N 2. N pyrolysis. NH 3 N black liquor NO N char N smelt. N green liquor 4.6-1

Transcription

1 THE FATE OF NITROGEN IN THE CHEMICAL RECOVERY PROCESS IN A KRAFT PULP MILL: Part 1: A general view M.Kymäläinen, M.Forssén and M.Hupa Åbo Akademi University, Turku, Finland Published in Journal of Pulp and Paper Science 25 (12): , 1999 ABSTRACT About a third of the black liquor nitrogen remains in the smelt in a recovery boiler and continues in the rest of the recovery cycle. The smelt nitrogen was found to be in a form which slowly converted to ammonia under the process conditions: the green liquor entering the slaker contained clearly more ammonia than the green liquor leaving the dissolver, about 7 % and 2 % of their total nitrogen, respectively. The nitrogen of weak wash entering the smelt dissolver was mostly in the form of, and at least part of it was released in the vent stack gases of the dissolver. The green liquor nitrogen, originating both in smelt and weak wash, corresponded to almost 4 % of the black liquor nitrogen. The major part, about 6 %, of the green liquor nitrogen entered the cooking process with the white liquor. A minor part, less than 2 %, continued to the weak wash, thus returning to the smelt dissolver. The rest, about 2 % of the green liquor nitrogen, was assumed to exit the process as gaseous around the slaker and the causticizers. The nitrogen in the white liquor was mostly. It was released mainly during the black liquor evaporation, and to a minor extent during the cooking. This was mostly dissolved into the foul and secondary condensates. The stripping of the foul condensates was the main exit point for ammonia in the recovery process. The amount of this ammonia represented about two thirds of a typical NO x emission of a recovery boiler. NTRODUCTION Reactive nitrogen enters the pulping process and the chemical recovery system of a kraft pulp mill mainly with the wood chips. During digesting the wood nitrogen has been found to transfer primarily to black liquor [1]. The nitrogen content of black liquor is typically % by weight, and it varies somewhat depending on wood species, being slightly higher in birch than in pine liquors [2]. There is hardly any quantitative data on the form of nitrogen in black liquor and the chemical behaviour of it is therefore not fully understood. The nitrogen has been found to occur mostly in organic compounds, both in straight chains and in heterocyclic ring structures, presumably mostly bound in lignin. Nitrogen-containing organic compounds, like several amino acids, pyrroles, pyridines, pyrimidines, and indoles have been found in black liquors [1,3,4]. In addition, Martin et al. [4] have reported a significant amount of nitrogen (up to one third of the total black liquor nitrogen), to be in the inorganic form, mostly as nitrate, in a southern pine black liquor. On the other hand, Martin and Malcolm [5] have found later, based on an analysis of lignin that the liquor nitrogen is mainly associated with the lignin, further suggesting little to no nitrogen to be associated with the inorganics. In these studies on the liquor nitrogen composition it has not been clearly mentioned whether a weak or a strong black liquor has been analyzed or, further, if the liquors analyzed were oxidized or not. The behavior of the liquor nitrogen in the recovery boiler has been extensively studied during recent years, by laboratory tests and NO x measurements in the boiler flue gases, as well as by kinetic modelling. In previous experimental work at Åbo Akademi University (ÅAU) by Aho et al. [2] and by Forssén et al. [6], the fate of nitrogen during black liquor burning has been clarified and the key routes for nitrogen in black liquor combustion have been suggested as shown in Figure 1 [6]. These studies showed that about one third of the liquor nitrogen will be bound in the char residue and may remain in the salt residue leaving the furnace along with the smelt. The present work is a continuation of the research at ÅAU on the behavior of black liquor nitrogen. In this work the main focus was to elucidate the fate of the smelt nitrogen further on in the recovery process. N pyrolysis N 2 N black liquor NO N char N smelt NO N 2 NO N green liquor Figure 1. Suggested fuel nitrogen pathways in black liquor combustion [6]. Prior to the present work, little information has been published concerning the nitrogen elsewhere in the recovery cycle than in black liquor burning. In general, the odour of has been noticed in the causticizing plant in many mills, but little work has been done to explain these emissions. Some recent studies by 4.6-1

2 Tarpey et al. [7,8] have focused on the cause of the ammonia release, especially in particulates, from a smelt dissolving tank. Tarpey [7] has also published nitrogen and ammonia analysis results of several liquor samples from the recausticizing plant, but no explanation for these results nor any mass balance for nitrogen has been given. On the other hand, Thompson et al. [9] have reported, based on a mass balance of a recovery furnace, that only about 9 % of the black liquor nitrogen entering the boiler leaves with the smelt. This amount of smelt nitrogen was found to be far too low to explain the nitrogen content of the green liquor leaving the smelt dissolver, and the reason for this large difference was unknown. The purpose of the present work was to clarify the fate of the smelt nitrogen in green liquor and in the rest of the recovery cycle. It was of particular interest to find out in which forms nitrogen is found in the process and in which forms it finally leaves the recovery cycle. The potential main exit points for nitrogen in the process were also of interest. This paper is the first part in a series of papers concerning the fate of nitrogen in the chemical recovery process. A general view of the nitrogen flows in one kraft pulp mill is given in this paper, and in the second part the effect of some process variables on the fate of nitrogen in another pulp mill is described. EXPERIMENTAL Sampling The measurements were conducted at a Finnish kraft pulp mill producing 2 tons pulp/day. The mill utilizes a batch digesting process, and both hardwood and softwood kraft pulps are produced. A simplified general flow sheet of the recovery process, excluding lime mud dewatering and calcining, is shown in Figure 2. The recovery cycle consists of a one-line evaporation and a causticizing plant, a recovery boiler with a design capacity of 28 tds/d. Noncondensible Gases (NCG) from the recovery cycle are collected and burned: high concentration, low volume (HCLV) gases from digesting and evaporation, including stripper gas as well as methanol are burned in a separate. Low concentration, high volume (LCHV) gases from the evaporation plant are also conducted there, whereas LCHV gases from the causticizing plant are burned in the lime kiln. The main sampling points (1-16) are shown in Figure 2, and the corresponding process streams are specified in Table 3 in the Results section. Liquid samples, like black, green and white liquors, and condensates, were collected into plastic containers (.5-1 liter), which were filled to the top and sealed immediately after sampling. Smelt samples from the recovery boiler smelt spouts were taken in two different ways: either with a steel mug or with a quartz glass tube. When sampling with a mug (about 1 cm 3 ), the sample was wrapped in aluminum foil and kept in a plastic bag. The sample for the analysis was taken from the middle of the original sample piece. Contact of the smelt samples with air was minimized, and an unoxidized part of the sample piece was used for the analysis. Most of the samples were analyzed immediately at the mill. Some of the samples were also stored for a certain period of time at room temperature and analyzed in order to compare the results with the results obtained from the samples analyzed promptly. The sample containers were kept closed during storage. The time elapsed between sampling and analysis varied from a few hours to a few months. As shown in Table 1, the Kjeldahl nitrogen content of the liquors was found to remain constant during storage, within the accuracy limits of the method of analysis. Table 1. Storage of two liquor samples, and their Kjeldahl nitrogen content. Storage time Kjeldahl before analysis nitrogen mg/l A few hours 18 Green Liquor 3 months 19 A few hours 84 White Liquor 6 weeks 82 The purpose of the sampling campaign during 1997 was to study possible variations of the nitrogen level in the recovery cycle during different sampling periods (Table 2). Here, the effects of varying process conditions were not studied in detail, but that has been studied during one sampling campaign in 1998 and will be reported in our second paper of this work. Table 2. Sampling periods. Period # Month in '97 Boiler load tds/d Hardwood % in cooking 1 Jan Feb. 255 *) 6 3 March Sep Oct *) in the beginning of the period, 212 tds/d Analyses Some details and comparisons of different analysis methods for total nitrogen are provided in our earlier 4.6-2

3 papers [6,1]. In the present work, the Kjeldahl method was used because the appropriate analysis facilities were available at the mill, thus making it possible to analyze the samples directly after the sampling. Initial tests showed that the Kjeldahl method gave analysis results comparable with the other methods [1]. The reproducibility of the Kjeldahl method determined by duplicate analyses for the studied samples was found to be good, within some 4 %. The ammonia ( ) content was determined by titration, where the is first liberated from alkalized samples by steam distillation and then absorbed into an acid solution followed by titration for the. The dry solids content of black liquor was determined by weighing the sample before and after drying with sand at 15 C for 18 to 24 hours (SCAN-N 22:77). RESULTS AND DISCUSSION Kjeldahl Nitrogen and Ammonia Results Kjeldahl nitrogen results of the samples (Figure 2) taken during one sampling period are given in Table 3, to provide data for mass balance calculations. In addition, ammonia results and the ratio of ammonia nitrogen to Kjeldahl nitrogen are given. The results are also shown in Figure 3. Ammonia Formation in Green Liquor Hardly any ammonia nitrogen was found in the smelt samples (Figure 3). The nitrogen of weak wash coming to the smelt dissolver was mostly in the form of ammonia. The green liquor entering the slaker contained clearly more ammonia than the green liquor leaving the dissolver, about 7 % and 2 % of the Kjeldahl nitrogen, respectively. Based on these data, it seems evident that the nitrogen coming with the smelt to the dissolver and continuing further with the green liquor to the causticizing converts to ammonia in the green liquor. Conversion takes place mainly during green liquor processing, in surge tanks and in clarifiers. Hardly any conversion takes place in the dissolver. A few laboratory tests, briefly described below, also support these findings. Strong black liquor was practically free from ammonia. Weak black liquor, instead, contained a significant amount of ammonia. This indicates that the ammonia nitrogen of weak black liquor volatilizes during the evaporation process. The nitrogen of the foul condensate, a combined stream from evaporators and batch digesters to the stripper, was solely in the form of ammonia. This ammonia was efficiently removed through steam stripping of the condensate, resulting in a treated condensate free of nitrogen. The stripper gases, including ammonia, are taken to the methanol condenser. Analysis results of weak and strong black liquor, smelt, and green liquor samples taken during different sampling periods are given in Figure 4. Every result here is an average value of two or more parallel determinations. Minor variation in the sample nitrogen content from one sampling period to another can be seen. A conversion of the nitrogen in green liquor into the form of ammonia is evident (Figure 3). So far, the origin of this ammonia has been unknown. In this work, laboratory measurements were done to clarify the formation of in green liquor. First, the temperature dependence of the formation in mill green liquor was studied. One result of those tests is presented below. Secondly, the key reactions responsible for ammonia formation were studied by studying the formation with well controlled additions of certain nitrogen compounds to both mill and synthetic green liquor. Those results and the description of the batch reactor system used in the experiments will be published in near future. A mill green liquor (the sample taken right after the dissolver) was kept at a certain temperature in a laboratory reactor and the concentration of the liquor was followed as a function of time (Figure 5a). The formation rate of in the studied mill green liquor was found to be fairly slow: a residence time of several hours at the process conditions is required for this green liquor to increase its ammonia content to correspond to the content of the green liquor entering the slaker, i.e. around 7 mgn/l. Furthermore, as shown in Figure 5b, the formation reaction is suggested to be a first order reaction with respect to the -forming compound, based on the following rate data analysis and the following assumptions: - Compound A is the only -forming compound in the liquor, and the initial concentration of A equals the difference between the total nitrogen and the nitrogen. - Compound A selectively forms as a sole nitrogen-containing product. In other words, r A = r NH3 = k C A 4.6-3

4 Based on a literature survey done in this work, one potential nitrogen compound which forms ammonia in alkaline solutions is a cyanate (OCN - ) ion, the formation reaction kinetics of which has been studied, for instance, by Jensen [11]. Our comparison between experiments measuring formation rates in liquors containing no aditional OCN - and in liquors containing additional OCN - supports that the -forming ion in the green liquor would be OCN -. These experimental results will be published in near future. Ratio of Nitrogen to Sodium in Various Liquors The molar ratio of nitrogen to sodium (N/Na) in four liquors is presented in Table 4. The ratio of N/Na in the input stream to the causticizing plant (green liquor) was clearly higher compared to the ratio in the white liquor leaving the causticizing plant. (The input of sodium with lime at the causticizing was, as calculated, negligible.) This means that nitrogen must leave the liquor more readily than sodium in the causticizing plant (during slaking and causticizing), the loss, most likely, being about 15 % of the nitrogen entering the slaker. Weak wash, instead, clearly had the highest N/Na ratio. This was caused by the nitrogen content of the secondary condensate used for the lime mud washing. Nitrogen Mass Balance The mass balance for nitrogen around the chemical recovery cycle was performed on the basis of the Kjeldahl analysis results shown above in Table 3 and the available flow data. Most of the flow data is based on the process measurements and/or were, in addition, calculated based on the density or active alkali content values of the liquor streams. The smelt flow to the dissolver was estimated by assuming that all sodium and sulfur in virgin black liquor, 19. and 5.4 weight %, respectively, end up in the smelt, the sulfur reduction being 96 %. Here, the load to the recovery boiler was 32 kgds/s, which corresponds to the green liquor flow of 98 l/s to the causticizing plant. In Figure 6, the amount of nitrogen is expressed as grams per second. Values estimated but not measured are marked with an asterisk (*). During one sampling period, a few vent gas measurements for detection of ammonia were made by absorbing (both gaseous and particulates) into an acid impinger solution (according to the standard method VDI 2461). The concentrations were very high in all measurements: The content, given mg /m 3 n (dry gas), varied between 3-7 in the vent gases from the smelt dissolver, between 3-4 in the combined vent gas from the green liquor tank and the clarifier, whereas it was around 7 in the vent gas from the slaker. Because of the difficulties encountered in gas flow measurements, no reliable mass balance data could be obtained. Thus, all the nitrogen flow values of the vent gases presented in Figure 6 are estimated based on the balance around each process unit. Nitrogen enters the recovery cycle with the wood chips. The amount of the wood chip nitrogen was estimated based on the balance around the digester system, assuming that the vent gas nitrogen flow equals the amount of the nitrogen found in the condensate from the turpentine recovery. The major part of the white liquor nitrogen entering the digester is in the form of ammonia, representing a nitrogen flow of 5.1 gn/s (84 % of its total nitrogen flow). About 2% of the weak black liquor nitrogen was found to occur in ammonia, representing a nitrogen flow of 6.2 gn/s. This suggests that most of the weak black liquor ammonia nitrogen originates in the white liquor, but some ammonia nitrogen also originates in the secondary condensate and the water used for pulp washing. The ammonia nitrogen of weak black liquor was found to be released mostly at the beginning of evaporation and it probably ended up in the condensates, mostly in the foul condensate, which, in turn, was treated by stripping, thus releasing the ammonia to the gas phase which, further on, was subject to methanol stripping. The stripped condensate and the condensate from the last stages of the evaporator were combined and further used for lime mud washing, thus introducing some ammonia nitrogen to the weak wash. The secondary condensate from the first stages of the evaporator, which is the cleaner side of the evaporator, was collected separately from the dirtier condensates and further used for the pulp washing, thus introducing a minor quantity of ammonia to the weak black liquor. According to these measurements, about 25 % of the black liquor nitrogen entering the recovery boiler leaves with the smelt. However, based on the mass balance around the smelt dissolver, the smelt nitrogen entering the dissolver was not quite enough to explain the amount of the green liquor nitrogen leaving the dissolver. The smelt nitrogen should have been about one third of the black liquor nitrogen to have a balance around the smelt dissolver. So far, the reasons for this difference are unknown, but one explanation could be that the large variation in the nitrogen content that was found between different smelt sample pieces. This is under further investigation

5 The vent stack gas from the dissolver was first scrubbed with weak wash, then led to the flue gas scrubber of the recovery boiler, which was operated at a ph of about 7.1. The scrubber solution continued to the dissolver, but its input to the nitrogen mass balance around the dissolver was, however, found to be quite insignificant. The significance of the nitrogen flow of the scrubber solution from the NCG was even less. The green liquor nitrogen leaving the dissolver corresponded to almost 4 % of the black liquor nitrogen introduced to the recovery boiler. Based on the balance calculations as well as on the N/Na ratio results, a significant loss of nitrogen was found to occur around the slaker and the causticizers, the loss being 15-2 % of the green liquor nitrogen. This indicates that some ammonia was released to the gas phase, which in this case was led as LCHV gases to the lime kiln. The major part, about 6 %, of the green liquor nitrogen entered the cooking process with the white liquor. A minor part, less than 2 %, continued with the lime mud and further mostly with the weak wash, and was thus returning to the smelt dissolver. DISCUSSION Forssén et al. [6] have recently suggested that the char nitrogen fixed in the smelt would be the source of the ammonia earlier reported to be present in the green liquor and in the vent gases from the dissolver [7,8]. This assumption is still valid, and based on this work it seems evident that the smelt nitrogen is in a form, which slowly converts to ammonia under process conditions. How the ammonia itself behaves under varied process conditions is briefly discussed below. In this paper, the term ammonia has been used to refer to both ammonium ions (NH 4 + ) and ammonia ( ) in a solution. An equilibrium exists between ammonium ions and un-ionized or free ammonia, depending on the ph of the solution, according to Figure 7 (left). In addition to the equilibrium between ammonium and ammonia in the solution, another important equilibrium between aqueous and gaseous is formed which, in turn, is very sensitive to temperature. The temperature sensitivity of this equilibrium can be seen on the right in Figure 8, which shows the equilibrium concentration of the gaseous above the solution containing various amount of aqueous. Due to the high ph-value of most of the recovery process liquors, the ammonia exists in these liquors as aqueous. Assuming that this is in equilibrium with the gaseous, various amounts of may escape from the liquor phase, depending on the temperature and openness of the process step. This study suggested, however, that quite vigorous and thorough treatment of the liquor, for instance with steam, is needed to strip out all the from the liquor. Only the evaporation step and steam stripping seemed to be effective ways to release all the aqueous. Once the ammonia is stripped from the foul condensate, it may either stay in the gas phase or dissolve into water/methanol during the methanol condensing stage. Supposing that there is no methanol condensing stage, i.e., the ammonia stays in the gas phase, it is interesting to examine the fate of this ammonia. In this case, the ammonia follows the HCLV gases to the. Assuming that the total amount of the HCLV gases is around 4 m 3 n/ton pulp, calculations show the concentration to be about one volume percent in these HCLV gases. After burning the HCLV gases, this ammonia would result in a NO x emission of around 19 ppm(v)no 2 (dry f.g., O 2 =3%) assuming a 1 % conversion of to NO x. Typical measured values of NO x in the flue gases of the at the mill have been around 8-15 ppm(v)no 2 (dry f.g., O 2 =3%), which indicates that the conversion of the to NO x would have been around 4-8 %. The NOx values given here originate in some earlier measurements during the years The fact that some ammonia may also be absorbed by the condensing methanol does not change these conclusions, because in this case, both the methanol and the HCLV gases were conducted to the same. CONCLUSIONS The primary purpose of this work was to study the fate of the smelt nitrogen in the green liquor and further down the recovery cycle. Much effort was put on the sampling at the pulp mill and on the analysis work. In addition, laboratory measurements were done to study the ammonia formation in green liquor in detail. The present nitrogen analysis results of mill samples support some findings of Tarpey [7] at another mill. The total nitrogen concentration of several liquors, like black liquor, green and white liquor and weak wash, are very much at the same level in both studies. In this work, the nitrogen mass balance of the recovery cycle was closed for the first time, and some very interesting findings were revealed. The smelt nitrogen was found to be in a form which slowly converted to ammonia under the process conditions. The green liquor leaving the dissolver 4.6-5

6 contained clearly less ammonia than the green liquor entering the slaker, about 2 % and 7 % of their Kjeldahl nitrogen, respectively. The total amount of the green liquor nitrogen, originating both in the smelt and the weak wash, corresponded to almost 4 % of the black liquor nitrogen. This means that about one third of the black liquor nitrogen had to remain in the smelt after burning. The major part, about 6 %, of the green liquor nitrogen left the recovery process and passed into the cooking process along with the white liquor. The rest of the green liquor nitrogen was partly released, preferably in the form of in the causticizing plant, part of it continued with the lime mud further into the weak wash, and thus returning to the smelt dissolver. The white liquor ammonia was transferred into the weak black liquor. It accounted for some 2 % of the total nitrogen in the weak black liquor. This is based on the assumption that no significant amount of ammonia was formed from the wood chips during cooking. This weak black liquor ammonia was found to be released from the liquor in the evaporator plant and it ended up mostly in the foul condensates. A minor part was found in the secondary condensate which was used for the lime mud washing. The nitrogen in this flow thereby passed the recovery boiler and mainly ended up in the weak wash. Otherwise all the white liquor nitrogen, which once left the causticizing never entered it again. So, in general, nitrogen did not accumulate in the recovery cycle, but it was effectively removed from the liquor system as gaseous ammonia. The following potential exit points for ammonia from the recovery process were revealed: - First, a significant amount of nitrogen, in the form of, was bled from the cycle to HCLV gases at the stripping of the foul condensates. In the mill studied, the HCLV gases were treated by condensing the methanol, after which the HCLV gases as well as the methanol were burned in a separate. - Secondly, another significant release from the liquor cycle was found to occur in the causticizing plant, around the slaker and the causticizers, to LCHV gases. The amount of the release was about half of the above mentioned release to the HCLV gases in the stripper. In this mill, the LCHV gases from the causticizing plant were burned in the lime kiln. - In addition, a minor amount of was found to be released in the smelt dissolver. It was led to the flue gas scrubber of the recovery boiler. Based on this work, an overall picture of the fate of nitrogen in the chemical recovery process of the studied pulp mill is presented in Figure 8. The overview figure is based on the values given in Figure 6, and on the assumption that the missing part of the green liquor nitrogen originates in the smelt, as explained above. For the picture, a NO x release of about 1 ppm(v) in the recovery boiler was assumed. A value of 1 represents the nitrogen stream for the black liquor introduced into the recovery boiler, so the other numbers associated with the arrows indicate the ratio of each nitrogen stream to that nitrogen input to the boiler. The authors wish to emphasize that this work was performed at one pulp mill only, and the results may not be directly generalized and transposed on another mill because of possible process variations between the mills. ACKNOWLEDGEMENTS The authors wish to extend their special thanks to the mill staff for their cooperation. M.Sc. Hanna Malm is acknowledged for doing the laboratory measurements. This work is part of the Combustion and Gasification Research Program LIEKKI 2 in Finland. The support from the Finnish Recovery Boiler Committee, Ahlstrom Machinery Oy, and Kvaerner Pulping Oy is gratefully acknowledged. REFERENCES 1. VEVERKA, P., NICHOLS, K., HORTON, R., ADAMS, T., On the Form of Nitrogen in Wood and its Fate During Kraft Pulping, 1993 TAPPI Environmental Conference, Boston, MA, (1993). 2. AHO, K., Nitrogen Oxides Formation in Recovery Boilers, Lic. Tech. Thesis, Åbo Akademi University, NIEMELÄ, K., Low-Molecular Weight Organic Compounds in Birch Kraft Black Liquors, Ph.D. Dissertation, Technical University of Helsinki, MARTIN, D., MALCOLM, E., HUPA, M., The Effect of Fuel Composition on Nitrogen Release During Black Liquor Pyrolysis, Eastern States Section, The Combustion Institute, Fall Technical Meeting Proceedings, Clearwater Beach, FL, MARTIN, D. and MALCOLM, E., The Impact of Black Liquor Composition on the Release of 4.6-6

7 Nitrogen in the Kraft Recovery Furnace, 1995 TAPPI Engineering Conference, Dallas, TX, (1995). 6. FORSSÉN, M., HUPA, M., PETTERSON, R., MARTIN, D., Nitrogen Oxide Formation During Black Liquor Char Combustion and Gasification, J. Pulp Paper Sci, 23 (9):J439-J446 (1997). 7. TARPEY, T., Adressing Ammonia and Particulate Emissions from a Kraft Smelt Tank, 1995 TAPPI Environmental Conference, Atlanta, GA, (1995). 8. TARPEY, T., TRAN, H. and MAO, X., Emissions of Gaseous Ammonia and Particulate Containing Ammonium Compounds from a Smelt Dissolving Tank, J. Pulp Paper Sci, 22 (4), J145- J15 (1996). 9. THOMPSON, L, MARTIN, D., EMPIE, H., MALCOLM, E., WOOD, M., The Fate of Nitrogen in a Kraft Recovery Furnace, 1995 TAPPI Chemical Recovery Conference, Toronto, Canada, B (1995). 1. KYMÄLÄINEN, M., FORSSÉN, M. and HUPA, M., The Fate of Nitrogen in the Chemical Recovery Process in a Kraft Pulp Mill, 1998 TAPPI Chemical Recovery Conference, Tampa, USA (FL), (1998). 11. JENSEN, M., On the Kinetics of the Decomposition of Cyanic Acid, II: The Carbonate Catalysis, Acta Chem. Scand., 13 (4), (1959). Table 3. Kjeldahl nitrogen and ammonia analysis results. Sampling point / Kjeldahl-N, % Sample Kjeldahl-N mgn/l mg N/l 1 GL1, Green liquor from dissolver GL2, Green liquor from clarifier to slaker WL1, White liquor from causticizers WL2, White liquor to digesters WW, Weak wash SS1, Scrubber solution from flue gas scrubber of rec. boiler SS2, Scrubber solution from flue gas scrubber of NCG-inciner WBL, Weak black liquor from digesters 976 *) 198 *) 2 9 SBL, Strong black liquor, virgin (before mixing tank) 84 *) 11 *) 1 1 S, Smelt from a smelt spout 448 **) 9**) 2 11 FOUL,Untreated contaminated ( foul ) condensate to stripper COND, Condensate from digesters TC, Treated condensate from stripper to storage tank 3 14 SC1, Secondary condensate to pulp washing 1 n.a. 15 SC2, Secondary condensate to lime mud washing WATER. Warm water to pulp washing 4 n.a. presented as n.a., not analyzed *) mg N/kgds **) mg N/kg Table 4. The molar ratio of N/Na in green and white liquors and weak wash. Green liquor White liquor to slaker from causticizers White liquor to digesters Weak wash N/Na ratio (x 1-3 )

8 NCG to turpentine 12 TURPENTINEcondensate RECOVERY FOUL COND. NCG to NCG to stripper gasmethanol 11 STRIPPER 13 CONDENSER MeOH to LCHV-gas WHITE LIQUOR FILTER WHITE LIQUOR STORAGE 3 lime mud 4 LCHV-gas CAUSTICIZIERS wood chips white liquor lime SLAKER DIGESTING + WASHING 5 pulp 2 16 water GREEN LIQ. CLARIFIER 8 weak black liquor secondarycondensate LCHV-gas dregs 14 GREEN LIQ. weak wash EVAPORATION strong black liquor 9 MIX RECOVERY BOILER smelt 1 1 green liquor SECONDARY COND. dirty clean 6 SMELT DISSOLVER scrubber sol. from NCG-inciner. FLUE GAS SCRUBBER VENT GAS SCRUBBER 7 LIME MUD MIXER LIME MUD FILTER lime mud to dewateringand calcining secondarycondensate 15 Figure 2. The main sampling points (1-16) in the recovery process of a kraft pulp mill. Black liquor and smelt Green and white liquor system Condensates Scrubber solutions mg N / kg DS mg N / l Kjeldahl-N -N mg N / l mg N / l WBL SBL S *) 4 2 GL1 GL2 WL1 WL2 WW FOUL TC SC 2 1 SS1 SS2 *) mg N / kg Figure 3. Kjeldahl nitrogen and ammonia analysis results. (Sample codes as in Table 3.) 4.6-8

9 mg N / kg DS mg N / kg mg N / l weak BL strong BL smelt green liquor Sampling period Sampling period Sampling period Figure 4. Variation of nitrogen concentration in black liquor, smelt and green liquor samples. 1 1., mgn/l ln([a]/[a] o ) Total-N Time, min.2 Fitted line Data Time, min Figure 5. a. Experimental data on the formation of in a mill green liquor at 96 C. b. The first-order reaction test for the data shown in Fig.a

10 .4* WHITE LIQUOR FILTER 1.7 WHITE LIQUOR STORAGE lime mud white liquor CAUSTICIZERS.4* wood chips LCHV-gas 1.9* lime SLAKER 2.4 NCG to turpentine TURPENTINE.4 RECOVERY condensate 23.9* 6.1 FOUL COND. 3.8* DIGESTING WASHING weak black liquor.6.2 secondary condensate water pulp 1.1 NCG to NCG to METHANOL gas 6.7* CONDENSER STRIPPER 6.7 MeOH to 1.6 EVAPORATION strong black liquor MIX 1.6* SECONDARY COND. dirty clean.6* 26.9 FLUE GAS LCHV-gas RECOVERY SCRUBBER BOILER.3*.2 1.5* VENT GAS smelt 6.8.3* SCRUBBER GREEN LIQ. GREEN LIQ. 1.9 CLARIFIER 1.2 SMELT green DISSOLVER liquor dregs.14 scrubber sol. from NCG-inciner..2 weak wash LIME MUD MIXER 1.6 LIME MUD FILTER lime mud to.5 dewatering and calcining secondary condensate Figure 6. Nitrogen flows in the recovery cycle of a mill. Units: gn/s. * : estimated (not measured) values log(c i /M) ph (.1 M) NH 4 + Equilibrium concentration of gaseous, ppm C 9 C 8 C 25 C Concentration of ammonia, mg/kg H 2 O Figure 7. Left: /NH 4 + -equilibrium as a function of ph in aqueous solution (25 C). Total concentration.1 M. Right: Equilibrium concentration of gaseous over aqueous solution as a function of aqueous concentration and temperature 4.6-1

11 DIGESTER (+WASHING) EVAPORATION RECOVERY BOILER SMELT DISSOLVER CAUSTICIZING White Liquor NO x 28 N 2 Wood Chips Water Weak Black Liquor 2 Virgin Black Liquor 1 38 Smelt 34 Wash Solutions Secondary Condensate 6 6 Green Liquor 38 1 Weak Wash 8 White Liquor Lime Mud Lime Mud Figure 8. An overview of the nitrogen flows in the recovery cycle