Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
METHOD FOR PREPARING SAMPLES IN CONTROLLING A PROCESS OF TREATING LIGNOCELLULOSIC BIOMASS PARTICLES
Document Type and Number:
WIPO Patent Application WO/2022/084574
Kind Code:
A1
Abstract:
A method is presented for preparing samples for measuring a lignin content in a mass of pretreated lignocellulosic biomass particles. The pretreatment process comprises subjecting the lignocellulosic biomass particles to a hemihydrolysis reaction. The method comprises washing out soluble reaction products from a sample taken from said mass after said hemihydrolysis reaction, mechanically crushing the structure of particles in said sample, and using a measurement to determine a value indicative of a lignin content of said sample after said mechanical crushing.

Inventors:
TURUNEN SAMI (FI)
Application Number:
PCT/FI2020/050694
Publication Date:
April 28, 2022
Filing Date:
October 21, 2020
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
UPM KYMMENE CORP (FI)
International Classes:
C07H1/08; C08H8/00; D21C1/02
Domestic Patent References:
WO2001032715A12001-05-10
Foreign References:
EP2759597A12014-07-30
US20190309002A12019-10-10
Other References:
WANG GUANHUA ET AL: "Enhanced lignin extraction process from steam exploded corn stalk", SEPARATION AND PURIFICATION TECHNOLOGY, vol. 157, 2 December 2015 (2015-12-02), pages 93 - 101, XP029362136, ISSN: 1383-5866, DOI: 10.1016/J.SEPPUR.2015.11.036
MAHAJAN RISHI ET AL: "Effect of pretreatments on cellulosic composition and morphology of pine needle for possible utilization as substrate for anaerobic digestion", BIOMASS AND BIOENERGY, vol. 141, 22 August 2020 (2020-08-22), pages 105705, XP055813164, ISSN: 0961-9534, DOI: 10.1016/j.biombioe.2020.105705
JOSEFSSON T ET AL: "Steam explosion of aspen wood. Characterisation of reaction products", HOLZFORSCHUNG, vol. 56, no. 3, 2002, pages 289 - 297, XP009528046, ISSN: 0018-3830, [retrieved on 20020601], DOI: 10.1515/HF.2002.047
Attorney, Agent or Firm:
PAPULA OY (FI)
Download PDF:
Claims:
28

CLAIMS

1 . A method for preparing samples for measuring a lignin content in a mass of pretreated lignocellulosic biomass particles , wherein said pretreatment process comprises subj ecting the lignocellulosic biomass particles to a hemihydrolysis reaction, characterized in that the method comprises :

- washing out soluble reaction products from a sample taken from said mass after said hemihydrolysis reaction,

- mechanically crushing the structure of particles in said sample , and

- using a measurement to determine a value indicative of a lignin content of said sample after said mechanical crushing .

2 . A method according to claim 1 , wherein taking said sample comprises taking the sample from a process stream of said pretreatment process after said hemihydrolysis reaction but before any further intermediate storing of said process stream .

3 . A method according to claim 1 , wherein :

- the pretreatment process comprises separating solids from liquids in the reaction products of said hemihydrolysis reaction, and

- taking said sample comprises taking said sample from the stream that comprises said separated solids .

4 . A method according to any of claims 1 to 3 , wherein said washing out of soluble reaction products comprises mixing the sample with water .

5 . A method according to any of the preceding claims , wherein said washing out of soluble reaction products comprises exclusion washing that comprises vacuum filtering .

6 . A method according to any of claims 1 to 4 , wherein said washing out of soluble reaction products comprises exclusion washing that comprises overpressure filtering .

7 . A method according to any of the preceding claims , wherein said mechanical crushing comprises mixing said sample in a blender .

8 . A method according to any of the preceding claims , wherein said preparing of said sample comprises pressing the sample into a solid form after said mechanical crushing .

9 . A method according to any of the preceding claims , wherein said measurement comprises subj ecting the prepared sample to a Kappa number determining method .

10 . A method according to claim 9 , wherein said Kappa number determining method is a titration method .

11 . A method according to any of claims 1 to 8 , wherein said measurement comprises subj ecting the prepared sample to infrared spectroscopy .

12 . A method according to any of claims 1 to 8 , wherein said measurement comprises subj ecting the prepared sample to an optical measurement of particle surfaces .

Description:
METHOD FOR PREPARING SAMPLES IN CONTROLLING A PROCESS

OF TREATING LIGNOCELLULOSIC BIOMASS PARTICLES

FIELD OF THE INVENTION

The invention concerns the general technical field of converting biomass into chemical bioproducts in industrial scale . In particular the invention concerns the technology of preparing samples for use in measurements that facilitate controlling the process .

BACKGROUND OF THE INVENTION

The production of biomass-based chemicals may use for example wood particles or other lignocellulosic biomass particles as the main raw material . In a biomass-to-sugar process the lignocellulosic biomass particles may be subj ected to various kinds of pretreatment such as washing, impregnating with water and/or other l iquids , and heating, in order to prepare them for the later stages of the process .

A pretreatment process may involve soaking the wood particles in steam or hot water, then soaking them in dilute acid, and subsequently taking the acid- impregnated wood particles into a hemihydrolysis reactor where hemicellulose is at least partly hydrolysed and at least partly solubili zed . After hemihydrolysis reactions solid material is broken down in a steam explosion reaction . For the purposes of this description the steam explosion may be considered to form part of the hemihydrolysis reaction, because reaction products of the actual hemihydrolysis are only available for further processing and possible sampling after the steam explosion . The outcome , i . e . the reaction products of the hemihydrolysis reaction ( to which the steam explosion is also considered to be included) , contains solid particles containing cellulose and lig- nin, and a soluble hemicellulose fraction in a water solution ( so-called C5 sugars ) . Mechanical conveyors such as screw feeders transfer the impregnated wood particles between the stages of the pretreatment process .

The quality of the mass that comes out of pretreatment depends on a number of factors , such as the chemical conditions and the retention time in the various pretreatment stages for example . The nature and implementation of the pretreatment process should be such that it allows continuous running for extended periods , so that the use of process chemicals and other process parameters can be controlled continuously with safe and robust methods . I f any feedback-based control method is used to control the factors that affect the quality of the produced mass , these should be applicable to continuous operation with as little delay as possible between making the measurement and implementing the control measures , if needed, on the basis of the measurement results . The preparation and handling of samples in such a control method is of utmost importance if accurate and useful measurements are to be made .

SUMMARY

According to a first aspect there is provided a method for preparing samples for measuring a lignin content in a mass of pretreated lignocellulosic biomass particles , wherein said pretreatment process comprises subj ecting the lignocellulosic biomass particles to a hemihydrolysis reaction . The method comprises washing out soluble reaction products from a sample taken from said mass after said hemihydrolysis reaction, mechanically crushing the structure of particles in said sample , and us ing a measurement to determine a value indicative of a lignin content of said sample after said mechanical crushing . According to an embodiment , taking the sample comprises taking the sample from a process stream of said pretreatment process after the hemihydrolysis reaction but before any further intermediate storing of said process stream . This involves the advantage that delays in using the measurement to control the pretreatment process can be minimi zed .

According to an embodiment , the pretreatment process comprises separating solids from liquids in the reaction products of said hemihydrolysis reaction, and taking said sample comprises taking said sample from the stream that comprises said separated solids . This involves the advantage that at least part of soluble reaction products may already become removed in the separated liquids .

According to an embodiment , said washing out of soluble reaction products may comprise mixing the sample with water . Thi s involves the advantage that further reduction in the relative amount of soluble reaction products can be achieved in a relatively straightforward way .

According to an embodiment , said washing out of soluble reaction products comprises exclusion washing that comprises vacuum filtering or overpressure filtering . This involves the advantage that very effective removal of soluble reaction products can be achieved .

According to an embodiment , said mechanical crushing comprises mixing said sample in a blender . This involves the advantage that a relatively simple mechanical device can be used to achieve an important result .

According to an embodiment , said preparing of said sample comprises pres sing the sample into a solid form after said mechanical crushing . This involves the advantage that the sample is relatively easily handled in subsequent measurements . According to an embodiment , said measurement comprises subj ecting the prepared sample to a Kappa number determining method . This involves the advantage that methodology and terminology that are trusted and well known among pulp processing engineers can be used to express the measured lignin content .

According to an embodiment , said Kappa number determining method is a titration method . This involves the advantages that the Kappa number can be determined in a very systematic and reliable way .

According to an embodiment, said measurement comprises subj ecting the prepared sample to infrared spectroscopy . This involves the advantage of being able to utili ze the good penetration capability of infrared radiation in the measurement .

According to an embodiment , said measurement comprises subj ecting the prepared sample to an optical measurement of particle surfaces . This involves the advantage that more versatile characteri zation of the mass may be achieved than with a more limited set of measurement methods .

BRIEF DESCRIPTION OF THE DRAWINGS

The accompanying drawings , which are included to provide a further understanding of the described embodiments and constitute a part of this specification, illustrate various advantageous features and examples of their combinations . In the drawings :

Figure 1 illustrates a chemical refining process on a general level ,

Figure 2 illustrates an example of process stages in pretreatment ,

Figure 3 illustrates an example of a pretreatment process ,

Figure 4 illustrates a method, Figure 5 illustrates graphically a dependency between total lignin and the Kappa number, and

Figure 6 illustrates graphically a dependency between the extent of washing and the Kappa number .

DETAILED DESCRIPTION

Numerical attributes such as first , second, third, and so on are used in this description and the appended claims for the purpose of giving unambiguous names to concepts . They do not refer to any particular order, unless otherwise explicitly stated .

In the context of this description the term wood particles refers to a material that consists mainly of pieces of wood formed by cutting or chipping larger pieces of wood such as trees , branches , logging residues , stumps , roots , and wood waste . The si ze of the wood particles may vary in a wide range from a few millimetres to a few centimetres , so the wood particles meant here are typically larger than those meant with the term sawdust . The wood used to make wood particles may be debarked or it main contain bark . For a wood-to-sugar process the preferred raw material is hardwood or broadleaf wood due to its relatively high inherent sugar content , but the use of other kinds of wood is not excluded . The terms wood chips , wooden chips , or j ust chips can be used to mean the same thing as wood particles . The term chips i s used in the appended drawings because it is short .

Fig . 1 illustrates schematically how in a method and arrangement for treating wood particles the wood particles may go to pretreatment , schematically illustrated as 101 . The purpose of the pretreatment 101 is to prepare the incoming wood particles for efficient use in the process , by removing some unwanted impurities , by compensating for some of the natural fluctuations in the characteristics of the material , and by breaking down the natural structure of the wood material . Soluble hemicellulose (C5 sugars ) can be separated from the pretreatment 101 , and cellulose rich solids ( or lignocellulosic material ) can be taken further to a hydrolysis 102 to produce carbohydrates of desired kind .

Fig . 2 illustrates an example of a product flow through various stages that all may belong to the pretreatment 101 of fig . 1 . The incoming wood particles are washed in washing 201 , which is done with water, removing some mainly inorganic impurities such as sand . Washed wood particles are taken to steam treatment 202 for the purpose of removing air from inside the wood particles and to preheat them to an elevated temperature .

Steam-treated wood particles may be taken to dilute acid treatment 203 for impregnating them with a dilute acid solution . The aim of the dilute acid treatment 203 is to make the dilute acid solution penetrate into the wood particles as evenly as possible .

Acid impregnation is not necessarily needed, because similar effects can be achieved through so- called autohydrolysis , i . e . hydrolysis of water- impregnated wood particles , where the increased temperature makes the organic acids naturally contained or formed in wood take the role of any externally appl ied acid . By us ing acid impregnation it i s possible , however, to catalyse the desired reactions so that they take place at lower temperatures than what is needed for autohydrolysis , and with better efficiency .

The wood particles , which may be acid- impregnated, are taken to hemihydrolysis where they are under elevated pressure and temperature . At the output of the hemihydrolysis the wood particles undergo a steam explosion that breaks their structure . The hemihydrolysis and steam explosion stages are shown together as stage 204 in fig . 2 . The output stream from stage 204 goes through steam separation (not sep- arately shown) to mixing 205 where water is added and the resulting mass is homogeni zed mechanically to break up agglomerates . Solids and liquids may then be separated at 206 for feeding into later process stages .

An alternative example where autohydrolysis is used on beech may go as follows . The wood chips with moisture content of at least 30 % are heated to 205 degrees centigrade and kept there for a retention time of 10 minutes . The material is released from this high temperature as a steam explosion ( instant pressure release ) resulting in essentially solid material having a measured Kappa number 58 . 6 with a coefficient of variation 2 . 1 % in three parallel analyses . Compared to fig . 2 this would involve either omitting step 203 or using only water as impregnation liquid in step 203 , and relying on autohydrolysis in step 204 .

Fig . 3 illustrates some parts of an arrangement that can be used to implement the pretreating of wood particles that was described above in process form with reference to f ig . 2 . A pre-steaming s ilo 301 may be provided for treating the wood particles with steam, i . e . implementing at least part of the stage shown as stage 202 in fig . 2 . One or more conveyors 302 may be provided for transferring steam-treated wood particles from the pre-steaming silo 301 to one or more impregnating vessels 303 in which the dilute acid treatment stage 203 of fig . 2 is implemented, if such a stage is included in the pretreatment process . One or more of the conveyors 302 may be compressing conveyors , in which case there may be a side stream of pressate out of the conveyor ( s ) 302 .

Downstream from the one or more impregnating vessels 303 there may be provided one or more silos 304 for giving the dilute acid solution more time to penetrate into the inner parts of the wood particles , which may be described as evening out the dilute acid content of the wood particles . I f autohydrolysis is used instead of an acid-catalysed reaction, the silos 304 may be used to maintain the wood particles at an elevated temperature for a longer time . One or more further conveyors 305 may be used to transfer the wood particles into a reactor 306 . Also here it is possible but not mandatory to use compres sing conveyor ( s ) . In the reactor 306 the hemihydrolysis reaction takes place and at the output of the reactor 306 the steam explosion takes place . Depending on the exact implementation of the proces s there may be a side stream of liquid out of the reactor 306 , but this is not the case in all processes of this kind .

The following description and the appended claims use the singular form for all parts of the hardware for the sake of linguistic clarity and simplicity of expression . This should not be taken as a limitation, because in many cases an arrangement for implementing a process may comprise two or more pieces of some hardware unit for example to achieve redundancy in certain critical stages of the process , or for other reasons .

The performance of a biomass-to-sugar process can be measured, and consequently optimi zed, in various ways . One important factor in the performance of the whole process is the capability of the pretreatment process of the lignocellulosic biomass particles in producing pulp at a high and even quality . Here the severity of the hemihydrolysis reaction plays an important role . I f the reaction is too mild, the hydrolysis level of hemicellulose remains insufficient and the constitution of the resulting cellulose-containing output is not optimal , making it more difficult to efficiently produce the desired sugars later in the process . I f , on the other hand, the hemihydrolysis reaction becomes too severe , it begins to produce excessive amounts of unwanted by-products such as furfural , which may be disadvantageous to the further reactions downstream in the process , for example by having an inhibitory effect . A reaction that is too severe also begins to hydrolyse cellulose , which is not wanted because it decreases the glucose yield in subsequent process steps .

In hemihydrolysis the aim is to hydrolyze hemicellulose , where the main product molecule is xylose . At the same time the relative content of lignin in water-insoluble solids (WI S ) will increase . Soluble reaction products of the hemihydrolysis reaction comprise C5 carbohydrates like xylose , C6 carbohydrates like glucose , soluble lignin, furfural , and organic acids like acetic acid, formic acid, uronic acids and the like .

According to a novel insight , the method for controlling values of one or more process parameters of the pretreatment process comprises measuring a lignin content in at least one stream of reaction products after the hemihydrolysis reaction, and controlling the value of at least one process parameter of the pretreatment process on the basis of the measured lignin content .

Measuring the lignin content of an intermediate product of a wood refining process is known as such for example from the production of pulp for use in making paper . However, the l ignin levels encountered in a conventional pulping process are far below those that may be encountered in a process of the kind described here . This difference reflects the fundamental differences in what the processes aim at . A conventional pulping process aims at removing the lignin early in the process by cooking the wood chips in a digester, the result being a mass with only some few per cent of lignin . The Kappa number ( I SO 302 ) , which is used to indicate lignin content in conventional pulp, is between 1 and 70 and is typically in the low- er end of this range . To the contrary, the pretreatment process of lignocellulosic particles meant here aims at removing the hemicellulose , so that the l ignin remains in the mass and can constitute some 30 to 35 percent of its dry weight or even more , like up to 60 percent . The Kappa number measured from the output of the hemihydrolysis and steam explosion steps of the process may be 70 or even more .

As such, the exact value of the Kappa number is not important to the present discussion, because the Kappa number range to be operated in can be affected by e . g . selecting certain coefficients in the calculations . For the present di scussion it is more important to note that the numeric results of Kappa measurements may be used as an indicator of the lignin content also in the products of the pretreatment process of lignocellulosic particles meant here , even if - as mentioned above - the lignin contents are at completely different levels than in conventional pulp . As long as the measurement method is kept the same , variations in the determined Kappa number give a reasonably reliable indication of corresponding relative variations in the lignin content of the analysed reaction products .

Measuring the lignin content in a stream of reaction products after the hemihydrolysis reaction but before any enzymatic hydrolysis involves the significant advantage that the measurement gives a relatively reliable indicator of the current severity of the hemihydrolysis reaction . The basic trend is that the higher the measured lignin content , the more severe the reaction, and vice versa . The severity of the reaction may depend on a variety of factors . As the pretreatment process may comprise impregnating the lignocellulosic biomass particles in an acidic solution before subj ecting the acid-impregnated lignocellulosic biomass particles to the hemihydrolysis reac- tion, a number of affecting factors may relate to the impregnation step : factors such as retention time of the lignocellulosic biomass particles in the impregnation in acidic solution, as wel l as acid concentration of said acidic solution . Another plurality of affecting factors relate to the hemihydrolysis reaction itself : factors such as retention time of the acid- impregnated lignocellulosic biomass particles in the hemihydrolysis reaction, pressure during the hemihydrolysis reaction, and temperature during the hemihydrolysis reaction . Changes in temperature and pressure typically go hand in hand, when saturated steam is used .

As indicated at the bottom of fig . 3 , the pretreatment process may comprise separating steam at step 307 and separating solids from liquids in the reaction products of the hemihydrolysis reaction at step 308 . The step of measuring lignin content may comprise measuring the lignin content in the stream 309 that comprises the separated solids . This involves the advantage that the measurement becomes more reliable , because it is particularly the lignin content of the water-insoluble solids that has been found indicative of the severity of the hemihydrolysis reaction . Soluble reaction products could interfere with the measurement or at least weaken its accuracy .

Measuring the lignin content in stream 309 may comprise taking a sample of a stream that comprises the separated solids , preparing the sample for the measurement of lignin content , and performing the measuring on the prepared sample . This involves the advantage that the measurement can be made even more accurate , because the preparing of the sample for the measurement can be tailored appropriately .

However, between steps 307 and 308 there may be an additional step where water is added to the reaction products and the resulting mixture is pumped into a storage tank where it may spend some time , like 45 minutes , an hour, or even more . Such intermediate storing for an additional retention time in a storage tank is useful in ensuring that soluble compounds in the reaction products dissolve in water . Additionally or alternatively, the nature of the process may be such that it is advantageous to have some buffering at this stage .

Here it must be remembered that the aim of measuring lignin content is to use the measurement results as a kind of feedback information to control the values of process parameters earlier in the process . Taking the sample only after step 308 means that the retention time in the storage tank between steps 307 and 308 causes considerable delay in controlling the process . For this reason it may be advantageous to make the measurement of lignin content earlier, like before step 307 or between steps 307 and 308 , or in particular without waiting for said additional delay .

Taking a sample for measurement before steam separation may be challenging, because the reaction products exit the reactor 306 at considerable speed, but it is not excluded either . Immediately after steam separation the reaction products have the appearance of relatively dry floc, in which form they are relatively difficult to move forward in the process by pumping for example . The materials involved have a difference to pulps encountered in paper-making : conventional pulp with a 30 percent solids content is already quite dry in appearance and feel , while in the process described here a solids content of 30 % means that one can still squeeze water out of the mass by hand .

A sample may be taken immediately after the steam separation at step 307 , but as there is typically the process step of adding water for making the reaction products easier to pump, the sample may be tak- en also after such a step . Such a case may be characteri zed as taking a sample of a stream of reaction products of said hemihydrolysis reaction that comprises both solids and liquids , preparing the sample for the measurement of lignin content , and performing the measuring of lignin content on the prepared sample .

In addition to measuring the lignin content from a separate sample , it is also possible to measure the lignin content directly from a production flow in a pipeline .

As an example , preparing the sample may comprise one or more preparation steps such as dispersing, homogeni zing, crushing, washing the sample , dewatering, pressure filtration, vacuum filtration, concentrating solids through whatever means , and pressing the washed sample into a solid form . This involves the advantage that measurement devices that accept lignincontaining samples in such concentrated or solid form may be utili zed . Such measurement devices have been known from the technology of conventional pulping, although - as indicated above - such known measurement devices were typically optimi zed to measure much lower lignin contents .

Preparing the sample has certain importance in ensuring that the lignin content of the mass after hemihydrolysis can be determined accurately . The pulp, also frequently referred to as the hemihydrolyzed pulp, has been found to exhibit many characteristics that make it difficult to get reliable results simply by applying e . g . standardi zed Kappa number determination in the form used in more conventional pulp processing industry . The relative amount of soluble reaction products may be high, in the order of 20 to 35 % w/w . Also the lignin content may be much higher ( like 30 to 45 % w/w in solid fraction ) than in the masses of conventional pulp processing industry . There is a relative high amount of soluble lignin, which may ( in favourable conditions ) precipitate or solidify on the surfaces of the mass particles . The physical si ze of particles in the mass may vary considerably, from very coarse to very fine , which is an important finding because the physical si ze of particles has an effect on e . g . Kappa measurement through characteristic surface area . Chemical reactions in a measurement may remain incomplete at the larger particles because of their large physical si ze .

A novel insight provides for an improved sample preparation process that comprises washing out soluble reaction products from a sample taken from the mass after the hemihydrolysis reaction, and mechanically crushing the structure of particles in said sample before using a measurement to determine a value indicative of a lignin content of the sample .

Washing out soluble reaction products ensures that their interfering effect mentioned above can be avoided or at least minimi zed . It may comprise substeps such as mixing the sample with water, and/or using exclusion washing that in turn may comprise e . g . vacuum filtering or overpressure filtering .

The order of washing and crushing is not important , and they can be accomplished in any order . Filtration may be faster if washing is done before crushing .

The mechanical crushing results in significant reduction in the variation of physical si ze of particles in the mass . It can be accomplished e . g . by mixing the sample in a blender, or by other suitable mechanical method .

The measurement may comprise e . g . subj ecting the prepared sample to e . g . a Kappa number determining method, to infrared spectroscopy, to near infrared spectroscopy, and/or to an optical measurement of particle surfaces . Of these exemplary methods , e . g . infrared spectroscopy and near infrared spectroscopy in- volves the advantage that the relatively good penetration of the radiation into the sample is achieved, and the requirements for sample preparation may remain more modest than in some other measurement methods .

Alternatively or additionally, measuring the lignin content may comprise subj ecting the prepared sample to spectroscopic measurement methods based on ultraviolet radiation .

Alternatively or additionally, the measuring of lignin content may comprise subj ecting the prepared sample to Klason lignin determination, gravimetric lignin determination, and/or titration of the prepared sample . Each measurement method has its own advantages and drawbacks . For example , while gravimetric measurement gives relatively accurate results , it is typically a slow and laborious method . On the other hand, a gravimetric measurement can be used for lignin content measurements even for masses that have undergone too severe a reaction, while Kappa measurements with e . g . near infrared might be confused by the large fraction of breakdown products of lignin in the mass .

As such, the Kappa number i s a well-known and relatively unambiguous way of expressing the measured lignin content . A measured lignin content in the form of a Kappa number may result in e . g . controlling of the value of at least one process parameter of the pretreatment process on the basis of the magnitude of the Kappa number . An optimal Kappa number to be aimed at , and/or other suitable ways in which the magnitude of the Kappa number is taken into account in controlling the process , may be found by experimenting, for example by producing a calibration curve with measured total lignin ( including both acid insoluble and acid soluble lignin) .

Currently it is believed that the optimal value of the Kappa number for a pulp after hemihydrolysis reaction and steam explosion should be above 50 . It should be noted, however, that the sample preparation protocol as well as method details such as chemicals used or dosing of the sample may affect the general way in which a measured Kappa number correlates with the actual lignin content . As a consequence , if the Kappa number is used to represent the lignin content , the optimal value or optimal value range for the Kappa number should be selected on a case by case basis through experimenting, where case by case means taking into account the exact nature and details of the selected method of measuring the Kappa number .

Fig . 4 illustrates an advantageous principle for use in the controlling of values of one or more process parameters of the pretreatment process . In one way or another, the measurement indicates the lignin content in at least one stream of reaction products after the hemihydrolysis reaction . Here in particular the lignin content in solids is considered, as illustrated by block 401 in fig . 4 . There is assumed to be a predetermined lignin content range that corresponds to optimal severity of the hemihydrolysis reaction ; the limits of such a predetermined lignin content range may be found by experimenting . The correlation between a lignin content range that corresponds to optimal severity of the hemihydrolysis reaction may depend on the raw material , like tree species , that is fed into the pretreatment process .

I f the measured lignin content is towards or over a higher end of the predetermined lignin content range , it is determined that the hemihydrolysis reaction may be too severe as illustrated by block 402 in fig . 4 . In response to such a finding, the method may comprise changing the value of at least one of the process parameters in order to decrease the severity of the hemihydrolysis reaction . Fig . 4 shows how changing the value of a process parameter may comprise decreasing for example a retention time 411 , a pres- sure 412 , a temperature 413 , and/or acid content 414 . Of these , the retention time 411 may be the retention time of the acid-impregnated lignocellulosic biomass particles in the hemihydrolysis reaction and/or the retention time of the lignocellulosic biomass particles in the impregnation in acidic solution before the hemihydrolysis reaction . The pressure 412 may be the pressure during the hemihydrolysis reaction . The temperature 413 may be the temperature during the hemihydrolysi s reaction, and the acid content 414 may be the acid concentration of the acidic solution used for the impregnating .

I f the measured lignin content is towards or below a lower end of the predetermined lignin content range , it is determined that the hemihydrolysis reaction may be too mild as illustrated by block 403 in fig . 4 . In response to such a finding, the method may comprise changing the value of at least one of the process parameters in order to increase the severity of the hemihydrolysis reaction . According to fig . 4 , that may mean increasing the value of at least one of the process parameters 411 to 414 described above .

Like in many other multi-variable control systems , the rules according to which the values of the process parameters are changed may involve a varying degree of complexity . Some of the process parameters may have mutual dependencies : for example , the temperature and pressure in the hemihydrolysis reaction often need to be handled together because a change in one causes a change in the other . The temperature and pressure also have certain windows of allowable values , so that the control algorithm is not allowed to change their values out of the respective window . The process parameters may have some priority order in which their values are controlled . For example , one of the retention times ( retention time in impregnating, or retention time in hemihydrolysis ) may be a preferred process parameter so that changes in its value are made first and only if that is not enough ( and/or if the retention time is already at an allowable limit ) , changes to other process parameter values are considered . Additionally or alternatively, if simultaneous changes in the values of selected process parameters are considered, their order of preference may be taken into account so that a more significant change is made to the value of a more preferred process parameter while a less significant change is made to the value of a less preferred process parameter .

Concerning at least some of the process parameters there may be also other factors than j ust the measured lignin content that must be taken into account in deciding about changes to their values . As an example , there may be minimum and/or maximum limits to the acid concentration in the reactor ( adj usted by acid concentration in impregnating) . Indications of where the actual acid concentration is in relation to such limits may be obtained by measuring the relation of monomers to total sugars in the soluble liquid phase of reaction products . I f the acid concentration is too low, also said relation is low ( in other words , the relative concentration of oligomers in the liquid phase of reaction products is too high) . I f such a measurement indicates that the acid concentration in impregnating is already low, it may be advisable not to lower it further even if the basic control principle shown in fig . 4 would suggest decreasing the severity of the hemihydrolysis reaction through decreasing acid content 414 .

A simple control algorithm may have a target value or target value range of the measured lignin content , so that changes in process parameter values are constantly made in order to drive the measured lignin content towards the target value or target val- ue range . A more elaborate control algorithm may operate according to a higher level of intelligence , for example by taking into account al so the recent history of measured lignin content . For example , if there has been a period of somewhat low lignin content the process could be deliberately driven into a state in which the measured lignin content is higher than normal for a certain duration of time , before eventually returning to a state in which the measured lignin content is at the normal value .

The control system may be based on a selflearning artificial intelligence unit such as a neural network or a fuz zy logic unit . In that case the history of previously obtained measurement results of lignin content , the measures taken on the basis thereof , and the observed results of those measures may affect the way in which the control system will handle a subsequent measurement result . A self-learning artificial intelligence unit of this kind may eventually optimi ze the strategy of controlling the process parameter values so that every encountered situation may be responded to in the best possible way .

In order to ensure best response to any encountered measurement of lignin content it would be advantageous to minimi ze all delays that may occur in the control loop that is schematically illustrated in fig . 4 . In particular, the delay from obtaining a sample from the stream of reaction products to the moment when its lignin content has been measured should be made as short as possible . The length of this delay depends heavily on the selected measurement method . An advantageous embodiment could involve using a so- called online Kappa measurement device to continuously measure the value of a Kappa number from the stream of reaction products .

As a practical example , we may consider a pretreatment process of lignocellulosic biomass parti- cles in which hemicellulose is hydrolysed and dissolved, and thus the lignin content of the sol id material in the output stream i s increased . As a first assumption we may assume that a first sample is taken after the pretreatment process , and in this case more than 90 per cent of the original xylose content of the lignocellulosic biomass particles had been removed . In order to prepare the pulp sample for measurement it i s dispersed with water, and all water-soluble components are washed away . The lignin content is then determined, and the value of at least one proces s parameter of the pretreatment process is controlled on the basis of the measured lignin content . I f , for example , the measured lignin content indicated that actually more than 95 per cent of the original xylose content had been removed, this could be interpreted as a sign that there is a risk of the reaction being too severe already . As a consequence the retention time of the acid-impregnated lignocellulosic biomass particles in the hemihydrolysis reaction and/or the retention time of the lignocellulosic biomass particles in the impregnation in acidic solution could be decreased by some minutes , like by two minutes for example .

Example 1 :

The following is a practical example of preparing a sample and performing a Kappa measurement with a titration method based on the standard I SO 302 : 2004 but particularly adj usted for hemihydrolysed pulp . In this example seven samples of the hemihydrolyzed pulp were taken . First , a sample taken from the pulp coming out of the hemihydrolysis reactor was dispersed with water and soluble compounds were washed away . Washing of this kind can be accomplished for example by filtrating, reslurring the solid, and filtrating again ; thus cycle being possibly repeated for a desired number of times . The dry matter content of the washed pulp was measured by drying in an oven of temperature 60 ° C overnight , resulting in a dry matter content of 33 . 5 % w/w . Pure water was added and the mixture was strongly dispersed and crushed with a Bamix hand blender for 30 seconds at full intensity to have good dispersion and to cut down a part of the solid particles . I f needed, the dispersed sample may be further mixed with more water, like with 360 ml of pure water, again using a suitable tool of which the use a Bamix hand blender for 30 seconds at highest agitator power is an example .

The disintegrated pulp was allowed to react with KMnCg for 10 minutes , after which the reaction was stopped with potassium iodide . Free iodide content was titrated with sodium thiosulphate solution, wherein a 20 millilitres burette for sodium thiosulphate was used . Measured Kappa values , as well as their average , standard deviation, and coefficient of variation are presented in the following table .

This example shows that the modified Kappa method can be used in analysis of hemihydrolyzed pulp with very low variation between parallel measurements . When the relationship between this Kappa-number and lignin content is known or the relationship between the Kappa-number and pulp quality is known, the pro- cess can be adj usted based on the results of thi s relatively simple analysis .

It was noticed that if thi s method gives a relatively high Kappa value of , say, 82 the retention time in the reactor may be s lightly decreased, for example by one minute .

Example 2 :

As a further practical example , the lignin content of the stream of reaction products after the hemihydrolysis reaction was measured by determining chlorine consumption in accordance with I SO 3260 : 1982 . Samples were taken of said reaction products , washed, screened, and dried in an oven . Before weighing the samples were conditioned for ( at least ) 20 minutes in the atmosphere near the balance . Samples of 0 . 5 g ±0 . 05 g were weighed for measurement , and the dry matter content was measured according to I SO 638 . After adding 250 millilitres of distilled water at temperature of 25 ° C to each weighed sample , disintegration of bigger particles was performed until the sample was free from bigger fibre cloths and large bundles . The mixed sample portion was transferred into a reaction flask and 135 millilitres of water was added to rinse the disintegrator . The flask was placed on a magnetic stirrer and water bath to maintain an accurate temperature .

A separating funnel was connected, and the flask was evacuated with a vacuum pump . The stop-cock of the funnel was closed, and 10 millilitres of hydrochloric acid was added into the solution in the funnel . The acid was sucked down without admitting in air, and a stopwatch was started . The funnel was rinsed with 10 millilitres of di stilled water, which was sucked down . After this , 15 millilitres of sodium hypochlorite solution was pipetted and sucked down within 2 minutes . The funnel was rinsed with 5 milli- litres of distilled water. The stopwatch was not stopped at this point.

Then 20 millilitres of potassium iodide was added into the solution and sucked down within 17 minutes after adding the hydrochloric acid. The funnel was again rinsed with 50 millilitres of distilled water, which was sucked down. The flask was shaken to dissolve gaseous chlorine. Again, 50 millilitres of distilled water was added into the funnel and sucked down. Now the stop-cock was left open, and the funnel was removed. The solution was titrated with sodium ti- osulphate using starch as indicator. The consumption was preserved as VI (millilitres) and a blank test was performed with consumption V2 (millilitres) .

The fraction r of added chlorine not consumed was calculated in the determination as the ratio V1/V2, where VI is the volume (in millilitres) of standard volumetric sodium thiosulphate solution consumed during titration with test sample, and V2 is the volume (in millilitres) of standard volumetric sodium thiosulphate solution consumed during blank titration test. If the calculated value of r was less than 0.5, the determination was repeated with a smaller sample portion. If the calculated value of r was bigger than 0.5, a correction factor f was obtained from the following table.

The chlorine consumption X can be calculated with the following equation and expressed as percentage by mass.

X = [3.545*f* (V2-V1) *c] /m where c is the concentration in mol/1 of standard volumetric sodium thiosulphate solution, and m is the mass in g of the test sample calculated on an oven-dry basis.

In the examples above, when analysing samples it is advantageous to ensure that all agglomerates are properly broken down and particle surfaces are exposed. A high rotating mixer can be used for such ensuring, if needed. It was found that intensive mixing, like with the Bamix hand mixer in the examples above, breaks down agglomerates like fibre bundles, lignin aggregations, and fibre-lignin aggregations. It also reduces fibre size and "opens" the particle surfaces so that the chemicals may affect the particles more effectively, which is highly advantageous for the success of the method.

With samples with relatively low chlorine consumption, a diluted solution of NaClO can be used. The blank test is carried out with the same solution and titration with lower concentrated solution thiosulphate. The correction factor f can be calculated with the following equation, which has been derived on the basis of certain assumptions that are accepted in the theory of pulp chlorination, but proven valid experimentally only if r is larger than 0.5: f = (1/2) * [1 + V2/ (V2-V1) + ln(V2/Vl) ] .

It was noticed, that the chlorine consumption of reaction products after the hemihydrolysis reaction varied from 7% to 20%, depending on the lignin content. A chlorine consumption of 11,4% corresponds to a lignin content of 36%.

Example 3 :

In a further example, the lignin content of hemihydrolyzed pulp was determined by titration with standard Kappa analysis (ISO 302:2004) which was adjusted for hemihydrolysed pulp. Five different pulp samples were taken after the pretreatment process, where xylan has been hydrolysed and removed. These samples were used for determining the correspondence of Kappa-results to total lignin results (acid insoluble and acid soluble lignin) determined by gravimetric and ultraviolet measurements. The pulp samples were prepared as explained further below.

Seven parallel Kappa-measurements and two parallel total lignin measurements were done for each sample. The samples were prepared according to the procedure explained further below. The prepared sample was let to react with KMnCg for 10 min and the reaction was stopped with potassium iodide. Free iodide content was titrated with sodium thiosulphite solution. The burette of sodium thiosulphate was selected to be suitable for the liquid consumption, having the exemplary volume of 20 ml. Also other parameters like minimum and maximum dosing volume, dosing rate and frequency as well as filling rate were adjusted based on need.

An average of each set of the seven Kappa measurements were calculated. The five samples gave Kappa numbers 55, 59, 49, 58, and 64, and their total lignin contents (the two parallel measurements averaged) based on gravimetric (acid insoluble lignin) and ultraviolet spectrophotometer (acid soluble lignin) measurements were found to be 35.6%, 36.9%, 33.0%, 35.9%, and 39.7% respectively. Based on the results it was found that the total lignin content could be calculated based on the formula

Total lignin = 0.433 * Kappa + 11.6

Fig. 5 illustrates graphically the dependency between total lignin and the Kappa number according to this formula. This formula and the results can be used to define a set-point point value for the Kappa number . I f a target lignin content is 35% , the set-point of Kappa-value is 54 . When the process sample is taken, the Kappa number may be measured to be for example 58 , indicating a lignin content of 36 . 7 % . The process may then be adj usted so that first the residence time in the reactor is decreased by 4 min . I f a low limit of residence time is reached, then additionally the acid concentration i s decreased 5 g/ 1 for reaching the setpoint Kappa-value 54 .

This example shows that modified kappa method with certain sample preparation method can be used in analysis of hemihydrolyzed pulp . When the relationship between kappa-number and lignin content is known the process can be adj usted based on the results of this relatively simple analysis .

In preparing the samples in the example above , the pulp samples were dispersed with water and soluble compounds were washed away . Dry matter content of washed sample was measured in oven 60C overnight resulting in dry matter 30 . 6 - 33 . 6 % w/w . The seven parallel Kappa-measurements were done so that 70 ml of pure water was added on wet pulp sample , 0 . 59 - 0 . 65 g as such, and these were strongly dispersed, and crushed with a Bamix hand blender ( 30 s with full intensity ) to have good dispersion and to cut down a part of solid particles .

The total lignin content determination based on gravimetric and ultraviolet spectrophotometric measurements were done by hydrolysing the samples with H2SO4 , filtering off the remaining solid lignin, drying it in an oven and determining the amount of solid lignin by weighing the lignin . The soluble lignin in the solution after filtering was measured with UV spectroscopy . Total lignin is calculated by adding up the results of the gravimetric and UV measurements . Fig. 6 illustrates the effect of washing when the sample is prepared. It represents a series of measurements where four samples, identical in solid lignin content, were taken from a mass that was initially washed completely clean of soluble reaction products. Controlled amounts of soluble reaction products were then added to each sample. As a result, the samples were representative of mass washed to different extent so that their content of soluble reaction products compared to solid were 0%, 2%, 30%, and 100%. Their respective Kappa numbers were measured, the results being 67, 69, 77, and 88 respectively. The linear regression line in fig. 6 gives an approximation of how the resulting Kappa number depends on the extent of washing in this method of measuring the lignin content of a sample.

It is obvious to a person skilled in the art that with the advancement of technology, the ideas explained above may be implemented in various ways. The claimed scope is thus not limited to the examples described above.