INOM
EXAMENSARBETE TEKNISK KEMI, AVANCERAD NIVÅ, 30 HP
STOCKHOLM SVERIGE 2019,
Steaming of wood chips
Experimental determination of heating times and effect of different parameters
HENRIK BROBERG
KTH
SKOLAN FÖR KEMI, BIOTEKNOLOGI OCH HÄLSA
———————————————————————————————————————————–
The Royal Institute of Technology
School of Engineering Sciences in Chemistry, Biotechnology and Health
Master thesis:
Steaming of wood chips; experimental determination of heating times and effect of different parameters
Author: Henrik Broberg1, henbro@kth.se
Supervisors: Raquel Bohn Stoltz2, Anders Hjort3
Examiner: Monica Ek4
2019-06-22
———————————————————————————————————————————-
1Master student at the department of Fibre and Polymer Technology, KTH
2RD Process Specialist, Valmet
3Engineer, Valmet
4Professor and head of the division of wood chemistry and pulp technology, KTH
Contents
1 Introduction 4
1.1 Project aim . . . 5
2 Background 6 2.1 Wood . . . 6
2.2 Kraft Pulping . . . 7
2.3 Batch Kraft Process . . . 8
2.4 Continuous Kraft Processes . . . 8
2.5 Presteaming and Impregnation . . . 9
2.6 Chemical modification during steaming . . . 11
3 Experimental 12 3.1 Materials . . . 12
3.2 Steaming experiments . . . 12
3.3 Condensate analysis . . . 13
4 Results & Discussion 15 4.1 Analysis of handmade chips . . . 15
4.2 The effect of chip thickness for pine . . . 16
4.3 Comparison between pine heartwood and sapwood . . . 17
4.4 The effect of higher pressure . . . 21
4.5 Comparison between pine and birch . . . 22
4.6 The effect of thickness for birch . . . 23
4.7 Comparison between birch heartwood and sapwood . . . 23
4.8 Comparison with older data, simulations and industry . . . 24
5 Conclusion 26
6 Future outlook 27
7 Acknowledgements 28
A Appendix 1 - Heating curves for all the experiments 31
B Appendix 2 - Chromatograms from the condensate analysis 34
Abstract
The presteaming of wood chips is an important step in the chemical pulping industry. It removes the air from within wood chips, allowing the cooking liquor to better impregnate wood chips, which leads to a more uniform cooking process, and lowers the amount of rejects. When steaming at atmospheric pressure, it is important that the temperature of the wood chips reach 100◦C, as otherwise there will be an equilibrium leaving some air left inside. Having poorly steamed chips in a process could cause severe problems when it comes to reaching the targeted kappa number, or having the adequate retention time in the digester. There are a few different ways in which the wood chips are presteamed within the industry, however, there is little experimental data regarding the heating time of wood chips that can be used when designing these systems. Most studies have mainly focused on the air removal, or improvement of the impregnation step, and the few studies that have included the heating of the wood chips were limited to only one type of wood chip, or failed to specify the experimental details.
Therefore, handmade wood chips pine and birch, two tree species commonly found in Sweden, were steamed in an ATEX designed digester with a steam jacket. The wood chips had thermocouples inside them and the temperature and time was recorded, and the effect of different parameters on the heating could thus be studied.
The results revealed that there could be more than a minute in average time difference between wood chips of different thicknesses, both for birch and pine, although the difference in heating time was more linearly correlated to thickness for the birch chips. Pine chips of different thickness were also studied when the pressure inside the digester was allowed to build up, which showed that it is mainly thicker chips that have reduced heating time under such circumstances, as the thinner chips stop heating for a while when the steam condensates on colder surroundings. When comparing heartwood and sapwood chips, it was noted that the difference in heating time could be around 1 minute at most for pine, but only a few seconds for birch. This was most likely due to the pine heartwood and sapwood having distinct moisture contents, 25 % and 58 % respectively, while it was 41
% and 42 % in birch heartwood and sapwood. Birch and pine chips wee also steamed together, however, the difference in heating time was only a few seconds on average.
When comparing these experimental results with simulation data of the steaming of wood chips, it fit rather well when it came to the general heating time. However, the effect of increased moisture content had a much larger impact in the simulations, which predicted that more moist wood chips would need several minutes more steaming time, while the experiments only showed at difference of, at most, around 1 minute. When comparing with old experimental data, that has been the basis for the design of older steaming processes, it gave very distinct results, where the effect of thickness did not have as big of an impact as in the old data. No further comparison could be made, however, as the experimental conditions for the old experimental data were not known. Based on these results, it was noted that a steaming time of at least 5 minutes would be needed to ensure that even the larges and more moist chips could reach 100◦C in this system.
Finally, the condensate from the handmade birch and pine chips was analyzed. It revealed the presence of low molecular weight compounds like methanol, formic acid and acetic acid. Common metal ions were also present, although the amount of sodium ions clearly surpassed the rest. The pH of the pine condensate was measured and it was very high, which implies that the condensate was contaminated.
Sammanfattning
Basning av flis är ett viktigt steg inom kemisk massaindustri. Det avlägsnar luft från flisens insida vilket gör att impregneringen av luten blir bättre, vilket i sin tur leder till en jämnare kokning och färre rejekt. När basningen sker vid atmosfärstryck är det viktigt att flisen når en temperatur på 100◦C, annars kommer det finna ett jämviktstillstånd där lite luft blir kvar på insidan. Att ha otillräckligt basad flis i en process skulle kunna orsaka stora problem när det gäller att nå önskade kappatal, eller att ha en önskad retentionstid i kokaren.
Basningen görs på ett par olika sätt inom industrin, men det finns väldigt lite experimentel data tillgänglig angående flisens upvärmning, som skulle kunna användas när dessa system designas. De flesta vetenskapliga studier har fokuserat på luftborttagningen eller på förbättringar av impregneringssteget, medan de få studier som inkluderat mätningar av temperaturen ofta varit begränsade till ett slags trä, eller så har de inte inkluderat detaljer kring experimentet.
Därför basades handgjorde flisbitar av björk och tall från Sverige i en ATEX-designad kokare med en ångjacka.
Flisen hade termoelement inuti och temperaturen samt tiden kunde avläsas, vilket gjorde det möjligt att studera effekten av olika parametrar. Resultaten visade att det kunde skilja mer än en minut i uppvärmningstid mellan flisbitar av olika tjocklekar, både för tall och björk, även fast skillnaden i uppvärmningstid var mer linjärt relaterad till tjockleken för björkflisen. Tallflisen studerades också när trycket inuti kokar tilläts stiga vilket visade att det de tjockare flisbitarnas uppvärmningstid som kortas ned mest, eftersom de tunnare flisbitarna slutar värmas upp när ångan börjar kondensera på kallare ytor runt omkring. När flis av splintved och kärnved jämfördes visade det sig att skillnaden i uppvärmningstid kunde vara omkring 1 minut för tall, men endast ett par sekunder för björk. Detta beror troligtvis på att kärnveden och splintveden i tall hade stora skillnader i fukthalt, 25 % respektive 58 %, medan det för björk var 41 % och 42%. Björkflis och tallflis basades även tillsammans men det skillde bara ett fåtal sekunder i genomsnitt i uppvärmningstid.
När den experimentella datan jämfördes med data från simuleringar visade det sig att de stämmer väl överens när det gäller uppvärmningstiden i allmänhet. Å andra sidan förutspådde simuleringsdatan att en ökad fukthalt skulle leda till flera minuters skillnad i uppvärmningstid, medan endast 1 minuts skillnad uppmättes. När jämförelser gjordes med gammal experimentell data som använts som grund för tidigare processers design, var resultaten ganska olika eftersom den gamla datan visade en större effekt av ökad tjocklek än den som uppmättes.
Tyvärr kunde inte fler jämförelser göras eftersom detaljerna kring experimentet bakom den gamla datan inte var kända.
Slutligen analyserades även kondensatet från de handgjorda flisbitarna av tall och björk. Det visade att det fanns små mängder av små polära organiska ämnen, såsom metanol, myrsyra och ättiksyra. Vanliga metalljoner detekterades också, där mängden natrium var klart större än övriga metalljoner. Tallkondensatets pH mättes och det visade sig vara väldigt högt, något som tyder på att det troligtvis var förorenat.
1 Introduction
The global production of pulp is mainly based on chemical pulping processes, with 77 % of all wood fiber materials coming from chemical pulp [1]. The main chemical process is the kraft pulping process, which dominates the overall chemical pulp production [1]. During kraft pulping processes, it is of great desire to achieve uniform delignification of all wood chips. It has been shown that uniform delignification results in better pulp yield and pulp strength, compared to non-uniform delignification [2]. However, differences in thickness, porosity, uneven distribution of lignin within the fiber, as well as temperature and chemical concentration gradients in the process, make it very hard to achieve total uniformity [2]. A study by Gullichsen et al (1992) showed that differences in thickness alone can lead to large differences in delignification, even at laboratory conditions [2].
One of the main prerequisites to achieve homogeneous pulping of wood chips, is however, that the pulping chemicals can enter well inside the wood chips through the voids present in their structure [3][4]. Furthermore, efficient impregnation of the wood chips and diffusion of chemicals can reduce cooking time, and potentially lead to lower process pressure, increasing process capacity and thus lead to higher pulp production [5]. The main hindrance for this is the presence of air inside the wood, which make up roughly 1/3 of the wood chip volume [4][6]. This is illustrated in figure 1. Air present within the wood chips will be compressed as cooking liquor enters the chip, and will thus prevent the penetration of liquid and also the complete impregnation of the chips. Furthermore, it has been shown that this also can lead to problems for continuous cooking processes. Too much air within the chips will lead to lower densities, which in turn causes problem when the density of the chips is lower than the density of the cooking liquor. At that point, the chips will not move down the column at the desired rate [7].
Figure 1: The composition of a wood chip. Reproduced with permission from Valmet.
Furthermore, when designing new digesters, simulations are often used to predict the flow inside the digester and the kappa number. In order to acquire accurate simulations, many parameters must be well known, including the amount of air inside the wood chips.
The air is usually removed by presteaming the wood chips before cooking, which is a very common process in the pulp industry today. Modern systems in the industry tend to have an atmospheric presteaming step in the chip bin, at the beginning of the process, and sometimes this is followed by an additional steaming step in a pressurized vessel. In addition to having the air within removed, the chips are preheated up to 100-120 ◦C [4]. Simulations have been made on the presteaming and air removal of chips, where variables such as thickness, moisture, steaming temperature and steam velocity have been altered. These simulations, together with the experiment presented by Malkov et al. (2002), indicate that the desired chip temperature can be reached within 5 minutes of presteaming, however in order to get good enough air removal it can be necessary to steam up to 30 minutes, depending on which type of wood is used [4][8].
Several independent studies have shown the positive effects of presteaming, using different types of wood to achieve better impregnation [3][4][9][10][11]. All indicate, or state, that wood chips have a quite fast heating time, compared to the time that is required until the air removal becomes satisfactory. However, these sources are quite specific, with
the article by Malkov et al. (2002) only taking samples from pine (Pinus Silvestris) without varying the thickness, and the report by Oulie (1992), which presents experimental data from the pioneering work of Karin Wilson, giving few details on the wood chips which were analyzed, and how the experiments were carried out [4][10].
1.1 Project aim
Overall, there appears to be little data available that focus on the heating of wood chips. The simulations that have been made regarding the heating of wood chips during presteaming, are based on theory from mass and energy transfers, and use average values for wood density, and calculated values for the heat conductivity and heat capacity of wood [8][12]. Comparing these with experimental data could provide valuable insight regarding their accuracy. Morevoer, since most of the studies mentioned before focus solely on the degree of air removal, and do not present detailed graphs from the heating of the wood chips, with the exceptions being the reports by Oulie (1992) and Malkov et al. (2002), it could be useful to provide more experimental data with detailed descriptions of the experimental setup. Furthermore, the experiment conducted by Karin Wilson were based on old technology, and it might therefore give different results compared to the modern equipment that can be used to measure temperature.
Thus, it is of interest to study the heating of a wider variety of wood species, including both hardwood and softwood, as well as studying the differences when steaming chips of different sizes, and comparing chips from heartwood with chips from sapwood. These should be steamed in a bed of chips inside a digester to more accurately resemble the conditions encountered within the industry. Moreover, it is of interest to study the chemical changes that may take place during presteaming, as components extracted into the condensate will be a part of the process further down. In this thesis, a method for acquiring heating data during steaming, from several chips at the same time, was developed, and wood chips from pine and birch were steamed under different conditions, and their temperature profiles were recorded. In addition to that, the condensate was analysed in this report, with particular focus on metal ion content, which can influence future processes. A GC-MS analysis was also performed to identify the types of resins that could be found.
2 Background
2.1 Wood
Wood is a complex material that can be divided into several parts depending on the wood tissues. Firstly, there is the dead bark that serves as protection. Under the bark, there is the phloem which consists of living cells and that transports nutrients and storage products. Moreover, there is the vascular cambium, which is a thin layer of cells that divide and produce cells for the phloem and to the xylem on the inside. The xylem itself constitutes the main part of the wood and it can be divided into the sapwood, which contain both living and dead cells, and heartwood, which is closer to the center and that consists of almost exclusively dead cells and that contains a higher amount of resin. In many trees, like pine for example, the difference between sapwood and heartwood can be seen through a change of color, as in fig. 2. Finally, there is the pith in the center which consists of the tissues from the tree’s first few years of growth. The wood material as a whole is made up of cells that have ordered themselves in a particular way, with cells along the longitudinal direction of the wood providing mechanical support and transport, and ray cells in the radial direction that primarily stores and distributes storage material such as starch [13][14].
When it comes to the chemical composition of wood, it consists of three major components, cellulose, hemicellulose, and lignin. Cellulose consists of β-1,4 linked glucose units that can form sheet due to hydrogen bonding, and these sheets can stack, forming fibrils which eventually build up the complex wood structure. The hydrogen bonding is present between every unit, making cellulose very stiff and thus give the mechanical strength to the wood [15].
The lignin is present as a large, crosslinked 3D-network based on mainly three types of monolignols. The structure of lignin can be varied quite a bit but some types of bonds are particularly common. The lignin holds all the cells together while at the same time making the cell wall hydrophobic and stiff, and at the same time protecting the wood from microbial attacks [16]. The hemicelluloses are polysaccharide chains that contain more types of monosaccharide units than just glucose, and they most likely act as an interface between the cellulose and lignin.
The composition of hemicelluloses vary depending on the type of tree. In softwoods, glucomannan consisting of glucose and mannose units, and arabinoglucuronoxylan consisting of xylose, arabinose and glucuronic acid units, are the main hemicelluloses that are present. In hardwoods it is glucuronoxylan consisting of xylose and glucuronic acid units, and glucomannan that are the most common [17].
Figure 2: Cross section of the pine tree used in this thesis. A clear difference can be seen between the darker sapwood and lighter heartwod.
Alongside with the main components, there are smaller amounts of resins, that can be divided into three groups:
aliphatic compounds, terpenes and phenolic compounds. The aliphatic compounds are mainly different types of fatty acids, where oleic acid and linoleic acid are the most common ones [14]. The terpenes contains a variety of molecules based on isoprene units that are mostly cyclic. Monoterpenes and sesquiterpenes, which are built up from two isoprene units and 3 isoprene units respectively, are characteristic compounds in softwood resins, which creates the typical pine aroma. Some of them also contains oxygen and these are referred to as terpenoids.
Furthermore, three terpenes, namely α-pinene, β-pinene and ∆-3-carene, make up the major constituents of turpen- tine, which has a wide variety of applications such as pesticides, insect repellants, fragrances and more [14][18]. The phenolic compounds are almost exclusively present in the bark or heartwood, and they usually acts as fungicides.
The heartwood’s distinct color can in many cases be attributed to the polyphenols that have been formed. Some common types of phenolic compounds include lignans, flavanoids and tannins [14]. There are also some inorganic elements present within wood, such as sodium, potassium, calcium, magnesium and silicone. The ratios between these components vary between softwoods and hardwoods, however cellulose is the main components in both and makes up around 40 % of the wood [14][19].
Figure 3: Difference in structure between softwood (to the left) and hardwood (to the right), as seen from a microscope [15].
In addition to this, there is also a difference in structural morphology between softwoods and hardwoods. The structure of softwood is considered to be more simple compared to hardwood, as softwood has fewer cell types and the cells in softwood are more uniform. Softwood consists of long tracheid cells, ray parenchyma cells and ray tracheid cells, while hardwoods consists of vessels, fibers, tracheids and parenchyma cells. These differences can be noticed by examining wood pieces under a microscope, and they do have an impact on mechanical and properties of the pulps produced from hardwood and softwood [13]. An example of the visual differences is shown in Fig. 3.
Morevoer, there are also differences in void structures and porosity between different types of wood and between wood species [1]. The porosity of pine has been measured to be roughly 65 %, and the porosity of birch is around 57 % [20]. Special reaction wood will also have a different structure, as these are much denser. Finally, the moisture content is another parameter that can vary significantly, even within the same piece of wood. Generally speaking, the sapwood of softwoods has a higher moisture content than heartwood, while in hardwoods that can differ. All of these properties will have an impact on the ability of liquids to penetrate inside and impregnate the wood [1].
2.2 Kraft Pulping
Kraft pulping is a chemical pulping process, in which lignin is removed through the degradation of its chemical bonds, thus liberating the cellulose fibers from the wood. It has become the main chemical pulping process, as it accounts for 89 % of the produced chemical pulps [1]. In this process, hydroxide and hydrosulfide ions react with the lignin, and they are introduced in an aqueous solution of sodium hydroxide and sodium sulfide, known as white liquor. It is the cleaving reaction of the β-aryl bonds of the lignin which is the main reaction during kraft pulping. However, several side reaction can occur, such as the peeling reaction and alkaline hydrolysis which are
undesirable as they lead to the degradation of cellulose and hemicellulose, and thus a yield loss. The kraft cooking can be divided into three phases, defined by the degradation of carbohydrates. In the initial phase, the reaction is not particularly selective, however, after the loss of around 20 % of both lignin and carbohydrates, the reaction enters the second phase and becomes quite selective. The main part of lignin is degraded during this phase while only smaller amounts of carbohydrates are affected. The final phase, is when there is little lignin left and thus the carbohydrate loss will increase more drastically again. It is therefore necessary to stop the cooking at the right time, when the right kappa numbers have been reached [1][21].
2.3 Batch Kraft Process
Batch cooking is one of the two main kraft pulping processes, the other one being the continuous processes. In batch processes, a set of individually loaded digesters are used during cooking. The chips are added from the top through an automatic capping valve, and discharged through the bottom where there is a blow valve. At the bottom, a portion of the liquor can usually be extracted, heated and reintroduced both at the top and the bottom to achieve more uniform heat and cooking liquor distribution. There is also a vent at the top that allows air and volatile gases to escape [22]. Batch digesters are designed to have volumes between 60-300 m3, and they are usually filled in around 30 minutes [22].
The chips can be presteamed inside the digester to increase the temperature up to 100 ◦C and to remove the air. The digester is then filled with cooking liquor, which is circulated inside the digester and heated up to the cooking temperature using heat exchangers and steam. The heating process can take 90-150 min, and the cooking can continue for another 45-60 min after the desired temperature has been reached. The cook can then be quickly stopped by degassing the digester and thus lowering the temperature by up to 20◦C. The contents inside the digester are blown out to an atmospheric blow tank and the generated flash steam is passed through a steam separator and a flash heat collecting system. All in all, the entire batch cooking process can take around 4 hours, and a pulp mill can use several digesters to have a smooth pulp delivery, by sequencing them in an appropriate way [22].
However, since the 1980s an alternative batch cooking process, known as "displacement batch cooking", has become more popular. This is due to the fact that it is more energy efficient, as it displaces the waste, known as the black liquor, from a cook and stores it, making it possible to capture the heat and residual chemicals and reuse them for a subsequent cook. Furthermore, these displacement batch cooking systems eliminated the need for complex blow heat recovery systems [22].
2.4 Continuous Kraft Processes
Modern kraft pulping processes use continuous digesters to a very large extent, and the development is constantly driven by the desire to increase the production capacity, and to improve the quality of the pulp[23]. These continuous systems were able to replace older batch processes, as continuous systems are inherently more efficient and only requires a single, vertical reactor vessel, in contrast to the many reactor vessels needed for batch processes. In addition to that, the operation becomes more stable overall, since there are no peaks in flow rates, contrary to batch processes. Finally, lower process energy requirements, more efficient energy recovery and process flexibility can be achieved by the specific design of continuous digesters [23].
The continuous digester system can be described, in a simplified manner, as starting with chips entering a chip bin, where atmospheric presteaming can take place up to around, from which the chips are subsequently discharged by a chip metering device. The chips continue through a low pressure feeder into a pressurized steaming vessel, which also functions as a screw conveyor, that sends the chips into a vertical chute. In the chute, there is process liquor that circulates through a high pressure feeder. The high pressure feeder then expels the chips, along with the liquor, to the top of the digester, where a separator screw separates the chips from the liquor. The liquor is then returned to the high pressure feeder, while the chips enter the digester and forms a chip column, that moves down the digester. Here, the cooking liquor is also added [23].
The reactor is essentially a solid-liquid system, as the amount of gas present is very low, and the liquid that occupies the free volume of the chip column is called unbound liquor. As the chips move down, they pass through several sections in the digester with different retention times. The first section is the impregnation zone, where temperatures
usually are around 115-125◦C and the cooking chemical can diffuse in during 45-60 min. The following section is the cooking zone, which is separated by a screen section that enables the extraction of unbound liquor from the column.
In the cooking zone, the temperature is raised to between 140-170◦C, and the chip retention time is between 1.5-2.5 h. Moreover, there is the extraction zone, where unbound liquid is extracted and removed from the system. The liquor is then used to acquire flash steam for the steaming vessel and chip bin, as it is brought down to atmospheric pressure and saturation temperature. The extraction generates a counter current of unbound liquor underneath the extraction zone. This is used to acquire a counter current washing zone beneath the screen that ends the extraction zone. Near the bottom there is a cooling zone, where wash filtrate at a lower temperature is introduced, and a part of that will move upwards in the digester. Finally, the cooled chips and filtrate mixture is discharged by an outlet scraping device into a blow-valve, where the chips are converted to pulp through depressurization [23].
Some major developments have taken place regarding continuous kraft processes in the last couple of decades.
One of these is the CompactCookingTM, which was developed by Valmet. In the CompactCookingTM process, the chips first enter an atmospheric impregnation vessel where they are presteamed. The impregnation liquor is then fed into the impregnation vessel, from the top screen of the digester, and then white liquor is added both to the impregnation vessel and the digester. The impregnation takes place at around 102◦C during 60 minutes, and thus the conditions can be considered milder compared to the general process. Once the impregnation is complete, the chips are fed to the top of the digester and steam is used to increase the temperature. The cooking will then be normally carried out at a temperature around 150◦C, however it depends on the raw pulping material that is used [24]. The impregnation in the CompactCookingTM process is an improvement compared to other processes, and thus it will ideally result in more uniform cooking and lower amounts of rejects when higher kappa numbers are targeted [25].
2.5 Presteaming and Impregnation
The main effect of the presteaming step, and the eventual pressurized steaming step, is to raise the temperature and expel the air from the inside of the wood chips. The pressure from the gases inside can be viewed, in a simplified manner, as the sum of the partial pressures of vapour and air [4]. As the temperature increases inside the chips and steam enters, the pressure inside the chips will increase, causing the air to be pushed outwards. Having the surrounding environment saturated with water vapor also causes the air to diffuse out from the chip, as a partial pressure gradient will be present [23]. On the other hand, some steam will condensate inside the wood chips, which can prevent the air from coming out [5]. Fig. 4 illustrate the process of air expulsion.
Figure 4: The expulsion of air caused by heating of the wood chips as steam condensates.
At atmospheric pressure, it is necessary that the chips reach a temperature of 100◦C, otherwise there will be an equilibrium in which the gas phase inside the wood chips contains some air together with the water vapour [8]. A
paper by Sergey Malkov et al. (2004) indicated, however, that the air removal also depends on the morphology of different woods. It was shown that the air removal was faster for pine (softwood) sapwood compared to eucalyptus (hardwood), which in turn had a faster air removal than pine heartwood. This can be explained by considering the fact that the pits between pine tracheids are larger than the intervessel pits of eucalyptus. However, the pine pits in heartwood are subjected to resinification, which makes the air transport through them less effective [7].
Simulations have been used to highlight how changing different parameters can affect the heating time and air removal. One of the most noticeable results is, that the simulations show that increasing the chip size and moisture has a significant impact on the time required to reach 100◦C, with thicker chip requiring more time [8]. Higher initial moisture content would also lead to increased heating time, as it will increase the chip density and heat capacity [8]. The expected time to reach the desired temperature for thicker chips is close to 5 minutes, however, during industrial conditions the heating time might be even longer than simulations predict, as the heat transfer might be more limited [26]. Altering the steam temperature between 110-130◦C also has a noticeable effect, however it was not considered to be that significant. Steaming at 110 ◦C would require around 100 seconds more than steaming at 130◦C [12]. Regarding the air removal, it was shown in the simulations that the moisture content will have an effect of the air removal speed, as the moisture takes up space, thus lowering the overall air content inside the chips, and lowering the air permeability factor. According to the calculations made during the simulations, chips with 30
% moisture would have more than 4 times as much air inside, compared to chips with 60 % moisture. However, the differences in air content become rather small once the system reaches a temperature near 100◦C [8].
The impregnation of wood chips is an important step, as mentioned previously, and it can be assessed through various methods. It consists essentially of two mechanisms: the penetration of liquor into the voids of the chips, and the diffusion of ions caused by the concentration gradient [27]. The penetration of water can be used as a way to determine the efficiency of steaming on air removal [27]. If certain simplifying assumption are made, such as that the chip volume is constant, density of wood is constant, water is incompressible, and that losses of wood substance can be neglected, the degree of penetration can be calculated by measuring the weight changes during impregnation. This requires that the theoretical maximum amount of water that can enter is calculated as well, and this can be done by measuring the initial weight of the chips, the dry matter content and the basic density of the chips [27]. It has been shown, however, that measuring the penetration degree is difficult, when liquors other than water are used. The reason is that chemicals diffuse inwards, simultaneous to the dissolution of wood material [28]. It is necessary to take into consideration the density of the liquor and the dissolution of wood material.
There are two main methods used in order to gain information regarding those properties of liquor that has impregnated wood chips. These are known as the pressing technique, where the entrapped liquor is pressed out of the wood at several MPa, and the leaching technique, where the alkali is allowed to diffuse out into water followed by titration to determine the amount. However both suffers from certain drawbacks and in a study by Määttänen and Tikka (2012), an experimental method was developed which could give more realistic results for different impregnation conditions [28].
Moreover, it has been shown by Inalbon et al (2005) that the impregnation process can be seen as a "reaction zone"
that moves further in towards the center of the chips [9]. In their study, which used alkali as a reagent, this reaction zone could be identified as the zone in which the alkali concentration dropped sharply, and the concentration of acetyl groups increase sharply. Beyond this zone, an unreacted internal zone can be identified if the impregnation process still isn’t complete. However, even before reaching the reaction zone, the alkali concentration slowly drops further inside the chips from dried woods, indicating there is a restriction to the diffusion of alkali [9].
The steaming and impregnation have a large impact in chemical pulping processes, as mentioned previously, and this can be seen in simulations made by Valmet, using a software called Crossim, developed by Valmet. By giving values to variables, attained from real pulp mills, it is possible to calculate the values for the chip compaction (the amount of chips that can fit in a unit of volume), chip pressure (the driving downwards force of the chip column), and the kappa number, throughout the length of the digester. An example of a simulation based on a real mill can be seen in fig. 5, which shows the effect of achieving perfect presteaming of the chips, compared to having poorly presteamed chips. As can be seen, the chip compaction and chip pressure drops noticeably for the case with poorly presteamed chips, which is represented by the blue lines. This would imply that the volume of each chips is larger, since the compaction is lower, and thus the retention time of each chip will be lower.
Figure 5: Simulation of the chip compaction (upper graph) and chip pressure in kPa (lower graph) as a function of the distance in meters, in a real digester based on two scenarios, good presteaming (red) and bad presteaming (blue). Unpublished data by Anders Hjort, Valmet.
2.6 Chemical modification during steaming
In addition to expelling air, steam treatment of wood has been shown to change the chemistry of the wood [29]. Most articles deal with conditions not encountered in a pulping process, however there are studies indicating reactions under conditions closer to those of a continuous pulping process. A study by Bäckström et al (2016), as an example, gave evidence suggesting that steaming at 5 bar for 60 min caused autohydrolysis and an increase amount of reducing end groups. It was shown that the samples steamed under those conditions had a higher alkali consumption, and that the yield was considerably lower, compared to samples steamed under atmospheric conditions [3]. Moreover, reports indicate that even steaming under atmospheric pressure can have an effect on wood. In the dissertation by Hultholm (2004), it was reported that a 45 min steaming lead to higher wood loss, compared to chips that had air removed via vacuum suction [3]. Furthermore, studies have shown that the fiber wall becomes more accessible to liquid after presteaming, and a study by Matsumura et al (1999) in particular, showed that the steam can remove resins inside, and that it can damage bordered pits and ray cells [3][30]. However, within the time frame of presteaming, only minor chemical effects on wood can be noticed, although it is yet not well known the extent to which resins are extracted [12].
3 Experimental
3.1 Materials
Handmade wood chips were prepared from two tree species commonly found in Sweden, namely pine and birch.
The pine tree used in this study was 20 cm in diameter and its age was determined to be 45 years by counting the annual rings. The birch tree had a diameter of 25 cm, and by attempting to count the annuals rings, it was estimated to be around 50 years The pine and birch chips were created with the same length and width, 35 mm and 25 mm respectively, but with the thickness being either 5 mm, 8 mm or 10 mm. The length direction of the chips was in the same direction as the tree growth. Finally, some chips with thicknesses of 5 mm or 10 mm, were made purely from sapwood or heartwood. A TC-08 data logger and K-type thermocouples from Pico Technology were acquired and used to measure the temperature inside wood chips in real time, recording the temperature once every second. The steaming took place in a circulatory digester that was specifically built for Valmet, known as CK5. The CK5 system was ATEX designed, and consisted of one digester with a steam jacket, three accumulators, three circulation pumps, one hydraulic diaphragm metering pump, and five heat exchangers. The digester vessel was 850 mm high and had a volume of 47.7 L, and was designed to withstand pressures up to 40 bar. The CK5 system is illustrated in fig. 6. The thermocoules were inserted to the CK5 by removing one of the bolts on the side of the digester vessel, and replacing it with a barrier gland which the cables could go through, rubber was then used to seal it.
Figure 6: The full CK5 system (left) and a close up on the digester vessel where the steaming took place (right).
Illustration reproduced with permission from Valmet.
3.2 Steaming experiments
Small holes, with a diameter of 1,2 mm, were drilled in the longitudinal direction of the chips, and the K-type thermocouples were inserted. Thinner chips would have been difficult to use, since it would be increasingly more difficult to drill holes in them. The rubber surrounding the cables fit rather tightly inside the drilled hole, thus it was decided that no additional measures to plug the holes would be taken. The moisture content of the chips was also measured. Some primary tests were conducted, by having chips with a thermocouple inside being hung over boiling water. After that, the experiments were conducted inside the CK5 digester, with steam being introduced
by manually opening a valve, letting in steam generated from a boiler at a pressure of around 11 bar. The pressure was controlled manually by opening and closing the valve to a certain degree, thus controlling the flow of steam into the digester, and by using an automatic valve that would release steam if the pressure would rise above a given value.
The handmade chips with thermocouples inside were placed in a basket with roughly 1 kg (dry weight) of mill chips, in order to resemble the condition encountered within in real mill processes as much as possible. However, in order to minimize the potential effects of steam channeling, the chip layer above the handmade chips should not be too thick, and therefore they were placed in the upper part of the basket with only a few centimeters of chips above. For each steaming experiment, a maximum of eight chips can be measured at the same time, however it was decided to always have one thermocouple that was not inserted into any chip as a reference during steaming.
The effect of thickness was studied by steaming pine wood chips with the different thicknesses at the same time.
Since only seven thermocouples were inserted into wood chips, there would be at least two samples of each chip thickness during the experiment, and for one of the chip thicknesses there would be three chips. The experiment was carried out six times with the pine chips, and half of the time it was done at atmospheric pressure, and half of the time the pressure was allowed to build up until it reached 2 bar. New chips were used every time, and it was arranged so that there would be in total seven temperature curves for each chip thickness.
Moreover, the difference between sapwood and heartwood was studied by steaming three chips of each wood type at the same time. Sapwood and heartwood chips with a thickness of 5 mm were steamed together, and afterwards sapwood and heartwood chips with a thickness of 10 mm were steamed together. Both experiments was also repeated three times, giving a total of nine curves for each type of wood, all of which were at atmospheric pressure.
Finally, the difference between thicknesses, and between sapwood and heartwood was studied in the same way for birch chips. Comparisons were made directly between pine and birch chips as well, by having 3 chips of each species with a thickness of 5 mm steamed together, followed by having 3 chips of each species with a thickness of 10 mm steamed together. These experiments were also repeated three times giving a total of nine curves for each wood type and chip thickness. During the experiments involving only pine chips, or involving pine and birch chips steamed simultaneously, the handmade chips were placed in a basket containing fresh pine mill chips. For the experiments with only handmade birch chips, they were placed in a basket with birch mill chips.
The experiments were numbered based on the order in which they were performed. Table 1 shows which how the experiment numbers and the respective descriptions of these experiments.
3.3 Condensate analysis
The CK5 digester was washed with hot water a few times to remove any remaining compounds from previous experiments. Roughly 138 g (dry weight) of the handmade pine chips were steamed, without the presence of any mill chips. After roughly 4 minutes, a tube was removed from the bottom of the digester to create an opening from which steam and condensate could be collected. Roughly half a 50 ml flask was filled with condensate. This experiment was repeated one more time, and the two flasks with condensate were stored in a refridgerator for 2 weeks, before being sent to MoRe Research AB (Örnsköldsvik, Sweden) for analysis. The condensate from birch chips with a dry weight of 86 g was collected in the same way, and it was stored for a week before being sent to MoRe Research.
Table 1: Experimental numbering and description.
Experiment number Description
1-3 Comparison between pine wood chips of the three dif-
ferent thicknesses, steamed at atmospheric pressure to- gether with fresh pine chips from a mill. A total of 21 chips were steamed during these three experiments, with 7 from each thickness.
4-6 Comparison between 5 mm thick pine wood chips of
pure sapwood and pure heartwood, steamed at atmo- spheric pressure together with fresh pine chips from a mill. A total of 18 chips were steamed during these three experiments, half from sapwood and half from heartwood.
7-9 Comparison between 10 mm thick pine wood chips of
pure sapwood and pure heartwood, steamed at atmo- spheric pressure together with fresh pine chips from a mill. A total of 18 chips were steamed during these three experiments, half from sapwood and half from heartwood.
10,12,13 Comparison between wood chips of the three differ-
ent thicknesses, steamed at over-pressure together with fresh pine chips from a mill. An automatic valve was used, in addition to the manual valves, to start releas- ing steam if the pressure surpassed 1.6 bar. A total of 21 chips were steamed during these three experiments, with 7 from each thickness.
14-16 Comparison between 5 mm thick pine and birch chips,
steamed at atmospheric pressure together with fresh pine chips form a mill. A total of 18 chips were steamed during these three experiments, half from pine wood and half from birch wood.
17-18 Repetition of experiments 4-6. A total of 12 additional
chips were steamed.
19-21 Repetition of experiments 14-16 .
22-24 Comparison between 10 mm thick pine and birch chips,
steamed at atmospheric pressure together with fresh pine chips form a mill. A total of 18 chips were steamed during these three experiments, half from pine wood and half from birch wood.
25-27 Same as experiments 1-3 but with birch chips.
28-30 Same as experiments 4-6 but with birch sapwood and
heartwood chips.
31-33 Same as experiments 7-9 but with birch sapwood and
heartwood chips.
4 Results & Discussion
4.1 Analysis of handmade chips
All the chips that were used in the experiments during this thesis had their moisture content measured. The results from these measurements can be found in table 2. The chips were divided into four groups, namely the mill chips, the average handmade chips, the pure sapwood handmade chips and finally the pure heartwood handmade chips.
One of the most noteworthy results was the fact that there is a big difference between the moisture content of pure pine heartwood chips from pine and the rest of the chips. This was expected, as there were clear differences between the heartwood and sapwood of the pine tree used, as shown in 2. Birch, on the other hand, does not have as clear of a distinction between heartwood and sapwood, and it can be seen as the moisture content was roughly the same for all the birch chips.
Table 2: Moisture content of wood chips used during the experiments.
Wood chips Moisture content Pine Birch Mill chips 53-55 % 43 % Handmade (average) 49 % 42 %
Sapwood 58 % 42 %
Heartwood 25 % 41 %
The condensate analysis was performed by MoRe Research, and the results are presented in table 3, and it shows that the amount of sodium is quite high, and it is the most common ion present both for pine and birch. Moreover, the second and third most common ions were potassium and calcium, respectively, for both tree species. Overall, the birch condensate contained more of each extractive that was detected, except for iron. According to the literature, potassium and calcium are the most common metal ions present in wood, as these both play important roles during the cell wall formation, and cellulose synthesis, respectively [14][19]. The high amount of sodium could potentially be due to the environment that these trees grew in. Regarding organic compounds, there were low amounts detected.
Small polar compounds could be noticed via HPLC, but these were not identified specifically, except for methanol.
Among common acids with carbon chains between C2-C7, only formic acid and acetic acid were detected. The pH was not measured for these samples, but according to literature it should be expected that the pH should be somewhere around 3-5 [31][32].
Table 3: Result from the analysis of the condensate collected from the CK5.
Extractive Unit Method Pine Birch Aluminium mg/kg SCAN-N 38 <0.2 <0.2
Iron mg/kg SCAN-N 38 0.8 0.7
Magnesium mg/kg SCAN-N 38 0.5 0.6 Manganese mg/kg SCAN-N 38 <0.2 <0.2
Calcium mg/kg SCAN-N 38 1.6 2.4 Potassium mg/kg SCAN-N 38 2.1 2.9 Cupper mg/kg SCAN-N 38 <0.1 0.1
Sodium mg/kg SCAN-N 38 168 223
Methanol mg/L KA 80.314 10 10
Formic acid mg/L KA 80.309 25 30 Acetic acid mg/L KA 80.309 35 40
4.2 The effect of chip thickness for pine
From the data recorded by the thermocouples, it was possible to plot graphs for how the temperature changed within each chip. All of the graphs from every experiment can be found in the appendix in numerical order. When steaming at atmospheric pressure, the temperature of the chips rose sharply in the beginning, which was followed by the curves reaching a plateau at a few degrees above 100◦C. However, in some cases the pressure and temperature of the system could start to rise sharply above 100 ◦C, before the valves were adjusted, which resulted humps on the curves. When steaming at higher pressure, it was more complicated to reach a stable plateau, however the shape of the curves was similar since they also had a sharp increase in the beginning followed by the curves levelling out.
Fig. 7 and fig. 8 are good examples of how the curves can look when steaming at atmospheric pressure. As could be expected, the chips with 5 mm thickness heat up faster than the 8 mm and 10 mm thick chips in all cases. However, when it comes to the 8 mm and 10 mm thick chips, there are exceptions where thicker chips heat up faster than thinner ones. In addition to those observations, it can be seen that the curves overall look a bit different between experiment 1 and 2, and the same is true when comparing the entire raw data curves for experiments 1-3, which can be found in the appendix. This is due to the fact that, unfortunately, it was difficult to achieve the exact same conditions in every experiment. The manually controlled valve was rather sensitive, and thus there most likely were small differences in the amount of steam which was let in during each experiment which had an effect on the heating time.
Figure 7: Graph from experiment 1, in which the effect of thickness on the heating time was studied.
Figure 8: Graph from experiment 2, in which the effect of thickness on the heating time was studied.
Furthermore, there was a difference in the time it took for the steam to reach the different chips. Some chips appear to start heating several seconds earlier than others. Therefore, in order to get a better estimate for the heating time, the point where the temperature had reached 30◦C was used as a starting point for each chip. The heating time was thus calculated as the amount of time each chip required to go from 30 ◦C up to 100 ◦C in the case of atmospheric steaming, and up to 120◦C in the case of steaming at over-pressure. The results are presented in bar graphs, and it can be seen that there are clear shifts in time between each steaming.
The bar graphs, which are shown in fig. 9 help assert the fact that the 5 mm chips continue to be the fastest in almost all cases (only one 5 mm chip heated up more slowly than one 8 mm chip), but that the results are more mixed when comparing the 8 mm chips and 10 mm chips. According to theory, and the simulations by K. Kovasin [8], thicker wood chips should always be slower, however, if the difference in thickness is only around 2 mm, it might not be enough to prevent other factors from having an impact on the results. In this case, structural differences between individual wood chips, such as the thickness between annual rings for instance, could be the reason behind some of the unexpected results. In addition to that, according to the simulations by Kovasin [8], the moisture content can have a significant impact on the heating time, with more wet chips requiring more time. Therefore, this could also have affected the results in this experiment, as separation between heartwood and sapwood was not
considered when these chips were prepared. Thus, it is possible that some of the chips that required more heating time were, to a larger extent, made up of sapwood, compared to some of the ones that heated up faster.
Figure 9: Heating time (in seconds) required to go from 30◦C to 100◦C for handmade wood chips of different thicknesses during the 3 repetitions of the experiment (denoted number 1-3).
During the experimental procedure, it was noted that it appeared that some wood chips stayed slightly below 100
◦C for significantly longer time compared to others. Thus, to get a better understanding of the heating pattern, average curves were generated from all of the measurements. Here, the time it took to go from 30◦C up to each temperature between 50◦C and 100◦C was recorded and an average was taken for each chip thickness. The results are presented in 10, where the temperature is on the X-axis, and the time it takes to reach that specific temperature is on the Y-axis. From these curves it is easy to see the general trend that the 8 mm and 10 mm chips do require a bit more time, and on average the difference between them is less than 10 seconds. The 5 mm clearly heat up faster, but the curve drastically increase its slope when it approaches 99◦C. This is most likely due to the fact that the surrounding is still much cooler than the 5 mm chips at the time that they reach 99 ◦C and that the vapour pressure still hadn’t become high enough. Thus, it took more time for them to reach 100 ◦C since the incoming vapour will condensate and spread its heat onto the surrounding first. By the time the thicker chips reached 99
◦C, the surrounding was at a higher temperature and thus the heating didn’t slow down as much, although it is possible to see an increase. From fig. 10 it can be seen that the average time it takes for a 5 mm chip to reach 100
◦C is roughly 140 s, while for the 8 mm and 10 mm chips it is 184 s and 205 s respectively.
4.3 Comparison between pine heartwood and sapwood
The results from the experiments 4-9, including only the pure sapwood and pure heartwood chips appear to fit well with the simulations by Kovasin [8], as it is the more moist sapwood chips that take more time to heat. It can be noted, however, that the difference was clearly more prominent for the 10 mm thick chips, while the results from the 5 mm thick chips are more varied. One example of the temperature curves for each chip thickness can been seen in fig. 11 and 12, and the rest can be seen in the appendix. The exact heating time for the 5 mm thick chips can be seen in the bar graph in fig. 13 and the average heating curve can be seen in fig. 14. Both of these graphs were calculated in the same way as for the experiment with chips of different thicknesses.
Figure 10: Average heating time (in seconds) required to reach the temperatures within the interval of 80 ◦C to 100◦C for wood chips of different thickness. (Experiments 1-3)
Figure 11: Graph from experiment 5, in which the difference in heating time between 5 mm thick pine sapwood and pine heartwood chips was studied.
Figure 12: Graph from experiment 7, in which the difference in heating time between 10 mm thick pine sapwood and pine heartwood chips were studied.
Figure 13: Heating time (in seconds) required to go from 30◦C to 100◦C for 5 mm thick handmade wood chips of pure heartwood and pure sapwood during the 3 repetitions of the experiment (number 4-6).
Figure 14: Average heating time (in seconds) re- quired go from 30◦C to temperatures within the in- terval of 50◦C to 100◦C for 5 mm thick handmade wood chips of pure heartwood and pure sapwood.
(Experiments 4-6).
It can be seen in fig. 13 that, despite the results being a bit varied, the general trend appears to be that the sapwood chips require more time in most cases. Interestingly enough, there is one bar from the heartwood chips that is higher than the rest in all three cases, and it corresponds to the same thermocouple in each case. This was unexpected as the heartwood chips were more dry, and they have larger space between the annual rings in this case.
A test was made to see if there was any error in measurements, in which all the thermocouples were heated together in a beaker with water, however no anomalies could be detected. There could perhaps be other factors that affect the result, probably related to differences in structure between individual chips, same as for the experiment with chips of different thicknesses.
When examining all the data it can be seen that some chips took a long time to go from 99 ◦C to 100 ◦C, while others reached the limit relatively fast. The thermocouple that recorded the longest heating time in all repetitions needed 47 s, 83 s, and 63 s to go from 99◦C to 100 ◦C in experiment 4, 5 and 6 respectively. In comparison, the other thermocouples only required between 10-20 s on average for the same temperature change. It was seen that it lingered between 99.5-99.7 ◦C during this period and it is difficult to specify the exact cause of this. To verify that this weird result was indeed linked to the thermocouple that was used, it was replaced and the experiment was repeated two additional times.
When examining the curves in fig. 14 it is possible to see that the average heating times fit with what could be expected. The sapwood chips take longer time to reach each temperature in the interval, except for 100◦C where the heartwood chips’ average heating time become larger due to the aforementioned long heating time of one particular thermocouple. The difference between these chips does not appear to be particularly big. It is slightly less than 10 s on average between the average heartwood curve and the average sapwood curve, however, when reaching 100
◦C the difference was 13 s. In the additional data from experiments 17 and 18, the sapwood required more heating time for all temperatures, and when this data was included in the calculations, the average difference was reduced to 0.8 s at 100◦C, which does help to certify that the the previous results were distorted by a faulty thermocouple.
The average curves with the additional data included can be seen in fig. 15.
In the case of 10 mm thick chips, the results are much more aligned with the expectations, and they are presented in fig. 16 and fig. 17. The heartwood chips heat up faster in all but one case. It is likely that at a larger thickness, the effects of moisture and larger annual rings are enough to prevent other smaller factors to distort the results.
Furthermore, it can be noted that there is a large increase with time, quite similar to that which can be seen in fig.
9. The time needed goes from around 100-150 s for the 5 mm chips, up to well above 200 s in some cases. It can be noted here that there no longer is any thermocouple that requires an abnormally large amount of time to reach 100 ◦C, even though the same thermocouples were used.However, there was one instance where one chip appears
to have been heated up much faster relative to the others in experiment 8. The chip in this case did not have any apparent damages, thus this could be the result of this particular chip having a more porous structure compared to the others. Another potential cause could be an error during the drilling of the hole, leading to a slightly tilted hole which would put the tip of the thermocouple closer to the surface. This could happen if the tip of the drill is bent, however that was not observed when drilling.
Figure 15: Average heating time (in seconds) required to reach the temperatures within the interval of 50 ◦C to 100 ◦C for wood 5 mm thick wood chips of pure heartwood and pure sapwood. (Experiments 4-6 together with experiments 17-18. Additional data added to compensate for a faulty thermocouple)
When looking at the average curves in fig. 17, it can be noted that the average difference is also much larger compared to the difference between 5 mm chips. At most, the difference in heating time is close to a minute, which can be seen in the time required to reach 96◦C.
However, near the end the heating time for the heartwood chips start to rise, and the difference shrinks noticeably down to 34 s. The rise in heating time near the end for heartwood chips is, thus, something that can be observed in both cases. It is most likely due to the fact that since they reach temperatures close to 100◦C, while the surroundings are at a lower temperature, the incoming steam will not condensate on them but on the cooler surroundings instead.
It is the same as for the 5 mm chips when the different thicknesses were compared.
The heating times were, on average, 103 s and 116 s for the 5 mm sapwood and heartwood chips without the data from experiments 17 and 18, and 143 s and 144 s with the additional data. For the 10 mm chips they were 240 s and 206 s for sapwood and heartwood respectively. There is, therefore, a difference between the average heating time for 5 mm chips and 10 mm from experiments 1-3, which was 140 s and 205 s respectively, and the average heating time from the sapwood and heartwood chips. Regarding the 5 mm chips, it is likely that the difference was caused by different conditions during the experiments. These different conditions, in turn, most likely arose because of the sensitivity of the manual valve. The valve was probably opened a bit more during experiments 4-6 which lead to a faster heating, while experiments 17-18 most likely had a more closed valve which caused them to heat slower.
On the other hand, it does not appear as if it was any major changes in the conditions other than the faster heating time. The system stabilized itself slightly above 100 ◦C in experiments 4-6 and 17-18, just as they did in experiments 1-3. When it comes to the 10 mm thick chips, the average heating time from experiments 1-3 seems to match the average heating time of the heartwood chips. However, since it is not particularly likely that all of the 10 mm chips used in experiments 1-3 were of pure heartwood, it is likely that the conditions during experiments
Figure 16: Heating time (in seconds) required to go from 30◦C to 100◦C for 10 mm thick handmade wood chips of pure heartwood and pure sapwood during the 3 repetitions of the experiment (number 7-9).
Figure 17: Average heating time (in seconds) re- quired go from 30 ◦C to temperatures within the interval of 50 ◦C to 100 ◦C for 10 mm thick hand- made wood chips of pure heartwood and pure sap- wood. (Experiments 7-9).
7-9 were slightly different, in the sense that valve might have been slightly more closed. Although, it appears as if the difference in conditions were not that big or significant for the results in this case as well.
Figure 18: Graph from experiment 13, in which the effect of pressure on the heating time of chips with different thickness was studied. The pressure was allowed to reach a maximum value of 2.0 bar.
Figure 19: Average heating time (in seconds) re- quired to reach the temperatures within the interval of 50◦C to 120◦C for wood chips of different thick- ness when steamed at over-pressure.
4.4 The effect of higher pressure
Experiments 10, 12 and 13 were conducted to study the difference in thickness, but this time an over-pressure of steam was allowed to build up inside the digester. An example of how the curves can look when steaming with over-pressure is shown in fig. 18, while the rest can be found in the appendix. Here the experiments were conducted a similar manner as for experiment 1-3, but with more steam coming in which generated a system where the temperature curves started to plan out at slightly above 130◦C. The average heating curves are shown in fig. 19, calculated in the same way as for the previous cases. The main thing that can be noticed is that the difference in
heating time between chips of different thickness is smaller than when they were steamed at atmospheric pressure.
This is mostly due to the heating of the 5 mm chips reaching temperatures close to 100 ◦C much faster than the rest, thus reaching a point where the steam no longer condensates on them, but on other colder areas until all of the surrounding are closer to 100 ◦C. After that point, the heating rate seems to be similar for all of the chips, and thus, the difference in heating time will have been diminished. The overall time that was required to reach 100 ◦C was reduced for all of the chips, as can be expected. The reduction in heating time is most noticeable for thicker chips, as these aren’t slowed down by the need to have the surroundings in the right temperature before.
For 8 mm and 10 mm chips the reductions in time are, on average, 52 s and 62 s respectively, while for 5 mm chips it’s 27 s. For even thinner chips, which are common in industrial pulping processes, the difference might be even smaller.
Figure 20: Average heating time (in seconds) re- quired to reach the temperatures within the interval of 50◦C to 100◦C for 5 mm thick wood chips made from pine and birch. Results from experiments 14- 16.
Figure 21: Average heating time (in seconds) re- quired to reach the temperatures within the interval of 50◦C to 100◦C for 5 mm thick wood chips made from pine and birch. Results from experiments 19- 21.
4.5 Comparison between pine and birch
The results from experiments 14-16 and 19-24, comparing the pine chips and birch chips, are presented in the average temperature curves in fig. 20 - 22. There is a small difference in heating time between these two wood species, with birch chips heating up slower in general. When 5 mm thick chips were used, the birch chips had around a 5 s longer heating time throughout the interval, except for at the end, as can be seen in fig. 20. The heating time was roughly the same at 90◦C and at 100 ◦C the pine chips had around 40 s longer heating time.
Considering that the pine chips had a higher average moisture content, it should slow down their heating compared to the birch chips. However, birch wood has a higher basic density than pine wood [33], and there is most likely also a difference in heat capacity and thermal conductivity. It doesn’t appear to be many sources that have studied the difference in heat capacity between individual tree species, as the general consensus is that the heat capacity mainly depends on moisture content and temperature. However, there have been studies that indicate that there are differences in heat capacity between softwoods and hardwoods, and between tree species [34]. If birch wood does have a higher heat capacity it could explain the slower heating time in general, and the fact that the heating of pine slows down more when approaching 100◦C could be due to the presence of more water.
Since the difference between the curves in fig. 20 was quite small, it was decided to repeat the experiment to verify that the difference was not caused by any random factor. Furthermore, the heating time appeared to have been much slower compared to previous experiments involving pine chips. This was probably due to some valve being a bit more open during these experiments, compared to the previous ones, and it also served as an incentive to repeat them. The result from the three new repetitions are shown in 21, and it can be seen that the shape of the curves are quite similar, although the heating time is lowered with around 40 s. The main thing to notice is that the pine
chips still heat up slightly faster, until around 95 ◦C where the curves intersect. At 100 ◦C the pine chips have around 20 s longer heating time. Thus, the difficulty of recreating the exact same conditions in every experiments is noticeable, although it does not appear to affect the general trends that are observed.
For the 10 mm thick chips, the difference is a bit more noticeable, as can be seen in fig. 22. The birch chips heat up slower, with the difference being slightly above 10 s for the majority of the interval. The difference is biggest at 95◦C, where the birch chips require 20 s more on average to reach that temperature, but the difference diminishes to 13 s at 100◦C. From these results, it appears as if the difference between pine and birch wood chips is rather insignificant for most cases. Although, the fact that the difference between the thin birch and pine chip grew notably near 100◦C during experiments 14-16, shows that it likely that the steaming conditions has an effect on the relative heating times. Further investigation regarding the steaming of wood chips from different trees, could provide valuable information regarding if there are any wood chips that do take considerably more time to steam.
Figure 22: Average heating time (in seconds) re- quired to reach the temperatures within the interval of 50◦C to 100◦C for 5 mm thick chips made from pine and birch. Results from experiments 22-24.
Figure 23: Average heating time (in seconds) re- quired to reach the temperatures within the interval of 50◦C to 100◦C for birch wood chips of different thicknesses. Results from experiments 25-27.
4.6 The effect of thickness for birch
The results from experiments 25-27, where birch chips of different thicknesses were steamed, are shown in fig. 23.
Compared to the results for pine chips of different thicknesses, the curves are much more evenly distributed, and throughout the majority of the temperature interval the difference is around 30 s between each curve. This is close to what would be expected from a material with a constant thermal conductivity, which would give a linear increase in heating time with thickness. Considering that the fact that both the sapwood and heartwood birch chips had roughly the same moisture content, while the pine sapwood and heartwood chips had large differences, it minimizes the risk of factors other than the thickness affecting the heating time. It should be noted, however, that as the chips approached 100◦C, the results resembled more the results from experiments 1-3. The final heating times were very close to the ones from experiments 1-3, with experiments 25-27 showing 147 s for the 5 mm chips, 195 for the 8 mm chips and 209 s for the 10 mm chips. This fit well with the observation that there doesn’t appear to be any major difference between birch and pine. It should also be noted that, the difference between the thickest and thinnest birch chips was around 1 min on average, which is the same as for the thickest and thinnest pine chips.
4.7 Comparison between birch heartwood and sapwood
The results from the final experiments, numbered 28-33, are shown in fig. 24 and 25. The differences between birch sapwood and birch heartwood were quite small, around 10 s at most, both for the 5 mm chips and the 10 mm chips.
