You are on page 1of 11

Fuel 165 (2016) 526–535

Contents lists available at ScienceDirect

j o u r n a l h o m e p a g e : w w w.e l s e v i e r.c o m / l o c a t e /f u e l

Co-processing potential of HTL bio-crude at petroleum refineries –
Part 1: Fractional distillation and characterization
Jessica Hoffmann a,⇑, Claus Uhrenholt Jensen a,b, Lasse A. Rosendahl a
Department of Energy Technology, Aalborg University, Pontoppidanstræ de 101, 9220 Aalborg, Denmark
Steeper Energy ApS, Sandbjergvej 11, 2970 Hørsholm, Denmark

a r t i c l e i n f o a b s t r a c t
Article history: This study presents detailed chemical and thermophysical analysis of bio-crude from a continuous
Received 16 March 2015
hydrothermal liquefaction research plant. Current research on bio-crude focuses mainly on specific bio-
Received in revised form 18 October 2015
mass feedstocks and conversion process conditions and resulting yields rather than on bio-crude proper-
Accepted 22 October 2015
Available online 29 October 2015
ties for downstream processing. This study contributes to the next level of research, where HTL bio-
crude is evaluated as a potential refinery co-processing feedstock with regard to bulk and fractional
properties. The bio-crude used in the current work has been produced from a hardwood feedstock.
Hydrothermal liquefaction
Bio-crude assays, adapted from conventional crude oil assays, have been obtained, including
Fractional distillation fractionation of the bio-crude through 15:5 vacuum distillation. The bio-crude and its fractions have
Co-processing been analyzed with respect to heating value, elemental composition, density and oxygen-
Biomass containing functional groups. Results show a highly promising bio-crude quality, with a higher heating
Bio-crude value of 40.4 MJ/kg, elemental oxygen content of 5.3 wt.%, a specific gravity of 0.97 and a distillation
recovery of 53.4 wt.% at an atmo- spheric equivalent temperature (AET) of 375 C, . Results show that
only minor upgrading is needed to achieve co-processing properties and to obtain a refinery bio-feed.
This forms the basis for subsequent assessment of drop-in and co-processing potential and parametric
upgrading trials of the bio-crude pre-
sented in part 2 of this work.
2015 Elsevier Ltd. All rights reserved.

1. Introduction streams such as agricultural and forestry waste and other non-
food biomass are of significant interest. Only a process covering
The concurrent increase in global primary energy consumption these attributes has the ability to produce a drop-in or refinery
by an annual 2.3% (2013) [1] and depletion of conventional co-processing feed that is commercially competitive to conven-
resources, combined with climate issues and the desire for tional fuels and therefore has the capability to endure on the
national/regional energy independence, has led to an urgent need market.
for renewable as well as sustainable energy sources and processes. Hydrothermal liquefaction (HTL) is one such promising feed-
Due to their potential carbon neutrality, liquid fuels from non-food stock flexible thermochemical conversion processes converting
biomass are essential to meet the imposing challenges of energy low-value, high moisture or wet biomass feedstock to a high value
supply and climate impact [2]. In 2013, fossil fuels still accounted bio-crude. It works at near critical water conditions with
for 87% of global and 77% of EU primary energy consumption [1]. tempera- tures of 360–450 C and pressure from 250 to 400 bars.
Since biomass will also become primary feedstock for chemicals, HTL offers numerous advantages compared to other
nutritional and pharmaceutical products, it will in future become thermochemical conver- sion processes: high moisture feedstock,
a high-cost commodity. Therefore it is of great importance to conversion efficiency and product quality in terms of delivering a
develop a sustainable and cost-effective process for the conversion bio-crude with high refin- ery biofeed and co-processing potential.
of biomass, which is feedstock flexible, energy efficient and offers Such co-processing offers a number of advantages, the first of
high conversion efficiency. Whereas some processes need special- which is the opportunity to reuse existing refinery and
ized feedstocks, processes that run on multiple low value feed downstream infrastructure and thus significantly shortening
implementation time scales and reducing CAPEX. Secondly, it
⇑ Corresponding author. Tel.: +45 21370482; fax: +45 9815 141. contributes to cost-effectiveness in terms of processing to
E-mail address: (J. Hoffmann). synthetic, chemically equivalent drop-in products with a
reduced environmental food print.
0016-2361/ 2015 Elsevier Ltd. All rights reserved.

Mechanically the CBS1 plant has been designed for . and a gas stable compared to e. If only parts of the bio-crude comparison between petroleum and bio-crude and thus fractions con. Hence for hydrocarbons have higher boiling points compared to their conventional petroleum refineries. Based on the distillation results and the subsequent analysis. research paper. wide range of process conditions and feedstocks. In literature only very standard methods for an assay characterization have not yet been few works on fractional distillation of HTL bio-crude that also devel. kinematic viscosity. Fractions with temperature intervals of 25 C have flash point. This was used. higher heating value and density. Methods co-processing of the bio-crude needs to be developed. Crude oil assays manure. density and FT-IR TAN (total acid number). In fact. et / al.g. HHV. Focus is hereby set on the oxygen removal prior to co-processing since conventional refineries are not 2. Comparing the heteroatom distribution stream. also gives insight into the yield for each fraction and taining oxygenated hydrocarbons to a higher degree. Hoffmann et al. which is well beyond the critical temperature an oil sample. High oxygen contents and The bio-crude used in this study has been obtained from pro- high TAN num.ric study on the removal of oxygen from the 4–10 wt. pre- 1. Bio-crude optimized to handle oxygen rich oils. Since refineries are refinery modifications. with the challenging difference of con. highly optimized to the crude feed as well as to the markets they serve. fractional properties include elemental analysis. Designed and built by Stee- refinery. mal cracking is observed by [8] through char formation during dis- e. The distillation plex hydrocarbon mixtures. a water phase containing soluble organics. Fuel/ Fuel 165 165 (2016) (2016) 526–535 526–535 52 52 Naturally. In 1980 crude quality. selected conventional crude oil assay parameters Thus. whereas fractional properties specify characteristics of ing and destructive distillation.1. an upgrading pathway prior to 2.g. It is virtually consequently to the value of the bio-crude. These (310 C) suggested by ASTM 2892 to avoid thermal cracking. Compared to conventional crude with an oxygen considered prior to co-processing. data on individual bio-crudes must be available for the crude by distillation is possible and thereby fractional refineries to consider introducing these into the processing properties can be obtained. The lack of analytical knowledge on bio-crudes is Elliott and Elliott [7] included a similar distillation study on HTL mainly due to the high diversity of bio-oil and crudes and the lack bio-crude from a woody feedstock. Ther- properties are determined according to standardized methods. the necessity for upgrading could evaluation of co-processing fea. carbon residue. no work will address this and try to initiate a procedure for bio-crude works have been published on the particular subject up until very assays for refining operations.2. As part of this crude and its properties.5% [5]. This information ‘heavier’ crude boiling point curve [6]. distillation of these products respective to the crude feed. the ASTM D1160 of focus on the downstream side of the conversion process. of 360–450 C.sibility and market value. Bulk and the bio-crude by decomposition of any degree. Such distillation changes the bio- the different fractions obtained by distillation [4].processing of bio-feed and versatile research platform to investigate continuous HTL under a conventional feed. Oxygenated impossible to identify every single molecule in crudes. Just as is true for fossil crudes of different origins. ASTM (American Society for Testing and Materials). metal content in the crude.oped.05–1. potentially even exploring chemical synergetic effects per Energy. bio-crude from HTL typically contains includes a paramet. economic upgrading process could be designed. the Continuous Bench Scale 1 (CBS1) facility provides a that may arise through co. The overall target should hereby be to introduce the critical water at 400 C and 320 bar in a continuous HTL bio-feed with as little detrimental impact as possible to the research facility at Aalborg University. Materials at a refinery stage. [6] complex chemical plants. include analysis of the collected fractions can be found. The second part of this paper content of 0. At pressures in This study will focus on the bulk as well as fractional assay the range of 280–350 bar and reaction temperatures in the range properties of HTL bio-crude produced from a hardwood feedstock. it has been of high priority in the current work not to affect are determined to facilitate evaluation of the bio-crude. where [8] characterized the physical and chemical proper- ties of bio-crude distillates from glycerol assisted HTL of swine 1. which makes it hard to compare and assess the bio. fractionation of phase.bers would lead to corrosion problems during cessing ligno-cellulosic biomass (hardwood 20 kg/h) with super- refining.1.1. For bio-crude such bio-crudes has not been studied extensively. tillation. Further analysis on the bio-crude should also include One of the most apparent differences between crude and bio. Bulk properties include plate column. be limited to these fractions and a more sustainable and Bio-crudes like conventional petroleum crudes are highly com. In this work. The goal is to meet conventional crude oil requirements prior to introduction 2. which results in an incorrect oil assay. simulate the Although significant work has been published on hydrothermal refining process and get insights to the refining process and end liquefaction of biomass to produce bio-crudes. standardized crude oil assays non-oxygenated hydrocarbon group. since metals can poison refinery crude is the oxygen heteroatom content and subsequently the catalyst severely and removal of those would need to be TAN number. and 5 fractions were included for analysis. pour point and been collected and elemental analysis. recent. wet biomass slurry can be processed into a crude As HTL bio-crude has the major advantage of being thermally oil phase. bulk properties and fractional properties [3]. H/C and O/C Fractional distillation in this study is done in a 15:5 theoretical ratios. HTL as function of boiling in the HTL bio-crude to that characteristic bio-crudes will differ depending on feedstock and process for petroleum crude oil yields important information on the parameters.1. Individual bio-crude assays allow a direct design of the downstream stages. J. which impacts the boiling have been developed to get an understanding of the chemical point curve and may shift it to a misleading and thermophysical properties of the crude. have been evaluated and related to the boiling point distribution.tain heteroatoms.% oxygen [4] depending on the biomass feedstock and bio-crude to enable co-processing without the need for major process conditions used during conversion. bio-oil from pyrolysis. but the fact that their fractional viscosities decrease as Bulk properties give an understanding of the overall crude oil function of boiling point above 400 C also indicate thermal crack- sample. Bio-crude assay and co-processing challenges treatment of the HTL bio-crude prior to co-processing will be pro- posed. Hoffmann J. enables modern refineries to characterize crude oils. The fractional distillation by [8] was carried out using a so-called advanced distillation setup under atmospheric pressure A typical crude oil assay contains two types of information for and up until 500 C.

Light ends from the cold trap are kept for further analysis. (3) Fractionation receiver. (4) Vapor temperature. the Maxwell and Bonnell derived equation described in ASTM D5236.1 torr. (1) Product condenser. The standard suggests operating pressures in the range from atmospheric to 2 torr. (2) Fraction- ation spider. A ¼ ð6:761559 0:987672 log 10 PÞ=ð3000:538 43:00 log 10 PÞ for 0:1 < P < 2 torr ½ ð2Þ A ¼ ð5:994295 0:972546 log 10 PÞ=ð2663:129 95:76 log 10 PÞ for P > 2 torr ½ ð3Þ 2.3 wt. For pressures above 2 torr the correlation between observed temperature and AET is linearized. The present set-up is able to reach an operating pressure of 0. Picture of the distillation column.1 and 2 torr. This is done using Fig. (2) if the pressure is above 2 torr. low distillation efficiency and control difficulties during distillation. 2. (2) Reflux valve. As the pres- sure is decreased below 2 torr. 2).1 torr. (9) Vacuum pump.1 torr. (4) Thermocouple pot-flask. It has been observed from previous experiments that water in the sample leads to unsteady boiling. (7) Pressure transducer. Before vacuum distillation the set-up has been leak tested accord- ing to ASTM D 5236 [10]. The standard does not specify at which vacuum pres- sures the distillation should be performed. (6) Pot flask with heating mantel. The pot flask in the bottom of the column has a volume of 2 liters. The set-up has a total of 8 fractionation collectors with a volume of 250 ml respectively (Fig. ASTM 2892 allows 0. The intermediate pres- sure is selected according to the course of the distillation process. (1) Reflux condenser. How- ever. Experimental set-up The distillation set-up used in this study is shown in Fig. 1999 [10]. the maximum boiling temperature reached in the present study is approximately 400 C at 0. 1). (8) Cold trap. (10) between 50 and 150 C due to the risk of steam distillation [11]. Presence of Fig. and Eq. The ASTM standard D2892 [9] used as guideline in present study is capable of producing fractions with a final cut temperature of 400 C. During vacuum distillation the observed vapor column head temperature must be converted to atmospheric equivalent temperatures to be able to obtain a TBP curve. 2.1. (5) Distillation column. Product receivers. 1.% of water in the sample to achieve clean cuts of the fractions. The equations applied in order to convert the observed temperatures to AET are given in Eqs. To protect the vacuum pump from light gases a dry ice cold trap is used. (3) applies for pressures between 0. The vacuum pump is connected to the top of the upper condenser as well as to the fractionation arm. 1.maximum operating pressures up to 400 bar and temperatures up to 550 C.2. Dehydration has been done at pressures of 100 torr. To avoid heat loss and provide adiabatic conditions the column is surrounded by a tem- perature controllable heating jacket (not shown in Fig. this linearized correlation becomes inaccurate and another correlation must be applied from 2 to 0. (3) water in the crude oil will lead to changes in the TBP curve Column head. Close-up view of the product receiver. and a feed capacity up to approximately 30 kg per hour.2. The set-up consists of a 15 theoretical plate column with an inner diameter of 25 mm and a Pro-Pack filling. The standard . (1)–(3) [10] AET ¼ 748:1=ð½1=ðT þ 273:1Þ þ 0:3861 A 0:00051606Þ ½C ð1Þ The factor A is determined through Eq. Dehydration Prior to distillation the bio-crude has to be dehydrated.

Distillation has been run at total reflux collected at all operating pressures and is treated as one light frac- for approximately 15 min. due the fact that some oxygenates tend to crack at lower From AET temper.3. Elliott [1] describes that the decomposition point for bio-crude from wood is approxi. The dehydration decanting at procedure is described in ASTM standard D2892. com- dehydrated. Cut temperature ( C) The higher heating value (HHV) has been measured using an Cold trap <100 C IKA bomb calorimeter. Elemental analysis b By difference.% corresponding to 53. pot flask Kinematic Carbon viscosity residue at 40 C (cSt) (wt. pressure.%) H (wt.3 1 Heating value (MJ kg ) 40. the pot flask tem.6 C (kg m ) 970.069 approximately every 2 min. Fourier transform infrared (FTIR) spectroscopy analysis The functional group composition of the bio-crude and the Table 2 fractions were investigated by FTIR spectroscopy 1 using a Thermo Elemental composition of dehydrated a bio-crude.41 0.8 vol.3. 2. The water from oil is collected for further analysis. This error is soon as no increase in vapor temperature is noticed anymore.3. approximately 15 g boiling pared in Table 3.3. 11 fractions have been collected in total.tion the bio-crude has been cooled down Through vacuum distillation the dehydrated bio-crude has been and fresh boiling stones have been added prior to initiating successfully divided in its fractions. Vacuum distillation into the oil sample.65 wt. All material not collected below 375.% water has been removed from the bio- mately 55 C below that experienced for petroleum crude.1 torr. Total sediment (wt.ture) vapor temperature to ensure adiabatic 375. Density analysis 5 200–225 Density has been measured in an Anton Paar density meter at 6 225–250 15.lected product contains an organic and a water phase.3. This is crude sample by dehydration at 100 torr to an AET of 130 C.3.5. perature of 330 C at 100 torr is reached and a maximum pot flask temperature of 250 C at 0. The HHV has been done in triplicates.2.%) 0.2. A total of 3.%) 11. Flash point ( C) 39 Vapor temperature.%) <0.047 C (wt. which has been reintroduced to the crude after dehydration and distillation of HTL bio. 1 100–125 2 125–150 3 150–175 4 175–200 2.4 C is been kept at 0–5 C below AET (atmospheric equivalent considered as residue.4 wt.1 torr.2.37 skin temperature and volume of collected fraction are recorded Ash content (wt.4.1.dration the fractional distillation crude is at AET of 87 C. Results and discussion been found that when initiating the distillation at vacuum condi- tions the risk of cracking is decreased (observed as white smoke 3.97 3. and 4000 cmSpectra . 7 250–275 8 275–300 9 300–325 2. The heating jacket temperature has tion boiling <100 C.%) O (wt.crudes.% of the initial feed For distillation approximately 1200 ml of bio-crude have been has been distilled. The cold trap fraction includes the condensate stones have been added. be stopped and the vapor temperature should not exceed Analysis Result 210 C [9].atures than the corresponding hydrocarbons and thus <100 C organic material could be observed in the fraction a lower pressure has been found to be advantageous in the receiver. . Therefore the residue fraction also includes the fuel pressure P10 torr and a ratio of 2:1 6 10 torr.3. Simulated distillation 10 325–350 Simulated distillation has been carried out according to ASTM D 11 350–375 2887 [13] by an external certified oil testing lab. Table 3 Temperature intervals for collected fractions. To be quantified avoid cracking.49 H/C 1. To ensure even boiling.1.6 C according to ASTM standards D5002.78 3 2.4 C at 0.%) eff S (wt. The col. however it has 3.43 2.perature intervals compare Table 3. Fractional distillation In total 56. After dehy.4 wt. During distillation a maximum skin tem- has been initiated.%) a Below calibration limit. After dehydra.2. No signs of cracking have been observed. A mass balance shows that the collected in 25 C tem. Elemental composition has been determined in a Perkin Elmer 2400 CHNSO analyzer in CHN mode.4 5. Dehydration in the pot flask at previous distillation runs). The organic phase is decanted at 10 C and reintroduced 3.88 10. the due to some distillation hold-up in the system that could not system is cooled to be able to lower the operating pressure. [12]. 2. a b wavelength between380600spectrometer. In the current set-up no fractionation above tempera. 2. As distillation error is 2.perature and the skin temperature should not exceed temperatures of 310 C Table 1 respectively for more than an hour before the distillation has to Bio-crude analysis (dehydrated). Higher heating value (HHV) analysis Fraction no. suggests dehydrating at atmospheric pressures. A reflux of 5:1 has been kept during distillation for apparatus. The fractions were oil part of the bio-crude.% of the distillation yield.%) N (wt.39 O/C 0.01 Pour point ( C) 12 1 Total acid number (mg KOH g ) 36. sample is heated up until a vapor temperature of 130 C is reached. The initial boiling point of the vacuum distillation.4 C could be achieved due to limitations in the distillation conditions.31 Scientific Nicolet were recorded at H/C 1. The bio-crude 10 C. The highest vapor AET reached was 375. Bio-crude and fraction characterization Density at 15. 83. pot flask temperature.

fraction 1 and residue is shown in been done for all samples. Elemental analysis. fractions 1–11 and residue (from left to right). 4. of the trend which can also be recognized in the higher HHV leading to a reflux ratio of 0.4 c c Capunitan et al. . Furthermore.2. to get an insight into functional Fig. together with simulated distillation results and distillation 3. hydrogen and oxygen content data on conventional crudes of 2 different qualities. have been determined for the bio-crude./vol. which is desirable in terms of market Athabasca (extra heavy) 200 Brent (light) pull for fuels in this range. observed at the very end of the distillation procedure. Crude equivalent Gasoline equivalent Diesel equivalent HHV (MJ/kg) HHV (MJ/kg) HHV (MJ/kg) a Crude oil 45. cold-trap fraction.2 a [16]. generally decreases with increase in boiling point and increase in ing with conventional crude standards. The molar hydrogen to carbon (H/C) ratio allows mass curve contains less data points since the cumulative amount conclusion on the chemical carbohydrate structure in the samples. Fraction 8 falls slightly out The reflux valve has been opened too harshly during operation.43 Bio-crude fractions 1–3 43.3.1. 3.8 FAME 40. 600 This results in a separation and distribution error for fractions 2–4.3 Fraction (180–250 C) 37. b Weighted average.-%] for HTL bio-crude and its fractions (see Tables 1 and 2). The which does not match with the sim-dist curve at temperatures residue FTIR also shows large absorption in the 1600 cm 1 wave- between 140 and 180 C can be explained by operational errors. cold-trap fraction. The TBP residue. Due to its high volatility and resulting inconsistent results presented in the following section.2 Fraction (266–382C) 35.2. This waxy product has been elemental analysis could not be done for the cold trap fraction.0 d Elliott Wood based d d d HTL oil 36. d [7]. The uneven TBP curve aromatic structure with the lowest ratio for the residue.-%] Bio-crude sim-dist [wt. 3. suggesting C@C absorption. True boiling point and simulated distillation curve The true boiling point curve (TBP) on a cumulative mass and volume basis obtained from vacuum distillation is shown in 3.7 Bio-crude fraction 7–10 42. The results are presented as an assay in Table 5 and discussed up in the glassware after cool down of the system. Table 4 Comparison of different crudes and the equivalent gasoline and diesel fractions. A picture of the feed.5 Conventional gasoline 46. cold trap.4 Fraction (100–180 C) 37. HGO = Heavy gas oil) removal would lead to an improved TAN number. HHV and density Fig.6 b b b HTL bio-crude 40.-%] 100 Naphth a 3. fractions 1–11 and the mark Brent crude oil and Athabasca extra heavy oil [14].3. groups present and will be discussed in Section 3. Fourier transformed infrared spectroscopy (FTIR) has bio-crude. which resulted in a high amount of compared to the lighter fraction 7. Analysis of bio-crude and its fractions 0 0 10 20 30 40 50 60 70 80 90 100 A selection of bulk and fractional properties has been measured Distillate [wt. length. Carbon.-%] middle distillate region. Elemental analysis. Picture of bio-crude feed.1. The Fraction 1 since it is the lightest has the highest H/C ratio.5 Low sulfhur diesel 45. the flatness of the distillation LG curve compared to these indicates a large fraction in the 300 O Bio-crude [wt. The operating reflux valve has been tightened immedi- Temperature [°C] 700 Residu e ately after a flush out of the distillation head had been noticed. Distillation curves from simulated and TBP distillation (LGO = Light gas oil. as a hold. Fig. Bio-crude [vol. density and distillation yield Fig. 800 distillate volume being flushed out of the distillation head at the same time. A picture of the below. of distillate has only been logged in volume during distillation. Regarding bio-crude quality it can be seen that the TBP 500 HGO curve lies between the curves of extra heavy and light 400 benchmark crudes. c [17].1. Elemental analysis. light bench. The high viscosity of the residue at room temperature made it impossible to measure the standard and therefore has not been included in the mass balances density. like shown with FTIR analysis. higher heating value (HHV). The H/C even though simulated distillation curves are obtained by calibrat.% have been measured for the bio-crude. cold trap and fractions 1–11 and the residue is given in Fig.3. 4.9 a a NA Bio-ethanol 29. in wt. 3.8 Fraction (up to 138 C) 37. fractions 1–11 and residue. suggest- TBP curve and the simulated distillation curve match very well ing short chained alkanes. 3. /Corn stover PO c c 35.3.1.

48 n. HTL bio-crude contains 5.12 1.54 1.3.44 1. The n.4.59 and paraffinic fraction since the FTIR of the cold trap fraction sug- 2. These factors have to 12.772 2.789 3. The HHV is rather constant throughout the frac- 1. 5 also shows the Bio-crude and fraction analysis.%.53 lower effective H/C ratio compared to the other fractions.95 can also be seen in the FT-IR spectra through alcohol groups in n. To enable comparison of energy con- 0.47 84. Higher heating value. Fractions 1–3 have oxygen contents c between 1.% resulting in a slightly lower HHV.64 85.88 1.17 2.% oxygen in the residue.a c two different petroleum crude oils found in [15].61 ported by the FTIR results.3. This is also sup- 83.a c 3. Table 5 Dry basis.2 n. A high acid Fraction number will lead to corrosion problems during refining and it is C)8 (275–300 expected that oxygen.08 n.195 9.% in fraction 10 and maximum of 9.15 1. 11. These Fraction deviations can be explained by the missing analysis of the cold trap C)5 (200–225 12. corn stover pyrolysis a oil [17].3 wt.3 wt.769 1.31 10. Ash (%) Proximate from other conversion processes. The higher 6. Density.87 87.9 n.a crude fractions would need more extensive refinery processing in c Cold trap Fraction 1 (100–125 C) regard to decreasing the density compared to processing of con- ventional crude. the lower the oxygen 10. Fig. Fig.3. fractionation arm) and no analysis Fraction could be done on it.29 n.82 86.4 wt.06 5.17 84.72 1.322 5.a compare Fig. 6 shows the elemental and energy Fraction balance of the collected distillation product compared to the bio- C)7 (250–275 crude that has been distilled. state of the art literature data on HHV has been collected and listed in Table 4. As expected the highest oxygen content can be found in the residue fraction with 9. Distillation balance. The lower H/Ceff.85 85.13 1.a these specific fractions.% (fractions 8–11) C)9 (300–325 11.a n.667 2.1.a 0.44 n. wood based HTL oil from the literature [7] and the HHV of their equivalent gasoline and diesel fractions. Fig. compare 3. the more severe 10 (325–350 the upgrading has to be. with the highest for fraction 6 with (>375 C) Res 3.a gest that it contains oxygen rich aliphatic ketones. 5.1.09 n.05–1.6 and 2.49 1.13 n.61 Compared to conventional crude with an oxygen content of 7. Fractions 4–7 contain n.37 ties of HTL bio-crude fractions with the corresponding fractions of n.99 would not need deoxygenation prior to co-processing.914 2.a n. No fraction boiling b .385 97.48 88.a 3. it can be seen that the densities Fraction Fraction 2 (125–150 of the bio-crude fractions are higher than specifications.786 1.728 3.% overall and a mini- c FractionC) mum of oxygen 0.2.a tions with the residue having the lowest HHV as expected and the cold-trap fraction having the highest. 12. Elemental analysis on the cold trap fraction C)6 (225–250 could not be obtained as described earlier.80 hold up could not be obtained since it stuck to the glass parts of the 6. 12.88 100 content the higher the HHV.07 tent of the HTL bio-crude to conventional fuels.2 wt. It contains the HHV literature data of petroleum crude oil [16].6–1. The amount of paraffinic distillation 11.724 2. bio-fuels and -oils a Calculated by difference. c The highest error can be seen in the hydrogen balance.9 wt. The higher oxygen 43.a n.%.3. In general the densities are higher than those of pet- C)3 (150–175 roleum crude boiling in the same range.12 1. The figure also gives standard fuel specifications.3.34 1. 13. Approximately 25 wt.57 84.1. and oxygen has been calculated by difference a slight error C)4 (175–200 in the elemental balance is expected as well.89 8.27 1. Oxygen is present in all fractions. 5 shows the elemental oxygen content of the HTL bio-crude Balance and its fractions.%.425 1.21 5.2.a c effective H/C ratio is generally used to estimate the upgrading/ deoxygenation effort needed.a c distillation set-up (condensers. 7 visualizes the densi- 12.a loss sums up to 2.8 mg KOH g 1. The density increases with increasing boiling point Fraction as expected.41 83. n.09 2.86 n. Fig.18 due to oxygenates present throughout every boiling fraction.%.34 wt.a c 3.9 wt.11 n.634 0.6 wt. Bio-crude dependency of HHV to the oxygen content. Bio- 9.29 85.a n. The higher heating value (HHV) has c been measured for the HTL bio-crude as well as for its fractions.34 be considered when interpreting the balances. which is 12.a higher amounts of oxygen.64 0.27 88.535 0.54 84.5 wt.16 81.% resulting in a 1. Since nitrogen c and sulfhur during elemental analysis were below the calibration Fraction value. The total distillation 5. Fractions 8–11 have a very low oxygen FractionC)11 (350–375 C) content of 0.a oxygen content compared to conventional crude also leads to a c higher total acid number (TAN) of 36.

04 38 0. 9 and 10 the oil and the Elliott wood based HTL oil and their fractions are signif.10 HHV 44 O/C 0.32 C [wt. but with limitations on drop-in ability and mileage.38 O [wt. O. The HHV of pyrolysis the functional groups present in the crude.-%] 83.3.80 Dist. C Fig. In Figs.66 83.08 O/C [mol mol -1] HHV [MJ kg-1] 42 0.06 Error [%] H/C_eff 0. The overall bio-crude FTIR is hard to inter- . 6.05 39 0.49 5. pret and analyzing the fractions as well gives a clearer idea of so the 180–250 C is used for comparison.047 O/C [mol/mol] 1. 3. FTIR analysis Interpretation of the main bands in the FTIR absorption spectra was based on [18. Elemental oxygen content and HHV over the fractions.07 41 0.46 H/C [mol/mol] 1. 5.31 2. H/C 0 1 2 3 4 0.39 2.43 HHV [MJ kg-1] Biocrude 1. O/C. unlimited refinery drop-in ability is believed to be achievable by deoxygenation prior to co- processing. Balance 40. The HHV of bio- ethanol and FAME are also given for comparison with the HTL bio-fractions.79 1. in the diesel range was distilled by Capunitan and Capareda [17].01 40. whereas the HTL bio-crude and its fractions are slightly lower than the conventional crude oil and the respective gasoline and diesel equivalent.44 10.-%] 5. H.-%] 10. Considering the HHV only.00 Biocrude Fraction 1 Fraction 2 Fraction 3 Fraction 4 Fraction 5 Fraction 6 Fraction7 Fraction 8 Fraction 9 Fraction 10 Fraction 11 Residue Fig.2.37 1. Energy and elemental balance.88 0 10 20 30 40 50 60 70 80 90 100 HHV.icantly lower. the HTL bio-crude shows potential as a sustainable fuel pro.02 36 0.03 37 0.06 40 0.01 35 0. 45 0. Error. Furthermore. These biofuels are presently used to a great extend worldwide.16 H [wt. Oxygen.viding a longer mileage than bio-ethanol and FAME.05 0.19].09 43 0.41 0.

the peaks at 1720 and 1 1 cm for the overall bio-crude suggest the presence of alcohol 1745 cm also suggest and carbonyl functional groups. 8 indicates with peaks at 2900. all fractions show 3 medium-strong sharp C-H peaks around The FTIR of fraction 1 also indicates 5-membered cyclic ketones 2853–2955 cm 1 along with the CAC peak at 1450 cm 1 and due to a sharp peak at 1745 cm 1. 1450. The presence of aliphatic 1 1 absorptions at 1375 cm and 720 cm which indicate esters is underlined in fraction 1 by the presence of the 1 paraffinic hydrocarbons. 8 fractions 1 and 2 and tions 1 and 2 mainly contain aliphatic compounds with a distinct 1 the CBS27 bio-crude are shown. 720 that frac- refinery cuts are being presented. 1375. In Fig.FTIR spectra of the bio-crude fractions compared to conventional Fig. and underline the presence of oxygenates in the crude. . The stronger intensity of the 3300–3600 carbonyl (C@O) stretch at 1745 cm together with a CAO cm 1 band (AOH stretch) and the CAO stretch at 1260–1050 stretch between 1200 and 1300 cm 1. The bio-crude and consequently peak at 1715 cm indicating aliphatic ketones for fraction 2.

6 Fuel Specifications 0. USA Bulk density: 0. FT-IR diagram of fractions 4–6 compared to kerosene.840 g/ml 1 Marine RMA <0. Marine Residual Fuels (RMA): fraction > 10. 7. unsaturated esters for fraction 1. Fig.720-0. jet-fuel: fractions 4–7. Fig.775 g/ml Jet-Fuel: 0. It is suggested that distinct.5 Cold Trap 1 2 3 4 5 6 7 8 9 10 11 Residue Fig. CA.840 g/ml Diesel: 0. The . It is apparent that stretches as described earlier. Comparison of fraction densities to fuel specifications. Table 6 shows a summary of pos. which is probably due to the paraffinic and pure nature all 3 fractions contain cyclic compounds and more detailed of kerosene. 5 seems to have a stronger intensity in the overall bands and peaks. TX. Fig. 8. The CAO stretch also suggests the FTIR diagrams of fractions 4–6 look quite similar.8 0. USA Bulk density: 0.775-0.970 g/ml Hasting Field. The kerosene FTIR diagram is less noisy and more CAO absorption between 1200 and 1300 cm 1. whereas fraction presence of ethers in fraction 1.7 HTL Biocrude Bulk density: 0. The fractions also contain aliphatic compounds as well as cyclic.9 0. the density of the residue has been estimated. 9. sible functional groups in the fractions. diesel: fractions 7–10. 1.1 Fuel Specifications: Gasoline: 0. 9 compares the kerosene fractions of the bio-crude to a Alcohol functional groups are suggested due to OAH and CAO con.820-0.920 g/ml Density [g/ml] 0.867 g/ml Elk Hills. Gasoline: fractions 1–3. Fractions 1 and 2 and bio-crude feed.896 g/ml 0. Esters and ethers are indicated by differences can mainly be seen as oxygen containing carbonyl 1 absorption between 1735 and 1740 cm (C@O) and due to the stretches.ventional kerosene spectrum.

4 wt.5 wt. Compared to conventional small at 12. 1 the strong broad peak at 1130 cm indicates CAO bonding along Fractions 8–11 Alkanes (less) Esters/ethers Substituted aromatics Aromatic ketones with the carbonyl peak around 1745 cm 1 esters/ethers can be indicated. Summary of FTIR analysis. Conclusion ing a carbonyl group sitting on a 7-membered ring or larger [18]. Fig.18 is proposed to Fractions 4–6 Alkanes Alcohols mainly contain aromatics. Table 6 summarizes the expected compounds of the fractions and the residue according to the IR spectra. Fig. Substituted aromatics Esters/ethers Aromatic ketones The high O/C ratio indicates oxygen containing functional groups.% at AET 375 C. 10. The oxygen distribution along the temperature profile. Residue Substituted aromatics Esters/ethers Carbonyl groups (less) 3.6 wt. which becomes obvious looking at the IR spectra. indicat- 4.2. FT-IR diagram the cold trap and the distillation residue. It has been seen that HTL bio-crude is thermally stable and fractional separation under vacuum possible. The significant peak at 1550 cm 1 along Cyclic (less) Esters/ethers 1 7-Membered cyclic ketones with the different peaks around 700–850 cm indicate substi- tuted aromatics. 11. 7-membered ketones due to the sharp peak at 1700 cm 1. The additional peak at 1375 cm 1 5-Membered cyclic ketone indicates CH3 methyl group. 1 and the carbonyl C@O stretches with peaks at 1708 cm of the Table 6 Interpretation of FT-IR spectra of the various fractions. Fractions 2 and 3 Alkanes Aliphatic ketones The residue fraction with an H/C ratio of 1. Only major absorptions are bio-crude fractions make the main difference. groups due to the OAH stretch between 3300 and 3600 cm and The equivalent gasoline fraction of the bio-crude is relatively 1 the CAO stretch 1050–1150 cm .3 and 16. FT-IR diagram of fractions 7–11 compared to diesel. Die- successfully with an overall mass recovery of 97. evaluated. Furthermore absorptions 1 A major outcome of this study is the distillation profile and the at 1500 to 1600 cm indicate some aromatic compounds. 1450 and Fraction 7 Alkanes Alcohols 1 1375 cm are probably related to substitution on aromatic rings. Fraction 7 also contains alcohol functional 1 found in literature.7% and a yield sel has a paraffinic nature. Functional groups in the cold trap Esters/ethers 1 are mainly saturated aliphatic ketones (C@O stretch at 1715 cm ).% respectively.%. Previous work identified peak around 700–850 cm 1 suggests substituted aro- on HTL non-destructive fractionation and analysis could not be matic compounds [18]. peaks at 2900. 10 shows the FTIR diagram of diesel boiling range fractions Fractional distillation on HTL bio-crude has been carried out (7–11) of the bio-crude compared to conventional diesel fuel. 1450. The clear CAH peaks around 2900. whereas the diesel and jet-fuel fraction are rel- 1 diesel fuels CAO (ester) stretches at wavenumbers <1300 cm atively large at 25.1. . The spectra of the 4 fractions indicate alkanes with of 53. 1375 and 720 cm 1. 11 shows the FT-IR spectra of the cold trap and distillation residue. Fig. Compounds in the cold trap have alkane structure due to Sample Hydrocarbon Structure Functional Groups CAH bond absorption from 2850 to 2960 cm 1 and methylene Cold trap Alkanes Aliphatic ketones Fraction 1 Alkanes Aliphatic ketones CH2 scissoring at 1450 cm 1. In terms of product Fig.3.

Jensen PA. 2005. Knudsen KG. to prevent corrosion issues of processing equipment dx. The overall HTL bio-crude would [6] Jensen CU. Elemental analysis and FTIR show that the oxygen content is distributed to a higher content in the gasoline and jet.php?&iback=3&rub= The authors are grateful to Steeper Energy ApS for 11&expand=3&image=europe>.doi. Appl Catal A 2011.totsa. Nankodo Company work by C3BO DSF Grant No. Davis SC. 1999. 2015. their petroleum counterparts. 10-094552. [17] Capunitan JA. 4th the Shell Refinery in Fredericia. Part 2: A parametric hydrotreating Study. <http://www.86 wt. D2887 standard test method for boiling range distribution of petroleum fractions by gas chromatography. in the mostly diesel range (fractions 8–10). D5002 standard test method for density and relative density of crude oils by digital density analyzer.112:60–73. No. 2007.S. energy information. Fuel 2014. Xiu S. Acknowledgements [13] ASTM. Shahbazi A. BP statistical review of world energy 2014. 2014. It would allow a straight forward the glycerol-assisted liquefaction of swine manure.fuel.1016/j. Refinery engineering. Characterization and separation of corn stover bio- oil by fractional distillation. D2892 standard test method for distillation of crude introduction to the market and keeping in mind the current over petroleum (15-theoretical plate column). Denmark for supplying ed. at petroleum refineries.2015. bio-crude characteristics need to be similar to catalytic upgrading of bio-oil to engine fuels. Rosendahl [9] International A. which are even expected to [10] ASTM. Bench Scale Unit. Weinheim: Wiley-VCH. Office of the Biomass Program. fuel fraction (fractions 1–7) and to a lower content (0. administration. Wright L. ASTM International. [8] Cheng D. Fuel 2013. Refinery of the future. The authors would also like to acknowledge funding of this [18] Nakanishi K. 2003. Department of [14] TOTSA. Process development for biomass liquefaction.% O) [2] International energy outlook. Petroleum refining. Diegel SW. Distillation and vaor pressure measurement in petroleum even more appealing. 1980. [7] Elliott D. 1305-00030B and FlexiFuel Grant Limited.407:1–19. during co-processing of the bio-crude to existing petroleum refin. technology and economics. [3] Chang A-F. 1962. D5236 standard test method for distillation of heavy hydrocarbon rise over the next years. Wang L. ASTM International.08. [11] Montemayor RG. products. [5] Speight JG. 2013. Prentice Hall. Infrared absorption spectroscopy: practical. 2006.S. co-processing of HTL bio-crude becomes mixtures (vacuum potstill method). capacities existing modern refineries. Oxford: Elsevier. Pashikanti K. we acknowledge with appreciation [16] Boundy B. [12] ASTM. U. [19] Wade LG. To ensure compatibility 2012. 2014. Characterization of the physical Nevertheless the overall co-processing potential of bio-crude and chemical properties of the distillate fractions of crude bio-oil produced by from HTL is very promising. 2008. . Biomass energy databook. Zhang B. Handwerk GE. Capareda SC. Elliott DC. hydrotreating catalyst and a sample of North Sea fossil crude oil. Co-processing potential of HTL bio-crude need light hydroprocessing to remove oxygenates prior co.047. [4] Mortensen PM.value.130:251–6. which will be studied in the 2nd part of this publication. A review of ery infrastructures. 2011. U. Grundwaldt J. Additionally. [1] BP. Liu YA. 2011. the middle distillate region is expected to dominate future References fuel markets. 2010. Hoffmann J. CRC Press. ASTM International. 8th ed. providing state-of-the-art HTL bio-crude from the Continuous [15] Gary JH. http:// processing. Jensen AD. Kaiser MJ. Organic Chemistry volume 1.