Mechanism of Formation of Soft Particles in Biodiesel Fuel Blends
ROBERTA MARIA FIORENZA
The large environmental impact related to the use of fossil fuel has driven the shift toward renewable sourced alternatives. Fossil diesel can be nowadays replaced by biodiesel, obtained from vegetable oils and fats, mostly used as biodiesel blends. However, some drawbacks are related to the use of this bio- fuel, among which the formation of deposit in injectors and filters causing a reduction of engine performances or engine failure.
This thesis project focuses on the analysis of the mechanism of formation of soft particles in biodiesel deposits. These particles are constituted mostly of carboxylic metal soaps, and were found in biodiesel engines after using aged biofuel with contaminants, such as engine oil. The role of short chain fatty acids (SCFAs) has been investigated, together with the use of three different calcium sources, to analyse the formation mechanism of calcium soaps. Arti- ficial ageing of B10 and B100 test fuels was performed, and in some cases an inert gas was bubbled to remove the formed SCFAs. Calcium sources, namely calcium oxide, calcium carbonate and engine oil, were added to investigate the formation of soft particles. Ion chromatography, pH measurements, NMR spectroscopy and oxidation stability tests have been performed on the liquid test fuel to verify the presence and effect of SCFAs, while FTIR spectroscopy and GC/MS analyses were used to verify the presence of calcium soaps in the deposit and in solution.
In contrast to the expectation, it was found that the presence of SCFAs in the fuel is not fundamental for the formation of carboxylic soap. Moreover, the different calcium sources result in different amounts and textures of metal soaps in deposits and in solution.
Den stora miljöpåverkan relaterad till användningen av fossilt bränsle har dri- vit övergången till förnybara alternativ. Fossil diesel kan numera ersättas av biodiesel, härledd av vegetabiliska oljor och fetter, oftast används som biodie- selblandningar. Vissa nackdelar är emellertid relaterade till användningen av detta biobränsle, bland annat bildandet av avlagring i injektorer och filter som orsakar en minskning av motorprestanda eller motorfel.
Detta avhandlingsprojekt fokuserar på analysen av mekanismen för bild- ning av mjuka partiklar i biodieselavlagring. Dessa partiklar består huvudsak- ligen av karboxylmetall tvålar och hittades i biodieselmotorer efter användning av åldrad biobränsle med föroreningar, såsom motorolja. Rollen för kortkedji- ga fettsyror (SCFA) har undersökts, tillsammans med användning av tre olika kalcium komponenter, för att analysera bildningsmekanismen för kalciumtvå- lar. Experimentellt åldring av testbränslen B10 och B100 på labbet utfördes, och i vissa fall bubblades en inert gas för att avlägsna de bildade SCFA. Kal- cium komponenter, nämligen kalciumoxid, kalciumkarbonat och motorolja, tillsattes för att undersöka bildningen av mjuka partiklar. Jonkromatografi, pH-mätningar, NMR- spektroskopi och oxidationsstabilitetstester har utförts på det flytande testbränslet för att verifiera närvaron och effekten av SCFA, medan FTIR-spektroskopi och GC/MS-analyser användes för att verifiera när- varon av kalciumtvålar i sedimentet och i lösning.
I motsats till förväntningarna visades det att närvaron av SCFA i bränslet inte är grundläggande för bildandet av karboxyltvål. Dessutom resulterar de olika kalciumkomponenterna i olika mängder och strukturer av metall tvålar i sediment och i lösning.
This Thesis work is dedicated to everyone who helped and supported me dur- ing these months.
I would like to thank my supervisors at Scania CV AB, Henrik Hittig and Mayte Pach, for their help and support during this journey. Thanks to them as I learned more than I expected and I enjoyed working on this project. I also thank my supervisor at KTH, Cláudio Lousada, for his support and useful tips to complete the project successfully.
I thank the Doctoral student Botond Csontos for his help and suggestions throughout the project. Thanks also to the other thesis students of the YTMC group at Scania CV AB, their presence made my long travels, work and break times more enjoyable.
Thanks to my friends, the lone survivors, for being with me during the en- tire Masters and for helping me to survive the cold Swedish winters and enjoy the warmer times. Special thanks to Sabrina, with whom I enjoyed having our
"girl talks" and sharing the best views in the city.
My heartfelt thanks to my Shibboleth friends, in particular, Arianna, Cate- rina, Alfredo and my vdrm: Charlie and Luisa. Thanks for your friendship during all these times and without you I wouldn’t be the same person whom I am today.
Thanks also to my friend Lucia for sharing with me all the good and bad moments, and for letting me know that she will be there.
My warmest thanks goes to Arun. Your support during these months was fundamental for me to overcome the hard times. Thanks for spending these moments with me and for being my quarantine.
Last but not least, special thanks to my family. Without your support, I wouldn’t be here today. Thanks for always believing in me and for helping me to be more confident and making me aim for the best.
Stockholm, May 2020 Roberta Maria Fiorenza
B0 Fossil diesel
B10 10% biofuel diesel blend B100 100% biodiesel
FAME Fatty acid methyl ester
FTIR Fourier transform infrared spectroscopy GC/MS Gas Chromatography – Mass spectrometry GHG Green house gasses
IC Ion chromatography
IDID Internal diesel injector deposit IP Induction period
KTH Kungliga Tekniska Hogskolan, Royal Institute of Technology LCFA Long-chain fatty acid
MK1 Standard fossil diesel
NMR Nuclear magnetic resonance RME Rapeseed methyl ester SCFA Short-chain fatty acid TAN Total Acidic Number XRF X-ray fluorescence
1 Introduction 1
1.1 Aim . . . 2
1.2 Scope . . . 2
1.3 Project Boundaries . . . 3
2 Background 4 2.1 Biodiesel . . . 4
2.2 Literature Review . . . 6
2.2.1 Biofuel Oxidation . . . 6
2.2.2 Deposit Formation . . . 6
3 Methodology 9 3.1 Strategy . . . 10
3.2 Ageing Method . . . 10
3.3 SCFA Removal and Addition . . . 11
3.4 Calcium Addition . . . 11
3.5 Analytical Methods . . . 12
3.5.1 Samples Preparation . . . 12
3.5.2 pH Measurements . . . 14
3.5.3 Rancimat . . . 14
3.5.4 Chromatography . . . 16
3.5.5 NMR . . . 17
3.5.6 FTIR . . . 18
4 Results 20 4.1 SCFA Measurements . . . 20
4.1.1 pH . . . 20
4.1.2 IC . . . 21
4.1.3 NMR . . . 22
4.1.4 Rancimat . . . 24
4.2 Soft Particles Measurements . . . 24
4.2.1 Aged Fuel . . . 26
4.2.2 Fresh Fuel . . . 28
4.2.3 Nitrogen Bubbling . . . 30
5 Discussion 33 5.1 Aged Fuel . . . 33
5.2 Bubbled Fuel . . . 34
5.3 Fresh Fuel . . . 35
5.4 Common Findings . . . 35
5.5 Ethics and Society . . . 36
6 Conclusions 37 7 Future Works 38 8 References 39 Appendices 45 A Experimental Setups 46 B Calculations 48 B.1 Calcium Addition . . . 48
B.2 Formic Acid Addition . . . 49
C NMR Spectra Results 50
D Rancimat Results 53
E FTIR Spectra Results 57
F GC/MS Spectra Results 64
Introduction
The need to switch into a more sustainable society is becoming more signif- icant in the recent years. The automotive industry is contributing by imple- menting the use of renewable sourced fuels and electric vehicles [1], due to the high environmental impact related to the use of fossil fuels.
Biodiesel, a biofuel derived from vegetable oils and fats, has great poten- tial as a substitute of fossil diesel thanks to the redution in emissions related to its use [2]. Moreover, when this biofuel is blended with fossil diesel, it can be used in diesel vehicles without any engine modification. However, in spite of the many advantages it presents, it may also cause reduction in engine perfor- mances [3]. The most important drawbacks are the result of the formation of deposits in vehicles engines filters and injectors, due to biodiesel ageing and contamination [4].
The objective of this project is to achieve a better understanding of the mechanism of formation of soft particles in the deposit formed in biodiesel fuel blends. In particular, the role of short chain fatty acids (SCFAs) generated by biodiesel oxidation.
Although different kinds of molecules have been detected in the deposit [4–9], this project focuses on metal soaps, namely carbon chain molecules ending with a carboxylic metal group. In a previous study conducted at Sca- nia on deposits found in engine filters and injector deposits [10], calcium soap was found to be the main component of the deposits. The source of calcium in diesel engines is suspected to be the engine oil used to lubricate the en- gine pumps. A small amount of engine oil can contaminate the biodiesel and catalyze the reaction of soft particle formation [10].
Different types of biodiesel and biodiesel blends are used worldwide. How-
ever, the same kind of particle deposits have been found in engines using dif- ferent biodiesels [4]. In this project, Rapeseed Methyl Ester (RME) was used since it represents the most used biodiesel in Europe [11]. In particular, the biofuel that has been tested consists of a blend of fossil diesel with 10% RME biodiesel, labelled as B10. It has been observed in previous studies that this blend has generated the largest amount of deposits [12].
1.1 Aim
The aim of this project is to achieve a better understanding of the mechanisms of formation of soft particles found in deposits of biodiesel fuel blends. In particular, the main focus is on the formation of calcium soaps, namely carbon chain molecules terminated by carboxylic groups bonded with calcium. This kind of deposit has been chosen since it is one of the main components of the deposits formed in biodiesel engines [10]. Moreover, the formation mecha- nism has been studied for FAME (Fatty Acid Methyl Ester) biodiesel blend B10, consisting of a mixture of 10% of RME and fossil diesel. The role of SCFAs, generated by biofuel oxidation, has been investigated by monitoring and manipulating their concentration. In order to achieve this, the biofuel is first artificially aged and subsequently the studies on calcium soaps formation mechanism are performed.
1.2 Scope
Understanding the deposit formation mechanisms is of high relevance for im- proving the performance of biofuel based diesels. To have a clearer idea of the particles formation mechanism implies to know about their structures. For this to be achieved, it is essential to develop a reliable protocol in order to artificially reproduce their formation in well-controlled laboratory conditions.
This will also lead to a better understanding of the mechanisms of deposition and adhesion of these particles in engines injectors and filters.
A good knowledge of the physical-chemical features of soft particles is essential to be able to design more robust injectors and filters. This in turn will improve the performance of biodiesel engines. Its direct outcome is that it will increase the deployment of biodiesel engines for transportation, with a consequent decrease of GHGs emission [1]. The replacement of fossil diesel with biobased fuel will hence lead to a more sustainable transportation system.
1.3 Project Boundaries
This project has been designed taking into consideration the following condi- tions:
• The time limit for the project is 20 weeks.
• The used biodiesel is FAME fuel, in particular, B10 RME.
• The effect of contamination, other than engine oil, should not be cov- ered.
• Fuel test rigs, such as injector and filter rigs, are not relevant for this project.
• The mechanisms of adhesion of soft particles in injectors and fuel filters should not be investigated.
Background
The usage of biodiesel as fuel has a series of environmental benefits [2, 13, 14]. However, many drawbacks related to biodiesel engines performance have been observed all around the globe (Fig. 2.1). In this chapter, biodiesel and its properties will be introduced. This will be followed by a literature investiga- tion on previous studies of the formation of deposit on fuel filters and injectors.
Figure 2.1: Biodiesel fuel drawbacks, where HFRR stands for High Frequency Re- ciprocating Rig, a measure of lubricity of diesel fuel [15].
2.1 Biodiesel
The term biodiesel refers to FAME, a fuel compatible with diesel engines and that derives from biological sources [16]. It is obtained from the transesteri-
fication of vegetable oils or animal fats. During this process, the triglyceride contained in these fats reacts with methanol, ethanol, propanol or butanol, through the use of a catalyst, to give an ester along with glycerol as byproduct [13, 14] (Fig. 2.2). This kind of biofuel, designated as B100, is a combina- tion of long chain fatty acids (LCFA) to which a series of additives is added in order to increase its performance and mitigate some of its disadvantages [14].
Figure 2.2: Biodiesel production, transesterification reaction [5].
In this study it has been used a biodiesel fuel blend containing fossil diesel and 10% of rapeseed methyl ester (RME), namely biodiesel derived from rape- seed oil, since it represents the most commonly used biodiesel in Europe [11].
It is designated as B10. Detailed information about the composition of biodiesel and biodiesel blends are not available to the public. From the research by Hoekman et al. [3], it is possible to know the main fatty acid components of different biodiesels. For RME, the three main fatty acids are: oleic, linoleic and linolenic with respective relative amounts of 59.5% (±7.8), 21.5% (±2.8) and 8.4% (±1.3) [3].
Biodiesel has many advantages over its fossil alternatives, but also has some drawbacks. B100 is classified as non-flammable and it is biodegrad- able, which makes it safer to transport and handle. Besides, it has a higher combustion efficiency than fossil diesel [14]. The most important advantages are its non-dependence on fossil resources and the consequent reduction of net emissions that accompanies its usage. The use of this renewable fuel leads to a significant decrease in GHG emissions of CO, N2O, SO2 and also particu- lates. The emission reduction of B20, compared to fossil diesel, of hydrocar- bons, particulate matter and carbon oxyde, is estimated to be 21,1%, 11,0%
and 10,1%, respectively. However, a 20,0% increase of NOxemission has been detected [2, 14]. This is not the only drawback related to biodiesels. Among the most significant ones it is important to mention the following: higher vis- cosity and engine wear, lower energy content, lower engine speed, lower power and torque [14], lower performances at cold temperatures and, the focus of this
project, clogging of injectors and filters [16, 17]. These pose constraints to the usage of B100 as vehicle fuel. Therefore, in order to minimise these problems and still draw benefits from the desirable biofuels properties, biodiesel blends are preferably used [14, 17].
2.2 Literature Review
Many studies have been conducted on biodiesel oxidation and on the prob- lems related to it. This project focuses in particular on internal diesel injectors deposit (IDID) and filters clogging. In this section previous results will be presented and discussed.
2.2.1 Biofuel Oxidation
One of the problems associated with biodiesel is its susceptibility to oxida- tion when exposed to the atmosphere. This is referred to as the ageing pro- cess. Upon oxidation, short acidic chains are formed in the biodiesel making it chemically unstable and decreasing its quality [18]. Biodiesel oxidation is also strictly related to deposit formation and to clogging of filters and injectors [13].
During the oxidation process, peroxides and hydroperoxides are formed.
These will degrade to aldehydes, ketones and other shorter chains products, like SCFAs, which increase the acidity of the biofuel [19]. Moreover, longer chain products, which result in insoluble compounds, are also formed due to polymerization reactions promoted by the presence of double bonds [18]. The oxidation rate, which is an index of formation of organic acids, was found to be related to the formation of these products, as well as to the number of double bonds observed in the biodiesel’s unsaturated chains [12, 20].
2.2.2 Deposit Formation
Formation of deposits in biodiesel fuel filters and injectors has been previously studied [5, 7, 8, 12]. In these studies it was observed the formation of soft and hard particles, the first being mostly water soluble carboxylic metal soaps, the latter being polymeric particles that tend to precipitate [7].
In a previous study conducted by Csontos et al. [10], the deposit collected from Scania fuel filters was analysed. With FTIR spectroscopy it was found
that the deposit includes mostly carboxylic soaps, formed due to the interac- tions between biodiesel oxidation products and contaminants. Moreover, from XRF analysis of the particles, calcium was identified as the main component (Fig. 2.3). The presence of P and Zn, together with Ca, indicates that the soft particles are generated due to the contamination of the biodiesel by lubricant oil [10, 21].
Figure 2.3: XRF measurement of three different biodiesel fuel filters. The filters were collected from three different trucks, operating under similar conditions [10].
Similar particles have been identified by Lacey et al. [7]. Carboxylate soaps were observed in the biodiesel deposits, and the presence of Na, Zn, Ca and P has been identified as an index of contamination from corrosion inhibitors, lubricity additives and residues from biodiesel production. The hy- pothesis of formation of these particles was identified by the reaction of the fuel contaminants together with the acids generated by biodiesel oxidation.
An analogous assumption was made by Fang and McCormick [5] and Omori et al. [12]. Iron ions are assumed to react with acidic oxidation products in order to form carboxylic iron soaps, as showed in Fig. 2.4.
The formation of acidic compounds with biodiesel oxidation has been stud- ied and confirmed by previous studies conducted both on fuel collected from the market and artificially aged [15, 22]. Formic, acetic, propionic, and caproic acid have been identified through the use of ion chromatography. However, there is still an ongoing discussion on the role of these SCFAs on the deposit formation [8].
During previous projects conducted at Scania CV AB the formation of soft particles was mimed by adding CaO to B10 and by simulating an accelerated
Figure 2.4: Assumed deposit formation and adhesion mechanism [12].
ageing by heating the fuel together with stirring and air bubbling [23, 24]. This procedure will be described more in detail in the methodology section.
The biodiesel blend B10 has been used since it was found that, together with B20, it tends to form the larger amount of deposit [12], as showed in Figure 2.5.
Figure 2.5: Deposit solubility with biodiesel content in biofuel blends [12].
Calcium oxide has been chosen both because it has been proved to be able to react with the biofuel to form calcium soaps and also because of its availabil- ity within the Scania’s laboratory. Through this method it is possible to obtain calcium soaps with a similar composition to the one that has been found in the trucks’ engine [25].
Methodology
Experimental tests and analysis have been performed in order to evaluate the impact of different parameters in the mechanism of formation of metal soaps.
The main focus of this project is to understand the role of the short-chain fatty acids (SCFAs) in the formation of these soaps. In order to confirm their in- volvement in the reaction, the biodiesel employed in these studies was subject to different test conditions in order to obtain biodiesel samples with different SCFAs content.
The different test fuels used for this project are represented in Fig. 3.1.
As previously mentioned, the biodiesel used for the experimental studies is FAME B10 because it represents the biofuel used by the trucks in the field.
The B10 biodiesel blend consists of 10% RME biodiesel and 90% fossil diesel.
Different samples of this fuel have been prepared. In some samples the fuel was aged in order to trigger the formation of SCFAs while in other samples the SCFA formic acid (CH2O2) was added to the fresh fuel. The B10 fuel was also produced by ageing the B100 biodiesel and by mixing it in the right amount (10% B100) with the fossil fuel B0, MK1.
Figure 3.1: Biodiesel test fuels and experimental steps.
Three calcium sources have been used for the reaction, namely CaCO3, CaO and engine oil. The reactivity of these compounds has been tested with all the different test fuels and the results have been analysed. A summary of the test fuels preparation steps and analyses is showed at the end of this chapter, in Table 3.1.
3.1 Strategy
The experiments and analyses that have been run for this project aimed at eval- uate a hypothesis for the mechanism of formation of soft particles. The first hypothesis is that this is a two step mechanism that can be schematically rep- resented as:
SF CA + CaO => Ca+2 Ca+2+ LCF A => M etalsoap CaO + LCF A 6=> M etalsoap
where SCFA and LCFA are respectively short-chain and long-chain fatty acid.
The SCFA are not present in the biofuel before ageing, therefore they are prod- ucts of the biofuel degradation process [19].
This hypothesis has been tested by adding the different calcium sources to the biofuel in different conditions, with and without the presence of SCFAs, in order to be able to analyse their role in the deposits formation.
3.2 Ageing Method
The formation of deposit of soft particles in biodiesel blends has been detected after fuel ageing. Therefore, it is essential to perform an artificially accelerated ageing in order to mimic the deposit formation process through laboratory ex- periments. The experimental setup is shown in the Appendix in Fig.A.1. The ageing procedure for the B10 fuel was developed in a previous study by Swarga [23]. In this method, 200 g of fuel were placed in a round flask and stirred with an elliptical magnetic stirrer at 550 rpm at a temperature of 110◦C for 72 h. An air flow was applied to assure that the fuel was constantly in contact with oxy- gen, and a condenser was placed on the flask to avoid evaporation of volatile components. The same setup has been used to age the B100, but in this case the ageing was performed at a temperature of 80◦C for 96 h, as performed in a previous internal study at Scania CV AB.
3.3 SCFA Removal and Addition
After ageing the fuel, in order to test the involvement of SCFAs in the soft particles formation, a procedure aimed at removing them from the test fuel was developed. The set-up of this new experiment is shown in Fig.A.2 in the Appendix. The fuel was placed in a round flask, stirred at 550 rpm and at a temperature of 90 ◦C for a time period that varies between 6 and 12 h, depending on the measured pH. To avoid that oxygen comes in contact with the test fuel and continues the oxidation, which would lead to the formation of new SCFAs, the smallest opening of the flask where the thermometer was placed was sealed while a washing flask with distilled water was connected to the flask’s largest opening. The fuel was bubbled with nitrogen gas in order to cause the evaporation of the SCFAs. The nitrogen gas flow was kept at a constant rate of 20 l/h, controlled with a flow-meter. The flow rate employed has been previously used for ageing and oxidation stability tests [18, 26, 27].
The pH of the water in the washing flask has been measured in time periods of 1 h to control the pH decrease due to the presence of SCFAs that evaporated from the fuel. The evaporation was stopped when the decrease of pH reached a predetermined value that varies for different fuel samples.
In some other cases, to be sure that SCFAs were present in the test fuel they have been added as formic acid (CH2O2, 98-100%) before ageing. The amount of formic acid to be added was calculated in order to meet the same Total Acidic Number (TAN) of the aged test fuel, according to previous studies conducted at Scania CV AB.
3.4 Calcium Addition
After preparing the test fuel, different compounds of calcium have been added in order to verify if it is possible to activate the mechanism of formation of the soft particles under study. Calcium has been added as:
• Engine oil is a contaminant in the biodiesel engines that causes the for- mation of Ca soaps [10]. Standard engine oil 10W-30 was used.
• CaCO3powder (95+%), this is a calcium compound found in the engine oil [28];
• CaO powder (97+%), this calcium compound was easily available in the laboratory fo usage in the experiments.
The test fuel was separated into three beakers, which were placed on heat- ing plates. The three calcium sources have been added to the fuel while stir- ring at 350 rpm at a temperature of 60◦C overnight. The amounts of calcium carbonate and calcium oxide to be added were calculated according to the ex- pected TAN value of the test fuel. For 100 g of test fuel, 0.076 ±0.001 g of CaCO3 and 0.040 ±0.001 g of CaO have been added. The calculations used for this purpose can be found in Appendix B. The amount of engine oil that has been added is equal to 1 wt% of the test fuel, as performed in previous studies by Swarga [23] and Abdel Alim [25].
3.5 Analytical Methods
Before and after the soft particles formation, both the test fuel and the deposit have been analysed with different techniques. The test fuel has been analysed to verify the presence of SCFAs and the deposit has been analysed to control if the calcium soaps were forming.
3.5.1 Samples Preparation
When analysing the formed deposits, fuel contamination must be minimised in order to be able to detect the calcium soaps. Depending on the analyti- cal technique to be used, different sample preparation steps were previously performed.
Filtration
In some cases, in order to analyse the particles formed in solution, after the calcium addition, it was necessary to filter the test fuel in order to concentrate the particles. This procedure has been performed by using a 1,2 µm glass fiber filter and a vacuum flask. The filtration set up is shown in Fig. 3.2.
Cleaning
The solid deposit and the filters after filtration have been cleaned with cyclo- hexane in order to remove the residual test fuel [29] to avoid any interference with the performed analysis. The method was developed by Swarga in a pre- vious study [23]. The procedure can be summarized as follows:
1. Place the sample or the filter into a centrifuge tube;
2. Add 5 ml of cyclohexane and mix in the vortex mixer for 1 minute;
3. Let the sample stand for 30 minutes;
4. Centrifuge for 2 minutes at 2500 rpm and remove the remaining liquid with a Pasteur pipette;
5. Repeat 3 times from step 2.
The samples prepared through this procedure have been then analysed by FTIR, or by GC/MS after esterification.
Figure 3.2: Filtration setup.
Esterification
The filter samples contained in the centrifuge tubes need to undergo an ester- ification process in order to be analysed by the GC/MS [23]. To do so, the following steps have been followed:
1. Add 3 ml of 0,5 vol% H2SO4in methanol;
2. Shake in the shaking table for at least 24 h;
3. Add 4 ml of 200 mM NaHCO3(aq);
4. Add 2 ml of cyclohexane;
5. Mix in the vortex mixer for 1 minute and shake in the shaking table for 30 minutes;
6. Centrifuge for 2 minutes at 2500 rpm;
7. Extract the hexane phase (upper part) with a pipette and transfer into a GC/MS vial for analysis.
3.5.2 pH Measurements
For every step of the different test fuels preparation, the pH has been measured with a pH-meter Mettler Toledo SevenEasy. The pH-meter has been calibrated daily before starting the measurements. In order to measure the fuel’s pH, 6 ml of fuel were placed into a glass vial together with 6 ml of distilled water.
The vial was sealed and shook in a vortex mixer for 90 seconds. After that the fuel separates from the water, the latter was extracted with a Pasteur pipette and its pH was measured. For every test fuel 2 samples have been analysed.
For each sample three different pH measurements have been performed and their mean value has been calculated. The accuracy of the measurements is equal to ±0.01.
3.5.3 Rancimat
The oxidation stability of all the different test fuels was measured with the 893 Professional Biodiesel Rancimat (Metrohm AG). A schematic representation of its set up is shown in Fig.3.3. The oxidation stability can be defined as the time at which, under defined aging conditions, oxidation products start to form. Within the Rancimat method, the oxidation stability is identified by the induction period (IP), namely the inflection point of conductivity given by the rapid increase of organic acids formed in the fuel [18, 20].
The Rancimat test method for the determination of biodiesel blends oxida- tion stability is specified by different standards, such as the American ASTM D6751 and the European EN 14213 and EN 14214. The most common method (EN 14112) in compliance with the previous mentioned standards, requires a temperature equal to 110◦C and a minimum IP of 6 h [18]. During this test, 7.5 g of biodiesel were kept inside a sealed glass reaction tube and heated at a constant temperature (110 ±0.3◦C) while air with a flow of 10 ±0.75 l/h was introduced into the sample. The formed acidic volatile products will evaporate and be collected into distilled water while its conductivity is constantly mea- sured. An abrupt increase in conductivity is the index for the induction period.
Two measurement methods of the induction period are described by the stan- dard EN14112: a manual and an automatic method; both derived from the
Figure 3.3: Rancimat setup.
conductivity-time curve measured by the instrument. In the manual method, the IP is identified by the intersection of the tangent of the first part of the curve with the optimal tangent in the inflection point of the second part of the curve (Fig.3.4(a)). In the automatic method the IP is determined by the peak of the second derivative of the curve (Fig.3.4(b)). The latter is mostly used since it avoids the problem of manually finding the optimal tangents, however in some cases the second derivative can have different maxima and the manual method is preferred [30].
Figure 3.4: Rancimat methods for the identification of the induction period: manual method (a) and automatic method (b) [30].
This method of measuring biodiesel oxidation stability is characterised by a good repeatability and accuracy [18].
3.5.4 Chromatography
Two chromatography techniques have been used: Gas Chromatography, cou- pled with Mass Spectrometry (GC/MS) and Ion Chromatography (IC). Their principles and applications are explained in the sections below.
GC/MS
The presence of calcium soap in solution in the test fuel has been verified with an Agilent 5943 Mass Spectrometer. The Gas Chromatography Mass Spec- troscopy (GC/MS) is a combination of two analytical techniques, also called a hyphenated technique. After cleaning and esterification, the samples are intro- duced into the GC column. Different compounds are here separated according to their absorption affinity since they reach the end of the chromatography col- umn at different times. These different compounds are then introduced into the MS. They are here ionized and separated according to their mass to charge ratio (M/Z), and then analysed. The mass spectrometer works in vacuum conditions in order to minimize collisions between molecules. The sensitivity of the de- tectors used for this analysis lies between 10-8and 10-15 g solute/s [31]. The obtained spectrum is then compared to the ones contained in the instrument library in order to identify the different molecules in the sample.
During this work, GC/MS has been used to detect the formation of soft par- ticles in solution through the analysis of the filters, after filtration of the test fuel. In particular, the presence of calcium monomethyl azelate has been in- vestigated [23]. The presence of soft particles was hence attributed to the pres- ence of a peak at 13.86 min, corresponding to the nonanedioic acid dimethyl ester molecule [32] (Fig.3.5), formed after esterification [10].
Figure 3.5: Nonanedioic acid, dimethyl ester [32].
IC
The Ion Chromatography (IC) analysis has been used in order to quantify the amount of SCFAs in the test fuel. The analysis was performed by an external laboratory, SGS Group, Germany.
This analytical technique is used for fast and accurate detection of ions and cations of different types of samples. The most common method is ion exchange chromatography, which uses an ion/cation exchange column and an eluent solution for the detection. First, in order to have a constant signal, the exchangeable ions are replaced by pumping the eluent solution into the col- umn. The sample is then injected. The sample ions replace the eluent ones which causes a change in the ionic concentration of the eluent detected as a peak of conductivity. The eluent solution is then pumped into the column and its ions replace the sample ones, which move down the column at different rates, according to their affinity with the exchange sites. The ions are then detected and concentration dependent signals are obtained [33].
The performed ion chromatography analysis is able to identify inorganic anions and organic acids in FAME and Blends, giving their concentration in mg/kg. It has been used in this study in order to quantify SCFAs, namely acetate, propionate, formate, butyrate and pentanoate (showed in Fig.3.6), in the test fuel samples.
Figure 3.6: SCFAs detected by IC: acetate (a), propionate (b), formate (c), butyrate (d) and pentanoate (e).
3.5.5 NMR
Nuclear Magnetic Resonance (NMR) spectroscopy has been used to verify the presence of formic acid in some test fuel samples. It was performed in the department of applied physical chemistry of KTH with a Bruker Advance III 500 MHz spectrometer using a Bruker BBI probe.
This analysis is used to obtain the structure of different molecules through the detection of1H or13C nuclear spin states intensity and energies of transi- tion [34]. When an electric field is applied the energies of these states separate,
giving information about the atom chemical environment, about the functional group the atom belongs to and also its vicinity.
The test fuel liquid sample is diluted with a deuterated solvent (deuterated chloroform has been used) and inserted into a NMR vial. The vial is placed into the spectrometer, where it is irradiated with radiofrequency radiation until it achieves resonance. The spin states are hence excited and return into equi- librium by emitting radiation, which is recorded by a detector. The shielding effect of neighboring atoms on the local atomic magnetic field generates in- formation of molecular structure [34]. The detection limit of 1H-NMR and
13C-NMR analyses on diesel fuel is equal to 0.01 mol% and 0.5 mol% respec- tively [35].
The test fuel samples were investigated to verify the presence of formic acid in the liquid, before the calcium addition. The obtained spectra were compared to database data in order to attribute the formic acid peaks, which occur at 8.06 ppm for1H-NMR and at 166.3 ppm for13C-NMR [36].
3.5.6 FTIR
Fourier Transform Infrared Spectroscopy (FTIR) was the first analysis used to confirm the presence of soft particles in the deposits. The analyses were done with a Perkin Elmer Spectrum 100 FT-IR with a diamond crystal plate.
The principle of FTIR is based on the ability of molecules to absorb light in the infrared region. When the frequency of the infrared radiation matches the vibrational frequency of the molecule, the radiation is absorbed causing the vibrational excitation of the molecular bonds [37]. The radiation absorption is detected by the spectrometer, which gives an infrared spectrum that it is sensitive to different types of chemical bonds. The used spectrometer is able to detect wavenumbers in the range 4000 - 650 cm-1. By comparison with a database, it is possible to identify the bonds and functional groups present in the molecule.
The FTIR analysis has been performed on the solid deposit samples to verify the presence of a carboxylic peak, which is indicative of the presence of calcium soap, at a wavenumber equal to 1550-1600 cm-1 [32].
Table 3.1: Schematic summary of the preparation steps and analyses performed on the different test fuel samples. pH measurements, Rancimat test, IC and NMR have been performed on the liquid fuel samples. FTIR and GC/MS have been performed on the particles formed after adding calcium. Moreover, IC and NMR analyses have not been performed on all the available samples, but on some representative ones.
Preparation Steps Analysis
Test fuel samples + Formic Acid
N2
bubbling pH Rancimat IC NMR FTIR GCMS
Fresh B100 - - Yes Yes - - - -
Aged B100 - - Yes Yes - - - -
Fresh B10 - - Yes Yes - - Yes Yes
Fresh B10 Yes - Yes - Yes - Yes Yes
Fresh B10 Yes Yes Yes Yes Yes - Yes Yes
Aged B10 - - Yes Yes Yes Yes Yes Yes
Aged B10 - Yes Yes Yes - Yes Yes Yes
Aged B10010%
+ B090%
- - Yes Yes Yes Yes Yes Yes
Aged B10010%
+ B090%
- Yes Yes Yes Yes Yes Yes Yes
Results
The three different calcium sources have been added to the test fuel after this was subjected to ageing and analyses. The pH values, the oxidation stabil- ity, Ion Chromatography and NMR spectrometry data were collected. Subse- quently, the calcium was added to the aged and fresh FAME biodiesels. The obtained deposits were analysed.
4.1 SCFA Measurements
In order to understand how the short chain fatty acids contribute to the forma- tion of soft particles in biodiesel blend deposits, it is crucial to quantify their presence in the test fuel.
4.1.1 pH
A preliminary investigation of the presence of short chain fatty acids in the test fuel was performed by measuring its pH, as described in section 3.5.2. The pH of all the different test fuel was recorded as shown in the bar chart in Fig. 4.1.
The pH of the fresh B10 and B100, represented in green, are 5.89 and 5.04 respectively. In both cases, after ageing it is possible to see a significant pH decrease, suggesting the formation of acidic compounds. Moreover, in both cases the pH values do not show a noticeable change after nitrogen bubbling.
However, the nitrogen bubbling technique showed to be able to remove formic acid from the test fuel when it was performed on fresh B10 with added formic acid, as shown in Fig.4.1 by the bars with black outline.
Figure 4.1: pH measurements results for the different test fuels.
4.1.2 IC
The ion chromatography results are shown in Fig.4.2, where the SCFAs are highlighted by a red outline. Within this technique is possible to quantify the inorganic anions and organic acids in the test fuel. The analysed test fuel samples are B10 aged, B100 aged added to B0 (to form B10) with and with- out nitrogen bubbling, B10 where formic acid was added and its N2bubbled equivalent.
The measured abundances of SCFAs are shown in Table 4.1. The most abundant acids are formic, acetic and propionic acid, with higher concentra- tion in B10 aged. Moreover, the IC results confirm that for fresh B10, the nitrogen bubbling is able to remove the formic acid in solution. A concen- tration reduction of 99.5% has been obtained. However, the reduction in the amount of SCFAs due to nitrogen bubbling is not as extensive for the aged fuel. The concentration difference of these compounds is much smaller for aged B10010%+ B090%with and without N2bubbling.
Moreover, it can be seen in Fig.4.2 that the concentration of the medium chain fatty acids hexanoate and heptanoate are significantly higher in the aged B10.
Figure 4.2: Ion chromatography results showing the amount of organic acids in the test fuel samples.
Table 4.1: SCFAs content in biodiesel test fuel from ion chromatography analysis.
B10 aged B100 aged + B0
B100 aged + B0 + N2bubbling
B10 + CH2O2
+ N2bubbling
B10 + CH2O2
Acetate mg/kg 69,7 36,9 12 <0,5 <0,5
Propionate mg/kg 56,6 64,4 26,9 <1,0 <1,0
Formate mg/kg 100,8 67,5 45,1 2,3 450,5
Butyrate mg/kg <1,0 <1,0 <1,0 <1,0 <1,0
Pentanoate mg/kg 94,6 8,8 6,9 <3,0 <3,0
Total mg/kg 321,7 177,6 90,9 2,3 450,5
4.1.3 NMR
NMR spectrometry has been performed on B10 aged, aged B10010%+ B090%
with and without N2 bubbling. The 1H-NMR and13C-NMR spectra are dis- played in Fig. 4.3, showing the section of the spectrum where the formic acid peak is predicted to appear. The complete spectra can be found in the Ap- pendix C.
In the1H-NMR the formic acid peak is visible on the aged B10 spectrum, around 8.06 ppm [36], while it cannot be detected on the other samples spectra (Fig. 4.3 (a)). The peak in the13C-NMR spectra, expected to be at 166.3 ppm [36], is not visible in all four cases (Fig. 4.3 (b)). This result can be related to the detection limit of 13C-NMR analysis, being considerably higher than
1H-NMR one [38].
Figure 4.3: (a)1H-NMR and (b)13C-NMR of biodiesel blends test fuel.
4.1.4 Rancimat
The oxidation stability test has been performed on various aged and fresh test fuel samples. The aim was to investigate the presence of acidic compounds and their effect on the oxidation stability of the different test fuels. The resulting data is shown in Table 4.2. It can be seen that there is an induction period equal to 0 h for the aged fuel, even after nitrogen bubbling, and an IP equal to 11,14 and 11,79 h for the fresh B10 and B100 respectively. It is worth noting that the fresh B10 + CH2O2after nitrogen bubbling shows a higher IP than the fresh B10 fuel blend.
Table 4.2: Induction period (IP) of biodiesel test fuels obtained through Rancimat method.
Test Fuel Induction Period (h)
B10 11,14 ±0,52
B100 11,79 ±0,01
B10 + CH2O2+ N2bubbling 21,54 ±0,35
B10 aged 0
B10 aged + N2 bubbling 0
B100 aged 0
B100 aged + B0 0
B100 aged + B0 + N2 bubbling 0
The graphs showing the Rancimat results can be found in Appendix D.
4.2 Soft Particles Measurements
After the test fuel was subjected to ageing and bubbling with nitrogen, calcium was added. Subsequently, the presence of calcium soaps in the liquid fuel and the formed deposits was studied. Whenever deposits were found, these were analysed with FTIR and GC/MS, as previously described in the Methodology chapter. The results of these analyses are shown in this section. Starting with the results of the aged test fuel, following with the fresh fuel data and finalizing with the results obtained for the nitrogen bubbled fuel.
It is important to notice that the textures of the deposits are different for different samples. It can be seen in Fig. 4.4 that for the aged B10, with and without nitrogen bubbling, the deposit was a mixture of solid and slimy matter after addition of CaO or CaCO3. After adding engine oil, only slimy deposits
were found. Moreover, the deposit obtained with B10010% + B090%, with and without N2 bubbling, has a slime texture (Fig. 4.5 (a)), while for the fresh fuels, when present it showed a powder texture (Fig. 4.5 (b)).
Figure 4.4: Deposit obtained with aged B10 + CaCO3.
Figure 4.5: Different texture of biodiesel deposit: (a) slimy texture, obtained with aged B10010% + B090% + CaO and (b) powder texture, obtained with fresh B10 + CaCO3.
In order to verify the repeatability of the experiments, almost all the follow- ing tests and analysis were performed at least two times. The only exception was for the aged B10010% + B090% after nitrogen bubbling. For this test fuel the GC/MS analysis has been performed only once.
4.2.1 Aged Fuel
The deposit obtained after adding the three calcium sources to aged B10 and B10010% + B090% were analysed with FTIR. The obtained spectra are shown in Fig. 4.6 and Fig. 4.7 respectively. For aged B10, small similar peaks are visible at 1600 cm-1 indicating the presence of a small amount of calcium soaps. In Fig. 4.7, however, the soap peak was obtained only with the addition of calcium oxide.
Figure 4.6: FTIR spectrum of aged B10 deposit.
Figure 4.7: FTIR spectrum of 10% aged B100 + 90% B0 deposit, first trial.
The presence of soft particles in solution was investigated by GC/MS anal- ysis of the filters obtained after filtration of the liquid fuel, after Ca addition.
As showed in Figures 4.9 and 4.10, in all cases the soft particles peak can be detected at 13.86 min. For both B10 aged and aged B10010%+ B090%, the peak obtained with CaO addition is considerably higher than the one resultant from the addition of calcium carbonate and engine oil. This is indicative of a higher concentration of metal soaps. However, in the first trial for aged B10 showed in Fig. 4.8, the largest peak is obtained with engine oil addition.
Figure 4.8: GC/MS spectrum of aged B10 filters, first trial. The soap peak is high- lighted
Figure 4.9: GC/MS spectrum of aged B10 filters, second trial. The soap peak is highlighted.
Figure 4.10: GC/MS spectrum of B10010% + B090% filters. The soap peak is high- lighted.
4.2.2 Fresh Fuel
The FTIR spectra of fresh B10 and B10 + formic acid in small amount, as described in section 3.3 (Fig. 4.11 and Fig. 4.12 respectively) show none or only moderate soap peaks. However, when a larger amount of formic acid was added to the fresh B10, equal to 1 wt%, a peak at 1600 cm-1was obtained for both CaO and CaCO3addition (Fig. 4.13).
Figure 4.11: FTIR spectrum of fresh B10 deposit, before cleaning.
Figure 4.12: FTIR spectrum of fresh B10 + CH2O2 deposit, after cleaning, first and second trial.
Figure 4.13: FTIR spectrum of fresh B10 + 1 wt% CH2O2deposit, before cleaning.
In all cases, soft particles have not been detected. The soap peak is not present in all the GC/MS spectra of fresh B10 and B10 + CH2O2 fuel filters, as displayed in Figures 4.14 and 4.15.
Figure 4.14: GC/MS spectrum of fresh B10 filters. The expected peak position is highlighted.
Figure 4.15: GC/MS spectrum of fresh B10 + CH2O2 filters. The expected peak position is highlighted.
4.2.3 Nitrogen Bubbling
The FTIR spectra of the deposits formed in the aged and nitrogen bubbled B10 show a significant soap peak after calcium oxide addition, and moderately large peaks in the other two cases (Fig. 4.16). The case of bubbled aged B10010% + B090%, leads to different results. In the first trial, CaO addition results in a significant soap peak (Fig. 4.17) and no deposit was obtained with CaCO3and engine oil, while in the second trial the peaks are not obtained with all three calcium sources (Fig. 4.18).
Figure 4.16: FTIR spectrum of aged B10 + N2bubbling deposit.
Figure 4.17: FTIR spectrum of aged B10010% + B090% + N2 bubbling, after CaO addition, first trial deposit.
Figure 4.18: FTIR spectrum of aged B10010%+ B090%+ N2 bubbling, second trial deposit.
Soft particles were identified in solution in the liquid fuel in both cases.
From the GC/MS spectrum of N2bubbled aged B10, it is possible to see simi- lar height soap peaks with all three calcium sources (Fig. 4.19). However, the peaks obtained with aged B10010% + B090% + N2bubbling show some differ- ences, CaCO3being the largest and CaO the smallest (Fig. 4.20).
Figure 4.19: GC/MS spectrum of aged B10 + N2bubbling filters. The soap peak is highlighted.
Figure 4.20: GC/MS spectrum of aged B10010%+ B090%+ N2bubbling filters. The soap peak is highlighted.
The results obtained after bubbling nitrogen to the fresh B10 + formic acid are similar to those obtained with fresh B10. Hence, no soap peaks were de- tected with both FTIR and GC/MS analysis (Fig 4.21 and Fig. 4.22).
Figure 4.21: FTIR spectrum of fresh B10 + CH2O2+ N2bubbling deposit.
Figure 4.22: GC/MS spectrum of fresh B10 + CH2O2 + N2 bubbling filters. The expected peak position is highlighted.
Discussion
Artificial ageing of biodiesel fuel blends and alteration of SCFAs amount in solution was performed during this project in order to test their effect on soft particles formation after calcium contamination. The liquid fuel was analysed together with the formed deposits and the particles in solution. The obtained results, shown in the previous chapter, are here analysed and discussed.
5.1 Aged Fuel
It is important to mention that some of the differences in the obtained results from different samples can be attributed to the errors associated with the mea- surement and preparation steps, and with variations in air flow used for the fuel oxidation. The flowmeter has shown fluctuations in the air flow espe- cially over long periods of time. In particular, when the ageing was performed over the weekend, the air flow decreased from 20 l/h to about 5 l/h. However, the presence of short chain fatty acids in the aged fuel was identified by differ- ent analysis. SCFAs concentrations of 321,7 and 177,6 mg/kg were detected with IC for aged B10 and aged B10010%+ B090%respectively.
After the addition of the three calcium sources, the amount of soft particles in solution was similar for both aged test fuels, as showed from the similar height of the soap peak in the GC/MS spectra.
It was unexpected to observe that the amounts of calcium soaps in the de- posits formed in the aged test fuels were different for both fuels. Small amount of soft particles were detected in the aged B10 with all three calcium sources, while for the aged B10010% + B090% a small amount of soap was identified only in the first trial after CaO addition. This result can be explained by the
hypothesis that to form soft particles with CaCO3 and engine oil addition, some compounds from aged B0 may be required [39].
5.2 Bubbled Fuel
After bubbling the test fuel with an inert gas, the same analyses were per- formed. From the IC results, it can be seen that the removal of SCFAs with nitrogen bubbling was not as effective as initially predicted due to their volatile nature [40]. For aged B10010%+ B090%the concentration of SCFAs decreases by 48.81% due to nitrogen bubbling. This decrease is lower than expected.
However, it is not possible to know if the IC analysis was performed on homo- geneous samples since this analysis has been handled by and external labora- tory and the results were obtained after almost one month from the collection of the samples. The amount of acid contained in the samples is expected to increase over a one month time period [41]. However, it is not possible to pre- dict the exact SCFAs amount of the test fuel used in the performed experiments without more accurate analyses.
Furthermore, the nitrogen bubbling time was not equal from one trial to another, as explained in the Methodology chapter. Based solely on pH mea- surements of the washing flask water, the determination of when most of the SCFAs have been removed from the test fuel has a large error associated. This occurs since small variations of their concentration can hardly be detected by pH measurements because SCFAs are weak acids [42].
For bubbled aged B10 and aged B10010%+ B090%, the soaps in the deposits and solution were analysed. Due to the lower SCFAs content, the amount of soft particles in solution was lower for bubbled aged B10, compared to the non-bubbled sample. The results from bubbled aged B10010%+ B090%cannot be confirmed since only one trial was performed. However, in this case the amount of soaps in solution is smaller than the previous cases.
The analyses of the deposits revealed that small amounts of soft particles were formed in all cases, with the exception of the second trial with aged B10010% + B090%. Furthermore, the soap peak obtained by CaO addition is considerably higher than the other cases. A possible explanation to this result is that N2bubbling is not able to remove all the formed SCFAs from the aged fuel due to high amount of oxygen in the formed products due to ageing. Hy- drogen bonds can form between the oxidation products, generating a matrix less prone to evaporate [42].
5.3 Fresh Fuel
From the analysis performed on fresh B10 after calcium addition, it is possible to state that the results show no formation of soft particles in both the deposits and the solution.
However, the addition of formic acid to fresh B10 leads to different results.
After the addition of CaO, a peak in FTIR spectrum can be detected from the analysis of the deposits. This peak is larger when a larger amount of CH2O2is used. In this case, a peak can be detected also after adding CaCO3. However, the soap peak is not present upon addition of engine oil, which is confirmed by the GC/MS spectra for all three cases. Hence, the obtained peaks are ex- pected to be an indication of the presence of calcium formate, generated by the reaction of formic acid with calcium oxide and calcium carbonate [43]. This hypothesis is supported by the fact that calcium formate FTIR peak is at 1600 cm-1, same as the soap one, and that no calcium soap is detected in solution [32, 43].
These results show that in order to form soft particles, fuel oxidation is fundamental.
When analysing the results of N2 bubbling of fresh B10 + formic acid, it can be seen that nitrogen bubbling can completely remove the added CH2O2. For this sample, all the analyses gave similar results to the fresh B10. More- over, from the Rancimat test results, the increased induction period can indi- cate that due to N2bubbling some compounds that are prone to oxidation were removed.
5.4 Common Findings
The overall results from calcium soaps analysis show that calcium oxide is more reactive and prone to form soft particles both in the deposit and in so- lution in comparison to calcium carbonate and engine oil, which often give similar results.
Moreover it is important to state that from the Rancimat and NMR analy- sis, no relevant results were obtained. The Rancimat method has been proven to not be appropriate to detect the effect of SCFAs in aged fuel, since the in- duction period of all these samples was equal to 0 h.
The spectra from NMR analysis show that the amount of formic acid in the samples, with the exception of the aged B10, is too low to be detected by the analytical method used in all cases.
5.5 Ethics and Society
This project has been performed in a controlled laboratory environment using adequate safety precautions and the chemicals involved have been safely dis- posed as it is typically done in the laboratories of Scania CV AB. From a sus- tainability perspective this work contributes to a reduction in the emission of GHG because its results can be used to accelerate the deployment of biodiesel fuel and to replace fossil diesel. Understanding the mechanisms of formation of particle deposits in biofuels leads to the development of better biofuels and to improved design of engine materials and engines, that can prevent the sig- nificant drawbacks related with the current usage of biodiesel. Solving the problems related with the reduction of engine performance due to the forma- tion of particle deposits can make biodiesel be preferred over fossil diesel at a large scale.
Conclusions
The main focus of this project is to generate new knowledge on the mechanism of formation of carboxylic soaps present in deposits found in biodiesel blends.
Different experiments and analyses have been performed, in particular, to in- vestigate the role of short chain fatty acids (SCFAs) generated by biodiesel oxidation. The presence of SCFAs was analysed by different techniques and in some cases they were added or removed through the use of heat and inert gas bubbling. Three compounds of calcium were added to the test fuel: CaO, CaCO3and engine oil, after which the formed deposit was analysed.
For the different fuels tested, SCFAs were identified in the aged biodiesel, as predicted, and the IC results showed that their removal with nitrogen bub- bling, even though possible, was not as effective as expected.
The formation of calcium soaps was identified both in aged B10 and aged B10010%+ B090%with and without nitrogen bubbling. Moreover, calcium ox- ide has shown to be more prone to react to form soft particles in the deposit, especially in N2 bubbled aged B10, and also in solution for most of the fuels tested.
In the fresh fuel, FTIR peaks similar to those attributable to the soaps were present after formic acid addition. However, the absence of soap peaks in the GC/MS analyses suggests that the formed compounds are most likely to be constituted by calcium formate instead of calcium soaps. No soft particles were detected in the fresh fuel when SCFAs are not present.
Overall, it was found in this work that the sole presence of SCFAs is not the decisive parameter for the formation of carboxylic soaps in the deposits.
The ageing and oxidation of biodiesel, on the other hand, are crucial.
Future Works
During the developing of this thesis work the effects of SCFAs on the forma- tion of soft particles in aged biodiesel fuel blends has been tested. To comple- ment the results here obtained, it would be important to evaluate other factors that can cause or affect deposits formation:
• test the effect of other fatty acids. Formic acid was the only SCFA used in this work. The effect of longer carbon chain acids has not been inves- tigated.
• test the effect of other fuel contaminants, such as water or glycerol.
Moreover, it would be useful to have a more accurate analysis of the formed soft particles structure and composition. NMR spectroscopy can be used, for this purpose, to obtain more information about the calcium soap molecular structure.
[1] Clovis Zapata and Paul Nieuwenhuis. “Exploring innovation in the au- tomotive industry: new technologies for cleaner cars”. In: Journal of Cleaner Production 18.1 (2010). The Roles of Cleaner Production in the Sustainable Development of Modern Societies, pp. 14–20. issn: 0959- 6526. doi: https:// doi.org/ 10.1016/ j.jclepro.2009.09.009. url: http:
//www.sciencedirect.com/science/article/pii/S0959652609002832.
[2] Jidon Janaun and Naoko Ellis. “Perspectives on biodiesel as a sustain- able fuel”. In: Renewable and Sustainable Energy Reviews 14.4 (2010), pp. 1312–1320. issn: 1364-0321. doi: https:// doi.org/ 10.1016/ j.rser.
2009.12.011. url: http://www.sciencedirect.com/science/article/pii/
S1364032109002913.
[3] S. Kent Hoekman et al. “Review of biodiesel composition, properties, and specifications”. In: Renewable and Sustainable Energy Reviews 16.1 (2012), pp. 143–169. issn: 1364-0321. doi: https://doi.org/10.1016/j.
rser.2011.07.143. url: http://www.sciencedirect.com/science/article/
pii/S136403211100390X.
[4] Naoki Kono et al. “FAME Blended Diesel Fuel Impacts on Engine/Vehicle Systems”. In: SAE International Journal of Fuels and Lubricants 5.1 (2012), pp. 163–179. issn: 19463952, 19463960. url: http : / / www.
jstor.org/stable/26272872.
[5] Howard L. Fang and Robert L. McCormick. “Spectroscopic Study of Biodiesel Degradation Pathways”. In: SAE Technical Paper. SAE Inter- national, Oct. 2006. doi: 10.4271/2006-01-3300. url: https://doi.org/
10.4271/2006-01-3300.
[6] Mitsuru Osawa et al. “Influence of Base Diesel Fuel upon Biodiesel Sludge Formation Tendency”. In: (Apr. 2009). issn: 1946-3952. doi:
https://doi.org/10.4271/2009-01-0482.
[7] Paul Lacey et al. “Internal Fuel Injector Deposits”. In: (Aug. 2011).
issn: 1946-3952. doi: https://doi.org/10.4271/2011-01-1925.
[8] Herbert Feld and Nadine Oberender. “Characterization of Damaging Biodiesel Deposits and Biodiesel Samples by Infrared Spectroscopy (ATR-FTIR) and Mass Spectrometry (TOF-SIMS)”. In: (Oct. 2016).
issn: 1946-3952. doi: https://doi.org/10.4271/2016-01-9078.
[9] Olivier Herbinet, William J. Pitz, and Charles K. Westbrook. “Detailed chemical kinetic mechanism for the oxidation of biodiesel fuels blend surrogate”. In: Combustion and Flame 157.5 (2010), pp. 893–908. issn:
0010-2180. doi: https : / / doi . org / 10 . 1016 / j . combustflame . 2009 . 10 . 013. url: http : / / www . sciencedirect . com / science / article / pii / S0010218009003058.
[10] Botond Csontos et al. “Characterization of Deposits Collected from Plugged Fuel Filters”. In: (Sept. 2019). issn: 0148-7191. doi: https : //doi.org/10.4271/2019-24-0140. url: https://doi.org/10.4271/2019- 24-0140.
[11] Ralph McGill et al. “Emission Performance of Selected Biodiesel Fu- els”. In: (May 2003). issn: 0148-7191. doi: https:// doi.org/ 10.4271/
2003-01-1866.
[12] Toshihiko Omori et al. “Biodiesel Fuel Effects on Injection System and Establishment of the Evaluation Method”. In: 16th Asia Pacific Auto- motive Engineering Conference. The Automotive Research Association of India, Oct. 2011. doi: https://doi.org/10.4271/2011-28-0057.
[13] Masjuki Hj. Hassan and Md. Abul Kalam. “An Overview of Biofuel as a Renewable Energy Source: Development and Challenges”. In: Pro- cedia Engineering 56 (2013). 5th BSME International Conference on Thermal Engineering, pp. 39–53. issn: 1877-7058. doi: https://doi.org/
10.1016/ j.proeng.2013.03.087. url: http:// www.sciencedirect.com/
science/article/pii/S1877705813004414.
[14] “Biofuels”. In: Biofuels: Securing the Planet’s Future Energy Needs.
London: Springer London, 2009, pp. 87–101. doi: 10 . 1007 / 978 - 1 - 84882-011-1_3. url: https://doi.org/10.1007/978-1-84882-011-1_3.
[15] Takahito Niwa et al. “Improvement of the Robustness of the Common Rail System for the Fuel Diversification”. In: JSAE/SAE 2015 Inter- national Powertrains, Fuels & Lubricants Meeting. SAE International, Sept. 2015. doi: https://doi.org/10.4271/2015-01-1967.
[16] “6 - Fuel Properties”. In: The Biodiesel Handbook (Second Edition). Ed.
by Gerhard Knothe, Jürgen Krahl, and Jon Van Gerpen. Second Edition.
AOCS Press, 2010, pp. 137–251. isbn: 978-1-893997-62-2. doi: https:
//doi.org/10.1016/B978- 1- 893997- 62- 2.50011- 5. url: http://www.
sciencedirect.com/science/article/pii/B9781893997622500115.
[17] Kamalesh A. Sorate and Purnanand V. Bhale. “Biodiesel properties and automotive system compatibility issues”. In: Renewable and Sustain- able Energy Reviews 41 (2015), pp. 777–798. issn: 1364-0321. doi:
https : / / doi . org / 10 . 1016 / j . rser . 2014 . 08 . 079. url: http : / / www . sciencedirect.com/science/article/pii/S1364032114007655.
[18] Rajesh Kumar Saluja, Vineet Kumar, and Radhey Sham. “Stability of biodiesel – A review”. In: Renewable and Sustainable Energy Reviews 62 (2016), pp. 866–881. issn: 1364-0321. doi: https://doi.org/10.1016/
j.rser.2016.05.001. url: http://www.sciencedirect.com/science/article/
pii/S1364032116301113.
[19] Ping Y. Furlan et al. “Investigating the Oxidation of Biodiesel From Used Vegetable Oil by FTIR Spectroscopy: Used Vegetable Oil Biodiesel Oxidation Study by FTIR”. In: Spectroscopy Letters 43.7-8 (2010), pp. 580–
585. doi: 10.1080/ 00387010.2010.510708. url: https:// doi.org/ 10.
1080/00387010.2010.510708.
[20] Tatsuki Sugiura and Yoichiro Nakamura. “Oxidative Deterioration Prop- erties of FAME-Blended Diesel Fuel”. In: (2018). issn: 0148-7191. doi:
https://doi.org/10.4271/2018-01-0924. url: https://doi.org/10.4271/
2018-01-0924.
[21] Botond Csontos et al. “Contaminants Affecting the Formation of Soft Particles in Bio-Based Diesel Fuels during Degradation”. In: (Jan. 2019).
issn: 0148-7191. doi: https://doi.org/10.4271/2019-01-0016.
[22] Gen Sugiyama, Akira Maeda, and Keiichi Nagai. “Oxidation Degra- dation and Acid Generation in Diesel Fuel Containing 5% FAME”.
In: JSAE/SAE International Fuels & Lubricants Meeting. SAE Inter- national, July 2007. doi: https://doi.org/10.4271/2007-01-2027.
[23] Shriharsha Swarga. “Development of a method that quantifies the fil- tration efficiency of soft particles in drop-in fuels”. MA thesis. KTH, School of Engineering Sciences in Chemistry, Biotechnology and Health (CBH), 2019, p. 62. url: http://urn.kb.se/resolve?urn=urn:nbn:se:kth:
diva-261143.
