the upgrading and hydrogen generation units is obtained

from the petroleum refinery.

4 Conclusion

HTL of wastewater algae in a pilot-scale CFR under 350 °C and 17 MPa produced bio-

crude oil yields of 28 wt.% with HHV of 39 MJ/kg. Low solids loading (5 wt.%) and high

calculated energy consumption ratio (3.57) necessitate further modification of the CFR

system to enable more concentrated slurries (>10 wt.%), to increase bio-crude oil yields

(>40%), and to prevent heat losses. Bio-crude oils produced in the CFR system were

comparable to those produced in a batch system. HTL reaction engineering is needed to

reduce O-content and inhibit aromatization of carbohydrate- and protein-derived fragments

in bio-crude oil.

Page 22 of 40
E-supplementary data of this work can be found in online version of the paper.

Acknowledgements

Funding for this work was provided by the NSF New Mexico EPSCOR Research

Infrastructure Improvement grant “Energize New Mexico” (#1031346), the NSF

“ReNUWIt” Engineering Research Center (#1028968), an NSF Major Research

Instrumentation grant (#1626468), and a USDA-NIFA grant, “Sustainable Bioeconomy for

Arid Regions” (2017-68005-26867). The authors would like to acknowledge assistance

from members of the Holguin, Khandan, Brewer, Jena and Herndon research groups during

algae production and harvesting, reactor operation, and product characterization.

Page 23 of 40
References

Albrecht, K.O., Zhu, Y., Schmidt, A.J., Billing, J.M., Hart, T.R., Jones, S.B., Maupin, G.,

Hallen, R., Ahrens, T., Anderson, D. 2016. Impact of heterotrophically stressed

algae for biofuel production via hydrothermal liquefaction and catalytic

hydrotreating in continuous-flow reactors. Algal Research, 14, 17-27.

Beal, C.M., Gerber, L.N., Sills, D.L., Huntley, M.E., Machesky, S.C., Walsh, M.J., Tester,

J.W., Archibald, I., Granados, J., Greene, C.H. 2015. Algal biofuel production for

fuels and feed in a 100-ha facility: a comprehensive techno-economic analysis and

life cycle assessment. Algal Research, 10, 266-279.

Billing, J., Anderson, D., Hallen, R., Hart, T., Maupin, G., Schmidt, A., Elliott, D.C. 2016.

Design, Fabrication, and Testing of the Modular Hydrothermal Liquefaction System

(MHTLS). TCS 2016 Symposium on Thermal and Catalytic Sciences for Biofuels

and Biobased Products, Chapel Hill, NC.

Bridgwater, A.V. 2012. Upgrading biomass fast pyrolysis liquids. Environmental Progress

& Sustainable Energy, 31(2), 261-268.

Chen, W.-T., Ma, J., Zhang, Y., Gai, C., Qian, W. 2014a. Physical pretreatments of

wastewater algae to reduce ash content and improve thermal decomposition

characteristics. Bioresource technology, 169, 816-820.

Chen, W.-T., Zhang, Y., Zhang, J., Yu, G., Schideman, L.C., Zhang, P., Minarick, M.

2014b. Hydrothermal liquefaction of mixed-culture algal biomass from wastewater

treatment system into bio-crude oil. Bioresource technology, 152, 130-139.

Page 24 of 40
Cheng, F., Cui, Z., Chen, L., Jarvis, J., Paz, N., Schaub, T., Nirmalakhandan, N., Brewer,

C.E. 2017. Hydrothermal liquefaction of high-and low-lipid algae: Bio-crude oil

chemistry. Applied Energy, 206, 278-292.

Cheng, F., Cui, Z., Mallick, K., Nirmalakhandan, N., Brewer, C.E. 2018. Hydrothermal

liquefaction of high-and low-lipid algae: Mass and energy balances. Bioresource

technology, 258, 158-167.

Cheng, F., Mallick, K., Gedara, S.M.H., Jarvis, J.M., Schaub, T., Jena, U.,

Nirmalakhandan, N., Brewer, C.E. 2019. Hydrothermal Liquefaction of Galdieria

sulphuraria Grown on Municipal Wastewater. Bioresource Technology, 292,

121884.

Demirbas, A., Demirbas, M.F. 2011. Importance of algae oil as a source of biodiesel.

Energy conversion and management, 52(1), 163-170.

Eboibi, B.E.-O., Lewis, D.M., Ashman, P.J., Chinnasamy, S. 2014. Hydrothermal

liquefaction of microalgae for biocrude production: improving the biocrude

properties with vacuum distillation. Bioresource technology, 174, 212-221.

Elliott, D.C. 2008. Catalytic hydrothermal gasification of biomass. Biofuels, Bioproducts

and Biorefining, 2(3), 254-265.

Faeth, J.L., Valdez, P.J., Savage, P.E. 2013. Fast hydrothermal liquefaction of

Nannochloropsis sp. to produce biocrude. Energy & Fuels, 27(3), 1391-1398.

Fortier, M.-O.P., Roberts, G.W., Stagg-Williams, S.M., Sturm, B.S. 2014. Life cycle

assessment of bio-jet fuel from hydrothermal liquefaction of microalgae. Applied

Energy, 122, 73-82.

Page 25 of 40
Frank, E.D., Elgowainy, A., Han, J., Wang, Z. 2013. Life cycle comparison of

hydrothermal liquefaction and lipid extraction pathways to renewable diesel from

algae. Mitigation and Adaptation Strategies for Global Change, 18(1), 137-158.

Grierson, S., Strezov, V., Shah, P. 2011. Properties of oil and char derived from slow

pyrolysis of Tetraselmis chui. Bioresource technology, 102(17), 8232-8240.

Herod, A.A., Bartle, K.D., Kandiyoti, R. 2007. Characterization of heavy hydrocarbons by

chromatographic and mass spectrometric methods: An overview. Energy & Fuels,

21(4), 2176-2203.

Hoffmann, J., Rudra, S., Toor, S.S., Holm-Nielsen, J.B., Rosendahl, L.A. 2013. Conceptual

design of an integrated hydrothermal liquefaction and biogas plant for sustainable

bioenergy production. Bioresource technology, 129, 402-410.

Huang, Y., Chen, Y., Xie, J., Liu, H., Yin, X., Wu, C. 2016. Bio-oil production from

hydrothermal liquefaction of high-protein high-ash microalgae including wild

Cyanobacteria sp. and cultivated Bacillariophyta sp. Fuel, 183, 9-19.

Jones, S.B., Zhu, Y., Anderson, D.B., Hallen, R.T., Elliott, D.C., Schmidt, A.J., Albrecht,

K.O., Hart, T.R., Butcher, M.G., Drennan, C. 2014. Process design and economics

for the conversion of algal biomass to hydrocarbons: whole algae hydrothermal

liquefaction and upgrading. Pacific Northwest National Laboratory (PNNL),

Richland, WA (US).

Lindemuth, T.E. 1981. Biomass conversion processes for energy and fuels. in: Biomass

Conversion Processes for Energy and Fuels, (Eds.) S.S. Sofer, O.R. Zaborsky,

Springer Science & Business Media. Boston, MA, pp. 187-200.

Page 26 of 40
Madsen, R.B., Bernberg, R.Z., Biller, P., Becker, J., Iversen, B.B., Glasius, M. 2017.

Hydrothermal co-liquefaction of biomasses–quantitative analysis of bio-crude and

aqueous phase composition. Sustainable Energy & Fuels, 1(4), 789-805.

Mecozzi, M., Pietroletti, M., Tornambè, A. 2011. Molecular and structural characteristics in

toxic algae cultures of Ostreopsis ovata and Ostreopsis spp. evidenced by FTIR and

FTNIR spectroscopy. Spectrochimica Acta Part A: Molecular and Biomolecular

Spectroscopy, 78(5), 1572-1580.

Minowa, T., Kondo, T., Sudirjo, S.T. 1998. Thermochemical liquefaction of Indonesian

biomass residues. Biomass and Bioenergy, 14(5), 517-524.

Minowa, T., Yokoyama, S.-y., Kishimoto, M., Okakura, T. 1995. Oil production from algal

cells of Dunaliella tertiolecta by direct thermochemical liquefaction. Fuel, 74(12),

1735-1738.

Mørup, A.J., Becker, J., Christensen, P.S., Houlberg, K., Lappa, E., Klemmer, M., Madsen,

R.B., Glasius, M., Iversen, B.B. 2015. Construction and commissioning of a

continuous reactor for hydrothermal liquefaction. Industrial & Engineering

Chemistry Research, 54(22), 5935-5947.

Mullen, C.A., Strahan, G.D., Boateng, A.A. 2009. Characterization of various fast-

pyrolysis bio-oils by NMR spectroscopy. Energy & Fuels, 23(5), 2707-2718.

Norman, J.H. 2001. Nontechnical guide to petroleum geology, exploration, drilling, and

production (2nd Ed.). Penn Well Corp., Tulsa, OK.

Ocfemia, K., Zhang, Y., Funk, T. 2006. Hydrothermal processing of swine manure into oil

using a continuous reactor system: Development and testing. Transactions of the

ASABE, 49(2), 533-541.

Page 27 of 40
Ou, L.W., Thilakaratne, R., Brown, R.C., Wright, M.M. 2015. Techno-economic analysis

of transportation fuels from defatted microalgae via hydrothermal liquefaction and

hydroprocessing. Biomass and Bioenergy, 72, 45-54.

Pearce, M., Shemfe, M., Sansom, C. 2016. Techno-economic analysis of solar integrated

hydrothermal liquefaction of microalgae. Applied Energy, 166, 19-26.

Peterson, A.A., Vogel, F., Lachance, R.P., Fröling, M., Antal Jr, M.J., Tester, J.W. 2008.

Thermochemical biofuel production in hydrothermal media: a review of sub-and

supercritical water technologies. Energy & Environmental Science, 1(1), 32-65.

Pierce_Biotechnology. 2013. Instructions of Pierce™ 660nm Protein Assay, (Ed.) P.

Biotechnology, Thermo Scientic. Rockford, lL, USA.

Posmanik, R., Cantero, D., Malkani, A., Sills, D., Tester, J. 2017. Biomass conversion to

bio-oil using sub-critical water: Study of model compounds for food processing

waste. The Journal of Supercritical Fluids, 119, 26-35.

Schenk, P.M., Thomas-Hall, S.R., Stephens, E., Marx, U.C., Mussgnug, J.H., Posten, C.,

Kruse, O., Hankamer, B. 2008. Second generation biofuels: high-efficiency

microalgae for biodiesel production. Bioenergy research, 1(1), 20-43.

Selvaratnam, T., Henkanatte-Gedera, S., Muppaneni, T., Nirmalakhandan, N., Deng, S.,

Lammers, P. 2016. Maximizing recovery of energy and nutrients from urban

wastewaters. Energy, 104, 16-23.

Slocombe, S.P., Ross, M., Thomas, N., McNeill, S., Stanley, M.S. 2013. A rapid and

general method for measurement of protein in micro-algal biomass. Bioresource

technology, 129, 51-57.

Page 28 of 40
Speight, J.G. 1999. The chemistry and technology of petroleum (3rd Ed., rev. and expanded

Ed.). Marcel Dekker, New York.

Sudasinghe, N., Dungan, B., Lammers, P., Albrecht, K., Elliott, D., Hallen, R., Schaub, T.

2014. High resolution FT-ICR mass spectral analysis of bio-oil and residual water

soluble organics produced by hydrothermal liquefaction of the marine microalga

Nannochloropsis salina. Fuel, 119, 47-56.

USEIA. 2017. Petroleum & Other Liquids — Supply and Disposition, Vol. 2017, U.S.

Energy Information Administration. Washington DC.

USEIA. 2018. U.S. Gasoline and Diesel Retail Prices, Vol. 2018, U.S. Energy Information

Administration. Washington DC.

Van Doren, L.G., Posmanik, R., Bicalho, F.A., Tester, J.W., Sills, D.L. 2017. Prospects for

energy recovery during hydrothermal and biological processing of waste biomass.

Bioresource technology, 225, 67-74.

Van Wychen, S., Laurens, L. 2013. Determination of Total Solids and Ash in Algal

Biomass: Laboratory Analytical Procedure (LAP). National Renewable Energy

Laboratory (NREL), Golden, CO.

Vardon, D.R., Sharma, B., Scott, J., Yu, G., Wang, Z., Schideman, L., Zhang, Y.,

Strathmann, T.J. 2011. Chemical properties of biocrude oil from the hydrothermal

liquefaction of Spirulina algae, swine manure, and digested anaerobic sludge.

Bioresource technology, 102(17), 8295-8303.

Vardon, D.R., Sharma, B.K., Blazina, G.V., Rajagopalan, K., Strathmann, T.J. 2012.

Thermochemical conversion of raw and defatted algal biomass via hydrothermal

liquefaction and slow pyrolysis. Bioresource Technology, 109, 178-187.

Page 29 of 40
Volk, R.-B. 2008. Screening of microalgae for species excreting norharmane, a manifold

biologically active indole alkaloid. Microbiological research, 163(3), 307-313.

Wang, F., Xiao, W., Gao, L., Xiao, G. 2016. Enhanced performance of glycerol to

aromatics over Sn-containing HZSM-5 zeolites. RSC Advances, 6(49), 42984-

42993.

Wychen, S.V., Laurens, L.M.L. 2013. Determination of Total Carbohydrates in Algal

Biomass. National Renewable Energy Laboratory.

Wychen, S.V., Ramirez, K., Laurens, L.M.L. 2013. Determination of Total Lipids as Fatty

Acid Methyl Esters (FAME) by in situ Transesterification. National Renewable

Energy Laboratory.

Yang, W., Li, X., Li, Z., Tong, C., Feng, L. 2015. Understanding low-lipid algae

hydrothermal liquefaction characteristics and pathways through hydrothermal

liquefaction of algal major components: crude polysaccharides, crude proteins and

their binary mixtures. Bioresource technology, 196, 99-108.

Zhang, C., Tang, X., Sheng, L., Yang, X. 2016. Enhancing the performance of Co-

hydrothermal liquefaction for mixed algae strains by the Maillard reaction. Green

Chemistry, 18(8), 2542-2553.

Zhu, Y., Biddy, M.J., Jones, S.B., Elliott, D.C., Schmidt, A.J. 2014. Techno-economic

analysis of liquid fuel production from woody biomass via hydrothermal

liquefaction (HTL) and upgrading. Applied Energy, 129, 384-394.

Page 30 of 40
Tables and Figures

Table 1. Composition of algal feedstocks.


Algal Species WWPC WWGS
Proximate Analysis
Initial Algal Solids wt.%a 13.6 ± 0.1 31.2 ± 0.8
Ash wt.%b 34.7 ± 0.1 42.0 ± 0.2
HHV MJ/kg b,e 15.7 ± 0.4 12.1 ± 0.4
Elemental Analysis wt.%b
Carbon 34.5 ± 0.2 29.0 ± 0.1
Hydrogen 5.3 ± 0.1 4.2 ± 0.2
Nitrogen 5.6 ± 0.1 3.4 ± 0.0
Sulfur 1.4 ± 0.0 1.1 ± 0.1
Oxygen d 18.6 ± 0.3 20.3 ± 0.3
Biochemical Analysis wt.%c
Lipids 1.7 ± 1.0 5.8 ± 0.2
Proteins 45.9 ± 2.4 41.0 ± 2.5
Carbohydrates 16.4 ± 0.5 10.5 ± 0.5
a
Wet basis. b Dry basis. c Dry ash free basis. d
By difference. e
HHV of feedstock was estimated by the
equation in (Eboibi et al., 2014).

Page 31 of 40
Table 2. The yield, quality, energy recovery (ER) and energy consumption ratio (ERC) of light bio-crude oil (LBO) produced from
HTL of microalgae in the CFR system and in the batch reactor. ± = standard deviations where n=3 for higher heating values (HHV)
and elemental contents of LBO.
HTL Reaction Reaction LBO Yield LBO HHV Elemental Content wt.%a ER ECR
Experiment Temp. °C Time min wt.%b MJ/kg C H N S O c %
Batch 5% 350 5 25.70 38.4 ± 0.2 80.7 ± 0.0 11.2 ± 0.2 4.4 ± 0.2 2.1 ± 0.2 1.6 ± 0.3 33.5 2.6
WWGSd
Continuous 325 3 22.56 38.8 ± 2.1 77.7 ± 0.6 11.2 ± 0.5 4.5 ± 0.5 1.4 ± 0.2 5.2 ± 1.0 41.8 7.2
3% WWGS
Continuous 350 6 28.12 38.2 ± 2.0 78.2 ± 0.9 11.4 ± 0.1 4.6 ± 0.1 1.6 ± 0.1 4.2 ± 0.9 51.3 3.7
5% WWGS
Continuous 350 9 2.31 38.9 ± 1.0 76.6 ± 0.1 10.7 ± 0.1 6.4 ± 0.1 4.9 ± 0.0 1.5 ± 0.2 3.8 37.8
5% WWPC
a
Dry basis. b Dry ash-free basis. c By difference. d Ref. (Cheng et al., 2019).

Table 3. Volatilization temperature distributions of algal biomass, WWGS- and WWPC-derived LBO produced under HTL at 325-
350 °C and 13-17 MPa.
HTL Operating Volatilization Temperature Distribution wt.% (dry basis)
Conditions IBP-70°C 70-100°C 100-180°C 180-240°C 240-360°C 360-540°C 540+°C
WWGS Feed 0.17 0.36 3.97 5.77 20.26 19.94 49.53
WWPC Feed 2.68 1.22 1.85 3.02 28.07 17.16 46.01
3%WWGS LBO 325 °C, 13 MPa 0.59 0.83 8.91 29.90 33.68 20.31 5.78
5%WWGS LBO 350 °C, 17 MPa 0.54 0.78 11.42 39.39 30.87 12.32 4.67
5%WWPC LBO 350 °C, 17 MPa 1.59 3.86 36.81 29.78 18.07 4.55 5.33

Page 32 of 40
Table 4. Proton distribution of light bio-crude oil (LBO) from HTL of WWGS and WWPC under different operating conditions.
Figures adapted from (Mullen et al., 2009)
Proton Type Aliphatic Unsaturated/ Alcohol/ Carbohydrate/ (Hetero-) Aldehyde
Heteroatom Methylene- Unconjugated Aromatic
Dibenzene Olefin

Range (ppm)
LBO 0.5-1.5 1.5-3.0 3.0-4.4 4.4-6.0 6.0-8.0 9.0-10.1
3% WWGS LBO 37.77 19.52 9.07 11.98 16.04 5.62
5% WWGS LBO 45.36 17.11 7.38 10.66 14.40 5.09
5% WWPC LBO 38.70 19.84 9.07 9.61 17.03 5.75

Page 33 of 40
Figure 1. Schematic diagram of the modified pilot-scale CFR system.

Page 34 of 40
Figure 2. Curves from (a) thermogravimetric analysis (TGA) and (b) differential thermogravimetric analysis (DTG) of WWGS and
LBO from HTL of 3% WWGS (325 °C and 3 min) and 5% WWGS (350 °C and 6 min).

Page 35 of 40
Figure 3. Heteroatom class distribution for LBOs from HTL of 3-5 wt.% WWGS in the CFR and 5 wt.% WWGS in the 1.8-L batch
reactor under 350 °C derived from a) positive-ion and b) negative-ion ESI FT-OF MS. WWGS: wastewater-treatment G. sulphuraria.
CFR: continuous flow reactor. LBO: light bio-crude oil.

Page 36 of 40
Figure 4. Color-coded abundance–contoured plots of DBE versus carbon number for heteroatom classes in LBOs from HTL of 3-5
wt.% WWGS in the CFR and 5 wt.% WWGS in the 1.8 L batch reactor under 350 °C derived from positive-ion and negative-ion ESI
FT-MS. WWGS: wastewater-treatment G. sulphuraria. DBE: double bond equivalence. SL: solids loading of feedstock (wt.%). LBO:
light bio-crude oil.

Page 37 of 40
Figure 5. Modified process flow diagram of conversion of wastewater algae into biofuel. Grey dashed lines represent system
boundaries.

Page 38 of 40
Highlights

 A continuous flow reactor was modified for producing algal bio-crude oil.

 Bio-crude oil (28 wt.% yield, 39 MJ/kg) was obtained from wastewater algae.

 Bio-crude oil quality from new reactor was comparable to oils from batch reactors.

Page 39 of 40
 The authors have no competing interests to declare.

Page 40 of 40

Вам также может понравиться