for testing the fuel system robustness against soft particles in biodiesels
Romain Couval
Materials Engineering, master's level 2021
Luleå University of Technology
Department of Engineering Sciences and Mathematics
i
Abstract
The future of fuels will most probably be a mixture of different fuels, called drop-in fuels. It is already known that these drop-in fuels will lead to solubility issues, with creation of deposit on crucial fuel system parts, due to the formation of soft particles.
The fuel system of the future should be robust against any type of soft particles. Today, there is no scaled up test fluid existing for testing full scale fuel systems. The objective of this thesis was to develop a scaled up test fluid which is a key element to the development of a test method to enhance the fuel system robustness against soft particles. A test fluid was achieved by a concentrate of calcium soap diluted two thousand times to reach a volume of 1000 litres with a concentration of 1,4 ppm. The concentration was measured by gas chromatography mass spectroscopy method following a derivatisation as sample preparation. The formation of the concentrate was established by changing the type of fuel, the level of aging, the amount of calcium and other counterions and eventually by addition of third elements. The concentrate was made of aged B100, calcium oxide powder and water. The test fluid was made by diluting the concentrate with fresh B7 and a protocol to characterise the stability of this test fluid was developed. This test fluid was tested under real condition in a filter rig giving homogeneous concentration all along the experiment, which confirmed the stability of the test fluid.
Sammanfattning
Framtiden för bränslen kommer säkert att vara en blandning av olika bränslen, så kallade ”drop-in fuels”. Det är redan känt att dessa ”drop-in fuels” leder till löslighetsproblem och igensättning av bränslesystemet på grund av bildandet av ”soft particles”. Framtidens bränslesystem bör vara robust mot alla typer av ”soft particles”.
Idag finns det ingen uppskalad testvätska för att testa fullskaliga bränslesystem. Syftet med denna avhandling var att utveckla en uppskalad testvätska som är en nyckelfaktor för utvecklingen av en testmetod för att förbättra bränslesystemets robusthet mot ”soft particles”. En testvätska uppnåddes genom ett koncentrat av kalciumtvål utspädd två tusen gånger för att nå en total volym på 1000 liter och en koncentration på 1,4 ppm.
Koncentrationen mättes med gaskromatografi i masspektrometri efter en derivatisering som provpreparering. Koncentratet utvecklades genom att ändra typen av bränsle, dess ålder, mängden kalcium och andra joner, och till sist genom tillstättning av tredje element. Det slutgiltiga koncentratet bestod av åldrat B100, kalciumoxidpulver och vatten. Den slutgiltiga testväskan gjordes genom att späda ut koncentratet med färsk B7 och ett protokoll för att karaktärisering testvätskans stabilitet utvecklades. Denna testvätska testades i en filterrigg som gav homogen koncentration under hela försöket, vilket bekräftade testvätskans stabilitet.
iii
Acknowledgement
I would like to thank first of all my two supervisors here at Scania, Mayte Pach and Henrik Hittig, for their immeasurable support throughout the whole thesis both on the technical side and on the human side. Working with such supervisors had been an awesome opportunity that I took advantage of and I wish the same for every students out there.
I would like to thank as well my examiner, Farid, for the precise feedback and the advices both on the presentation and on the final report, which helped me produce a better version of what you are going to read.
I would naturally like to thank Patrick and Saurabh, with whom I spent most of my working time talking about the technical challenges and way more, without forgetting Doctoral student Botond Csontos, whose help to understand the important aspects at the beginning of the thesis is not to be neglected.
I would like to thank at the same time the colleagues at YTMC for their support in the different machines, their explanations in the lab and their trust towards me. Without forgetting Pascal and Linus, two other master thesis students for their presence and proximity, whom I disturbed every thirty minutes when I was going to the lab.
I would especially like to thank all the people I met in the Värdsholmsgatan building (Chetan and Alex to cite only two of them) for making the Friday nights a bit more exciting, as well as some Saturday nights, and some random evenings during week days.
This journey wouldn’t have been such a life lesson without all the people I spent time with during lunch or weekend dinners.
Eventually, I would like to thank my parents for their life long support and for having pushed me into studies early, as well as all the friends I made at the EEIGM and at LTU that helped me push forward every day.
Table of content
1 Introduction ... 1
1.1 Aim and Scope ... 2
1.2 Limitations ... 2
2 Theoretical Background ... 3
2.1 Internal Combustion Engine (ICE) ... 3
2.1.1 Combustion Ignition Engine ... 3
2.1.2 Diesel Chemistry ... 4
2.2 Biodiesels ... 4
2.2.1 The Feedstocks ... 5
2.2.2 Fatty Acid Methyl Ester ... 6
2.2.3 Hydrotreated Vegetable Oils ... 7
2.2.4 The advantages and disadvantages of Biodiesel ... 7
2.2.5 Oxidation stability of Biodiesel ... 8
2.2.6 Formation of nonanedioic acid ... 10
2.3 Soft particles in Combustion Ignition Engines ... 10
2.3.1 Soft particles definition ... 11
2.3.2 Soft particles formation... 11
2.3.3 Soft particles related problems for engine operation ... 12
2.4 Literature review on test fluid ... 13
3 Materials & Methods ... 15
3.1 Fuel and chemicals ... 15
3.2 Dispersion tool ... 15
3.3 Rancimat ... 15
3.4 Fourier Transform Infrared Spectroscopy Attenuated Total Reflection (FTIR- ATR) 16 3.5 Gas Chromatography / Mass Spectroscopy (GC/MS) ... 16
4 Experiments & Procedure ... 17
4.1 Biodiesel production ... 17
4.2 Fuel aging... 18
4.3 Metal counterion addition ... 20
v
4.5 Stability protocol ... 23
4.6 Scaled up test fluid protocol ... 24
4.7 Injector rig screening test ... 24
4.7.1 Experimental setup... 25
4.7.2 Soft particles related FTIR graph ... 25
5 Results and Discussion ... 26
5.1 Concentrate creation ... 26
5.1.1 Soft particles formation... 26
5.1.2 Influence of fuel... 28
5.1.3 Evaluating the solution ... 29
5.2 Quest for stability ... 31
5.3 Scaling up: concentration in the filter rig tank... 38
5.4 Injector rig screening test ... 40
6 Conclusion ... 43
7 Future Work ... 44
8 References ... 45
List of acronyms
B10 10 % biodiesel and 90 % fossil diesel B100 Pure Biodiesel
CO Carbon Monoxide
DDS Dodecenyl Succinic FAME Fatty Acid Methyl Ester
HC Hydrocarbons
HVO Hydrotreated Vegetable Oil IBC Intermediate Bulk Container ICE Internal Combustion Engine IID Internal Injector Deposit ISTD Internal Standard
KTH Kungliga Tekniska Högskolan (Royal Institute of Technology) NOx Nitrogen Oxides
PM Particulate Matter
PIBSI Polyisobutylene Succinimides SCFA Short Chain Fatty Acid SIC Selected Ion Chromatogram ULSD Ultra-Low Sulphur Diesel
1
1 Introduction
The recent pandemic brought a new wave of awareness on the current global warming situation, showing that it was possible to reduce drastically our impact. However, only one year has passed and industries are making up lost time, inducing a peak in energy demand all over the world. According to the International Energy Agency (IEA), the energy demand for the first quarter of 2021 has already reached 2019 level, with global CO2 emissions heading for their second largest annual increase ever. (1)
The transport industry being responsible of 8 Gt of CO2 emissions worldwide per year is now the priority in terms of decarbonization. Stopping the increase in GHG emissions for the transport industry will be done thanks to the introduction of alternatives for fossil fuel (2). Alternatives are multiple and electrification is leading the way, but biofuels break out of the pack as the most sustainable solution when high density of power is combined to long distances. These parameters correspond to the transportation sector related to the heavy duty trucks specialised in long haulage, this is why Scania CV AB is interested in this topic.
Biodiesel is a renewable fuel based on vegetable oil leading to the formation of molecules similar to the one present in fossil diesel, called fatty acid methyl esters (FAME). Biodiesels are said to be carbon neutral as carbon emissions are absorbed during the feedstock production (3). Moreover, the presence of oxygen atoms in the biodiesel molecules increases the combustion which improves the efficiency and reduces the particulate emissions. In addition, no sulphur atoms are present in this type of fuel.
In spite of these advantages, biodiesels must overcome a few challenges before being fully introduced into the market. These main challenges are oxidation stability, cold flowability and material compatibility. A good summary diagram written by Lin et al.
is present in Figure 1-1 (4). There is one more challenge concerning the use of biodiesels, and especially biodiesel blends, which is the formation of deposit leading to loss of power and increase in consumption. These deposits are soft and sticky in nature, which gave the name soft particles (SP). The soft particles of interest in this work are carboxylate salts formed from oxidative products and metal counterions.
The soft particles problem will grow over the next decades as the future of fuels will be a mixture of different fuels, referred as “drop-in fuels”, increasing the solubility issues.
The main picture of this project is for SCANIA CV AB to create a robust fuel system being able to cope with the different soft particles present all around the world. To achieve this goal, a test method must be developed, including a test fluid.
This work focuses on the development of this test fluid which will be composed of calcium soap, or calcium methyl nonanediate in its full name, as soft particles, that are introduced as a water calcium oxide mixture in aged biodiesel. Biodiesel has been chosen to create the concentrate as it is very prone to soft particles formation.
1.1 Aim and Scope
This Master Thesis is about elaborating a protocol for scaling up a test fluid containing an homogeneous and known amount of soft particles in a 1000 litres recipient. Creating such amount of test fluid is not possible within laboratory scale, this is why the goal is to create a concentrate that could be diluted 1000 times and give a stable test fluid.
Reproducibility of the concentrate formation and stability of the test fluid are the two most important factors to take into account to succeed in this task.
1.2 Limitations
• The project will only focus on the test fuel preparation
• Deep understanding of filter and injector systems is out of scope
• Understanding the formation mechanism of sticky deposit is out of scope
• Only B10 and B7 are used as base fuel for the test fluid
• Only the nonanedioic acid is the molecule of interest for the soaps
• The duration of the project is 20 weeks
Figure 1-1: schema of challenges and opportunities for the biodiesels (4)
3
2 Theoretical Background
2.1 Internal Combustion Engine (ICE)
Internal combustion engines are powering most of the vehicles on land, on water and even in the air before the introduction of jet engines since the beginning of the XXth century. The development of these engines started back in the late XIXth century with two different mechanisms; a spark ignition and a compression ignition. Nowadays, these two modes are still widely used and spread in function of their application, with the spark ignition engine being known as the gasoline engine and the combustion ignition engine being known as the diesel engine (5).
This work will focus on the diesel engine as this mode of propulsion is more suitable for heavy-duty vehicles generally used for long hauls and requiring high density energy.
2.1.1 Combustion Ignition Engine
Diesel engines are proved to produce high torque with a low fuel consumption while being extremely reliable, this is why these engines are used in the heavy-duty market as well as on the construction and mining industry, as fixed electricity generator or even for endurance races.
This engine is based on the four stroke cycle model, visible on Figure 2-1. During the intake stroke, only air is aspired by the vacuum created by the moving piston and pushed by a turbocharger that increases the amount of air introduced. During the compression stroke, this air is compressed and its temperature reaches multiple hundreds of degrees Celsius. A few milliseconds before the end of this stroke, the diesel is injected (or atomised) via the injector at two thousand bars. The temperature of the air and dispersion of the fuel allows the mixture to auto ignite, giving rise to the expansion stroke. The expansion stroke is the stroke in which the combustion power is converted into mechanical power. The exhaust stroke allows the gases produced by the combustion to exit the chamber, a new cycle can start again (5).
Figure 2-1: four-stroke cycle of a diesel engine (7)
Once filled inside the tank, the fuel undergoes modifications, mainly filtration and compression, before arriving in the combustion chamber and igniting. The fuel path inside a combustion ignition engine is depicted in Figure 2-2. The fuel is filled inside the fuel tank. The feed pump sucks the fuel from the tank via the suction filter and the pressure filter before entering the high pressure pump. This pump pressurises the fuel up to 2000 bars, explaining the necessity for good filtration. The common rail stocks the pressurised fuel and distributes it to the injectors that atomise the fuel into small droplets following a specific pattern inside the combustion chamber where the air is pressurised. The air/ fuel mixture combusts and is expelled to the exhaust system for after treatment.
2.1.2 Diesel Chemistry
Diesel is a transformed product of crude oil in a refinery using an atmospheric distillation column. Nowadays, refineries are much more complex than previously to meet environmental and quality standards. To form diesel, crude oil has to undergo different treatments such as desalting, cracking or hydrotreating. The result product is a mixture of hydrocarbons generally ranging from C9 to C23 containing mainly paraffinic hydrocarbons (series of saturated hydrocarbons), cycloparaffinic hydrocarbons and aromatic hydrocarbons (6-membered rings with 3 apparent conjugated double bonds). In conjunction with these hydrocarbons, the fuel contains, in small quantities, organic compounds of sulphur, nitrogen and oxygen. Figure 2-3 gives a carbon number distribution in an all-purpose middle distillate fuel (6).
This carbon distribution affects the fuel properties as these molecules have different cetane number, boiling point, cloud point, etc. In consequence, every fuel is different as they undergo different additions, different distillations and are produced from different crude oil which composition can differ even from day to day (5) (7).
Figure 2-2: schema of the fuel path for combustion ignition engines
5
can be defined as a conversion of renewable sources of carbon, known as biomass, into liquid fuel (3). To keep the carbon balance of biofuels close to neutral, it is important to create it from local resources, thus the large variety of available biodiesels, coming with their advantages along with their disadvantages.
2.2.1 The Feedstocks
As aforementioned, biodiesels can be produced from a wide variety of feedstock. These feedstocks evolved over the years and can be divided into three generations. It is important to mention that according to the geographic location and politics in place, the last generation might not be the best one. Figure 2-4 shows a world repartition of biodiesel feedstock from 2011 (4).
Figure 2-3: carbon number distribution of No. 2 Diesel Fuel (6)
Figure 2-4: FAME feedstock repartition over the world (4)
The first generation of biodiesels, which are the most common both in the literature and in the current blends, are based on an edible biomass such as rapeseed, soy, palm or sunflower. However the use of this first generation of biodiesels is decreasing as it brings the ethic question of using arable land for fuel production.
The second generation biofuels are similar to the first one, to the exception that the product used are not edible, such as wood chips, or already used food grade product, such as cooking oil. This is a good way of valorising waste but the process will be more complex as it has to purify the product first.
The third generation biodiesels will be produced from algae. There won’t be any arable land utilisation, moreover the algae microorganism can produce the same amount of biofuels on a smaller land compared to lignocellulose biomass (3).
2.2.2 Fatty Acid Methyl Ester
Vegetable oils cannot be used directly as a fuel alternative because of their too high viscosity which is not suitable for the engine pump and injector. Transesterification process enables the transformation of triglycerides (main constituent of vegetable oils) into fatty acid methyl ester (FAME) and glycerol. This process occurs with addition of methanol in presence of a base alkaline as catalyst as visible in Figure 2-5. The length and number of unsaturation sites of the FAME depends on the nature of the feedstock.
The transesterification process is a three step reversible process converting the triglyceride into diglycerides, then into monoglycerides and finally into glycerol, leading to the production of FAME. To ensure complete transformation, excess of methanol is usually introduced (8). The reaction gives a two phase product that can be easily separated with the FAME phase being on top and the glycerol phase being at the bottom. The reaction time is about an hour, but separation of the two phases takes time, multiple days can be necessary to ensure good separation.
The FAME phase has to undergo further purification after the transesterification process with the European Standard for biodiesel fuel (EN 14214) setting the maximum values, regrouped in Table 2-1, for the different constituent of biodiesel to ensure high purity and protect the engine. A minimum oxidation stability index had been set to 6 hours according to standard EN 14112, providing a minimum limit for oxidation
Figure 2-5: general equation for transesterification
7
to reach the required purity, before introduction of some additives for oxidation stability and cold flowability mainly (9).
Table 2-1: EN 14214 standards maximum values
Biodiesel component Unit EN 14214 maximum
Methyl ester content % m/m 96,50
Free glycerol content % m/m 0,02
Monoglycerides content % m/m 0,80
Diglycerides content % m/m 0,20
Triglycerides content % m/m 0,20
Methanol content % m/m 0,20
Water content mg/kg 500
Acid value mg KOH/g 0,50
FAME are similar to fossil diesel to the exception that they contain oxygen which brings advantages along with disadvantages as it will be discussed later in the following sections.
2.2.3 Hydrotreated Vegetable Oils
Hydrotreated vegetable oils (HVO) are vegetable oils and waste fat transformed into bio-based hydrocarbons long chains by hydrotreating process which allows to get rid of the oxygen present in the FAME and thus having a molecule identical to the fossil diesel. HVOs are better biofuels than regular FAME as their cetane number is said to be at least 50 % higher and do not provide stability and cold flowability problems.
However, the process is much more complicated than the transesterification route and produces CO2 as a result of the decarboxylation. Moreover, mixing HVO with the current blend of B10 results in the phenomenon of “drop-in fuels” causing operating issues (10). As a stable biofuel, no further investigation will be done on the HVO in this work.
2.2.4 The advantages and disadvantages of Biodiesel (FAME)
Biodiesels were developed as an alternative fuel to fight against greenhouse gases. The advantages are not to be discussed as global warming is finally entering into the industrial goals but the use of biodiesels brings a lot of new constraints in the whole
“well to wheel” path. From these disadvantages, one can underline the material compatibility, meaning that slight modifications have to be brought to the current ICE, due to the lubricity and corrosive behaviour which is damaging parts made of cast iron, yellow metals and elastomers, and the poor cold flowability, meaning that the consumption will drastically increase at low temperatures (11). But the main problem in biodiesel resides in its oxidation stability, being responsible for storage and handling issues and for creation of a newly deposit in some parts of the engine; the soft particles, which will be discussed in section 2.3 (12).
Table 2-2 gives a summary of biodiesel advantages and disadvantages while Table 2-3 gives the damages on the engine according to impurities present in the fuel (9) (13).
Table 2-2: list of biodiesel advantages and disadvantages
2.2.5 Oxidation stability of Biodiesel
Oxidation stability, or instability, of biodiesels is the main issue of this alternative fuel as it requires a stricter handling and storage than fossil fuel but more importantly, it leads to the formation of soft particles that are responsible for filter plugging and nozzle fouling. This reduces the lifespan of engine components and increases consumption.
As mentioned before, FAMEs come from a large variety of feedstock which has for effect to introduce a large variety of chain length but more importantly of unsaturated sites. It is well established in the literature that the instability is due to the unsaturated sites of the fatty acid chains. These molecules can be mono or polyunsaturated with one, two or three carbon double bonds. However, the number of unsaturated sites alone is not enough to explain the instability, the number of bis-allylic sites have to be taken into account. An allylic site is defined as adjacent to an unsaturated carbon atom. A bis- allylic site is thus adjacent to two unsaturated carbons. For example, one of the most reactive FAME molecule is the methyl linolenate and it contains 3 double bonds and 2 bis-allylic sites (present in C11 and C14) (14). The molecule is shown in Figure 2-6.
Impurity Effect on fuel system
Methanol Corrosion
Storage and Use problems
Water Corrosion
Filter plugging
Free fatty acids Corrosion
Oxidation stability
Glycerides Injector and filter deposit
Metals Injector deposit
Filter plugging
Biodiesel benefits Biodiesel concerns
Reduces net CO2 on a life-cycle basis Oxidation stability
Higher cetane number Lower energy content
Improved lubricity Solvency effect may plug filters
Low sulphur content Poor cold flowability
Reduced HC, CO and PM emissions Increase NOx emissions
Zero aromatics Material compatibility
Non-toxic and biodegradable Microbial contamination from increased water content
Table 2-3: effect of fuel impurities on the fuel system (9) (13)
9
Figure 2-6: 3D representation of a methyl linolenate molecule (15)
An accepted oxidation mechanism of FAME is by free-radical attack; an hydrogen atom is separated from the fatty acid chain, creating an highly reactive radical which reacts with oxygen to from hydroperoxides. This is often referred as the primary oxidation.
These hydroperoxides then decomposes into a multitude of secondary oxidation products, in which are formed the short chain fatty acids (SCFA) (16) (17) (18) (19).
In addition to this primary and secondary oxidation mechanism, a 4-step mechanism can be drawn with the primary oxidation referring to the initiation and propagation &
chain branching while the secondary oxidation refers to the termination step. A simple view of this mechanism is presented in Figure 2-7. It is important to mention that this pathway is not the only one occurring, but it is the dominant one for most of the cases.
Figure 2-7: oxidation mechanism of hydrocarbon by free-radical attack (19)
Getting rid of the unsaturated sites is a solution to drastically improve biodiesel stability like in HVOs however, it has been proved that high proportion of saturated fatty acids lowers the cloud point. The challenge resides then in finding a compromise between oxidation stability and cold flowability (14).
2.2.6 Formation of nonanedioic acid
Previous work effectuated at Scania CV AB determined one oxidation product to be of interest for the formation of calcium soap as this molecule is present in the on-field deposit and provides good yields while being easily identifiable with the GC/MS graphs obtained. This molecule is the methyl nonanedioic acid, alternatively named azelaic acid. The formation mechanism won’t be developed in this report but a probable reactive molecule is the oleic acid that decomposes in the unsaturated site and reacts with oxygen atoms, as depicted in Figure 2-8 (20). Nonanedioic acid is an example of SCFA. Oleic acid is a mono-unsaturated fatty acid present in quantity in most of the biodiesels produced. It contains 18 carbons with the unsaturated site being situated in the middle of the hydrocarbon chain, in position 9. FAMEs present in biodiesel used in Sweden are mainly formed of Rapeseed Methyl Esters (RME) that comes from canola oil feedstock, which contains approximately 55 % of oleic acid (7) (21). A study on the influence of saturated fatty acid in biodiesels blended Palm oil, Coconut oil, Neem oil, Mahua oil, Jatropha oil and Pongamia oil in different possibilities and measured the oleic acid 18:1 to be present circa 30 % in weight in every blending, proving that this molecule is a common one in most of the FAMEs biodiesels (22).
2.3 Soft particles in Combustion Ignition Engines
Understanding the right type of particles is a crucial step towards the formation of a good test fluid.
Figure 2-8: possible nonanedioic formation from oleic acid 18:1 (20)
11 2.3.1 Soft particles definition
The term “Soft Particles” (SP) is an inherent term used at Scania to define these sticky deposits causing trouble in the filters and the injectors. In the literature, these deposit are labelled as “sticky deposits”. The term soft has been chosen in contradiction to the already existing hard particles whose impact on the engine is known and mastered.
These problems appeared for the first time with the introduction of biodiesel and mainly biodiesel blends. It has been found in studies that the formation of deposits occurs mainly for blend ratio situated between B10 and B20 and slowly decreases when the blend ratio increases over 20 %. It is important to mention here that higher blending ratio will admittedly reduce the deposit formation but increase the material compatibly issues. Figure 2-9 displays in red the precipitation time required to form a deposit according to the blending rate of FAME biodiesel (23).
In the literature, these soft particles are often referred as carboxylate salts, or as metal soaps. The definition of a carboxylate salt is a molecule composed of a carboxylate and a metal counterion. In this work the term calcium soap is widely used, it corresponds more precisely to a carboxylate salt composed of short chain fatty acids and a metal counterion.
2.3.2 Soft particles formation
Soft particles englobe a wide variety of sticky deposit and their exact nature and formation is yet to be fully determined despite that a decade has gone since the first studies. This work focuses on the metal carboxylate salts; insoluble particles made from a carboxylate body and a metallic counter ion causing injector deposit and filter clogging. The carboxylate body exists because of the presence of secondary oxidation product SCFAs or from lubricity enhancers when the fuel is an ultra-low sulphur diesel (ULSD) (24). The metal counter ion has numerous sources. Some metals, especially the “yellow ones” like copper and its alloys can directly be removed from the joints inside the fuel system due to the lubricity and polarity of biodiesels. Others can be introduced in the various fuel additives such as low molecular weight polyisobutylene succinimide (PIBSI), coming directly from the engine oil or even coming directly from the transesterification process of biodiesel (25). Or simply introduced in the tank during fuelling, due to issues in the handling and storage.
Figure 2-9: biodiesel deposit according to the blending rate (23)
2.3.3 Soft particles related problems for engine operation
The interest for soft particles takes its roots from the on-field feedback where loss of power and increase in consumption had been reported. First investigations rapidly determined the causes to be located in the filter and in the injectors. Further investigations combined with the growing interest for alternative fuels allowed to be more precise on the problems and classify them in function of their location, relative temperature of the environment or characteristics as presented in Table 2-4.
Table 2-4: simplified classification of diesel deposits according to Scania CV AB Filter
clogging
Internal injector
Nozzle fouling
Nozzle cooking
Location Fuel filter Inside the injector
Inside the nozzle hole
Inside the nozzle hole Temperature of
the Environment
Cold Hot Very hot Burning hot
Characteristic Slimy Sticky Brittle Very hard
Soft particles induce different types of deposit. The deposits considered as “slimy” or
“sticky” are included in this thesis work while the deposits characterised as “brittle”
and “very hard” are excluded.
2.3.3.1 Filter Clogging
Modern engines have a really tight tolerance towards contamination, this is why filters are more crucial than ever in the ICE. A filter is made of a housing and a filter media which ensures the filtration and can be easily changed once its lifetime is reached. It has been proved that filters in contact with blended fuel are clogging in a faster rate (26). A study on the on-field plugged filters from Scania trucks running with B10 demonstrated the dominance of calcium carboxylate salts in the deposit (27).
When a filter is blocked, the pumps have to produce more work to convey the fuel. This work being taken from the engine, less power is transferred to the wheels, resulting in a loss of power for the end user.
The soft particles observed in the fuel filters are located on the filter media, appear slimy and are formed under room temperature environment.
For further information, the reader is invited to have a look at a related master thesis on
“Development and scale up of a test protocol from pilot scale to full scale filter rig to investigate soft particles filtration in biofuel blends” (28).
2.3.3.2 Internal Injector Deposit (IID)
Injectors are complex components of the diesel engines but their role is simple; atomise the fuel inside the combustion chamber. The way fuel is injected affects drastically the combustion, one can note the ignition delay or spray pattern which influence engine
13
The soft particles observed in the injector are located in the interface between the lower plunger and the nozzle appear sticky and are formed under high temperature environment. Figure 2-10 shows the location of these sticky deposit.
For further information, the reader is invited to have a look at a related master thesis on “To develop a test method to run the injector rig to build internal injector deposits”
(29).
2.4 Literature review on test fluid
The test fluid is a commercial fuel answering the Euro VI norm which is doped with soft particles to be used in the different rigs to pursue the investigation on the deposits in biodiesels. A lot of work has been previously done internally at Scania through different master thesis and doctoral thesis working on formation and characterisation methods for calcium soap.
The role of SCFAs in the formation of SP was studied by adding or removing these acids. It has been concluded that the oxidation of biodiesel was a more important parameter than the presence of these acids. This thesis work demonstrated as well that calcium oxide (CaO) as extraneous calcium source was more prone to form SP than calcium carbonate (CaCO3) or engine oil (30).
A quick method was established to prove the formation of calcium soap as soft particles using the FTIR by looking at the presence of a peak around 1540 to 1590 cm-1. Another method was established to give an approximate quantification of the calcium soap using the GC/MS and looking at the nonanedioic acid peak. The GC/MS method has to be used when the presence of SP is confirmed as it requires a long sample preparation involving an esterification reaction that will be discussed in 4.4.1. This peak has been chosen as it is the most present one in the esterified sample when compared to an unesterified reference sample (31).
No other test fuel for deposit in biodiesel based on calcium soap were found on the literature in the automotive industry, however, the use of sodium carboxylate salts doped test fuel has some documentation. The sodium carboxylates are formed using sodium hydroxide and additives such as the dispersant PIBSI or the corrosion inhibitor Figure 2-10: schema of an injector tip showing the sticky deposit location. Courtesy of Cummins
dodecenyl succinic (DDS) (24) (32). Another study focusing on internal injector deposit concluded that addition of 1 ppm of sodium in B10 could lead to the formation of deposit on hot metal surface and highlighted the effect of the presence of succinic acid.
The article further conclude by emphasising the effect of fuel aging on the rapidity and quantity of deposit formation (33).
Other than calcium and sodium soaps, zinc soap can be used as doping agent for accelerated injector fouling test as presented by R.D. Burke et al. (34). The carboxylate salt used in this article is the zinc neodecanoate introduced at a 6 ppm concentration.
The author detailed that the doped fuel was used the same day it was blended.
Hypothesis can be made on the lack of stability of this test fluid, hypothesis reinforced by the fact that 90 % of the test samples were above 3 ppm according to Inductively Coupled Plasma Spectroscopy. It is mentioned in the article that the level of variability was not investigated.
It is interesting to mention that this zinc soap has been used recently in the filter rig present at KTH and experiments concluded that this soap was soluble in fuels.
15
3 Materials & Methods
3.1 Fuel and chemicals
The commercial fuels used for dilution in the lab and in the test rig are standard Euro VI fuels coming from external suppliers. At the beginning of the thesis, B10 was commercially furnished and the switch to B7 occurred midway through the work. B100 was commercial grade too.
Calcium oxide powder 97+ % for analysis from Acros Organic, methanol for analysis from Merck KGaA, sodium hydrogen carbonate 99,5 % for analysis from Merck KGaA, rapeseed coocking oil from ICA, cyclohexane for liquid chromatography from Merck KGaA, sodium hydroxide in pellets from Merck KGaA and sulfuric acid 98 % for analysis from Merck KGaA were the main chemicals used in this work.
3.2 Dispersion tool
The homogenizer T-18 digital from IKA, was used as a dispersion tool in the lab as the maximum volume is 1,5 L in water and should be used at room temperature.
3.3 Rancimat
The oxidation stability of the biofuels was observed via the 893 Professional Biodiesel Rancimat (Metrohm AG) in coordination with the StabNet software. This device is calibrated to follow EN 14112 and EN 1571 standards respectively for pure and blended FAME biodiesels, which are used to determine the oxidation stability by accelerated oxidation method. As explained in the schema on Figure 3-1 a), a biodiesel sample is heated at a precise temperature while air flow is introduced. Acids carboxylic are formed and flow in gas state towards the water recipient. This will increase the conductivity of the water. The electrode measures the conductivity in real time and plots it in function of time. The software will give the oxidation stability, or induction period, by calculating the inflection point of the conductivity curve. Figure 3-1 b) displays in green curve the conductivity and in the blue curve the acceleration, which the highest point gives the inflection point. Physically, this highest acceleration point corresponds to the moment when every anti-oxidant are consumed and the FAMEs start to disintegrate into SCFAs.
Figure 3-1: a) schema of the rancimat apparatus (30). b) screenshot of conductivity curve
3.4 Fourier Transform Infrared Spectroscopy Attenuated Total Reflection (FTIR- ATR)
A Perkin Elmer Spectrum 3 FTIR spectrometer mounted with a PIKE GladiATR tech sample holder including a monolithic diamond lens and a Perkin Elmer Spectrum IR software was the FTIR-ATR device used for this project.
Infrared spectroscopy is an analytical method allowing to characterise the functional groups present in the sample. Indeed, matter absorbs the infrared wavelength and vibrates according to the type of bond (sigma or pi bond, distance between atoms). Each vibration is matching a particular wavelength. The infrared spectrum can be given under absorbance or transmittance mode. The more a bond is present in the sample, the more it will absorb a wavelength and a bigger peak will be displayed in the spectrum.
Usual IR spectrometry uses a mid-infrared wavelength ranging from 4000 cm-1 to 400 cm-1 with the single bonds being present in the 4000 cm-1 to 2500 cm-1 region, the double bonds in the 2000 cm-1 to 1500 cm-1 region and the triple bonds in the 2500 cm-
1 to 2000 cm-1 region while the 1500 cm-1 to 400 cm-1 region is called the fingerprint of the sample (35).
Fourier transform is using a Michelson interferometer to give a better signal to noise ratio, with better accuracy while being faster to capture the data.
Attenuated Total Reflection corresponds to the type of crystal used to make contact between the IR source and the detector through the sample. The ATR crystal material must possess a higher reflective index than the observed sample. This will allow the light to interact with the first few microns of the sample in what is called the
“evanescent field”. ATR is the most common technique in IR spectroscopy because it allows to observe almost any kind of material regardless of their shape and size, as long as good contact between the crystal and the sample is assured (36).
3.5 Gas Chromatography / Mass Spectroscopy (GC/MS)
Two GC/MS apparatus were used during this thesis as a new machine was installed midway through. The old machine was an Agilent 6890 GC + 5973 MSD with the ChemStation software while the new machine was an Agilent 8890 GC + 5977B MSD with the OpenLab 2 software.
The gas chromatography is a technique used for separation of volatile compounds. It vaporises 1 microliter of the sample which goes into a 30 meters capillary column with the help of a carrier gas, helium in that case. This columns acts as a separatory column for the different chemicals present in the initial sample as they all have differences in affinity with the stationary phase present inside the column and thus they reach the end of the column at a different time, called retention time. This method only allows to separate the chemicals and plots the abundance in function of retention time.
The mass spectrometer allows to identify the chemical by matching the result with a data base, the NIST database in that case. The molecules entering the mass detector are hit with 70 eV electrons which is the optimum energy to break molecules, leading to the creation of cations. The cations are led to the analyser by the help of a positive repeller electrode. The analyser gives a mass to charge ratio spectrum for each
17
4 Experiments & Procedure
4.1 Biodiesel production
Homemade biodiesel was produced as a subsidiary fuel for soft particles production.
As explained in section 2.2, biodiesel can be produced from vegetable oil by transesterification mechanism. To produce 1 L of FAME, 1 L of vegetable oil is required alongside 220 mL of methanol and 4,0 g of sodium hydroxide. A good esterification will transform 100 % of the vegetable oil into FAME and will give the same volume of glycerol as the volume of methanol introduced. A biodiesel-o-matic excel sheet had been developed and helps to select the catalyst quantity in function of the type, purity and purity of the oil, especially if cooking oil had been used instead of fresh vegetable oil (38). A quick protocol for biodiesel production is to follow, along with Figure 4-1 which shows the required material.
Protocol for biodiesel production:
1. Heat 500 mL of vegetable oil at 50 °C.
2. In a separated recipient, mix 110 mL of methanol and 2.0 g of sodium hydroxide until complete dissolution.
3. Mix the Sodium Methoxide mixture into the warm vegetable oil in a separatory funnel.
4. Shake vigorously, open to release pressure.
5. Let it rest for few hours to overnight, the phase separation appears around 30 minutes.
The goal behind the production of homemade biodiesel instead of using commercially available biodiesel was to be able to tune the quality of it. Indeed, commercially available B100 contains anti-oxidants and more importantly in our case, has a complete reaction meaning that there should only be traces of mono and di-glycerides. Moreover, the washing step removed excess of polar components present in the fuel, which are of interest when it comes to deposit formation in the on field fuel system. This explains why the protocol does not include washing steps. To create a good quality B100, two parameters are required; complete reaction and good washing. The complete reaction
Figure 4-1: material necessary for transesterification reaction of homemade biodiesel from biomass
has already been explained, while a good washing must be done by removing excess of methanol by flashing it and not by wet washing as the amount to remove is too high.
Wet washing is then used to remove remaining polar components until reaching the levels indicated in the EN 14214 standard in Table 2-1.
Four different homemade biodiesels (HMBD) were created with different proportion of methanol and sodium hydroxide during this thesis and are represented in Table 4-1.
The results will be presented here as it was barely used for soft particles production and the few experiments can serve as a starting point for future research.
HMDB 1 corresponds to the standard amount for good biodiesel creation, the goal was to train producing a good biodiesel as a reference for the upcoming bad biodiesels. The glycerol phase was 20 mL.
HMBD 2 has half of the methanol required, the idea being to limit the reaction and shift the transesterification reaction back to the left of the equilibrium, meaning less conversion of the glycerides. The glycerol phase turned out to be 10 mL but more viscous than the first experiment.
HMBD 3 has a huge amount of NaOH. The idea was to introduce more sodium as it is a source of metal for the carboxylate salts formation. Unfortunately, these parameters are too bad to produce biodiesel as there was no phase separation in the separatory funnel even after 24 hours. In addition, the amount of sodium hydroxide introduced is close to its solubility limit in methanol, thus extra effort is mandatory to achieve perfect mixing.
HMBD 4 was realised with low amount of both transesterification agent and catalyst, to ensure incomplete reaction, resulting in an unclear separation of the phases, with the glycerol phase having very few amount (around 3 mL) and much higher viscosity.
Table 4-1: experiments effectuated on homemade biodiesel
Name Vegetable oil (in mL) Methanol (in mL) Sodium Hydroxide (in g)
HMBD 1 200 40 0,8
HMBD 2 200 20 0,8
HMBD 3 200 20 3
HMBD 4 200 15 0,6
4.2 Fuel aging
The production of oxidation products, which will be part of the soft particles once the metal counterions introduced, are formed during the fuel aging step. The experimental setup had been developed during a previous master thesis on this topic and did not
19
The heating plate with the round flask was set at 110 °C for every experiments in this work. The reasons are multiple; correspondence with the Rancimat apparatus that heats at this temperature, suitable start and end time and avoiding going too high in temperature and starting to evaporate critical elements. The air flow, coordinated with a constant air supply, was set to 20 L/min regarding of the quantity of fuel to age, which played a role in the oxidation products. The aging can be decomposed in two steps.
The first step corresponds to the oxidation stability. This is the time necessary to consume all the anti-oxidants present in the fuel, whether naturally occurring in diesel production or added as anti-oxidants. This oxidation stability, or induction period, can be easily determined by the Rancimat apparatus. This induction period decreases over time as the fuel is naturally oxidising. For example, good quality fossil diesel with additives can stay a couple of decades in a closed tank without sustaining severe degradation. On the other hand, pure biodiesel should be used with care after a year of storage, even in a closed tank with no air exchange with the exterior, constant room temperature and opaque walls because of microbial growth. Table 4-2 regroups the induction period related to different biodiesel blends over different periods.
Table 4-2: induction period related to the different biofuel batches
Fuel batch Date Induction period
B10 September 2020 45 hours
B10 January 2021 38,5 hours
B100 February 2021 12,5 hours
B100 May 2021 11 hours
B7 May 2021 44 hours
Figure 4-2: fuel aging experimental setup (30)
The second step corresponds to the formation of the oxidation products. The more time the more oxidized the fuel will be. In function of temperature, time can be unlimited.
In the first time, small short chain fatty acids will be formed like acetic acid, followed by small to medium chain fatty acids like octanoic acid. The more oxidized the fuel is, the more chance for these oxidised products to have contact between each other and start polymerizing the fuel, leading to a deposit. It is important to mention that this deposit is completely different to the deposit related with soft particles, as it does not contain any metal counter ion and thus cannot be considered as a soap. The oxidation time depends on temperature, air flow and fuel quantity and has been set as well during a previous work (39). Only the fuel quantity was increased to 500 mL during the thesis, giving repeatability and good quantity for the soft particles.
The formation of oxidation product for B10 was pre-defined to 56 hours while the time allocated for B100 was originally 48 hours but this number couldn’t match the working hours so it was decided to increase it to 53,5 hours.
To conclude on the fuel aging topic, B10 was aged for 38,5+56=94,5 hours while B100 was aged for 12,5+53,5 = 66 hours, which made it convenient as it was possible to start aging right before the weekend and to end it first thing on Monday morning.
4.3 Metal counterion addition
Once the fuel was aged and the nonanedioic acid formed, the metal counterion must be introduced to create the desired soft particles. Only calcium oxide will be used in this thesis (as a calcium counterion source) as calcium carboxylate showed poor results and engine oil was put aside. Preliminary tests determined that calcium oxide should be introduced as a mixture with lots of water into the aged biodiesel to trigger a change of colour and produce a hazy solution, full of soft particles.
Calcium oxide is slightly soluble in water (around 1,4 g/L) but highly hygroscopic, which means that it will absorb every small amount of water present to form a paste known as slaked lime. Introducing quick lime in the aged biofuel will result in a big chunk of soft particles moving around, as only the outside of the solid will react with the oxidised products while the inside will stay in form of solid calcium oxide. This explains why huge amount of water is required for the introduction of calcium oxide inside aged biofuel. Because of its low solubility and high hygrometry, it is important to add lots of water from the beginning of the mixture creation while agitating vigorously to avoid formation of slaked lime. A trivial protocol has been developed to ensure reproducible calcium introduction.
Protocol for a general calcium oxide and water mixture introduction (a more detailed protocol for concentrate production with exact values will be explained in 4.6):
1. Insert calcium oxide in a small volumetric flask.
2. Add 60 % of the water.
3. Shake vigorously, hold it upside down, pour into aged biofuel.
4. Add the remaining water.
5. Shake vigorously, hold it upside down, pour into aged biofuel.
Once introduced, this calcium oxide water mixture (made of dissolved Ca(OH) and
21
oxidised molecules act as an inversed micelle with its hydrophilic heads and hydrophobic “tail”. If there is no mixing, calcium oxide, the water phase and the oil phase while separate, with the oxides at the bottom separated to the oil phase by the water phase. However, due to the amount present, the water phase does not cover the whole surface, and there are contacts between calcium water and short chain fatty acids at some points in the beaker. Soft particles are created in these points and take a few days to develop, but the most of the calcium oxide powder is still present in its original form inside the beaker.
For better reaction, the use of a dispersion tool is required especially for higher volumes.
4.4 Quantification of soft particles
The method developed internally by Scania CV AB to measure the soft particles quantity uses the GC/MS graph along with an internal standard; Copper (II) 2- ethylhexanoate. However, it is not possible to observe directly the calcium soap as this component is not enough volatile. It is then necessary to derivatise the soap to increase its volatility by esterification method.
4.4.1 Sample preparation for GC/MS
The calcium soap soft particles are present in solution but undissolved in the fuel.
Filtration with a simple vacuum filtration setup is sufficient to get the soft particles.
The 1,2 micrometre glass fibre filter is placed inside a centrifuge tube. Washing is done with cyclohexane, a non-polar chemical with the advantage of being used as an eluent for GC/MS extraction. The washing step consists in introducing 5 mL of cyclohexane into a centrifuge tube, vortex mixed for 1,5 minutes before being gently mixed for 30 minutes. Cyclohexane with fuel and the particles are separated using a centrifuge for 2 minutes at 2000 rpm after which the liquid phase is poured away, with great care. This is repeated 3 times in total to ensure good background on the GC/MS graph. Once the soft particles assumed to be washed, the esterification reaction can start. 3 mL of a 0,5 𝑣𝑜𝑙⁄𝑣𝑜𝑙% sulphuric acid in methanol solution is added to the soft particles along with 0,2 mL of the internal standard. The reaction time is set to 24 hours and the acids have to be neutralised with a sodium bicarbonate in water solution. The esters, considered to be corresponding to the soft particles, are extracted with 2 mL of cyclohexane under vortex and gentle mixing, similar to the washing step. Extraction consists of sucking 1,5 mL of the cyclohexane phase into a GC/MS vial.
Consistency in the different steps is mandatory to reach reproducibility as all these steps introduce operator error, mainly in the washing steps where it is assumed that neglectable amount is trashed away.
4.4.2 Peaks measurement on GC/MS
GC/MS can be used as a quantification technique by the presence of an internal standard. In this work, only the concentration in nonanedioic acid di-methyl ester is of interest. Every concentration given in this report are measured with the same method and on this peak only as it gives a good representation of the total amount present.
Before proceeding to any calculation, it is important to know the amount and the concentration of the internal standard present in the sample. Moreover, the volume filtered with the fluid density are additional pre-requisites to be able to give a calcium soap concentration.
Two chromatograms are showed in Figure 4-3, one for each machine, and thus each software. The first peak of interest is the one related to the internal standard, the hexanoic acid peak, coming at 8,60 minutes in the old machine and coming at 5,885 minutes in the new. The second peak of interest is the one related to the targeted peak, the nonanedioic acid peak, coming at 13,85 minutes in the old machine and coming at 10,620 minutes in the new. This change of retention time from machine to machine is mainly due to the differences in the columns.
The area of the peaks are proportional to the molecule concentration. By knowing the concentration related to the area of the internal standard peak, one can use an equal ratio equation to calculate the concentration of the targeted peak. The peak area can be automatically extracted from the software.
Before starting the soft particles concentration calculation, it is important to assume that every nonanedioic acid present in the test fluid reacted with calcium cations to form a calcium soap made of Ca2+ and nonanediate. The esterification replaces each calcium ion-covalent bond with a methyl group, giving birth to the nonanedioic acid di-methyl ester.
First step (equation 4-1) in the calculation is to measure the hexanoate weight inside the sample, which had been introduced after the washing step and suffered the same esterification than the calcium soap. Second step (equation 4-2) is to calculate the Figure 4-3: chromatograms with the hexanoate peak (left) and nonanediate peak (right) for
the old GC/MS (top) and the new GC/MS (bottom)
