Thermal characterization of hydrated eco-friendly clinkers produced from pulp and paper mill waste

Caracterização térmica de ecoclínqueres hidratados produzidos a partir de resíduos da indústria de papel e celulose

Abstract

Industrial wastes from pulp and paper mills are potentially useful materials to use as alternative materials in clinker production. In this study, eco-friendly clinkers were obtained using only waste as feedstock. Eco-friendly clinker with different C₃S and C₂S contents were produced and hydrated for 7 days at 40 °C, or for 28 and 90 days at room temperature. The potentialities of these waste and the products obtained were revealed through chemical (XRF), thermal (DTA/TG) and surface area (BET) analysis. For comparison, a common ordinary Portland cement (OPC) was tested. As expected, the hydration rate of formulations presenting higher C₃S content was higher, generating higher amounts of Ca(OH)₂. Furthermore, the amount of Portlandite formed was lower when compared to OPC.

Keywords:
pulp and paper mill waste; clinker; hydration; portlandite

Resumo

Os resíduos do setor de papel e celulose são potenciais para serem usados como materiais alternativos na produção de clínqueres. Neste estudo, ecoclínqueres foram obtidos utilizando apenas resíduos como matéria-prima. Ecoclínqueres com diferentes teores de C₃S e C₂S foram produzidos e hidratados por 7 dias a 40 °C, ou por 28 e 90 dias em temperatura ambiente. As potencialidades destes resíduos e dos produtos obtidos foram reveladas através de análise química (FRX), térmica (ATD/TG) e de área superficial (BET). Para comparação, um cimento Portland comercial (OPC) foi testado. Como esperado, a taxa de hidratação das formulações apresentando maior teor de C₃S foi maior, gerando maiores quantidades de Ca(OH)₂. Além disso, a quantidade de portlandita formada foi menor quando comparada ao OPC.

Palavras-chave:
resíduos da indústria de papel e celulose; clínquer; hidratação; portlandita

INTRODUCTION

Industrial wastes from pulp and paper mills are potential replacements for raw materials (e.g. limestone and clay) used in clinker production ^{(1, 2)}. This is mainly due to the high concentration of major oxides (CaO ^3-6, SiO₂^7-10, and Al₂O₃^1,2) present in such wastes, which are required for the formation of desirable crystalline phases in the clinker (C₃S, C₂S, C₃A and C₄AF) ¹¹. Works ¹² demonstrated that this method of waste recovery is technically and environmentally feasible, and reported that high alkali content (Na₂O + K₂O) is a major issue in the end product. Those alkalis found in the waste, and consequently in the clinkers produced from the waste, arise from chemicals used in the digestion and separation of cellulose and lignin ¹³. Those alkalis may cause some problems in the durability of the final concrete. Literature reports indicate that high alkali content is harmful to concrete, causing alkali-aggregate reactions that result in delayed expansion and rupture of the material with time ¹⁴. According to regulatory standards, the acceptable alkali content in concrete is less than 0.6% ¹⁵, while it is typically between 2 and 3% in cement produced with waste from pulp and paper mills ¹⁶.

High proportions of alkalis may also negatively influence the ultimate strength of cementitious materials, even in the absence of alkali-aggregate reactions ¹⁶. Depending on how alkalis are present in cementitious binders, hydration and properties of the cement may be affected. In typical cements, alkalis may be incorporated into the clinker phases or in the form of sulfates ¹⁶. Some works ¹⁷ showed that alkalis in the form of sulfates do not alter the hydration of cement, but significantly reduce its tensile strength. In similar works ¹⁸, the increase of the amount of alkaline oxides in hardened cement pastes (from 0.15 to 2.8 wt%) reduced the compressive strength by ~56%. Smaoui et al.¹⁶ reported cases of loss of strength at any curing age, whereas Jawed and Skalny ¹⁹ reported that this attenuation occurs only at ages later than 28 days, having no effect during early curing.

The proportion of the crystalline phases C₃S (alite) and C₂S (belite) is another factor that directly influences the hydration of cement. Alite hydrates much faster than belite. After 28 days of hydration, 70 wt% of C₃S reacted (hydrated), while only 30 wt% of C₂S completed hydration ²⁰. The main hydration compounds of typical cements are calcium silicate hydrate (C-S-H) and portlandite (Ca(OH)₂), being the last expectably more abundant when C₃S/C₂S ratio is higher. When subjected to thermal characterization by differential thermal analysis and thermogravimetry (DTA/TG), those hydrated compounds suffer dehydration between 120 and 400 °C (C-S-H), and between 400 and 520 °C (Ca(OH)₂) ¹¹. The subsequent mass loss of Ca(OH)₂ dissociation might be used to quantify its amount ²¹. In the present work, hydrated samples of eco-friendly clinker were analyzed thermally to quantify the amount of portlandite present.

EXPERIMENTAL

Materials: the materials used in the present work were ashes from biomass combustion, wastewater treatment plant (WWTP) sludge, and lime mud from pulp and paper mills from Brazilian plants. Ordinary Portland cement (purchased in the Brazilian market) was used as a hydraulic binder reference (HBR).

Characterization: particle size distribution was determined by laser diffraction (Cilas 1064) for lime mud and WWTP sludge, and by sieving (5-0.03 mm) for biomass ashes. The chemical analysis was performed by X-ray fluorescence spectrometry (XRF, WDXRF Axios^mAX, PANalytical), while thermal analyses (DTA/TG, differential thermal analysis/thermogravimetry) were performed in a simultaneous thermal analyzer (SDT Q600, TA Instruments). The heating was conducted in synthetic air atmosphere at temperatures up to 1400 °C, at a heating rate of 10 °C.min^-1, and with samples of 15 mg mass in an alumina crucible. The hydrated eco-friendly clinkers were analyzed in nitrogen to ensure an inert and dry atmosphere, and at a heating rate of 5 °C.min^-1 up to 600 °C. Eq. A was used to estimate the portlandite (CH) content in hydrated eco-friendly clinkers from the mass loss (ML_CH) associated with its dihydroxylation ²¹:

$CH\left(\%\right)=M{L}_{CH}\left(\%\right)\times \frac{M{M}_{CH}}{M{M}_{H_{2}O}}$ (A)

wherein MM_CH and are the molar masses of portlandite and water, respectively. Specific surface area (BET Quantachrome Nova 1200e) measurements were conducted using nitrogen gas as adsorbate, for both eco-friendly clinkers and HBR. The mineralogical compositions were analyzed by X-ray powder diffraction (XRD, Shimadzu XRD-6000; CuKα radiation, 0.02º 2θ step-scan and 10 s.step^-1) and phase quantification was performed by Rietveld refinement using GSAS (general structure analysis system) with EXPGUI software. The internal standard was refined as a crystalline phase and used to determine the amorphous fraction of samples. The evaluation of the quality of refinement was based on least squares (χ²) and the graph.

Preparation and hydration of eco-friendly clinkers: the formulations (Table I) of eco-friendly clinkers were prepared based on the lime saturation factor (LSF) and Bogue equations ¹⁰. Compositional adjustments using the Bogue equations were performed in order to control the crystalline phases after thermal processing. In this sense, F1 and F5 had a greater tendency to form belite and alite phases, respectively ². The formulations were fired in a bottom loading furnace (ME 1700/V/Eg, Fortelab) in two steps (decarbonation and clinkering). In the first step, decarbonation was held at 1000 °C with 5 °C.min^-1 heating rate, and 120 min holding time. For clinkering, three different maximum temperatures were used: 1350, 1400 and 1455 °C, fixing the heating rate at 10 °C.min^-1, and the holding time at 90 min. These temperatures were chosen from the ternary diagram of CaO-SiO₂-Al₂O₃ and of a similar procedure related in the literature ^12,22. At the end of the firing cycle, the samples were rapidly (>50 °C.min^-1) cooled down to room temperature. After processing, the clinkers were ground and sieved (<45 μm) and 1 g mass was mixed with distilled water (w/c= 0.48) to form cementitious pastes. The cement pastes were placed to a plastic film and, after 24 h in air, bodies were immersed in water. Some samples were hydrated for 7 days in a climate chamber at 40 °C, while another set of samples were cured for 28 and 90 days in room conditions (25 °C).

Table I -
Investigated formulations (wt%).
Tabela I -
Formulações investigadas (m%).

Formulation	Lime mud	Biomass ashes	WWTP sludge
F1	66	34	0
F2	66	31	3
F3	68	30	2
F4	69	29	2
F5	70	30	0

RESULTS AND DISCUSSION

Characterization of wastes and clinkers: the results obtained by XRF confirmed that the wastes considered in this work contained high levels of CaO, Al₂O₃ and SiO₂, and are potential substitutes for traditional raw materials used in clinker production ¹. The generation of these wastes would not provide a clinker production only with these materials, but their use could complement the raw commercial materials. However, considerable alkali contents were found in all three wastes, highlighting the high K₂O content in biomass ashes (3.14 wt%) and Na₂O in lime mud (1.06 wt%), as listed in Table II. Mineralogical composition of the waste (Fig. 1) showed that WWTP sludge did not present intense crystal peaks. Lime mud and biomass ashes, in turn, showed crystal peaks of calcite and quartz, respectively. Particle size distributions ¹ showed d₅₀ of 14.6, 20.7 and 80.0 µm for lime mud, WWTP sludge and biomass ashes, respectively. These values can be considered suitable for obtaining clinker.

Table II -
Chemical composition (wt%) of the wastes.
Tabela II -
Composição química (m%) dos resíduos.

Component	Biomass ashes	WWTP sludge	Lime mud
Al2O3	9.94	13.24	0.22
SiO2	49.65	5.24	1.06
Na2O	0.33	0.67	1.06
K2O	3.14	0.30	<0.05
CaO	2.10	2.14	53.55
MgO	1.25	0.32	0.91
Fe2O3	4.07	0.75	0.16
MnO	0.24	0.29	0.10
P2O5	0.65	1.83	0.65
TiO2	0.95	0.07	<0.05
aLOI	26.97	74.84	42.25

• aLOI - loss on ignition at 1000 °C.

Figure 1:
XRD patterns of biomass ashes, lime mud and WWTP sludge.
Figura 1:
Difratogramas de raios X das cinzas de biomassa, lama de cal e lama de ETE.

All three wastes showed extensive mass losses (Fig. 2), either by burning of organic matter (in WWTP sludge -74%; and biomass ashes - 34%) or by decarbonation of compounds (in lime mud - 41%). The DTA and TG curves of lime sludge (Fig. 2a) clearly showed the endothermic peak at 745.9 °C corresponding to decarbonation of calcium carbonate. Thermography (DTA) of the ash sample (presented in detail in Fig. 2c) shows an exothermic peak at 499.2 °C, whereas the data for WWTP sludge shows peaks at 342.6 and 444.0 °C (Fig. 2b). These two exothermic peaks in the WWTP sludge may distinguish primary sludge from secondary sludge in the sample. The primary sludge has a high C:N ratio (150-250), while the secondary sludge has a low C:N ratio (5-30) ²³. These proportions suggested that the primary sludge had more complex molecules than the secondary sludge and that the two materials may, therefore, have different degradation temperatures. In this regard, it may be concluded that the higher temperature peak (444.0 °C) is due to the primary sludge, while the lower temperature peak (342.6 °C) corresponds to the burning of the secondary sludge. The energy released in these reactions is extremely interesting when clinker formulations are considered, since it may be used to reduce fuel consumption during clinkering.

Figure 2:
Thermograms (DTA/TG) of: (a) lime mud; (b) WWTP sludge; and (c) biomass ashes.
Figura 2:
Termogramas (ATD/TG) de: (a) lama de cal; (b) lama de ETE; e (c) cinzas de biomassa.

Estimated chemical compositions of the formulations (Table III) indicate a higher CaO content in F5 than in F1, mainly due to the more use of lime mud in F5. In general, the large amount of CaO in F5 resulted in a high C₃S content in the resultant clinkers, when compared with high C₂S content for F1 (Table IV), characterizing them as Portland and belite clinkers, respectively. The HBR sample showed an alkali (Na₂O+K₂O) content equal to 0.97 wt%. The eco-friendly clinker samples had alkali contents that were similar to each other (considering volatilization losses, Na₂O + 0.658×K₂O), but much higher than HBR, ranging from 2.31 wt% in F5 to 2.37 wt% in F2.

Table III -
Estimated chemical compositions (wt%) of the formulations.
Tabela III -
Composição química estimada (m%) das formulações.

Component	F1	F2	F3	F4	F5	aHBR
Al2O3	4.88	6.02	5.38	5.24	4.35	3.46
SiO2	24.33	22.91	22.06	21.40	21.69	19.65
Na2O	1.36	1.43	1.43	1.45	1.42	0.13
K2O	1.52	1.43	1.37	1.33	1.35	0.84
CaO	62.11	62.28	64.02	64.92	65.70	66.30
MgO	1.62	1.61	1.61	1.61	1.62	4.74
Fe2O3	2.08	2.00	1.92	1.87	1.87	3.65
MnO	0.23	0.25	0.24	0.24	0.22	0.10
P2O5	1.04	1.23	1.17	1.18	1.05	0.12
TiO2	0.50	0.47	0.45	0.44	0.45	0.31
Na2O+K2O	2.88	2.85	2.81	2.78	2.77	0.97
bNa2O eq	2.36	2.37	2.34	2.32	2.31	-

• a- Real composition obtained by XRF;bNa2O eq - Na2O+0.658xK2O.

Table IV -
Crystalline phases (wt%) determined by Rietveld method.
Tabela IV -
Fases cristalinas (m%) determinadas pelo método de Rietveld.

Sample		C3S	C2S	C3A	C4AF	χ2
HBR	-	59	26	4	11	4.264
F1	1350	10	69	9	11	4.746
	1400	31	49	3	17	4.692
	1455	4	63	7	26	4.708
F2	1350	10	71	4	15	3.840
	1400	22	46	16	16	6.529
	1455	19	44	20	17	7.791
F3	1350	12	59	14	15	5.935
	1400	49	31	6	14	7.102
	1455	34	33	13	20	6.489
F4	1350	29	44	13	14	6.052
	1400	39	37	8	16	5.772
	1455	46	32	11	11	4.704
F5	1350	23	52	13	12	6.538
	1400	36	35	6	23	5.502
	1455	49	22	8	21	5.164

Both the eco-friendly clinkers and the HBR were passed through a 45 µm mesh and show specific surface areas between 900 and 2500 m².kg^-1 (Table V). These results demonstrate that the samples had much higher specific surface areas than the required standards (240, 260, and 280 m².kg^-1 for cement classes 25, 32 and 40, respectively), according to NBR 05732 ²⁴. When compared with HBR, only F5 at 1350 °C had a lower specific surface area, but it was still higher than the NBR 05732 recommendation.

Table V -
Specific surface area (m².kg^-1) of the materials, after milling and sieving (<45 µm).
Tabela V -
Área de superfície especifica (m².kg^-1) dos materiais, após moagem (<45 µm).

Formulation	1350 °C	1400 °C	1455 °C
F1	1352	2499	2341
F2	1578	1575	2033
F3	2332	-	2391
F4	1668	1911	1196
F5	936	1964	1926
HBR		1107

Estimation of portlandite content formed upon hydration: DTA and TG analyses of the hydrated F1, F2 and F3 pastes at different ages and conditions showed the absence of dehydrating peaks of portlandite. Only decomposition of calcium silicate hydrates (C-S-H), between 120 and 400 °C, was evidenced by DTA (Fig. 3), as presented in Table VI. The same behavior was observed for pastes of F4 fired at 1350 and 1400 °C (Fig. 4). The absence of portlandite peaks in F1, F2 and F3 hydrated pastes may be explained by the relatively low amount of alite in such clinkers, as estimated by DRX (Rietveld method). In this sense, these formulations would take a longer time to hydrate. On the other hand, formulations F4 and F5 had higher alite content at 1455 °C, and therefore showed higher hydration rates and more expressive formation of portlandite as a by-product (see Figs. 4 and 5). The maximum amount of alite (and portlandite on hydrated paste) was observed for HBR (Fig. 6). Table VII gives estimations of portlandite (CH) contents formed in hydrated F4, F5 and HBR pastes, from TG analysis. In all cases, there was an increase in the portlandite amount formed with curing age due to a continuous hydration process (Table VII, Figs. 4 and 5). As expected, curing at 40 °C accelerates the hardening reactions, and F4 and F5 pastes tend to develop higher portlandite amounts, even when compared with samples cured for 90 days in room conditions. Estimations for HBR pastes (Table VII and Fig. 6) may be erroneous and difficult to interpret, probably because differences are within the experimental error of determination in such relatively portlandite-rich samples.

Figure 3:
Thermograms (DTA) of formulations F1, F2 and F3 at: (a) 7, (b) 28 and (c) 90 days.
Figura 3:
Termogramas (ATD) das formulações F1, F2 e F3 em: (a) 7, (b) 28 e (c) 90 dias.

Table VI -
Dehydration temperatures (°C) of C-S-H identified in the thermograms of formulations F1, F2, F3 and F4.
Tabela VI -
Temperaturas de desidratação (°C) do C-S-H identificadas nos termogramas das formulações F1, F2, F3 e F4.

Formulation		7 days	28 days	90 days
F1	1350 °C	224.34	214.08	220.20
	1400 °C	219.57	218.11	225.38
	1455 °C	211.61	222.68	227.98
F2	1350 °C	209.73	217.18	219.70
	1400 °C	210.42	222.16	225.84
	1455 °C	214.20	217.67	223.34
F3	1350 °C	217.58	222.85	222.24
	1400 °C	212.67	-	-
	1455 °C	219.52	218.86	237.38
F4	1350 °C	216.21	216.98	220.82
	1400 °C	212.10	-	-

Figure 4:
Thermograms (DTA/TG) of formulation F4 at: (a) 7, (b) 28 and (c) 90 days.
Figura 4:
Termogramas (ATD/TG) da formulação F4 em: (a) 7, (b) 28 e (c) 90 dias.

Figure 5:
Thermograms (DTA/TG) of formulation F5 at: (a) 7, (b) 28 and (c) 90 days.
Figura 5:
Termogramas (ATD/TG) da formulação F5 em: (a) 7, (b) 28 e (c) 90 dias.

Figure 6:
Thermograms (DTA/TG) of HBR at: (a) 7, (b) 28 and (c) 90 days.
Figura 6:
Termogramas (ATD/TG) de HBR em: (a) 7, (b) 28 e (c) 90 dias.

Table VII -
Quantification of portlandite (CH) from DTA-TG curves.
Tabela VII -
Quantificação de portlandita (CH) a partir de curvas de ATD-TG.

Formulation		Day	Wi (mg)	Peak (°C)	TG (mg)	CH (mg)	CH (wt%)
HBR		7a	15.9670	421.49	0.4290	1.7637	11.05
F4	1455		19.7990	413.68	0.1511	0.6212	3.14
F5	1350		16.6930	412.59	0.2478	1.0187	6.10
	1400		16.9530	410.95	0.1300	0.5344	3.15
	1455		15.0000	409.53	0.1141	0.4691	3.13
HBR		28	16.6540	431.31	0.5072	2.0852	12.52
F4	1455		16.8490	415.30	0.1096	0.4506	2.67
F5	1350		17.2550	410.11	0.0652	0.2680	1.55
	1455		17.7140	413.51	0.0510	0.2095	1.18
HBR		90	15.4430	431.75	0.3710	1.5252	9.88
F4	1455		18.7680	422.29	0.2011	0.8267	4.41
F5	1350		16.4310	418.15	0.1153	0.4740	2.88
	1455		15.9870	424.32	0.1623	0.6672	4.17

• Wi - initial weight; CH - calcium hydroxide (portlandite); a - in oven at 40 °C.

CONCLUSIONS

The results of this study showed that the thermal characterization of hydrated cements might be a useful technique for the study of cementitious materials. The major components (Al₂O₃, SiO₂ and CaO) required for clinkerization were identified in lime mud waste, WWTP sludge, and biomass ashes, which were used for the formation of clinker compounds (C₃S and C₂S). F1-F3 formulations had lower CaO contents and, as a consequence, the respective clinkers were belite-richer and had relatively small amounts of alite. Their hydration was slower and the formation of portlandite was incipient, being undetectable by DTA-TG. By contrast, F4 and F5 clinkers predictably showed alite as the main phase, as well as the reference material (HBR). Hydration reactions were faster, as proved by the stronger dehydration portlandite peaks. Major concerns regarding the eco-clinkers now processed are related to the higher amount of alkalis. Thus, this issue deserves a future characterization of the alkali phases formed and potential measures for mitigating their potentially harmful effects.

ACKNOWLEDGEMENTS

Authors are very grateful to National Council for Scientific and Technological Development (CNPq/Brazil) and Coordination for the Improvement of Higher Education Personnel (CAPES/Brazil) for funding this study.

REFERENCES

• ¹
  L. Simão, N. Lóh, D. Hotza, F. Raupp-Pereira, O.R.K. Montedo, in Proc. VI Forum Int. Resíd. Sól. (2015) 1.
• ²
  L. Simão, J. Jiusti, N.J. Lóh, D. Hotza, F. Raupp-Pereira, J.A. Labrincha, O.R.K. Montedo, Process Saf. Environ. Prot. 109 (2017) 106.
• ³
  Y. Jia, C. Maurice, B. Öhlander, Environ. Earth Sci. 72 (2013) 319.
• ⁴
  J. Qin, C. Cui, X. Cui, A. Hussain, C. Yang, S. Yang, Ceram. Int. 41 (2015) 5648.
• ⁵
  R. Sun, Y. Li, C. Liu, X. Xie, C. Lu, Chem. Eng. J. 221 (2013) 124.
• ⁶
  Y. Liu, C. Yan, Z. Zhang, H. Wang, S. Zhou, W. Zhou, Fuel 185 (2016) 181.
• ⁷
  M. Berra, T. Mangialardi, A.E. Paolini, Constr. Build. Mater. 76 (2015) 286.
• ⁸
  T.C. Esteves, R. Rajamma, D. Soares, A.S. Silva, V.M. Ferreira, J.A. Labrincha, Constr. Build. Mater. 26 (2012) 687.
• ⁹
  T. Ramos, A.M. Matos, J. Sousa-Coutinho, Constr. Build. Mater. 49 (2013) 343.
• ¹⁰
  R. Siddique, Resour. Conserv. Recy. 67 (2012) 27.
• ¹¹
  P.K. Mehta, P.J.M. Monteiro, Concrete: microstructure, properties, and materials, 3rd Ed., McGraw-Hill (2005).
• ¹²
  L.H. Buruberri, M.P. Seabra, J.A. Labrincha, J. Hazard. Mater. 286 (2015) 252.
• ¹³
  F. Castro, C. Vilarinho, D. Trancoso, P. Ferreira, F. Nunes, A. Miragaia, Int. J. Mater. Eng. Innov. 1 (2009) 74.
• ¹⁴
  P. Krivenko, R. Drochytka, A. Gelevera, E. Kavalerova, Cem. Concr. Compos. 45 (2014) 157.
• ¹⁵
  Am. Soc. Test. Mater., ASTM C 150-97 “Standard specification for Portland cement” (1997).
• ¹⁶
  N. Smaoui, M.A. Bérubé, B. Fournier, B. Bissonnette, B. Durand, Cem. Concr. Res. 35 (2005) 203.
• ¹⁷
  I. Odler, R. Wonnemann, Cem. Concr. Res. 13 (1983) 771.
• ¹⁸
  S.J. Way, A. Shayan, Cem. Concr. Res. 19 (1989) 759.
• ¹⁹
  I. Jawed, J. Skalny, Cem. Concr. Res. 8 (1978) 37.
• ²⁰
  H.F.W. Taylor, Cement chemistry, 2nd Ed., Thomas Telford (1990).
• ²¹
  A.M.G. Tavares, “Síntese e caracterização de pastas de cimento aditivadas com resinas epóxi - análises cinéticas, termodinâmicas e calorimétricas”, Tese Dr., Un. Fed. Sergipe (2010).
• ²²
  J.P. Baptista, R.F. Silva, Diagramas de fases, Aveiro (1998).
• ²³
  V.C.M. Barretto, “Resíduos de indústria de celulose e papel na fertilidade do solo e no desenvolvimento de eucalipto”, Tese Dr., UNESP (2008).
• ²⁴
  Assoc. Bras. Nor. Técn., ABNT NBR 5732, “Cimento Portland comum” (1991).

Publication Dates

• Publication in this collection
  Jul-Sep 2018

History

• Received
  16 Aug 2017
• Reviewed
  22 Oct 2017
• Accepted
  09 Dec 2017