Characterization of CO₂ precipitated Kraft lignin to promote its utilization

Abstract

Converting pulp mills into forest biorefineries to produce biopower and biomaterials can decrease their environmental impact and increase feasibility at the same time. One of the key challenges to reach this goal is the recovery of lignin from process streams for subsequent utilization in a variety of innovative green processes. This study examines the fundamental chemical structure of lignin recovered from Kraft pulping streams by an acid precipitation/washing methodology. Functional group analysis and molecular weight profiles were determined by NMR and SEC with promising results for future conversions; such as low hydroxyl (oxygen) contents and low molecular weights (∼3000 g mol⁻¹).

---

Continuously developing new technologies that allow for a more efficient utilization of resources including biomass is crucial in all industries for reasons like sustainability and higher profit. The pulp and paper industry is also constantly searching for alternative solutions to valorize its Kraft cycle – like the process detailed below – mainly by trying to find ways to recover lignin and sell it as a higher value material.

During chemical pulping, lignin is chemically degraded and extracted from wood in an aqueous environment in a pressure reactor at pH values of 13–14 and temperatures of 140–170 °C.^2,3 These conditions remove 85–93% of the lignin and approximately 56–71% of the hemicelluloses.^1–3 The approach when NaSH is used in the cooking process along with caustic to delignify wood is referred to as Kraft pulping. In the United States alone, the pulp and paper industry collects and processes ∼108 million tonnes of pulpwood for the production of pulp, paper and paperboard annually.¹ In turn, the paper industry produces over 50 million tons of residual lignin per year worldwide in a form of a caustic sidestream.¹ Currently, this material is burned in a low efficiency Thompson recovery furnace to recover energy and cooking chemicals. A continuing interest in this field is the desire to recover fractions of lignin from the Kraft cooking liquors for biopower, biochemical and biomaterial utilizations. Recently, a green process referred to as “LignoBoost” provides a viable separation of lignin from these cooking liquors by employing carbon-dioxide to precipitate lignin from alkaline solutions.^4–8 The application of this process involves acidification of an alkaline cooking liquor with CO₂, precipitation, filtration and washing.^9–11 Process integration and mill trials have tested this process from an engineering and economic point of view.^8,9,12

The lignin recovered by the LignoBoost process has been shown to be valuable green resource for biopower production.¹³ This method enables lignin to be exported in the form of a solid biofuel and also gives the opportunity to transform it into materials of higher value.^13–15

Recognizing the possibilities in recovering lignin from black liquor via CO₂ precipitation and washing, we became interested in the detailed chemical composition and structure of the isolated lignin; anticipating that this data could facilitate future applications of this bioresource. Herein, we wish to report the characterization of LignoBoost derived lignin in terms of molecular weight profiles and functional group properties.

Experimental

Lignin separation from a commercial Scandinavian softwood Kraft pulping liquor was accomplished following published methods.^5,10,16 In brief; Kraft cooking liquor (BL) was collected in a reaction vessel then treated with pressurized CO₂ (1500 kPa)⁴ until the solution's pH reached 10.5, causing some of the lignin to precipitate. The suspension was subsequently separated by filtration giving a precipitate (P) and a filtrate (F) part. This process was done on two separate occasions with two different final pHs: 10.5 and 9.5, resulting the following samples: P 10.5, P 9.5 (precipitates) and F 10.5, F 9.5 (filtrates). Solids were obtained from the F and BL solutions by evaporation, then all samples were freeze dried and kept frozen until purification.⁴

Detailed structural analysis of the lignin in the above samples required additional purification. The initial LignoBoost samples were diluted in distilled water to 5 wt% solid content. Next EDTA-2Na⁺ was added to the aqueous solution (5.00 g L⁻¹) to facilitate metal-ion removal and the pH was adjusted to a value of 6 with aqueous sulfuric acid (2 M) and stirred for 1 h. Subsequently, the pH was further lowered to a value of 3 facilitating lignin precipitation.^4,17 The resulting samples were frozen (−20 °C) overnight, thawed and filtered through a medium sintered glass funnel at 0 °C.^9,14,16 Retentates were dissolved in pH 3 aqueous sulfuric acid solution up to 5 wt% and the filtration process was repeated three times for effective salt removal. All filtrates were collected, the solvent was removed under reduced pressure and the remaining solid provided the salt fraction (SF). Retentates were air dried, extracted with pentane to remove free sulfur and dissolved in dioxane : water (9 : 1) solution (1 g L⁻¹).¹⁷ After filtration through a medium sintered glass funnel the solvent was removed under reduced pressure. The resulting solid provided purified lignin samples which were stored in a freeze dried form at (−20 °C) until further use.

All NMR spectral data reported in this study was recorded with a 400 MHz DMX Bruker spectrometer (Billerica, MA, USA) at 25 °C. Qualitative ¹H and ¹³C-NMR spectra were acquired on lignin samples (80–120 mg) dissolved in dimethyl sulfoxide (DMSO-d₆) (450 μL).¹⁸ Quantitative ¹H-NMR spectra were acquired on dry lignin samples (∼20 mg) dissolved in dry DMSO-d₆ (450 μL) using pentafluorobenzaldehyde (∼1 mg mL⁻¹) as internal standard.^18,1931P-NMR lignin spectra were acquired on dry samples (∼25 mg) that were derivatized with 2-chloro-4,4,5,5,-tetramethyl-1,3,2,-dioxaphospholane (TMDP) and analyzed by ³¹P-NMR.^18,20

For size exclusion chromatography (SEC), purified lignin (20 mg) was acetylated by stirring with acetic-anhydride : pyridine 1 : 1 (2 mL) at room temperature for 72 h. The solvent mixture was then removed under vacuum at 50 °C, the acetylated lignin was dissolved in chloroform (50 mL) and washed with deionised water (20 mL). The organic fraction was dried over anhydrous MgSO₄ and the chloroform was removed under reduced pressure. The dried acetylated lignin was dissolved in tetrahydrofuran (1 mg mL⁻¹) for analysis.^18,21

Results and discussion

After purification of the unwashed samples, the lignin content of the initial LignoBoost samples could be determined. The mass balance data showed that the initial unpurified dry samples had various amounts of lignin as the following (in wt%): 25.8% in BL, 73.3% in P 9.5, 71.3% in P 10.5, 11.1% in F 9.5 and 20.0% in F 10.5. This data clearly shows that treatment of the pulping liquor with CO₂ yields a lignin rich stream and a filtrate fraction that is enriched in salts. The pH 9.5 treatment condition resulted in a better separation of lignin. Results of the elemental analysis are consistent with the mass balance data and shows that the purified lignin samples contain only organic elements up to 97.1%.

To evaluate the primary components present in the initial LignoBoost fractions and in their purified samples, qualitative ¹H- and ¹³C-NMR measurements were conducted (Fig. 1) following literature methods.^17,18,22-24 Qualitative NMR data shows that during the purification step, sugars and fatty acids were removed with salts and salt-EDTA complexes from the initial LignoBoost samples and ended up in the SF. With the lignin purification method used in this study, the chemicals used earlier to bring lignin into solution can be separated from the crude samples. This allows the process chemicals to be returned to the main stream and reused, while it also produces lignin with 95–97% purity, resulting in a higher environmental and economic efficiency.

Fig. 1 Qualitative ¹H and ¹³C-NMR spectrum of purified P and F 10.5 LignoBoost samples, measured in dimethyl sulfoxide (DMSO-d₆) as solvent. Abbreviations: Sub. = substituted, Unsub. = unsubstituted.

To selectively follow the structural changes of the lignin biopolymer on the molecular level, quantitative ¹H-NMR measurements were conducted on the purified fractions. The differences between the process fractions and the changes in the distribution of selected lignin moieties at different final pHs are shown in Table 1. Quantitative ¹H-NMR shows that the F lignin has more carboxylic and phenolic groups, while P is enriched in methoxyl moieties. The hydroxyl group content of the lignin plays a crucial role in determining its solubility.^2,3 Selective phosphitylation of the hydroxyl groups on the lignin polymer with TMDP, followed by quantitative ³¹P-NMR measurement provides facile monitoring of the changes in the hydroxyl content throughout the process.¹⁸

Table 1 Partial hydrogen content [mol mol⁻¹]% of different lignin functional groups in the ratio of all H containing functional groups as determined by quantitative ¹H-NMR

Sample name	Hydrogen content of selected groups (mol mol^−1% relative to all H containing groups)
	Carboxylic acid	Formyl	Phenolic	Aromatic, vinyl	Aliphatic	Methoxyl	Aliphatic
	(13.50–10.50) ppm	(10.10–9.35) ppm	(9.35–8.00) ppm	(8.00–6.00) ppm	(6.00–4.05) ppm	(4.05–3.45) ppm	(2.25–0.00) ppm
	–C(O)OH	–C(O)H	=HC–OH	CH=CH	CH–O	–OCH3	C–CH2–C
				CH2=CH	C–CH2–O		C–CH3
Black liquor	1.26	1.50	6.73	20.22	8.44	45.84	14.42
P 9.5	1.06	0.90	4.19	18.75	5.92	52.42	16.70
P 10.5	0.81	0.98	3.66	19.69	8.22	49.16	16.04
F 9.5	1.71	1.61	6.16	19.82	7.89	41.25	17.35
F 10.5	1.22	1.67	5.73	19.58	6.78	44.58	16.87

---

The pK_a values of phenolic lignin groups fall between 9.4 and 10.8,⁴ hence at pH 9.5 and 10.5 these groups get protonated consequently affecting the solubility of the polymer by determining – lowering – its charge. A polymer with a higher total hydroxyl and carboxyl content will have more sites that can stay deprotonated causing better solubility. Our ³¹P-NMR results in Table 2 confirm the above logic by showing that both hydroxyl and carboxyl contents of F lignins are higher than in their analogous P lignins; this is also consistent with the results from ¹H-NMR analysis. F lignins have almost double the amount of carboxylic hydroxyl groups that contribute significantly to their ability to stay in solution despite the decreasing pH. Moreover, the quantities of guaiacyl phenolic groups are also almost doubled in F lignins meaning more free hydroxyls compared to their P analogs that most likely contain etherified units in most of these positions.

Table 2 Hydroxyl content of different LignoBoost fractions determined by quantitative ³¹P-NMR after derivatization with TMDP

Sample name	Total –OH content/μmol mg^−1	Hydroxyl content of selected groups/μmol mg^−1
		Aliphatic hydroxyl	Condensed phenolic	Guaiacyl phenolic	Carboxylic hydroxyl
	(149.0–133.8) ppm^a	(149.0–145.6) ppm^a	(144.4–140.4) ppm^a	(140.4–137.6) ppm^a	(136.0–133.8) ppm^a
a All samples were referenced to an internal standard of cyclohexanol at 144.9 ppm.
Black liquor	6.37	1.49	1.73	2.46	0.69
P 9.5	4.30	1.11	1.14	1.47	0.59
P 10.5	3.18	0.91	0.85	1.03	0.39
F 9.5	6.74	1.27	1.80	2.55	1.11
F 10.5	5.76	1.22	1.57	2.20	0.78

---

Changes in the polymer structure of the lignin were followed by molecular mass distribution analysis with SEC on the purified fractions using acetylated lignin samples.^4,8,18 Polystyrene equivalent weight average molecular weight (M_w), number average molecular weight (M_n) and polydispersity (PD) were determined using calculation strategies from Baumberger for whole curve integration.²¹M_w/M_n gives a PD that directly shows how accurate it is to evaluate a peak as one and not as a sum of multiple peaks. SEC data shows that P samples were enriched in the lower M_w fraction in the 200–300 g mol⁻¹ region, which represents a lignin DP of 1–2 units.^1,18,25 Their peaks were recognizable and easy to separate from the main peak resulting from the higher M_w lignin fractions.^18,21 While on the contrary, in the case of F and BL samples, no additional peaks were recognizable in the lower M_w region and their SEC curves were integrated as one peak, and as a result their respective PD's were 2–3 times larger than in case of the precipitates. These results are consistent with previous research data of Wallmo.⁴Table 3 shows all main peak data obtained by SEC.

Table 3 SEC results for purified and acetylated samples

Sample	Mw	Mn	PD
Polystyrene standards were used with Mw 1200–195 000 g mol^−1.
BL	3601	812	4.44
P 10.5	2939	1694	1.73
F 10.5	2718	795	3.42
P 9.5	2979	1795	1.66
F 9.5	2101	735	2.86

---

It is noteworthy that the polydispersity of all separated lower M_w peaks fell between 1.02 and 1.12 (data not shown), which confirms the validity of the calculation strategy used.

In conclusion, the entering BL separates into fraction P which is enriched in lignin and into fraction F which is enriched in salts and also contains some sugars and short chain acids. It is noteworthy that a lower final pH resulted in a better lignin separation. SEC data obtained on the purified P phase showed that the fraction is enriched in the ∼3000 g mol⁻¹– low degree of polymerization (DP) – and in the 200–300 g mol⁻¹ monomer regions. Quantitative NMR data showed that F contains almost two times the amount of carboxylic and phenolic groups causing its better solubility in water. Low DP together with low quantities of oxygen containing functional groups make both P 9.5 and P 10.5 viable starting feedstocks for future biofuel (or biomaterial) production. Ongoing pyrolysis experiments showed promising oil yields (∼40% based on dry weight of purified lignin from P 9.5 sample), however much more analytical data need to be obtained for the determination of these products, which is the goal in the continuation of this project.

Acknowledgements

Authors thank the support of the AtlanTICC Alliance and its member universities and the doctoral grant of MN and MK. AJR also wishes to thank the support of the Fulbright Fellowship program for the support of his Chair in Alternative Energy.

Notes and references

1. A. J. Ragauskas, M. Nagy, D. H. Kim, C. A. Eckert, J. P. Hallett and C. L. Liotta, Ind. Biotechnol., 2006, 2, 55.
2. J. Gullichsen, and C.-J. Fogelholm, Chemical Pulping 1999TAPPI Press.
3. C. W. Dence, and D. W. Reeve, Pulp Bleaching, Principles and practice, TAPPI Press, 1996.
4. H. Wallmo, T. Richards and H. Theliander, Paperi ja Puu, 2007, 89, 436.
5. F. Öhman, H. Wallmo and H. Theliander, Nord. Pulp Pap. Res. J., 2007, 22, 9.
6. F. Öhman, H. Wallmo and H. Theliander, Nord. Pulp Pap. Res. J., 2007, 22, 188.
7. A. Moosavifar, P. Sedin and H. Theliander, Nord. Pulp Pap. Res. J., 2006, 21, 180.
8. A. Moosavifar, P. Sedin, H. Brelid and H. Theliander, Nord. Pulp Pap. Res. J., 2006, 21, 493.
9. H. Luotfi, B. Blackwell and V. Uloth, Tappi Journal, 1991, 74, 203.
10. F. Öhman, H. Theliander, P. Tomani and P. Axegard, Patent WO 2006/031175 A1, 2006.
11. R. Alén, E. Sjöström and P. Vaskikari, Cellulose Chemistry Technol, 1985, 19, 537.
12. M. Olsson, E. Axelsson and T. Berntsson, Nord. Pulp Pap. Res. J., 2006, 21, 476.
13. U. Wising, J. Algehed, T. Berntsson and L. Delin, Tappi Journal, 2006, 5, 3.
14. F. Öhman and H. Theliander, Paperi ja Puu, 2006, 88, 287.
15. C. Laaksometsä, E. Axelsson, T. Berntsson and A. Lundström, Clean Technol. Environ. Policy, 2009, 11, 77.
16. F. Öhman, H. Wallmo and H. Theliander, Filtration, 2007, 7, 309.
17. P. Froass, A. J. Ragauskas and J. E. Jiang, Holzforschung, 1998, 52, 385.
18. M. Nagy, K. David, G. Britovsek and A. J. Ragauskas, Holzforschung, 2009, 63, 513.
19. T. M. Runge and A. J. Ragauskas, Holzforschung, 1999, 53, 623.
20. Y. Pu and A. J. Ragauskas, Can. J. Chem., 2005, 83, 2132.
21. S. Baumberger, A. Abaecherli, M. Fasching, G. Gellerstedt, R. Gosselink, B. Hortling, J. Li, B. Saake and E. de Jong, Holzforschung, 2007, 61, 459.
22. F. Chakar and A. Ragauskas, Ind. Crops Prod., 2004, 20, 131.
23. E. Sjöström, Wood Chemistry: Fundamentals and applications, Academic press, Orlando, 1981.
24. J. Gierer, Wood Sci. Technol., 1980, 14, 241.
25. L. B. Davin and N. G. Lewis, Curr. Opin. Biotechnol., 2005, 16, 407.

---

Footnote

† Contributed equally to this work.

---