NC State
Hubbe, M. A., Becheleni, E. M. A., Lewis, A. E., Peters, E. M., Gan, W., Nong, G., Mandal, S., and Shi, S. Q. (2018). "Recovery of inorganic compounds from spent alkaline pulping liquor by eutectic freeze crystallization and supporting unit operations: A Review," BioRes. 13(4), 9180-9219.


After the kraft or soda pulping of lignocellulosic materials to produce pulp suitable for papermaking, the spent pulping liquor typically has been recovered by multi-effect evaporation, followed by incineration in a recovery boiler.  This review article considers one unit operation, eutectic freeze crystallization (EFC), that may have potential to save some of the energy that is presently consumed in the evaporation step during recovery of inorganic chemicals from spent pulping liquor.  Based on a review of the literature it appears that EFC can be employed to obtain relatively pure sodium sulfate and sodium carbonate, along with relatively pure water (in the form of ice) from the spent liquor, under the assumption that lignin previously has been removed by acidification and precipitation.  Issues of inorganic scale formation, during the operation of an EFC process applied to lignin-free black liquor, will require research attention.  The chemical reactions to regenerate the active pulping chemicals sodium hydroxide and sodium sulfide from sodium carbonate, sodium sulfate, and other compounds isolated by EFC can be carried out either in a separate operation or by returning the materials to the feed of an existing recovery boiler.

Download PDF

Full Article

Recovery of Inorganic Compounds from Spent Alkaline Pulping Liquor by Eutectic Freeze Crystallization and Supporting Unit Operations: A Review

Martin A. Hubbe,*,a Emily M. A. Becheleni,b Alison E. Lewis,c Edward M. Peters,d Weixing Gan,e Guanzai Nong,f Sujata Mandal,g and Sheldon Q. Shi g

After the kraft or soda pulping of lignocellulosic materials to produce pulp suitable for papermaking, the spent pulping liquor typically has been recovered by multi-effect evaporation, followed by incineration in a recovery boiler. This review article considers one unit operation, eutectic freeze crystallization (EFC), that may have potential to save some of the energy that is presently consumed in the evaporation step during recovery of inorganic chemicals from spent pulping liquor. Based on a review of the literature it appears that EFC can be employed to obtain relatively pure sodium sulfate and sodium carbonate, along with relatively pure water (in the form of ice) from the spent liquor, under the assumption that lignin previously has been removed by acidification and precipitation. Issues of inorganic scale formation, during the operation of an EFC process applied to lignin-free black liquor, will require research attention. The chemical reactions to regenerate the active pulping chemicals sodium hydroxide and sodium sulfide from sodium carbonate, sodium sulfate, and other compounds isolated by EFC can be carried out either in a separate operation or by returning the materials to the feed of an existing recovery boiler.

Keywords: Kidney technologies; Pollution abatement; Black liquor; Sodium sulfate; Sodium carbonate; Indirect cooling crystallization; Incremental pulp mill capacity; Replacing the recovery boiler

Contact information: a: North Carolina State University, Department of Forest Biomaterials, Raleigh, NC, USA; b: Universidade Federal de Minas Gerais, Minas Gerais 31270-901, Brazil; c: University of Cape Town, Western Cape 7701, South Africa; d: KTH Royal Institute of Technology SE-100 44, Stockholm, Sweden; e: Guangxi University, Material Science and Engineering, Nanning, Guangxi, China; f: College of Light Industry Engineering, Guangxi University, Nanning, Guangxi, China; g: University of North Texas, Bioproducts Manufacturing, Mechanical and Energy Engineering Dept., Denton, Texas, 76207-7102; *Corresponding author:


Since about 1930 the pulp and paper industry has relied heavily on a process of evaporative concentration of spent pulping liquor followed by incineration under reducing conditions and recausticizing as a means of recovering pulping chemicals (Sjöström 1993; Bajpai 2017). The primary purpose of chemical pulping is to remove lignin from the woody material. As has been detailed in standard texts and review papers, the delignification can greatly enhance the value of the resulting fibers (Biermann 1996; Tikka 2008; Fardim and Tikka 2011). Chemical breakdown and dissolution of the lignin permits the fibers to be separated from wood without major damage or loss of fiber length. The pulped fibers generally have greater conformability, collapsibility (to form ribbon-like shapes), and more hydrophilic character of the surfaces, all of which contribute to stronger inter-fiber bonding (Young 1994; Fardim and Tikka 2011). The lignin-free fibers generally also have a more stable appearance, and they can be bleached to a high brightness with little tendency of yellowing (Dence and Reeve 1996; Suess 2010). In addition, delignified fibers can contribute superior absorbency to tissue, toweling, and highly absorbent disposable products (Parham and Hergert 1980; Lund et al. 2012; Hubbe et al. 2013).

The classical kraft recovery cycle, which depends on evaporation of water from spent pulping liquor (Venkatesh and Nguyen 1992), has been employed throughout the world as the primary means of converting spent pulping liquor back to its active form, suitable for reuse in the pulping process (Grace 1992; Empie 2009; Bajpai 2017). The capacity of a single recovery boiler can be as high as 6000 metric tons of dry solids in a 24 hour period (Haaga and Laitinen 2002). Parrish (1998) describes a range from about 340 (“small”) to 2700 (“large”) metric tons black liquor solids per day. Such a recovery cycle, in its time, has provided many advantages. The heat from the burning of low-valued lignin byproducts and some hemicellulose degradation products in the spent pulping liquor is used to supply the energy needed to evaporate water from the spent pulping liquor (Tikka 2008). Excess steam can be used for the drying process on paper machines. The conditions within the furnace can be controlled such that the sodium sulfate is reduced to sodium sulfide. Sodium carbonate within the mixture can be converted to sodium hydroxide by addition of lime (CaO), the active form of which is later recovered by use of a lime kiln (Venkatesh 1992; Tikka 2008; Empie 2009; Bajpai 2017).

In this article the term “recovery boiler” will be used when referring to the furnace, since the heat generated by the combustion of lignin and other organic materials is used to generate steam, which can then be used as a power source. Despite the many advantages and high level of maturity of kraft recovery technology, there are several reasons for the paper industry to consider either supplementing or replacing the recovery boiler and associated processes with something else. These reasons fall into the categories of hazards, capital expenses, operating expenses and problems, and the energy footprint. Recovery boiler explosions, though infrequent, cause major disruptions and sometimes loss of life when they occur at facilities throughout the world (Grace 1986; Burelle et al. 2016). Empie (2009) notes that there was an average of about two smelt-water explosions per year, somewhere in the world, during the period 1970 to 1995. A new recovery system of economically viable size can cost in the range of about 100 million to a billion dollars (Brewster 2007; Empie 2009; ETSAP 2015). Operation of such a system can have major expenses associated with corrosion of the high-grade metal tubing used to convert heat energy into steam (Singbeil and Garner 1989). The potential for severe scaling discourages the implementation of recovery boiler use in the pulping of such lignocellulose resources as bamboo and straw due to their relatively high content of silica (Hammett et al. 2001; Tutus and Eroglu 2003; Pekarović et al. 2006; Bajpai 2017). Evaporative removal of water, as used in the conventional kraft recovery technology, also has a relatively high energy footprint (Fernández-Torres et al. 2012); in other words, the usage of heat energy to evaporate water from pulping liquor can be regarded as wasteful based on current concepts of life cycle analysis. Alternatively, there is potential to employ EFC technology as a kind of kidney, selectively removing certain inorganic compounds while returning desirable inorganic compounds such as sodium sulfate to the recovery boiler system.



Among the technologies having potential to replace the evaporative concentration of black liquor, in the recovery system for kraft pulping, this review article focuses on eutectic freeze crystallization (EFC). The phase change from water to ice requires only about 15% as much energy as the corresponding transformation of water to steam (Osborne 1939; Osborne et al.1939). Water’s heat of fusion (at zero ºC) is 6.00 kJ/mole, whereas its heat of evaporation (at 100 ºC) is 40.65 kJ/mol. To put these numbers into perspective, the energy to raise the temperature of water from 0 to 100 ºC is 7.54 kJ/mole. Somewhat higher heat amounts are needed to change the temperature of salt solutions, depending on the details of their composition (Magalhaes et al. 2002). Even when taking into account such issues as inefficiencies of refrigeration processes and the need to cool hot water to the freezing point, there is clearly a motivation to consider EFC as a means of separating water from various industrial process water streams. When water freezes, the crystals tend to exclude impurities (Terwilliger and Dizio 1970; Stepakoff et al. 1974; Hasan and Louhi-Kultanen 2016; Szpaczynski et al. 2017). When saturated solutions of various compounds are super-cooled to a moderate degree, two solid phases – ice and an inorganic compound – will separate out of solution simultaneously. In a suitably designed device, the ice floats and the inorganic crystals settle, leaving behind an aqueous solution that is concentrated in ions associated with more-soluble inorganic compounds and also more concentrated in any organic compounds present in the initial mixture. In broad terms, EFC has been considered for treatment of relatively concentrated solutions, such as salty brines and the retentates from reverse osmosis membrane processes.

Though there is a real advantage, in terms of energy, when employing freezing rather than evaporation as a means to remove water from black liquor, it is important that such an advantage should not be over-stated. Typical evaporation systems are arranged as a series of about five to seven “effects”, each of which removes a corresponding portion of the water from the mixture (Tikka 2008; Bajpai 2017). A lot of the energy is saved because the water vapor obtained from black liquor evaporation in each effect is used as a heat source and recondensed at the subsequent effect. According to Bajpai (2017), one unit of steam is employed to evaporate 0.7 to 0.9 mass units of water in each effect. The overall economy, with about six units in series, has been estimated to be about 4.5 to 5 units of water evaporated per unit mass of steam (Tikka 2008; Empie 2009), though lower values (perhaps 3.6 to 3.8) may be more realistic. A variety of pumps and vacuum systems, all of which use power, have to be run in order to operate such an evaporator system. In the case of EFC, related cost-savings are possible too, since the ice can be allowed to melt in order to bring down the temperature of the stream entering the EFC system (Gerbasi et al. 1993). On the negative side, there can be inefficiencies associated with the needed cooling systems for EFC. The presence of impurities can lower the operating temperature, which may increase costs. Another concern is that an EFC process generally uses electrical energy, whereas the conventional kraft recovery process uses incineration of renewable organic matter as the main energy source. Based on the values given above, one still can expect a large energy advantage when using EFC in place of evaporative concentration before a recovery boiler. However, a detailed energy analysis of an optimized EFC system is needed, with an integrated strategy of heat recovery, and the results should be compared to a corresponding analysis for a typical 6-effect evaporator system for black liquor.

An early investigation of EFC was reported by Pangborn (1963), who showed that a combination of slow cooling and seeding with crystals was able to separate salt solutions into two solid phases – salt and relatively pure water. Stepakoff et al. (1974) showed that the EFC process could be optimized by initially separating just ice, until the eutectic point was reached, at which point both the salt and the water were at the point of incipient generation of new crystals from solution. Thereafter, conditions were optimized for more gradual isolation of ice and salt, followed by their gravity separation and washing. Barduhn and Manudhane (1979) showed that the EFC process could be optimized and run in a continuous manner. Advances in equipment by van der Ham et al. (1999) and Rodriguez-Pascal et al. (2010) achieved more efficient gravitational separation of the ice and salt crystals, since the ice tends to float due to its low density, and the inorganic compounds tend to settle, due to their higher density.

Applications of EFC

Application areas for which the EFC technology may be especially well suited appear to include not only spent pulping liquors from paper manufacturing, which will be the focus of this article, but also the handling of brines from reverse osmosis and ion exchange operations. Though it has been proposed to dispose of such brines by deep-well injection (Barduhn and Manudhane 1979; Afrasiabi and Shahbazali 2011; Kim 2011), such practices are wasteful of resources, expensive, and have the potential to cause contamination of groundwater. Desalination, which is widely practiced as a means of obtaining fresh water in arid regions, can benefit from EFC technology because the salt can be obtained as a relatively pure byproduct (Afrasiabi and Shahbazali 2011; Reddy et al. 2010). If the saline brine were merely discharged into the ocean, it would disrupt local salinity levels and interfere with marine life (Kwon and Yoon 2017). Contaminated brines from either reverse osmosis concentration or ion exchange treatment of industrial effluent streams have also been treated, at a research level, using EFC (Cob et al. 2014). Also, EFC can be used for the isolation of various chemical compounds (Chowdhury 1988).

For freeze-concentration in general, the most prominent industrial usage appears to be in the food industry (Liu et al. 1997; Aider and de Halleux 2009; Petzold and Aguilera 2009). In such applications the product has relatively high value, and the cold temperatures tend to minimize any decomposition. However, such systems do not involve eutectic mixtures, and no salts are precipitated from solution.

Some recent review articles dealing with the end-processing of brines either have mentioned EFC only briefly (Burbano and Brankhuber 2012; Subramani and Jacangelo 2014; Giwa et al. 2017; Tsai et al. 2017) or not at all (Khawaji et al. 2008; Elimelech and Phillip 2011; Kim et al. 2011; Morillo et al. 2014). Wiegandt and von Berg (1980) have attributed a lack of attention to freezing technologies to the widespread promotion of alternative technologies, for which more information has been provided to potential users. Chowdhury (1988) attributed the lack of investment in freeze concentration to capital costs, which were said to be as much as three times higher than simple evaporation and other crystallization techniques. Pronk et al. (2006) mentioned high operating costs in addition to capital costs. However, recently there have been reports of large-scale implementation of EFC (Creamer 2016; Chambers 2017; Lewis 2017, 2018; Steenkamp 2017). Therefore, in light of the high costs of conventional chemical recovery system for spent pulping liquors, the capital costs of EFC need to be evaluated.

Basic Principles of EFC

Theory related to the main steps of supersaturation, nucleation (where nuclei are born) and growth of crystals in solution has been reviewed (Mullin 2001; Lewis et al. 2015). The term supersaturation can be defined as a situation in which the solubility product of a specific mineral (e.g. NaCl) exceeds the value at which the crystal is known to be capable of being present as a solid phase. In industrial practice, a condition of supersaturation can be achieved by removal of water from a solution, e.g. by evaporation or other means. Nucleation means that new crystals appear, either due to high enough supersaturation, due to the nucleating effects of shear or surfaces, or in many cases due to the intentional feeding of “seed” crystals that match the composition and crystal habit of the desired product. The term crystallization, in the context of inorganic ionic matter, means that, although there may be imperfections, the substances that come out of solution are expected to be arranged in a highly regular pattern, often dictated by the sizes of the positive and negative ions that compose the compound. Finally, the term growth implies that material from the solution phase continually migrates to the surface of the solid particles, relieving the condition of supersaturation.

Nucleation has been shown to be a critical event during the EFC process (Rousseau and Sharpe 1980; Louhi-Kultanen 1996; Vaessen et al. 2003b; Petzold and Aguilera 2009; Peters et al. 2016; Quist-Jensen et al. 2017). The nucleation step has been found to depend on the degree of supersaturation and surface tensions between solid and liquid phases (Mullin 2001; Mersmann et al. 2002). There is always a competition between nucleation and growth. This process tends to transform a big number of small crystals into a smaller number of larger crystals. Vaessen et al. (2003b) observed the onset of nucleation during EFC, in a batch system, as a sudden presence of crystals throughout the crystallizer, with a simultaneous abrupt increase in temperature. According to Quist-Jensen et al. (2017), one of the keys to achieving favorable results with EFC is by establishing steady, relatively uniform conditions of nucleation and growth. In principle, the nucleation can occur either at a chilled surface, on the surface of seed crystals (see later), on impurities or particles that happen to be present within the mixture, or at the interface of droplets of chilled fluids that are being used for direct EFC systems.

Principles underlying the EFC process, as described in the literature (Barduhn and Manudhane 1979; Avram et al. 2004; Lewis et al. 2010; Becheleni et al. 2017), can be understood from the phase diagram in Fig. 1 (see Stepakoff et al. 1974). As shown, the phase equilibrium depends on both the temperature and the solute concentration. For solutions with concentration lower than the eutectic concentration, referred to as hypoeutectic solutions, sufficient cooling results in the crystallization of ice, which will co-exist with the solution. As more ice is removed, the solution becomes more concentrated, resulting in an increased freezing point depression. Thereby, the system tends towards the eutectic point, shown by the red dot. At this point, further cooling will cause simultaneous crystallization of both ice and salt. If the solute concentration is higher than the eutectic concentration, referred to as hypereutectic solutions, cooling will result in crystallization of salt, which reduces the solution concentration. Likewise, as more salt is removed, the system tends towards the eutectic point. When the process is carried out batchwise, the temperature of the system initially may be chilled to below the eutectic point, but upon nucleation of crystals, the temperature quickly rises and stabilizes near the eutectic point due to the release of heat of fusion (Himawan et al. 2006).

Fig. 1. Schematic diagram illustrating the phase behavior of mixtures of sodium sulfate and water, depending on concentration and temperature. The red circle corresponds to the eutectic point, as determined by Pronk (2007).

Under laboratory conditions it is possible to obtain salt crystals that are relatively free of other materials (Rodriguez-Pascual 2010). Hasan and Louhi-Kultanen (2016) determined that water obtained by the freeze crystallization process as ice contained only 38 ppm of nickel sulfate. Vaessen et al. (2003a,b) reported that the level of solute contamination (KNO3-HNO3) within ice from EFC was less than 5 ppm or 15 ppm in different cases. On the other hand, if high purity of the produced ice is not needed, then a higher level of supersaturation can be employed; such adjustments can allow for faster processing (van der Ham et al. 2004).

In principle, as the rate of growth of ice increases, it can become increasingly challenging to avoid entrapment of solute within the ice crystals. Especially in the absence of stirring, a concentration gradient will develop ahead of the advancing front of freezing ice (Burton et al. 1953; Terwilliger and Dizio 1970). In such circumstances, a high concentration of solute in the immediate neighborhood of the interface increases the likelihood of engulfment, as ice surrounds pockets of solution (Gerbasi et al. 1993). In addition, when the chilled crystallizer surface is sufficiently colder than the equilibrium freezing point of the adjacent aqueous solution, dendritic ice can form, and the growth rate of such ice can be two orders of magnitude higher than the rate of growth of a perfectly flat interface (Kapembwa et al. 2013). Butler (2002) reported the contamination of ice in a direct cooling EFC system with a certain type of antifreeze. In that case the ice contamination was attributed not only to a rapid advancing of the interface, but also to a crystalline similarity between the ice and the molecular structure of the solute. Also, relatively large molecules, having low rates of diffusion, will have a greater tendency to contaminate a growing crystal of ice in the absence of stirring.

Even in mixtures containing more than one solvent, various researchers have documented the recovery of relatively pure inorganic compounds from EFC processes. For instance, Rodriguez-Pascual et al. (2010) managed to separate pure ice and pure sodium carbonate from an industrial process sample. Vaessen et al. (2003a,b) isolated very pure potassium nitrate from a eutectic solution that also contained nitric acid.

Indirect Cooling EFC

Since there have been many publications dealing with so-called “indirect cooling eutectic freeze crystallization,” that type of system will be considered here first. The word “indirect” means that the system is cooled at a metal surface by circulating a coolant on the other side of the metal (Vaessen et al. 2002; Rodriguez-Pascual et al. 2010; Van Spronsen et al. 2010; Hasan et al. 2017a). An inherent advantage of such methods is that no coolant ever comes into contact with the processed materials, thus minimizing the chance of contamination. On the other hand, an indirect EFC system can suffer from problems related to build-up of ice layers and other scale on the chilled surfaces, in addition to ice and salt separation challenges, as described next.

Scale formation

Several authors have called attention to a problem of markedly reduced heat transfer following the development of a layer of ice on a cooled surface (Vaessen et al. 2002; Pronk et al. 2008; Hasan et al. 2017a). According to Hasan et al. 2017a), the formation of an ice layer is perhaps one of the key issues standing in the way of implementation of EFC technology at an industrial scale.

The rate of ice growth on a chilled surface can be complex, often involving a delay period before ice growth begins to interfere with EFC operations (Vaessen et al. 2003a). Vaessen et al. (2003a) noted that it can take several hours for ice to grow to a detectable level on a cooled wall of an EFC device. Hasan et al. (2017a) observed a time delay, before ice formation, that depended on both the temperature depression below the eutectic point and the rate of agitation. Delay times in the range of about 5 minutes to over an hour were observed. Engineers often have based their estimates of ice build-up on a square-root-of-time growth rate (Ashton 1989). For instance, Hasan and Louhi-Kultanen (2015) observed that the advancement of the thermal and concentration boundary layer at the ice-solution interface was proportional to the square-root of freezing time. However, based on inspection of data from several studies, the cited authors showed that the initial rate of ice growth is generally linear with time, and then there is a transition to a t0.5 dependency. As will be described later, the temperature plays a key role with respect to the onset of ice accumulation. The dynamics of ice growth during EFC has been predicted by simulation (Himawan et al. 2002; Lotfollahi 2009).


Several studies have been carried out to evaluate parameters related to the scraping of chilled surfaces during indirect EFC (Vaessen et al. 2002, 2003a,b; Roos et al. 2003; Genceli et al. 2005; Rodriguez-Pascual et al. 2010; Van Spronsen et al. 2010; Hasan et al. 2017a). For example, Vaessen et al. (2002) observed that either PTFE or silicon scrapers were effective in removing ice crystals from a chilled surface. The effect could not be attributed to mere refreshing of the surface due to flow. In the absence of using mechanical force, the ice built up after its nucleation.

Jooste (2016) investigated the effects of heat transfer driving force, scraper speed, solute type, and concentration of inorganic impurities on the induction time of ice scale formation, with longer induction times being favorable for the operation of an EFC process. An increase in heat transfer driving force resulted in a decrease in induction time due to the higher supersaturation generated at the wall, while an increase in scraper speed, solute concentration, and concentration of impurities resulted in an increase in the induction time of scale formation. From this, it can be generally inferred that real solutions will have lower scaling potential than pure binary systems due to the presence of contaminants in the system.

Little attention appears to have been devoted to smoothness or roughness of the chilled surface in the case of EFC. Hirata et al. (2000) studied the formation and removal of ice crystals on a chilled polyvinyl chloride (PVC) plate. Ice crystals that formed on the as-received PVC surface readily detached themselves. However, ice that formed on a PVC surface that had been sanded did not spontaneously separate from the PVC.

Separation by density

In principle, the ice particles and crystals of inorganic compounds formed during an EFC process can be separated by gravity. Such a separation can happen if both types of crystals are free from any chilled surfaces so that they are individually suspended in the saturated solution. For example, Rodriguez-Pascual et al. (2010) reported that only small amounts of sodium carbonate were dragged upwards with the rising ice particles and essentially no ice was dragged downwards with the settling sodium carbonate. Typical eutectic solutions of inorganic compounds have densities in the range of 1.1 to 1.2 g/cm3 (Chen et al. 1980; Vaessen et al. 2003b). Water ice, by contrast, has a density of 0.90 g/cm3 (Vaessen et al. 2003b), and various salt crystals have higher densities (Vaessen et al. 2003b). Specifically, the density of sodium sulfate, in its common decahydrate form (Thompson and Nelson 1954; Toghiani et al. 2008; Fernández-Torres et al. 2013), is about 1.464 g/cm3 (Becheleni et al. 2017).

Another key factor that affects separation by gravity is particle size. Studies of the EFC process have reported ice particle sizes in the range 50 to 200 m (Vaessen et al. 2003b; Reddy et al. 2010; Cob et al. 2014). Becheleni et al. (2017) observed ice crystals as large as 1 mm. However, it is clear that the ice particle size can be a function of such factors as the supersaturation at nucleation and the residence time (Chivavava et al. 2014). Increase in supercooling yielded larger ice particles. Himawan et al. (2002), in their simulation study, predicted sizes of ice particles in the range of about 50 to 500 m. The same authors predicted sizes of NaCl particles in the range of about 100 to 140 m. Reddy et al. (2010) and Peters et al. (2016) reported sodium sulfate decahydrate particles in the range of 20 to 400 m and 32 to 171 m, respectively.

If one assumes sphere-like particles having a diameter of 100 m, then the cited density values imply rates of ice floatation and Na2SO4 settling of about 6 cm/min and 12 cm/min, respectively, based on the Stokes equation (Batchelor 1967) and the viscosity of cold water,


where v is the terminal velocity, p is the density of the particle, f is the density of the fluid, is the shear viscosity of the fluid, g is the acceleration of gravity, and R is the particle radius. It is important to note that EFC may involve turbulent flow, due to the action of scrapers. Since the Stokes equation does not take such flows into account, the equation, taken alone, can predict only the net movements of particles relative to the local fluid surrounding them.

Because EFC can involve relatively high concentrations of particles in suspension, there can be deviations from the Stokes equation. Richardson and Zaki (1954) considered such systems in the case of uniform spheres. The equation that they derived can be regarded as a version the Stokes equation that contains a correction term. The correction was found to become significant for volumetric concentrations above about 5%.

The effectiveness of separation can also be affected by the morphology of the crystals, which depends on factors such as the degree of supersaturation and presence of impurities in the system. As discussed earlier, dendritic ice forms at high supercooling levels and tends to attach to crystallizer surfaces and also encapsulate some salt crystals, implying that the floating rate is impeded. Certain salts have been observed to crystallize in different shapes than their thermodynamically predicted shapes in the presence of impurities, which affects the settling rates of such particles (Buchfink 2011). Not only does morphology affect separation, but also it has an influence on the efficiency of downstream processes, such as washing and drying.

Hydrocyclones are known to be effective for separation of both light and heavy particles from suspensions (Rietema 1961; Narashimha et al. 2007). As reported by Stepakoff et al. (1974), hydrocyclones can be employed for effective separation of ice (which emerges from the large end of the hydrocyclone) and salt (which emerges from the small end of the hydrocyclone). The advantage of such a system is that the driving force for separation is greatly increased relative to ordinary gravitational acceleration. So although much of the ice and inorganic compound can be expected to separate themselves by gravity alone, there is potential to use hydrocyclones to divide the remaining mixture into one fraction that contains only salt (together with solution) and another that contains only ice (together with the rest of the solution).

Certain designs of EFC devices have been claimed to facilitate the gravity separation of ice and inorganic compounds, though the reasons often are not completely known. According to Vaessen et al. (2003b), their design, involving a stack of cooled disks, each provided with scrapers, exhibited different zones of rising ice and settling salt crystals.


Although gravity can be used to achieve a gross separation of ice and inorganic compounds during an EFC process, washing stages may be used to remove impurities from both solid phases (Stepakoff et al. 1974; Vaessen et al. 2003a,b; Chivavava et al. 2014; Cob et al. 2014; Becheleni et al. 2017). Stepakoff et al. (1974) employed a column in which both floatation and washing of ice were achieved. Peters et al. (2016) reported that washing of an antiscalant from sodium sulfate and ice was highly effective, suggesting that any impurities were mainly physisorbed on the outsides of crystals. Becheleni et al. (2017) removed essentially 100% of sodium sulfate from ice by use of three stages of counter-current washing with deionized water. However, it should be noted that optimal operational conditions in the EFC process have potential to enhance the washing efficiency. Contributions to better washing efficiency, by optimizing the EFC process, can include reduced incorporation of impurities into the solids and adjustments to achieve more effective separation of ice and salt. In other words, if the rising ice entrains large quantities of salt, then more washing is required to remove this salt from the ice with inevitable salt and ice losses.

Direct Cooling EFC

The words “direct cooling” imply that cooling is achieved by adding an insoluble refrigerant fluid to the system, which can then be recovered and recycled. The use of direct cooling in EFC systems was the subject of several studies (Stepakoff et al. 1974; Barduhn and Manudhane 1979; Chowdhury 1998; van der Ham et al. 1998). Chowdhury (1998) reported that butane, propane, and fluorocarbons, such as those used in refrigeration systems and air conditioners, can be suitable for use with EFC processes. Typically the refrigerant is injected into the system under pressure; upon absorption of heat energy, the refrigerant becomes vaporized and is recovered as a gas. As noted by van der Ham et al. (1998), direct cooling systems provide a way to avoid ice scaling. No more recent reports of direct cooling EFC systems were found in the present search of the literature.


Having considered, in the previous section, the principles of operation of EFC processes, the focus in this section will be on what has been published about tailoring or optimizing the process, with a focus on the type of separation that could be useful for recovering sodium sulfate, and possibly other solids, from spent pulping liquor.

Temperature Difference from Eutectic Point

Many studies have emphasized the importance of operating at low supercooling levels, that is, the temperature deviation from the eutectic temperature. Attention to the temperature of the coolant, so that the chilled surface is not excessively cooled, can prevent or at least reduce scale formation in an EFC process (Barduhn and Manudhane (1979; Vaessen et al. 2003a,b; van Spronsen et al. 2010; Chivavava et al. 2014; Kapembwa et al. 2014). Stepakoff et al. (1974) used a coolant temperature of -21 C for EFC processing of water containing NaCl, matching the eutectic point. Barduhn and Manudhane (1979) recommended that the coolant temperature of an EFC natural water system (Na, K, Ca, Mg, Cl, SO4, HCO3) be no lower than –25 C. According to these sources, depending on the eutectic point of the mixture, it is possible to choose a coolant temperature that is low enough to favor crystallization but not so low as to promote excessive scale formation on the chilled surfaces. Generally, low supersaturation favors steady crystal growth over nucleation, whereas high supersaturation results in domination of nucleation over growth, resulting in fine crystals and an impure product (Mullin 2001; Myerson 2002; Lewis et al. 2015).

Scraping Rate and Agitation

Some studies have varied the frequency with which chilled surfaces are scraped (Vaessen et al. 2002, 2003b). According to Vaessen et al. (2002) the way that scrapers impede the development of scale is by continually removing the initial crystals from which a contiguous layer of scale would continue to grow. The cited authors also found that the required rate of scraping tended to increase with increasing levels of supersaturation (decreasing surface temperatures) in the EFC system. An optimum scraping rate appeared to aid in the gravitational separation of the ice and salt crystals, with the development of separate zones for floating and settling (Vaessen et al. 2003b).

Some evidence suggests that at least part of the benefit of scraping action may be in reducing the boundary layer of stagnant fluid that tends to limit heat transfer at wetted surfaces. For instance, Vaessen et al. (2003b) stated that the fact that the heat transfer coefficient increased as the square-root of the scraping rate was consistent with the removal of a boundary layer (i.e. the penetration theory). The scraper action was reported to aid in convection (Vaessen et al. 2003a). Jooste (2016) also observed that increase in scraper speed delayed the onset of scale formation at constant heat transfer driving force. Direct agitation also can provide benefits. It was reported by van der Ham et al. (2004) that agitator use can minimize scale formation. Ahmad et al. (2017) reported benefits of agitator use in an EFC system; mechanical stirring, ultrasonic vibrations, and bubbling all were beneficial. Hasan et al. (2017a) noted that a higher intensity of agitation or a lower degree of cooling below the eutectic point both tended to delay the onset of scale formation on the chilled surfaces. When pushing the temperature of the chilled surface to relatively low levels, the benefits of agitation became pronounced. The heat transfer was increased.

Design of the EFC Unit

At the laboratory or pilot equipment scale, researchers have tested several design concepts in an attempt to make the EFC process work more effectively, focusing on indirect cooling by means of chilled surfaces. Table 1 categorizes such studies in terms of some essential design elements. Figure 2 (parts a, b, and c) provides some schematic sketches to show the essential geometries of the respective systems. The design used by Rodriguez-Pascual et al. (2010) incorporated an inverted cone at the top of the chilling chamber; this device cause the floating ice to drift to the edges of the chamber, where it could overflow and be collected.

Table 1. Design Concepts for EFC Equipment

A key theme that is common to the research work represented in Table 1 is that the scraping operation can be vital to effective performance of an EFC unit. Functions of the scrapers can include (a) repeatedly dislodging any crystals of ice or inorganic compounds from the chilled surface, which will favor further crystal growth in the bulk, rather than on the chilled surface; (b) physically encouraging the solids to move away from the chilled surfaces, such that separation according to density can take place; and (c) increasing of the thermal conductance near the interface of the highly conductive metal and the relatively low thermal conductance of a stagnant layer of aqueous solution (Himawan and Witkamp 2002; Vaessen et al. 2003a).

Fig. 2A. Schematic illustration of a cooled disk crystallizing unit for eutectic freeze crystallization, as described by van der Ham et al. (2004)

Fig. 2B. Schematic illustration of a cylindrical chilling module crystallizing unit for eutectic freeze crystallization, as described by Vaessen et al. (2003)

Fig. 2C. Schematic illustration of a concentric cylinders crystallizing unit for eutectic freeze crystallization, as described by Rodriguez-Pascual et al. (2010)

The orientation of the chilled surface is different in certain designs. The disc-type system pioneered by van der Ham et al. (1998, 2004) presents mainly horizontal chilled surfaces. From a practical standpoint, such a geometry would appear to be rather complex, providing possibilities for particles to either float or settle into positions where they could get stuck. By contrast, the designs that are here called “inner & outer wall of annular vessel” and “cylindrical chilling modules” present vertical walls, from which ice or inorganic compounds would be able to rise or fall without running into any obstructions. However, it would appear that there is an opportunity for further simplification of design, by incorporating the chilled surface as the outer wall of a cylindrical vessel, as illustrated in Fig. 3. Such simplicity would be obtained at the penalty for a lower area of chilled surface, in comparison to the system shown in Fig. 2B.

The reason that Fig. 3 depicts ice mainly rising in the central part of the crystallization unit is due to the mild centrifugal effect imparted by the rotating scrapers. Depending on the rate of rotation, it is likely that the centrifugal effect will be enough to impart a preference for sodium sulfate to settle near to the outside perimeter of the crystallization unit and for the ice to tend to channel upwards near the axis of the unit. A minor spray of water, from the top of the unit, can serve as a first step in removing inorganics from the ice particle surfaces.

Fig. 3. Simplified crystallizing unit for eutectic freeze crystallization

Avram et al. (2004) describe the innovation of tilting the orientation of an EFC device. Such a strategy was found to encourage the flow of ice upwards on one side of the device and downward flow of the crystals on the other side. Such convection was said to allow continuous operation at the laboratory scale.

Regarding the scraper, it is clear from the literature that researchers have been concerned not only about the durability of the scraper, but also the need to avoid damaging the chilled surface. Vaessen et al. (2003b) as well as Rodriguez-Pascual et al. (2010) employed high density polyethylene (HDPE) scraper blades, which appeared to be durable enough for their lab-scale work. In the publication of van der Ham and co-authors (1998, 2004), though no details of the scraper design are provided, the word “wipe” is used to describe their function, suggesting that the action could be described as gentle.



One of the great challenges of applying a technology such as EFC to the recovery of pulping chemicals from spent pulping liquor is that the systems are highly impure, containing a wide range of both organic and inorganic compounds. While it is likely possible to nucleate ice crystals within as-received pulping liquor, the tar-like character, due to the presence and enrichment of lignin and other organic compounds, can be expected to impede the separation of the ice and other newly formed crystalline materials, such as sodium sulfate.

Crude black liquor, immediately after it has been removed from the fibers in the brown-stock washers, typically has a solids content of about 15% (Tikka 2008; Kousini et al. 2016). According to Bajpai (2017), organic compounds typically comprise about 2/3 of the dissolved solids content of black liquor, and the other 1/3 are inorganic compounds. As noted by Empie (2009), the main inorganic components are sodium hydroxide, sodium sulfate, sodium carbonate, and any unreacted sodium sulfide. According to Kevlich et al. (2017), most of the reduced sulfur is likely to be present in black liquor as thiosulfate, e.g. Na2S2O3. Other inorganic components can include sodium chloride and various compounds related to silica.

Table 2 summarizes the types and amounts of inorganic compounds typically present in black liquor (Clayton et al. 1989; Niemelä and Alén 1999). The second and third numerical rows in the table provide estimates of the corresponding amounts of compounds following a hypothetical oxidation step (converting any reduced sulfur species to sulfate), or a combination of oxidation and neutralization of the pH. In addition, black liquor will contain poly-aromatic lignin byproducts, saccharidic compounds (mainly from the hemicelluloses), low mass organic acids, and extractives, which will include fatty acids and resin acids. The pH of typical black liquor is in the range 11.5 to 12.5. Bajpai (2017) notes that most of the carbohydrates solubilized in black liquor are monomeric, whereas lignin breakdown products have a wide range of molecular mass. It is also possible that a black liquor sample could contain NaCl (Kevlich et al. 2017); this is undesirable due to the corrosive nature of chloride toward stainless steel. Due to the high solubility of NaCl, it will tend to be among the last compounds to precipitate during EFC processing.

Table 2. Inorganic Composition of Typical Black Liquor (wt.% based on initial)

* data from Clayton et al. 1989 and Niemelä and Alén 1999; ** The higher number range for SiO2 pertains to grasses (Kevlich et al. 2017). See the Appendix regarding the assumed stoichiometry used to determine the values in the last two rows.

Lignin Removal

Though the topic of lignin removal from black liquor will be reviewed in more detail in a later article, some key aspects will be summarized here. It is well known that though lignin byproducts tend to be soluble at the very high pH values of spent pulping liquors, the solubility is markedly decreased when the pH is lowered. The reason to consider removing lignin by acidification, prior to the EFC process, is that it might precipitate out during EFC processing, thus contaminating the equipment. If it does not precipitate, it would get concentrated, thus raising the solution viscosity, hindering effective separation of ice and salt, and possibly contaminating the final products. Spent pulping liquor is known to become quite viscous as water is removed (Empie 2009), and the cold temperatures associated with EFC can be expected to favor a tar-like or solid-like character of the organic content of black liquor. Tikka (2008) lists marked increases in black liquor viscosity with decreasing temperature even when the lowest temperature considered was 90 C.

It has been well demonstrated that the majority of the lignin-related compounds in kraft liquor can be separated and recovered when the pH of the spent liquor is suitably reduced (Alén et al. 1979; Loutfi et al. 1991; Gilarranz et al. 1998; Stoklosa et al. 2013; Dominguez-Robles et al. 2016; Kousini et al. 2016). In particular, carbon dioxide, which is readily available from combustion processes at pulp and paper mills, can be used to reduce the pH to near-neutral or moderately acidic values, thus bringing about the precipitation of lignin (Alén et al. 1979; Wienhaus et al. 1990; Loutfi et al. 1991; Wallmo et al. 2009; Kousini et al. 2016). The use of carbon dioxide in this process presents great economic benefits due to its in-house availability. Under the hot conditions typically employed in such processing, the lignin phase can be separated and recovered as a water-insoluble liquid (Stoklosa et al. 2013).

Alternatively, the following cited studies represent considerable research work that has been carried out to concentrate lignin by ultrafiltration and related membrane methods (Holmqvist et al. 2005; Giummarella et al. 2016; Humpert et al. 2016; Kevlich et al. 2017). Relative to the just-described acidification and precipitation work, the use of membrane separation seems to have some important disadvantages. One of them is that the concentrated lignin remains in solution form; thus it is not truly isolated. The other problem is that other substances in the concentrate can include extractable materials (fatty acids, resin acids, unsaponifiable materials, etc.), as well as hemicellulose breakdown products. It is unclear whether such a mixture would have commercial value without further steps to isolate the components.

Optional Isolation of Extractives from Spent Pulping Liquor

In a conventional chemical recovery operation, in the course of evaporating the spent pulping liquor there is opportunity to separate out some of the relatively hydrophobic components that become released from the lignocellulosic material during pulping. The extractive fraction, which is collectively called “soap,” often comprises 30 to 50 kg per ton of black liquor solids (Foran 1992). This will include fatty acids, resin acids, and unsaponifiable materials (Pirttinen et al. 2007; Laxén and Tikka 2008). In a conventional recovery plant, such materials would be collectively called “tall oil” (Ren et al. 2016; Aro and Fatehi 2017). Typically, the separation is carried out after sufficient water has been evaporated from the black liquor to raise its dissolved solids content to about 20 to 30%; at that point the soap becomes insoluble and readily floats to the surface at near-boiling temperatures within a tank (Foran 1992; Aro and Fatehi 2017). According to Empie (2009) the lowest solubility of tall oil substances can be expected when the solids content is in the range of 25 to 28%. Crude soap skimmings will contain a wide range of components, including pulp fines and hemicellulose breakdown products (Ren et al. 2016).

Some of the extractive materials are likely to be precipitated along with the lignin, if the facility were to implement an acidification process (Stoklosa et al. 2013). According to Laxén and Tikka (2008), acidification of black liquor either with carbon dioxide or with other acids will tend to bring about precipitation of the fatty acids and resin acids. Any unsaponifiable compounds (neutrals) that have been emulsified by the soap compounds would then precipitate as well. Laxén and Tikka (2008) suggested that full separation of the tall oil fraction might require lowering the pH to a range of 2.6 to 3. However, as was noted earlier, there may be an opportunity to isolate a greater fraction of the extractives after an ultrafiltration operation before the black liquor pH is lowered; such treatment has the potential to retain most of the lignin, while allowing most extractives to pass through (Keyoumu et al. 2004; Manttari et al. 2015).

Foran (1992) noted that the separation of the extractive materials from black liquor can be promoted in various ways. If one depends only on gravity, then the equipment needs to be sized appropriately so that the slow creaming process can take place, making it possible to skim a floating layer of soap. Soap was reported as rising at a linear speed of about 1 to 6 meters per hour. Foran (1992) noted that the separation can be sped up by sparging the system with tiny air bubbles, leading to floatation of the hydrophobic droplets. In principle, one also could consider the use of hydrocyclones to speed up the density-dependent separation (Rietma 1961; Narasimha et al. 2007).

Already in 1998, McGinnis et al. had established that such separation of extractives could be enhanced by addition of surfactants. Pirttinen et al. (2007) showed that such a separation could also include neutral soluble species. Thus, the sterols, alkanols, and aliphatic compounds present in the pulping liquor will tend to be removed along with the more surface-active resin acids and fatty acids. Ren et al. (2016) were able to separate not only extractives from black liquor, but also hemicellulose byproducts. Kenny et al. (1992) employed dissolved air floatation (DAF) to separate resinous materials from the concentrate prior to freeze crystallization to separate ice from the mixture.

Even if steps are undertaken to remove extractives prior to EFC, it is reasonable to expect some level of those materials to enter the EFC process. Tests under realistic conditions will be needed to answer questions about the levels and effects of such compounds and to determine optimum conditions of operation.

Pre-concentration by Reverse Osmosis

Although EFC requires less energy to increase the concentration of solids in comparison to evaporative technologies (Fritzmann et al. 2007; Khawaji et al. 2008; Burbano and Brankhuber 2012; Morillo et al. 2014), in some cases even greater savings can be expected when using a combination of reverse osmosis (RO) followed by EFC (Stepakoff et al. 1974; Elimelech and Phillip 2011). RO has become the most widely used technology for desalination throughout the world (Elimelech and Phillip 2011). The RO process uses pressure to force a combination of water and solutes toward a very exclusive semi-permeable membrane, and only the water is able to pass through (Lee et al. 2011). Such a process is called “reverse” osmosis because the differences in electrolyte concentration on the two sides of the membrane ordinarily would induce flow in the opposite direction, i.e. a process of osmosis. One of the advantages of RO is that the process, if successfully implemented, can achieve highly pure water, having low ionic strength, and such water can have many uses within an industrial facility.

Past studies have shown that RO systems can be cost-effective for separation of relatively pure water from solutions that contain low to moderate concentrations of dissolved substances (Burbano and Brankhuber 2012). Various strategies to maximize the upper practical level of brine concentration that can be applied to an RO system have been discussed elsewhere (Burbano and Brankhuber 2012; Subramani and Jacangelo 2014). Eventually, however, one reaches a point at which the resulting brine is concentrated high enough to render further RO processing economically unattractive. The osmotic pressure that needs to be overcome to isolate additional fresh water approaches the capacity of the RO system. In addition, problems due to the fouling of RO membranes tend to become more severe at the limits of high pressure and high concentration of the brine. It has been estimated that RO can be used effectively to concentrate typical salts up to a concentration of about 7% (Tsai et al. 2017).

In the case of spent pulping liquor, even after most of the lignin has been removed by acidification and precipitation, the solution would most likely contain sodium sulfate, sodium carbonate, sodium thiosulfate, a variable amount of still-active sodium sulfide, hemicellulose decomposition products, wood extractives, and other inorganic ions including potassium, calcium, magnesium, manganese, iron, aluminum, oxalate, phosphate, chloride, and silicate ions (Venkatesh and Nguyen 1992; Engdahl et al. 2008; Empie 2009; Bajpai 2017).

EFC for Recovery of Selected Compounds

Sodium sulfate

Some of the major inorganic compounds that potentially can be precipitated from spent pulping liquor streams (possibly after removal of lignin, extractives, and some of the hemicellulose-related organics, as described above), include the sodium compounds sodium sulfate, sodium carbonate, sodium thiosulfate, sodium sulfite, sodium sulfide, sodium silicate, and sodium chloride.

Several authors have studied the eutectic freeze crystallization of sodium sulfate in the form of its hydrate with 10 parts of water (Thompson and Nelson 1954; Lewis et al. 2010; Reddy et al. 2010; Chivavava et al. 2014; Randall et al. 2014; Hasan et al. 2014; Hasan and Louhi-Kultnen; Hasan et al. 2017a,b; Becheleni et al. 2017). Some of the conditions of operations are shown in Table 3. As noted by Garrett (1998), the appearance and growth of crystals of sodium sulfate is a common occurrence in natural brine pools, especially when temperatures get cold in the winter. The ready formation of crystals of NaSO410H2O can be attributed partly to its lower solubility compared to other common salts such as sodium chloride. Thus, as shown in the table, the eutectic temperature for NaSO410H2O is only a little more than a degree below the freezing point of pure water.

The differences in the temperature at which sodium sulfate decahydrate was crystallized out can be attributed to different compositions of the starting solution. For a binary Na2SO4-H2O system, the eutectic temperature at which both ice and NaSO410H2O crystallize out is in the range -1.1 to -1.2 C. In the presence of impurities, the temperature can be different due to the effect of impurities on solubility and kinetics of the respective salt. It is known that the presence of common ions results in reduction of solubility, while some ions cause an opposite effect. The effect also depends on the concentration of impurities in the system. During the progress of a batchwise EFC operation, the temperature is expected to decrease when there is buildup of ions not participating in the crystal formation. For this reason, it is necessary to determine the relative nucleation temperatures of the salts from the spent pulping liquor to establish the energy requirements. This could be accomplished by thermodynamic modeling using software such as the OLI Stream Analyser and experimental verification.

Table 3. Conditions Employed for EFC Precipitation of Sodium Sulfate Decahydrate

EFC isolation of more than one salt

Several authors have written about the ability to utilize successive stages of EFC to separately precipitate different inorganic compounds, each in relatively pure condition (Lewis et al. 2009; Afrasiabi and Shahbazali 2011). From the standpoint of pulping technology, the ability to remove non-participating chemical species from the process stream offers a great advantage; in conventional chemical recovery system, employing evaporation and a recovery boiler, substances such as potassium can act as a dead load on the process, lowering the overall efficiency (Engdahl et al. 2008; Bajpai 2017). Such materials can be partially purged from a conventional chemical recovery operation at a pulp mill only by their disposal as dregs from green liquor clarification (Cornell 1992). In the early work of Thompson and Nelson (1954), sodium sulfate decahydrate was precipitated at -8.2 C, and this was followed by precipitation of sodium chloride at -22.8 C. Nathoo et al. (2009) confirmed the cited results and added a thermodynamic analysis of systems for the sequential isolation of sodium sulfate decahydrate and sodium chloride by means of EFC. Randall et al. (2011) separated both calcium sulfate (98% pure) and sodium sulfate (96.4% pure) from brine. van Spronsen et al. (2010) reported the simultaneous precipitation of two compounds, sodium carbonate and sodium bicarbonate. The needle-like shapes of the bicarbonate salt filled up the voids in the crystal bed, resulting in poor filtration and low product purity. Neilly et al. (2009) discuss earlier work by Arakel and co-authors, who sequentially precipitated sodium sulfate, then sodium chloride.

One of the complications of multi-salt systems is that the combined eutectic point can be markedly depressed. For instance, Reddy et al. (2010) observed a eutectic point of -21.2 C for Na2SO4 and water at high NaCl concentration, compared to a eutectic point of -1.24 C for the pure binary system containing just sodium sulfate and water. In this case, Na2SO4.10H2O crystallized first at about -10 C followed by ice at -21.2 C. If one wishes to crystallize Na2SO4 only, then the process will be maintained at the condition just before NaCl crystallizes, at the low temperature of -21 ºC.

In the case of the spent pulping liquor, the order in which salts are expected to crystallize out based on the eutectic temperatures of the binary salt-water systems published in literature is shown in Table 4. In each case, as a result of EFC processing, the concentrations of the ions associated with the material forming the crystals had risen to match the eutectic concentration. Table 4 shows the binary eutectic temperatures of the salts expected to crystallize. However, in the context of the spent pulping liquor, which is a multi-component solution, a thermodynamic model is required to determine the crystallization order of salts and their respective crystallization temperatures. In each case, as a result of EFC processing, the concentrations of the ions associated with a crystalline phase would have risen to its respective eutectic concentration. The presence of other salts and impurities can affect the nucleation temperatures of the salts.

Table 4. Anticipated Order of Crystallization of Salts Based on their Binary Eutectic Temperatures (van der Ham et al. 1999)

There is need to verify these values by conducting a thermodynamic model of the actual system composition, since the presence of other salts and impurities can affect the nucleation temperatures of the salts (Lewis et al. 2009). NaOH has not been included here since it has a binary eutectic temperature of about -25 ºC and the economic viability of recovering NaOH could be unfavorable. Also, if the pH has been lowered by treatment with CO2 to precipitate lignin, then the NaOH would already have been essentially converted to Na2CO3, as indicated in Table 2.

EFC precipitation of sodium carbonate

Garrett (1998) notes that sodium carbonate decahydrate (natron) is similar to mirabilite with respect to having a steep solubility curve, i.e. increasing the temperature does not rapidly increase its solubility. Thus it makes sense that natron can be the next main salt form to be precipitated in a sequential EFC process with a mixed brine. Rodriguez-Pascual et al. (2010) separated sodium carbonate and bicarbonate from aqueous solution by EFC. Coolant temperatures between -10 and -17.5 C were employed, resulting in an operational temperature of -3.8 C. The crystallizer was able to operate at a solids level as high as 36% by mass. Growth rates of ice were in the range 1 to 2 x 10-8 ms-1, and those for sodium carbonate decahydrate were 3.5 to 8.3 x 10 -9 ms-1. van Spronson et al. (2010) extended the research at a larger scale and demonstrated a heat transfer rate of 5 kWm-2. The preferred temperature of operation was -3.8 C, which was just high enough to avoid co-precipitation of the sodium bicarbonate.

EFC precipitation of sodium chloride

Because of its relevance to desalination operations, there have been several studies focusing on the EFC precipitation of NaCl (Stepakoff et al. 1974; Cob et al. 2014; Hasan et al. 2014; Ahmad et al. 2017). Such a process is highly relevant to pulp and paper industry applications because the chloride ion is corrosive to stainless steels, and it could be helpful to remove such ions from the system. Table 5 shows the temperatures that have been used for EFC precipitation of NaCl. The economic feasibility of operating at such very low temperatures on an industrial scale needs to be assessed.

Table 5. Conditions Employed for EFC Precipitation of Sodium Chloride

EFC precipitation of other inorganic compounds

Though sodium sulfate and sodium carbonate (as discussed above), are of particular relevance to the recovery of pulping chemicals, other compounds also have been isolated by EFC. These include calcium sulfate (Randall et al. 2011, 2013); magnesium sulfate (Genceli et al. 2005), nickel sulfate (Hasan and Louhi-Kultanen 2016), copper sulfate and mono-ammonium phosphate (van der Ham et al. 1999), sodium nitrate (van der Ham et al. 1998), and a mixture of potassium nitrate and nitric acid (Vaessen et al. 2003b).

Scale Management


Because the chilled metal surface of an EFC precipitation cell is the coldest point accessible to the suspension, it follows that generally there is a high likelihood that either ice or the inorganic compound or both will tend to scale the surface. Peters et al. (2016) investigated the effect of a phosphonate-based anti-scale agent at different concentrations on the nucleation and growth rates of both ice and Na2SO410H2O from a synthetic sodium sulfate solution. Both the ice and the salt precipitated during the EFC process exhibited measurable contamination by the antiscalant, though most of the antiscalant could be effectively washed off. Such washing would be attractive from the point of view of obtaining products of high purity. Generally, the nucleation rates of ice and salt were reduced in the bulk suspension. The study did not assess the effectiveness of the antiscalant in reducing scale formation on the crystallizer walls, but it can be generally inferred that antiscalants could reduce nucleation of scaling compounds on the chilled surface.

Ion exchange softening

A promising way to lower the likelihood or rate of problematic scale formation during EFC and related processes is to pre-emptively remove ions that have a propensity to form scale (Barduhn and Manudhani 1979; Subramani and Jacangelo 2014). Water softening, using ion exchange resins, is well known to be an effective way to remove Ca2+ and Mg2+. Such ions, which contribute to water hardness, are associated with a variety of inorganic compounds that cause serious scaling problems in industry. Examples include calcium sulfate, calcium oxalate, calcium carbonate, and calcium phosphate (Gal et al. 1996; Severtson et al. 1999; Yu and Ni 2006; Du et al. 2009).

Scale formation also can be minimized in some cases by chemical compounds that react with problematic anionic species, for example by adsorption or precipitation. For instance, Burbano and Brankhuber (2012) reported that activated alumina can be used to remove silica from water, prior to EFC processing. Empie (2009) and Bajpai (2017) discuss several technologies that have been implemented in the recovery boiler system to remove silica; such approaches also could be considered as a pretreatment before EFC processing.

Temporary reheating

Temporary heating of the chilled surfaces of an EFC device has been shown to be an effective means of periodically eliminating any build-up of the main solids being precipitated (van der Ham et al. 2004). According to the cited authors, such heating can quickly release accumulated build-up, since the local conditions momentarily favor dissolution of both ice and the inorganic compound. Subsequently the EFC process can be started again with no apparent adverse consequences. Such a system could be optimized with adequate and effective process control to ensure that the instantaneous heating is not excessive, thereby disturbing the system’s steady state. Kenny et al. (1992) likewise employed a “rapid heating” cycle to overcome problems related to ice buildup on chilled surfaces. The downside of such a strategy is that it will destabilize the crystallization operation. Every new cycle may need to be seeded, and a new steady state will need to be established.

Fluidized bed

By recirculating suitable insoluble particles within an EFC system, additional surfaces become available for potential nucleation of either ice or other inorganic compounds. Such an approach has been evaluated in the case of EFC (Pronk et al. 2006; Pronk 2007; Pronk et al. 2008). The cited work showed that the use of a fluidized bed approach was able to reduce the growth of an ice scale on the chilled wall during EFC. Steel balls were reported to remove crystals from the walls of the chilling device, and it was not necessary to utilize scrapers (Pronk et al. 2006). It also was reported that the fouling of the chilled walls was more severe under eutectic conditions, when both ice and salt were being formed, compared to when just one solid phase was being formed.


In light of the concerns expressed in publications about scale formation in EFC systems, it makes sense to examine procedural steps that tend to avoid or minimize scale buildup. In principle, it is less likely for crystals of either ice or inorganic compounds to form on a chilled surface if there are numerous crystals of the same type already in suspension (Rousseau and Sharpe 1980). Thus, rather than nucleating on the chilled surface, the ice or other compound merely builds upon the surfaces of the respective crystals, increasing their size.

Several authors have reported the use of seeding as part of an EFC process (Stepakoff et al. 1974; van der Ham et al. 1999; Randall and Lewis 2009; Reddy et al. 2010; Peters et al. 2016; Becheleni et al. 2017). Chivavava et al. (2014) chilled a sodium sulfate solution to -2 C and then induced nucleation of the supersaturated solution by adding crystals of ice. One can envision a system in which a part of the ice and inorganic crystals could be crushed and recycled as a means of greatly increasing the number and surface area of particles available for nucleation, without exceeding the optimum solids density beyond which the separation of ice and salt is interfered with.

Main Alternatives to EFC

While the main focus of the present article has been eutectic freeze crystallization (EFC), there are some related technologies that merit mention, if only to clarify how they differ from EFC. As noted by Subramani and Jacangelo (2014), there have been three main approaches used to minimize the volume of brines that result from reverse osmosis (RO) operations, and these are membrane-based, thermal based, and “emerging”. Of these three, the membrane-based approaches merely recover water and generate more concentrated brine, without any full resolution of the disposal problems associated with brines. Thermal approaches include, essentially, either evaporation or freezing as a means of separation of the phases. However, the two freezing-based approaches of EFC and membrane distillation are regarded by the cited authors as “emerging”.

Evaporative crystallization

In terms of capital equipment and the maturity of the technology, it may appear to be advantageous just to crystallize out salts by evaporation (Nathoo et al. 2009). However, in the cited work it was shown that, to achieve the same goals, EFC has the potential to reduce operating costs by between 80% and 85% relative to such a process. Further work will be needed to be able to make realistic estimates of energy savings. Attention needs to be paid to energy-saving methods, including multiple-effect evaporation. The cited authors also suggested that the relatively low capital cost of equipment for evaporative crystallization is consistent with its long period of scale-up and optimization within desalination plants around the world. Since EFC is an emerging technology, it is expected to undergo a similar scale-up and optimization cycle with significant cost reductions.

Membrane distillation and crystallization

The term “membrane distillation” refers to systems in which vacuum is applied to the accepts side of a membrane, allowing the liquid phase to evaporate and later be recondensed as a relatively pure liquid (Afrasiabi and Shahbazali 2011). According to Pérez-González et al. (2012), such operations are often carried out at 60 to 80 C with an energy consumption of 15 to 20 kW hm-3. The equipment has a relatively small footprint, which can be important, especially when installing operations in an existing facility. Based on the cited accounts, it appears that most of the work related to membrane distillation has been focused on systems in which the main goal has been to obtain relatively pure water and where there is less interest in whether or not inorganic compounds can be obtained, especially in their pure form. Fouling is expected to be a key challenge to the successful implementation of membrane distillation systems (Afrasiabi and Shahbazali 2011). A design has been proposed recently for membrane distillation crystallization that takes into account the solubility curve of the crystallizing compound, so as to reduce fouling and energy consumption (Pantoja et al. 2015, 2016).

If membrane distillation is continued to the point at which the concentration of dissolved substances reach and begin to exceed their solubility, then there is potential to precipitate solids within the brine. Some recent publications have described such an approach in detail (Chen et al. 2014; Quist-Jensen et al. 2017). Because the process of evaporation generally causes the temperature of a system to decrease, it is conceivable that membrane distillation and crystallization carried out by application of vacuum might entail conditions that could match those of EFC systems. However, Pérez-González et al. (2012) state that membrane distillation operations are typically carried out at ambient pressure in the temperature range of 40 to 50 C. Since the phase change involved in evaporation requires much more energy than the phase change involved with freezing, it seems unlikely that such membrane distillation methods will be able to approach the energy-efficiency of EFC. Another important difference is the dependency on use of a membrane, making the process susceptible to fouling. Future research, including scale-up efforts, will be needed in order to determine the relative seriousness of such challenges, especially in comparison to the challenges of scaling, throughput, and phase separation inherent in the EFC technology.

Curcio et al. (2010) describe an alternative approach in which membrane crystallization is combined with nanofiltration. Such as approach was evaluated for the removal of sodium sulfate from the effluent of a battery factory. A membrane crystallizer, positioned after the nanofiltration operation, produced the anhydrous form of sodium sulfate.

Anaerobic Treatment of Process Water

If some or all of the procedures already mentioned in this article were successfully combined together, bypassing usage of a furnace for the recovery of pulping chemicals and minimization of discharges to the environment, it is likely that there still would be an excessive level of contaminants in the liquid left over after removal of sodium sulfate, sodium carbonate, and possibly sodium sulfide, using a suitable number of stages of EFC processing. A major class of such contaminants is likely to be the organic breakdown products of hemicellulose. As has been reviewed recently (Hubbe et al. 2016), one of the most cost-effective ways to deal with such contaminated water is by use of an anaerobic reactor system. The up-flow anaerobic sludge blanket type of reactor is especially to be recommended, since it can readily convert hemicellulose-related organic compounds to methane, which is useful as a fuel. Since the bacteria responsible for anaerobic decomposition are often adversely affected by the inorganic components in the water (Isik 2004; Aslan and Sekerdag 2016), it is likely that the use of EFC to remove ionic compounds will have a positive effect on subsequent anaerobic digestion of the remaining material. Conditions of salinity may need to be optimized so that the combined system of EFC and anaerobic digestion works effectively. Although the contaminated water also could be treated using an aerated (aerobic) wastewater treatment system, such options tend to take up more space, produce more sludge, and there is no fuel byproduct. However, the details of potential upstream processing such as biological treatment techniques will not be dealt with in this article.


Reactivation of the Pulping Chemicals

Return of sodium sulfate to a recovery boiler

A great advantage of conventional technology, employing a Tomlinson boiler for the recovery of pulping chemicals, is the fact that sodium sulfate can be converted to sodium sulfide in a reducing zone of the furnace (Cameron and Grace 1983, 1985). Such an option can be considered in a hypothetical case where EFC has been used to separate Na2SO4 and Na2CO3 from pulping liquor. For instance, there might be a pulp and paper mill in which EFC has been installed in order to overcome limitations in the capacity of an existing recovery boiler system. In such as case, the Na2SO4 from EFC could be fed with ordinary black liquor to the furnace. The relatively dry and pure condition of the Na2SO4 would tend to be favorable for such systems, since the heat load would be low and one could avoid the recycling of non-process elements such as potassium to the recovery cycle (Bajpai 2017). Trials would need to be carried out to achieve efficient and safe operation of a conventional chemical recovery boiler that incorporates the processing of extra Na2SO4 that is being isolated by EFC.

Separate conversion of sodium sulfate to sodium sulfide

One of the challenges inherent in replacing the conventional recovery process with something else is that changes will need to be made in how the sulfate becomes reduced to sulfide. In principle, reduction can be carried out at high temperature using different reducing agents, including hydrogen (Birk et al. 1971) and carbon monoxide (Li and van Heiningen 1988, 1995). The mechanism by which reduction takes place in the described systems can involve adsorption of sulfate at an interface. According to Cameron and Grace (1983, 1985), the mechanism of conventional recovery in a Tomlinson boiler involves adsorption of sulfate onto carbon within the hearth, where the reduction takes place. The reduced sulfur species then desorb. Li and van Heiningen (1988) found evidence of a similarly catalytic and surface-depending reaction (on a TiO2 surface) when using CO as the reductant. Though the cited literature describes some procedures to carry out such reductions, it is unclear whether systems based on hydrogen or carbon monoxide would be cost-effective relative to conventional processing, in which carbon essentially acts as the reducing agent.

Electrolytic reduction of Na2SO4 and causticization of Na2CO3

The technologies just described, in which sulfate is reduced by adding carbon, CO, or hydrogen, all involve equilibrium reactions. By contrast, Wartena et al. (2000, 2002a,b) have investigated an electrolytic approach that can drive the reductive reaction essentially to completion. The process is carried out with a molten mixture of sodium sulfate and sodium carbonate, a combination that corresponds to what could be obtained using stages of EFC processing. In addition, any sodium carbonate present is simultaneously converted to sodium hydroxide, making it unnecessary to employ a lime kiln. Carbon dioxide is emitted as a byproduct. High conversion efficiency was observed at an operating temperature of 860 C. According to Wartena et al. (2002a), a key to achieving success of this technology would be to provide a barrier between the cathode and anode, such that the reduced sulfur compounds are not re-oxidized by oxygen that is released from the other electrode.

Scale-up Considerations

As was noted earlier, recovery boiler systems tend to be large, and there has been no good way to increase the capacity of an existing boiler if a pulp mill wishes to expand its production. Some gains in production, using existing recovery systems, have been achieved by precipitating lignin from black liquor (Alén et al. 1979; Kousini et al. 2016). Though the reduction in burnable material being returned to the boiler is reduced, thereby decreasing the generation of energy, such strategies can permit a higher production of pulp at the facility.

A convenient aspect of EFC technology, at least as far as it has been developed to this point, is that the units are small – mainly scaled to lab and pilot-scale evaluations. Accordingly, one might consider pilot-scale trials in which EFC units are run along-side of an existing conventional chemical recovery operation. If such trials were successful, then the EFC unit could have potential to decrease the amount of evaporation needed in the recovery operation. More EFC units could be added incrementally, depending on the need for more chemical recovery capacity. A potential side-benefit of such a system is that the EFC could function as a type of kidney, ideally returning only the valuable sodium sulfate, sodium carbonate, and possibly the reduced sulfur compounds to the cycle, while excluding such materials as potassium and manganese.

Concepts for a Combined Process

Preparative steps before EFC

Based on principles, materials, and unit operations discussed in earlier sections of this article, now it is possible to consider some possible scenarios by which EFC might be incorporated into a system to deal with black liquor. Figure 4 shows a possible sequence of steps that might precede EFC. In the envisioned process, the first step is an optional oxidation of the highly alkaline, hot black liquor. Such a treatment has been employed industrially as part of a lignin precipitation process in order to avoid later problems associated with reduced sulfur gases (Kouisni et al. 2016). The next step involves neutralization of the highly alkaline pH of the spent pulping liquor, thereby causing most of the lignin byproducts to precipitate from solution. Although such a step can be regarded as being unconventional, it has already been implemented in a few pulp mills as a means to achieve greater throughput of pulp production (Gellerstedt et al. 2013; Kouisni et al. 2016). Along with the lignin, some extractive materials can also be present in the precipitated material at this stage. The figure assumes that a system can be optimized to separately recover such extractives by a skimming process, while the greater density of the lignin phase would cause it to settle.

Fig. 4. Possible sequence of preparatory steps to remove components from spent pulping liquor in preparation for an EFC recovery of sodium sulfate and sodium carbonate

The next step shown in Fig. 4, reverse osmosis (RO), can be considered as a potentially cheaper way to separate relatively pure water from the aqueous mixture, depending on the concentrations of salts. It should be noted that RO operation may be subject to problems related to membrane fouling; relatively highly concentrated aqueous electrolytes and organic compounds can be expected to be in solution even after lignin removal. Evaluations could be carried out to optimize the target level of solids in the brine, before it is sent to EFC. On the other hand, if severe problems of membrane fouling are encountered with RO, there would be an option to omit the RO step in favor of complete reliance on freeze crystallization as a means of separation of water (as ice) from the other components of the mixture while concentrating it towards the eutectic conditions, at which point the operation becomes fully EFC.

The “softening” step listed in Fig. 4 implies usage of some sort of ion-exchange material. Higher-valence metal ions, especially Ca2+ and Mg2+, would replace Na+ ions on such resins. The objective of including such a step is to avoid problematic scaling of EFC equipment with such compounds as calcium carbonate, calcium oxalate, calcium sulfate, and calcium silicate. Such compounds also would have the potential to co-precipitate and contaminate the desired Na2SO410H2O and Na2CO310H2O during EFC operations. The topic of water softening has been well reviewed (Wachinski 2017).

After the EFC step (which will be discussed next), the last operation indicated in Fig. 4 is an anaerobic digester. The concentrated liquid phase left over after all of the EFC stages can be expected to contain a variety of water-soluble organic compounds. Such an aqueous mixture would be a good candidate for anaerobic digestion (Hubbe et al. 2016). The anaerobic digestion operation would generate methane, which can be used as a fuel or sold.

Multiple stages of EFC

Figure 5 envisions four potential stages of eutectic freeze crystallization. See Fig. 3 and the adjacent text for more detail regarding how such an EFC unit could incorporate scraping units and ice-washing operations. Due to the presence of other salts, it is not certain whether ice will precipitate when the temperature is cooled to near the eutectic point for Na2SO410H2O in the presence of other ions. The diagram shows a process of crushing or grinding a small portion of the precipitated Na2SO410H2O so that it can be immediately used as seed crystals, thus encouraging the formation of relatively pure Na2SO410H2O.

Fig. 5. Possible stages of eutectic freeze crystallization for continuous or batchwise separation of sodium sulfate decahydrate, sodium carbonate decahydrate, and possibly sodium sulfide from an aqueous concentrate (see previous figure)

Though Fig. 5 shows just four EFC units, it should be understood that the size and number of parallel units would need to be set according to the required production capacity and residence time in each unit. Though continuous operation is implied in the diagram, batchwise processing also can be considered. In a continuous operation, the liquid portion collected from the unit could pass to a subsequent unit, within which the temperature would tend to be lower, depending on the concentrations of ions present and the eutectic condition for the next compound to be collected.

The second EFC unit shown in Fig. 5 represents the precipitation of Na2CO310H2O. Though ice is shown in the figure as being generated from this unit, such ice precipitation would depend on the concentration of other electrolytes present. The feasibility of EFC to isolate Na2CO310H2O is supported by previous work (Rodriguez-Pascual et al. 2010).

No account was found in the literature for the precipitation of sodium sulfide or sodium thiosulfate by EFC; thus the third unit shown in Fig. 5 can be regarded as speculative. Studies need to be carried out to determine whether or not Na2S or Na2SO3, probably in a hydrated form, can be isolated by EFC. This is an important issue, since not all of the reduced forms of sulfur will have been consumed and converted to the sulfate form in a typical pulping cycle (Kevlich et al. 2017).

The final EFC unit depicted in Fig. 5 assumes that highly soluble inorganic compounds such as NaCl can be removed all together at a sufficiently low temperature, representing the eutectic point of the remaining mixture. This would allow the subsequent dry disposal and/or recovery and use of the precipitated salt mixture.

Reduction of the sodium sulfate and causticization of the sodium carbonate

Figure 6 depicts three main options (numbered 1, 2 and 3) for the conversion of sodium sulfate and sodium carbonate to the active forms (sodium sulfide and sodium hydroxide) required for pulping (i.e. white liquor). The options include (1) conventional processing in a recovery furnace, (2) separate reduction of the sodium sulfide, and (3) electroreduction. As stated earlier, if there is currently a recovery boiler in operation, then it makes sense to determine whether the additional sodium sulfate and sodium carbonate obtained from EFC can be just added into the black liquor fuel stream entering the boiler. The reduction of the sodium sulfate takes place within the reducing zone of the furnace, whereas the transfer of the alkalinity from sodium carbonate to sodium hydroxide (Cornell 1992) requires the use of a lime kiln (Venkatesh 1992) (see chemical process at the lower right within the figure).

Fig. 6. Options for reduction of sodium sulfate and causticizing of sodium carbonate

If there is no recovery boiler system, or if such a system is not able to handle extra materials obtained by EFC, then a separate reduction of the sodium sulfate could be considered. Though, as discussed earlier in this article, it is possible to use hydrogen or carbon monoxide as reducing agents for such a step (Birk et al. 1971; Li and van Heiningen 1988, 1995), it seems likely that carbon would still be the most cost-effective reducing agent. This line of reasoning is consistent with the running of a separate furnace that might not be much different from a conventional recovery system, except that the input material (sodium sulfate, without lignin, carbohydrates, or other inorganic materials) could be much purer. It is uncertain what would be employed as the main fuel in such a system. Due to the lack of recent research dealing with the reduction of sodium sulfate in a carbon-based furnace system, it seems unlikely that such an approach would be commercialized in the foreseeable future.

The middle option of electroreduction shown on the right within Fig. 6 appears to be the most likely scenario in cases where a recovery boiler is not an option. As is apparent from the term “electroreduction,” the envisioned process involves the use of electricity to bring about the desired chemical transformation (Wartena et al. 2000). As noted earlier, the advantages include essentially complete conversion of the sodium sulfate and sodium carbonate to sodium sulfide and sodium hydroxide, respectively. Also, there is no opportunity for smelt-water explosions.

Incomplete conversion of sodium sulfate to sodium sulfide in a conventional recovery furnace renders such processes much less efficient. According to Empie (2009), about 5 to 10% of the sodium sulfate fails to be reduced in a typical recovery plant at a pulp mill. Such unconverted sodium sulfate goes through the subsequent cycle of pulping as a “dead load”. Likewise, in a typical lime kiln operation about 15 to 20% of the slaked lime (Ca(OH)2) fails to be converted to NaOH. So along with such ionic materials as calcium, magnesium, iron, manganese, silicate, chloride, phosphate, and sulfate, the sodium sulfate and sodium carbonate will circulate in the reactivated (white) pulping liquor without making a contribution to pulping. Such considerations provide further incentive for researchers to run more evaluations of electroreduction (Wartena et al. 2000).

Whether or not electroreduction can ever achieve the cost-effectiveness of the conventional recovery systems used in the pulping industry, it might be ideal for relatively small-scale operations, filling a niche in cases where the construction of an economically viable recovery plant for spent pulping liquor cannot be justified.


The scenario outlined in this review article explores an important area of potential innovation for the pulping industry. As noted in the Introduction, there is potential for reduced costs as well as for increased production rates, depending on how the described technology might be implemented in existing pulp mills. Some theoretical information, as well as some technological experiences have been documented here, hopefully providing a sense of how a system could be put together to achieve some of the many simultaneous objectives in a pulp mill that needs to process black liquor and regenerate pulping chemicals. Parts of the envisioned process have been demonstrated, though some of the demonstrations have been at a relatively small scale and with less complex starting mixtures compared to genuine black liquor. Thus, it can be expected that a lot of practical experimentation will be needed, involving a range of scientists and engineers, to move the technology forward toward industrial practice.


The work of Martin A. Hubbe is supported by the Buckman Foundation, which supports papermaking chemistry research. The authors also warmly thank the following volunteers who read the text and indicated items in need of correction or clarification: Dr. Marjatta Louhi-Kuntanen, Aalto University, School of Chemical Engineering, POB 16100, FI-00076 Espoo, Finland; and Marcello M. Seckler, University of Sao Paulo, Department of Chemical Engineering, Polytechnical School, Av. Prof. Luciano Gualberto, Trav 3,380, BR-05508010 Sao Paulo, SP, Brazil. The authors also thank Raimo Alén of the University of Jyväskylä, Finland, for input related to the inorganic composition of black liquor.


Afrasiabi, N., and Shahbazali, E. (2011). “RO brine treatment and disposal methods,” Desalin. Water Treat. 35, 39-53. DOI: 10.5004/dwt.2011.3128

Ahmad, M., Al-Wazzan, Y., Al-Odwani, A., and Thomas, J. P. (2017). “Solid layer freeze crystallization processes for desalting saline waters,” Desal. Water Treat. 66, 60-71. DOI: 10.5004/dwt.2016.11141

Aider, M., and de Halleux, D. (2009). “Cryoconcentration technology in the bio-food industry: Principles and applications,” LWT – Food Sci. Technol. 42(3), 670-685. DOI: 10.1016/j.lwt.2008.08.013

Alén, R., Patja, P., and Sjöström, E. (1979). “Carbon-dioxide precipitation of lignin from pine kraft black liquor,” TAPPI 62(11), 108-110.

Aro, T., and Fatehi, P. (2017). “Tall oil production from black liquor: Challenges and opportunities,” Separation Purif. Technol. 175, 469-480. DOI: 10.1016/j.seppur.2016.10.027

Ashton, G. D. (1989). “Thin ice growth,” Water Resour. Res. 25(3), 564-566. DOI: 10.1029/WR025i003p00564

Aslan, S., and Sekerdag, N. (2016). “Salt inhibition on anaerobic treatment of high salinity wastewater by upflow anaerobic sludge blanket (UASB) reactor,” Desalination and Water Treatment 57(28), 12998-13004. DOI: 10.1080/19443994.2015.1059369

Avram, D., Gavriliuc, R., and Ghiaus, A. G. (2004). “Technologies for eutectic freeze crystallization,” Revista de Chimie 55(10), 769-772.

Bajpai, P. (2017). Pulp and Paper Industry Chemical Recovery, Elsevier, Amsterdam.

Barduhn, A. J., and Manudhane, A. (1979). “Temperature required for eutectic freezing of natural waters,” Desalination 28(3), 233-241. DOI: 10.1016/S0011-9164(00)82232-8

Batchelor, G. K. (1967). An Introduction to Fluid Dynamics, Cambridge Univ. Press, ISBN 0-521-66396-2.

Becheleni, E. M. A., Rodriguez-Pascual, M., Lewis, A. E., and Rocha, S. D. F. (2017). “Influence of phenol on the crystallization kinetics and quality of ice and sodium sulfate decahydrate during eutectic freeze crystallization,” Indust. Eng. Chem. Res. 56(41), 11926-11935. DOI: 10.1021/acs.iecr.7b02668

Biermann, C. J. (1996). Handbook of Pulping and Papermaking, 2nd Ed., Academic Press, San Diego.

Birk, J. R., Larsen, C. M., Vaux, W. G., and Oldenkamp, R. D. (1971). “Hydrogen reduction of alkali sulfate,” IEC Proc. Des. Dev. 10(1), 7-13. DOI: 10.1021/i260037a002

Brewster, J. (2007). “Economic impact of modern kraft recovery boilers,” International Chemical Recovery Conference.

Buchfink, R. (2011). “Effects of impurities on an industrial crystallization process of ammonium sulfate,” Verfahrenstechnik, Taschenbuch, thesis document, Shaker, ISBN-10: 3844002421, ISBN-13: 978-3844002423

Burbano, A., and Brankhuber, P. (2012). “Demonstration of membrane zero liquid discharge for 582 drinking water systems – A literature review,” WERFST10a,

Burelle, R., Borduas, P., Morrison, D., and Hughes, R. (2016). “Working towards a safer recovery boiler operation,” Pulping, Engineering, Environmental, Recycling, Sustainability Conference 2016, PEERS 2016, Vol. 1, pp. 370-376, ISBN-13: 9781510831445, TAPPI Press, Atlanta.

Burton, J. A., Prim, R. C., and Slichter, W. P. (1953). “The distribution of solid in crystals grown from the melt, Part 1. Theoretical,” J. Chem. Phys. 21, 1987-1991. DOI: 10.1063/1.1698728

Butler, M. F. (2002). “Freeze concentration of solutes at the ice/solution interface studied by optical interferometry,” Cryst. Growth Des. 2(6), 541-548. DOI: 10.1021/cg025591e

Cameron, J. H., and Grace, T. M. (1983). “Kinetic study of sulfate reduction with carbon,” Indust. Eng. Chem. Fund. 22(4), 486-494. DOI: 10.1021/i100012a021

Cameron, J. H., and Grace, T. M. (1985). “Kinetic study of sulfate reduction with kraft black liquor char,” Indust. Eng. Chem. Fund. 24(4), 443-449. DOI: 10.1021/i100020a008

Chambers, D. (2017). “Pure genius! UTS’s main wastewater solution could also aid desalination,” Times Live, 11 July 2017,

Chen, C. T. A., Chen, J. H., and Millero, F. J. (1980). “Densities of sodium chloride, magnesium chloride, sodium sulfate, and magnesium sulfate aqueous solutions at 1 atm from 0 to 50 C and from 0.001 to 1.5 m,” J. Chem. Eng. Data 25(4), 307-310. DOI: 10.1021/je60087a013

Chen, G. Z., Lu, Y. H., Krantz, W. B., Wang, R., and Fane, A. G. (2014). “Optimization of operating conditions for a continuous membrane distillation crystallization process with zero salty water discharge,” J. Membrane Sci. 450, 1-11. DOI: 10.1016/j.memsci.2013.08.034

Chivavava, J., Rodriguez-Pascual, M., and Lewis, A. E. (2014). “Effect of operating conditions on ice characteristics in continuous eutectic freeze crystallization,” Chem. Eng. Technol. 37(8), 1314-1320. DOI: 10.1002/ceat.201400094

Chowdhury, J. (1988). “C. P. I. warm up to freeze concentration,” Chem. Eng. 95(6), 24-31.

Clayton, D., Easty, D., Einspahr, D., and Lonsky, W. (1989). “Kraft liquors,” in: Pulp and Paper Manufacture, Vol 5, Alkaline Pulping, The joint Committee of the Paper Industry, TAPPI and CPPA, Atlanta, GA and Montreal, pp. 15-22.

Cob, S. S., Guner, F. E. G., Hofs, B., van Spronsen, J., and Witkamp, G. J. (2014). “Three strategies to treat reverse osmosis brine and cation exchange spent regenerant to increase system recovery,” Desalination 344, 36-47. DOI: 10.1016/j.desal.2014.03.009

Cornell, C. F. (1992). “Cooking liquor preparation,” in: Chemical Recovery Alkaline Pulping Processes, Green and Hough (eds.), TAPPI Press, Atlanta, pp. 119-152.

Creamer, T. (2016). “Eskom aiming to ‘freeze’ pollution from power station and mine wastewater streams,” Engineering News, 22 Jan. 2016,

Curcio, E., Ji, X. S., Quazi, A. M., Barghi, S., Di Profio, G., Fontananova, E., Macleod, T., and Drioli, E. (2010). “Hybrid nanofiltration-membrane crystallization system for the treatment of sulfate wastes,” J. Membrane Sci. 360, 493-498. DOI: 10.1016/j.memsci.2010.05.053

Dence, C. W., and Reeve, D. W. (1996). Pulp Bleaching: Principles and Practice, TAPPI Press, Atlanta.

Dominguez-Robles, J., Espinosa, E., Savy, D., Rosal, A., and Rodriguez, A. (2016). “Biorefinery process combining Specel (R) process and selective lignin precipitation using mineral acids,” BioResources 11(3), 7061-7077. DOI: 10.15376/biores.11.3.7061-7077

Du, K., Zhou, Y. M., Wang, L. Q., and Wang, Y. Y. (2009). “Fluorescent-tagged no phosphate and nitrogen free calcium phosphate scale inhibitor for cooling water systems,” J. Appl. Polymer Sci. 113(3), 1966-1974. DOI: 10.1002/app.30213

Elimelech, M., and Phillip, W. A. (2011). “The future of seawater desalination: Energy, technology, and the environment,” Science 333(6043), 712-717. DOI: 10.1126/science.1200488

Empie, H. J. (2009). Fundamentals of the Kraft Recovery Process, TAPPI Press, Atlanta.

Engdahl, H., Halinen, E., Jäntti, J., Kapanen, J., Kottila, M., Lankinen, M., Lintunen, T., Näsänen, H., Toropainen, T., and Parviainen, K. (2008). “While liquor preparation,” in: Chemical Pulping. Part 2, Recovery of Chemicals and Energy, Book ser.: Papermaking Science and Technology, Paper Engineering Association / Paperi ja Puu Oy, Helsinki, Finland, Chapter 4, pp. 124-193.

ETSAP (2015). “Pulp and paper industry,” Energy Technology Systems Analysis Programme, Technology Brief 107, May.

FAO (2017). “Pulp and recovered paper,” Forest Products 2015, Food and Agriculture Organization of the United Nations, FAO Statistics, YearBook, ROME.

Fardim, P., and Tikka, P. (eds.) (2011). Chemical Pulping, Paper Engineers Assoc., Paperi ya Puu Oy.

Fernández-Torres, M. J., Randall, D. G., Melamu, R., and von Blottnitz, H. (2012). “A comparative life cycle assessment of eutectic freeze crystallization and evaporative crystallization for the treatment of saline wastewater,” Desalination 306, 17-23. DOI: 10.1016/j.desal.2012.08.022

Foran, C. D. (1992). “Tall oil soap recovery,” in: Chemical Recovery in the Alkaline Pulping Processes, R. P. Green and G. Hough (eds.), 3rd Ed., TAPPI Press, Atlanta, Ch. 4, pp. 45-56.

Fritzmann, C., Lowenberg, J., Wintgens, T., and Melin, T. (2007). “State-of-the-art of reverse osmosis desalination,” Desalination 216, 1-76. DOI: 10.1016/j.desal.2006.12.009

Gal, J. Y., Bollinger, J. C., Tolosa, H., and Gache, N. (1996). “Calcium carbonate solubility: A reappraisal of scale formation and inhibition,” Talanta 43(9), 1497-1509. DOI: 10.1016/0039-9140(96)01925-X

Garrett, D. E. (1998). Sodium Sulfate—Handbook of Deposits, Processing, Properties and Use, Academic Press, London, UK.

Gellerstedt, G., Tomani, P., Axegård, P., and Backlund, B. (2013). “Lignin recovery and lignin-based products,” in: Integrated Forest Biorefineries – Challenges and Opportunities, C. Lew (ed.), Royal Soc. Chem., Cambridge, pp. 180-210.

Genceli, F. E., Gartner, R., and Witkamp, G. J. (2005). “Eutectic freeze crystallization in a 2nd generation cooled disk column crystallizer for MgSO4H2O system,” J. Crystal Growth 275, E1369-E1372. DOI: 10.1016/j.jcrysgro.2004.11.172

Gerbasi, B., Stuart, P. R., Arsenault, F., and Zaloum, R. (1993). “Technoeconomic assessment of several closed-cycle technology alternatives for an existing TMP-newsprint mill,” Pulp Paper Canada 94(12), 123-128.

Gilarranz, M. A., Rodriguez, F., Oliet, M., and Revenga, J. A. (1998). “Acid precipitation and purification of wheat straw lignin,” Separ. Sci. Technol. 33(9), 1359-1377. DOI: 10.1080/01496399808544988

Giummarella, N., Lindgren, C., Lindstrom, M. E., and Henriksson, G. (2016). “Lignin prepared by ultrafiltration of black liquor: Investigation of solubility, viscosity, and ash content,” BioResources 11(2), 3494-3510.

Giwa, A., Dufour, V., Al Marzooqi, F., Al Kaabi, M., and Hasan, S. W. (2017). “Brine management methods: Recent innovations and current status,” Desalination 407, 1-23. DOI: 10.1016/j.desal.2016.12.008

Grace, T. M. (1986). “Recovery boiler explosions,” TAPPI J. 69(6), 167-168.

Grace, T. M. (1992). “Chemical recovery process chemistry,” in: Chemical Recovery in the Alkaline Pulping Processes, R. P. Green and G. Hough (eds.), 3rd Ed., TAPPI Press, Atlanta, Ch. 5, pp. 57-78.

Haaga, K., and Laitinen, M. (2002). “The experiences of the world’s biggest recovery boiler in Yang Pu at Hainan Jinhai Pulp & Paper Co.,” Kvaerner Power Oy.

Hammett, A. L., Youngs, R. L., Sun, X. F., and Chandra, M. (2001). “Non-wood fiber as an alternative to wood fiber in China’s pulp and paper industry,” Holzforschung 55(2), 219-224. DOI: 10.1515/HF.2001.036

Hasan, M., Filimonov, R., Chivavava, J., Sorvari, J., Louhi-Kultanen, M., and Lewis, A. E. (2017a). “Ice growth on the cooling surface in a jacketed and stirred eutectic freeze crystallizer of aqueous Na2SO4 solutions,” Separ. Purif. Technol. 175, 512-526. DOI: 10.1016/j.seppur.2016.10.014

Hasan, M., and Louhi-Kultanen, M. (2014). “Determination of Pitzer parameters for 1-1 nitrate and 1-2 sulfate solutions from freezing point data,” Chem. Eng. Technol. 37(8), 1340-1346. DOI: 10.1002/ceat.201400202

Hasan, M., and Louhi-Kultanen, M. (2015). “Ice growth kinetics modeling of air-cooled layer crystallization from sodium sulfate solutions,” Chem. Eng. Sci. 133, 44-53. DOI: 10.1016/j.ces.2015.01.050

Hasan, M., and Louhi-Kultanen, M. (2016). “Water purification of aqueous nickel sulfate solutions by air cooled natural freezing,” Chem. Eng. J. 294, 176-184. DOI: 10.1016/j.cej.2016.02.114

Hasan, M., Partanen, J. I., Vahteristo, K. P., and Louhi-Kultanen, M. (2014). “Determination of the Pitzer interaction parameters at 273.15 K from the freezing-point data available for NaCl and KCl solutions,” Ind. Eng. Chem. Res. 53(13), 5608-5616. DOI: 10.1021/ie404042h

Hasan, M., Rotich, N., John, M., and Louhi-Kultanen, M. (2017b). “Salt recovery from wastewater by air-cooled eutectic freeze crystallization,” Chem. Eng. J. 326, 192-200. DOI: 10.1016/j.cej.2017.05.136

Himawan, C., Kramer, H. J. M., and Witkamp, G. J. (2006). “Study on the recovery of purified MgSO47H2O crystals from industrial solution by eutectic freezing,” Sep. Purif. Technol.50(2), 240-248. DOI: 10.1016/j.seppur.2005.11.031

Himawan, C., Vaessen, R. J. C., Kramer, H. J. M., Seckler, M. M., and Witkamp, G. J. (2002). “Dynamic modeling and simulation of eutectic freeze crystallization,” J. Crystal Growth 237(3), 2257-2263. Article no.: PII S0022-0248(01)02240-0 DOI: 10.1016/S0022-0248(01)02240-0

Himawan, C., and Witkamp, G. J. (2002). “Scale formation of ice from electrolyte solu-tions on a scraped surface heat exchanger plate,” J. Crystal Growth 237(3), 2172-2177. Article no. PII S0022-0248(01)02263-1 DOI: 10.1016/S0022-0248(01)02263-1

Hirata, T., Nagasaka, K., and Ishikawa, M. (2000). “Crystal ice formation of solution and its removal phenomena at cooled horizontal solid surface: Part I: Ice removal phenomena,” Int. J. Heat Mass Transfer 43(3), 333-339. DOI: 10.1016/S0017-9310(99)00168-4

Holmqvist, A., Wallberg, O., and Jonsson, A. S. (2005). “Ultrafiltration of kraft black liquor from two Swedish pulp mills,” Chem. Eng. Res. Des. 83(A8), 994-999. DOI: 10.1205/cherd.04204

Hubbe, M. A., Ayoub, A., Daystar, J. S., Venditti, R. A, and Pawlak, J. J. (2013). “Enhanced absorbent products incorporating cellulose and its derivatives: A review,” BioResources 8(4), 6556-6629. DOI: 10.15376/biores.8.4.6556-6629

Hubbe, M. A., Metts, J. R., Hermosilla, D., Blanco, M. A., Yerushalmi, L., Haghighat, F., Lindholm-Lehto, P., Khodaparast, Z., Kamali, M., and Elliott, A. (2016). “Wastewater treatment and reclamation: A review of pulp and paper industry practices and opportunities,” BioResources 11(3), 7953-8091. DOI: 10.15376/biores.11.3.Hubbe

Humpert, D., Ebrahimi, M., and Czermak, P. (2016). “Membrane technology for the recovery of lignin: A review,” Membranes 6(3), article no. 42. DOI: 10.3390/membranes6030042

Isik, M. (2004). “Efficiency of simulated textile wastewater decolorization process based on the methanogenic activity of upflow anaerobic sludge blanket reactor in salt inhibition condition,” Enzyme Microb. Technol. 35(5), 399-404. DOI: 10.1016/j.enzmictec.2004.04.018

Jooste, D. (2016). “Ice scaling in continuous eutectic freeze crystallization,” Masters of science thesis, University of Cape Town.

Kapembwa, M., Rodriguez-Pascual, M., and Lewis, A. E. (2013). “Heat and mass transfer effects on ice growth mechanisms in pure water and aqueous solutions,” Cryst. Growth Des. 14(1), 389-395. DOI: 10.1021/cg401428v

Kenny, R., Gorgol, R. G., Martineau, D., and Prahacs, S. (1992). “Freeze crystallization of Temcell’s BCTMP effluent,” Pulp Paper Canada 93(10), 55-58.

Kevlich, N. S., Shofner, M. L., and Nair, S. (2017). “Membranes for kraft black liquor concentration and chemical recovery: Current progress, challenges, and opportunities,” Separ. Sci. Technol. 52(6), 1070-1094. DOI: 10.1080/01496395.2017.1279180

Keyoumu, A., Sjodahl, R., Henriksson, G., Ek, M., Gellerstedt, G., and Lindstrom, M. E. (2004). “Continuous nano- and ultra-filtration of kraft pulping black liquor with ceramic filters – A method for lowering the load on the recovery boiler while generating valuable side-products,” 20(2), 143-150. DOI: 10.1016/j.indcrop.2004.04.017

Khawaji, A. D., Kutubkhanah, I. K., and Wie, J.-M. (2008). “Advances in seawater desalination technologies,” Desalination 221, 47-69.

Kim, D. H. (2011). “A review of desalting process techniques and economic analysis of the recovery of salts from retentates,” Desalination 270, 1-8. DOI: 10.1016/j.desal.2010.12.041

Kouisni, L., Gagne, A., Maki, K., Holt-Hindle, P., and Paleologou, M. (2016). “LignoForce system for the recovery of lignin from black liquor: Feedstock options, odor profile, and product characterization,” ACS Sust. Chem. Eng. 4(10), 5152-5159. DOI: 10.1021/acssuschemeng.6b00907

Kwon, D. Y., and Yoon, J. S. (2017). “Study on the assessment of environmental impact indicators on concentrated effluent from desalination facility,” Desalination Water Treat. 96, 201-210. DOI: 10.5004/dwt.2017.21036

Laxén, T., and Tikka, P. (2008). “Soap and tall oil,” in: Tikka, P. (ed.), Chemical Pulping. Part 2, Recovery of Chemicals and Energy, book series: Papermaking Science and Technology, book 6, Paper Engineer’s Assoc., Paperi ja Puu, Chapter 8, pp. 359-380.

Lee, K. P., Arnot, T. C., and Mattia, D. (2011). “A review of reverse osmosis membrane materials for desalination – Development to date and future potential,” J. Membrane Sci. 370, 1-22. DOI: 10.1016/j.memsci.2010.12.036

Lewis, A. E. (2017). “From crazy to conquest – A chilling waste solution,” Leadership Mag., Nov. 2017,

Lewis, A. E. (2018). “Turning skepticism into success: A chilling waste solution,” in: Green Mining: Beyond the Myth, Gangle Gold Ashanti/Minerals to Metals/Mineral Law in Africa, Jan. 2018,

Lewis, A. E., Nathoo, J., Thomsen, K., Kramer, H. J., Witkamp, G. J., Reddy, S. T., and Randall, D. G. (2010). “Design of a eutectic freeze crystallization process for multicomponent waste water stream,” Chem. Eng. Res. Des. 88(9A), 1290-1296. DOI: 10.1016/j.cherd.2010.01.023

Lewis, A., Randall, D., Reddy, T., Jinvanji, R., and Nathoo, J. (2009). “Worth its salt – How eutectic freeze crystallisation can be used to recover water and salt from hypersaline mine waters,” Water in Mining Conference, Perth, WA, pp. 1-7.

Lewis, A., Seckler, M., Kramer, H., and van Rosmalen, G. (2015). Industrial Crystallization. Fundamentals and Applications, Cambridge University Press. DOI: 10.1017/CBO9781107280427

Li, J., and van Heiningen, A. R. P. (1988). “Kinetics of sodium-sulfate reduction in the solid-state by carbon-monoxide,” Chem. Eng. Sci. 43(8), 2079-2085. DOI: 10.1016/0009-2509(88)87087-8

Li, J., and van Heiningen, A. R. P. (1995). “Kinetics of solid-state reduction of sodium sulfate by carbon,” J. Pulp Paper Sci. 21(5), J165-J173.

Liu, L, Miyawaki, O., and Nakamura, K. (1997). “Progressive freeze concentration of model liquid,” Food Sci.Technol. Int. 3(4), 348-352. DOI: 10.1016/S0924-2244(97)85552-5

Lotfollahi, M. N. (2009). “Design of optimal process flowsheet for fractional crystallization separation process,” Iran. J. Chem. Chem. Eng. – Intl. English Ed. 28(2), 63-73.

Louhi-Kultanen, M. (1996). Concentration and Purification by Crystallization, Acta Polytechica Scandinavica, Finnish Academy of Technology, Series 241, 112 pp. ISSN: 1239-0518, ISBN: 9525148033, 9789525148039

Loutfi, H., Blackwell, B., and Uloth, V. (1991). “Lignin recovery from kraft black liquor – Preliminary process design,” TAPPI J. 74(1), 203-210.

Lund, K., Sjöström, K., and Brelid, H. (2012). “Alkali extraction of kraft pulp fibers. Influence on fibre and fluff pulp properties,” J. Eng. Fiber Fabr. 7(2), 30-39.

Magalhaes, M. C. F., Konigsberger, E., May, P. M., and Hefter, G. (2002). “Heat capacities of concentrated aqueous solutions of sodium sulfate, sodium carbonate, and sodium hydroxide at 25 C,” J. Chem. Eng. Data 47(3), 590-598. DOI: 10.1021/je010314h

Manttari, M., Lahti, J., Hatakka, H., Louhi-Kultanen, M., and Kallioinen, M. (2015). “Separation phenomena in UF and NF in the recovery of organic acids from kraft black liquor,” J. Membrane Sci. 490, 84-91. DOI: 10.1016/j.memsci.2015.04.048

McGinnis, T., Svarz, J., and Gabel, R, (1998). “Additives for improving the separation of crude tall oil soap from black liquor and analytical methods for measurement of their performance,” TAPPI Pulping Conf., Montreal, Oct. 25-29, 1998, TAPPI Press, Atlanta, pp. 191-934.

Mersmann, A., Braun, B., and Loffelmann, M. (2002). “Prediction of crystallization coefficients of the population balance,” Chem. Eng. Sci. 57(20), 4267-4275. Article Number: PII S0009-2509(02)00343-3. DOI: 10.1016/S0009-2509(02)00343-3

Morillo, J., Usero, J., Rosado, D., El Bakouri, H., Riaza, A., and Bernaola, F.-J. (2014). “Comparative study of brine management technologies for desalination plants,” Desalination 336, 32-49. DOI: 10.1016/j.desal.2013.12.038

Mullin, J. W. (2001). Crystallization, Butterworth-Heinemann, Oxford.

Myerson, A. S. (2002). Handbook of Industrial Crystallization, Butterworth-Heinemann, Boston, 313 pp.

Narasimha, M., Brennan, M., and Holtham, P. N. (2007). “A review of CFD modelling for performance predictions of hydrocyclone,” Eng. Appl. Computational Fluid Dynam. 1(2), 109-125. DOI: 10.1080/19942060.2007.11015186

Nathoo, J., Jivanji, R., and Lewis, A. E. (2009). “Freezing your brines off: Eutectic freeze crystallization for brine treatment,” in: International Mine Water Conference, pp. 431-437.

Neilly, A., Jegatheesan, V., and Shu, L. (2009). “Evaluating the potential for zero discharge from reverse osmosis desalination using integrated processes – A review,” Desalin. Water Treat. 11, 58-65. DOI: 10.5004/dwt.2009.843

Niemelä, K., and Alén, R. (1999). “Characterization of pulping liquors,” in: Analytical Methods in Wood Chemistry, Pulping, and Papermaking, E. Sjöström, and R. Alén (eds.), Springer, Heidelberg, pp. 193-231.

Osborne, N. S., Stimson, H. F., and Ginning, D. C. (1939). “Measurements of heat capacity and heat of vaporization of water in the range 01 to 100 C,” J. Res. Natl. Bur. Stand. 23, 197-260. DOI: 10.6028/jres.023.008

Osborne, N. S. (1939). “Heat of fusion of ice—A revision,” J. Res. Natl. Bur. Stand. 23, 643-646. DOI: 10.6028/jres.023.043

Pangborn, J. B. (1963). “Observations on the eutectic freezing of salt solutions,” unpublished final report, Syracuse Univ., Dept. of Chem. Eng., May.

Pantoja, C. E., Nariyoshi, Y. N., and Seckler, M. M. (2015). “Membrane distillation crystallization applied to brine desalination: a hierarchical design procedure,” Ind. Eng. Chem. Res.54(10), 2776-2793. DOI: 10.1021/ie504695p

Pantoja, C. E., Nariyoshi, Y. N., and Seckler, M. M. (2016). “Membrane distillation crystallization applied to brine desalination: Additional design criteria,” Indust. Eng. Chem. Res. 55(4), 1004-1012. DOI: 10.1021/acs.iecr.5b03807

Parham, R., and Hergert, H. (1980). “Fluff pulp. A review of its development and current technology,” Pulp Paper 54(3), 110-115, 121.

Parrish, D. (1988). “Black liquor recovery boilers – An introduction,” National Board Bulletin, Winter 1998, 4 pp.,

Pekarović, J., Pekarovicora, A., and Fleming, P. (2006). “Two-step straw processing – A new concept of silica problem solution,” TAPPI Journal 5(11), 5-8.

Pérez-González, A., Urtiaga, A. M., Ibáñez, R., and Ortiz, I. (2012). “State of the art and review on the treatment technologies of water reverse osmosis concentrates,” Water Res. 46(2), 267-283. DOI: 10.1016/j.watres.2011.10.046

Peters, E. M., Chivavava, J., Rodriguez-Pascual, M., and Lewis, A. E. (2016). “Effect of a phosphonate antiscalant during eutectic freeze crystallization of a sodium sulfate aqueous stream,” Indust. Eng. Chem. Res. 55(35), 9378-9386. DOI: 10.1021/acs.iecr.6b02024

Petzold, G., and Aguilera, J. M. (2009). “Ice morphology: Fundamentals and technological applications in foods,” Food Biophys. 4(4), 378-396. DOI: 10.1007/s11483-009-9136-5

Pirttinen, E., Stenius, P., and Kovasin, K. (2007). “Surfactant phase equilibria and separation of amphiphilic extractives from black liquor in kraft cooking of wood,” J. Dispersion Sci. Technol. 28(1), 5-10. DOI: 10.1080/01932690600992852

Pronk, P. (2007). Fluidized Bed Heat Exchangers to Prevent Fouling in Ice Slurry Systems and Industrial Crystallizers, Ph.D. Dissertation, Delft University of Technology, The Netherlands.

Pronk, P., Ferreira, C. A. I., and Witkamp, G. J. (2006). “Influence of solute type and concentration on ice scaling in fluidized bed ice crystallizers,” Chem. Eng. Sci. 61(13), 4354-4362. DOI: 10.1016/j.ces.2006.02.019

Pronk, P., Ferreira, C. A. I., and Witkamp, G. J. (2008). “Prevention of crystallization fouling during eutectic freeze crystallization in fluidized bed heat exchangers,” Chem. Eng. Proc. 47(12), 2140-2149. DOI: 10.1016/j.cep.2007.11.006

Quist-Jensen, C. A., Macedonio, F., Horbez, D., and Drioli, E. (2017). “Reclamation of sodium sulfate from industrial wastewater by using membrane distillation and membrane crystallization,” Desalination 401, 112-119. DOI: 10.1016/j.desal.2016.05.007

Randall, D. G., and Lewis, A. E. (2009). “Seeding for selective salt recovery during eutectic freeze crystallization,” in: International Mine Water Conference, 19-22 October, Pretoria, South Africa, pp. 639-646.

Randall, D. G., Mohamed, R., Nathoo, J., Rossenrode, H., and Lewis, A. E. (2013). “Improved calcium sulfate recovery from a reverse osmosis retentate using eutectic freeze crystallization,” Water Sci. Technol. 67(1), 139-146. DOI: 10.2166/wst.2012.540

Randall, D. G., Nathoo, J., and Lewis, A. E. (2011). “A case study for treating a reverse osmosis brine using eutectic freeze crystallization – Approaching a zero waste process,” Desalination 266, 256-262. DOI: 10.1016/j.desal.2010.08.034

Randall, D. G., Zinn, C., and Lewis, A. E. (2014). “Treatment of textile wastewaters using eutectic freeze crystallization,” Water Sci. Technol. 70(4), 736-741. DOI: 10.2166/wst.2014.289

Reddy, S. T., Lewis, A. E., Witkamp, G. J., Kramer, H. J. M., and van Spronsen, J. (2010). “Recovery of Na2SO410H2O from a reverse osmosis retentate by eutectic freeze crystallization technology,” Chem. Eng. Res. Des. 88(9), 1153-1157. DOI: 10.1016/j.cherd.2010.01.010

Ren, H., Dai, X., and Omori, S. (2016). “Preparation of biodiesel and separation of hemicellulose from soap skimmings,” Cellulose Chem. Technol. 50(2), 247-255.

Richardson, J. F., and Zaki, W. N. (1954). “The sedimentation of a suspension of uniform spheres under conditions of viscous flow,” Chem. Eng. Sci. 3(2), 65-73. DOI: 10.1016/0009-2509(54)85015-9

Rietema, K. (1961). “Performance and design of hydrocyclones. 1. General considerations,” Chem. Eng. Sci. 15(3-4), 298-302. DOI: 10.1016/0009-2509(61)85033-1

Rodriguez-Pascual, M., Genceli, F. E., Trambitas, D. O., Evers, H., Van Spronsen, J., and Witkamp, G. J. (2010). “A novel scraped cooled wall crystallizer recovery of sodium carbonate and ice from an industrial aqueous solution by eutectic freeze crystallization,” Chem. Eng. Res. Des. 88, 1252-1258. DOI: 10.1016/j.cherd.2009.07.015

Roos, A. C., Verschuur, R. J., Schreurs, B., Scholz, R., and Jansens, P. J. (2003). “Development of a vacuum crystallizer for the freeze concentration of industrial waste water,” Chem. Eng. Res. Des. 81(A8), 881-892. DOI: 10.1205/026387603322482121

Rousseau, R. W., and Sharpe, E. E. (1980). “Freeze concentration of black liquor – Characteristics and limitations,” Indust. Eng. Chem. Proc. Des. Devel. 19(1), 201-204. DOI: 10.1021/i260073a037

Severtson, S. J., Duggirala, P. Y., Carter, P. W., and Reed, P. E. (1999). “Mechanism and chemical control of CaCO3 scaling in the kraft process,” TAPPI J. 82(6), 167-174.

Singbeil, D., and Garner, A. (1989). “Lower furnace corrosion in kraft recovery boilers. 1. A review of research findings to date,” TAPPI J. 72(6), 136-141.

Sjöström, E. (1993). Wood Chemistry: Fundamentals and Applications, Academic Press.

Steenkamp, T. (2017). “Squeezing the last drop out of mine water,” Sunday Times, 20 Aug. 2017,

Stepakoff, G. L., Siegelman, D., Johnson, R., and Gibson, W. (1974). “Development of a eutectic freezing process for brine disposal,” Desalination 15(1), 25-38. DOI: 10.1016/S0011-9164(00)82061-5

Stoklosa, R. J., Velez, J., Kelkar, S., Saffron, C. M., Thies, M. C, and Hodge, D. B. (2013). “Correlating lignin structural features to phase partitioning behavior in a novel aqueous fractionation of softwood kraft black liquor,” Green Chem. 15(10), 2904-2912. DOI: 10.1039/c3gc41182f

Subramani, A., and Jacangelo, J. G. (2014). “Treatment technologies for reverse osmosis concentrate volume minimization: A review,” Separ. Purif. Technol. 112, 472-489. DOI: 10.1016/j.seppur.2013.12.004

Suess, H. U. (2010). Pulp Bleaching Today, De Gruyter, New York. DOI: 10.1515/9783110218244

Szpaczynski, J. A., White, J. A., and Cote, C. L. (2017). “Separation of contaminants in the freeze/thaw process,” Chem. Proc. Eng. – Inyneiria Chemiczna I Procesowa 38(2), 249-264.

Terwilliger, J. P., and Dizio, S. F. (1970). “Salt rejection phenomena in the freezing of saline solutions,” Chem. Eng. Sci. 25(8), 1331-1349. DOI: 10.1016/0009-2509(70)80010-0

Thompson, T. G., and Nelson, K. H. (1954). “Deposition of salts from sea water by frigid concentration,” J. Marine Res. 13, 166; also Dept. Oceanography, Univ. Washington, Tech. Rept. 29.

Tikka, P. (ed.) (2008). Chemical Pulping. Part 2, Recovery of Chemicals and Energy, book series: Papermaking Science and Technology, book 6, Paper Engineer’s Assoc., Paperi ja Puu.

Toghiani, R. K., Phillips, V. A., Smith, L. T., and Lindner, J. S. (2008). “Solubility in the Na+SO4+NO3 and Na+SO4+NO2 systems in water and in sodium hydroxide solutions,” J. Chem. Eng. Data 53(3), 798-804. DOI: 10.1021/je700666t

Tsai, J. H., Macedonio, F., Drioli, E., Giorno, L., Chou, C. Y., Hu, F. C., Li, C. L., Chuang, C. J., and Tung, K. L. (2017). “Membrane-based zero liquid discharge: Myth or reality?” DOI: 10.1016/j.jtice.2017.06.050

Tutus, A., and Eroglu, H. (2003). “A practical solution to the silica problem in straw pulping,” APPITA J. 56(2), 111-115.

Vaessen, R. J. C., Janse, B. J. H., Seckler, M. M., and Witkamp, G. J. (2003a). “Evaluation of the performance of a newly developed eutectic freeze crystallizer – Scraped cooled wall crystallizer,” Chem. Eng. Res. Des. 81(A10), 1363-1372DOI: 10.1205/026387603771339573

Vaessen, R., Seckler, M., and Witkamp, G. J. (2003b). “Eutectic freeze crystallization with an aqueous KNO3-HNO3 solution in a 100-L cooled-disk column crystallizer,” Inust. Eng. Chem. Res. 42(20), 4874-4880. DOI: 10.1021/ie020946c

van der Ham, F., Seckler, M. M., and Witkamp, G. J. (2004). “Eutectic freeze crystallization in a new apparatus: The cooled disk column crystallizer,” Chem. Eng. Proc. 43(2), 161-167. DOI: 10.1016/S0255-2701(03)00018-7

van der Ham, F., Witkamp, G. J., de Graauw, J., and van Rosmalen, G. M. (1998). “Eutectic freeze crystallization: Application to process streams and waste water purification,” Chem. Eng. Proc. 37(2), 207-213. DOI: 10.1016/S0255-2701(97)00055-X

van der Ham, F., Witkamp, G. J., de Graauw, J., and van Rosmalen, G. M. (1999). “Eutectic freeze crystallization simultaneous formation and separation of two solid phases,” J. Crystal Growth 198(1), 744-748. DOI: 10.1016/S0022-0248(98)01003-3

van Spronsen, J., Rodriguez-Pascual, M., Genceli, F. E., Trambitas, D. O., Evers, H., and Witkamp, G. J. (2010). “Eutectic freeze crystallization from the ternary Na2CO3-NaHCO3-H2O system. A novel scraped wall crystallizer for the recovery of soda from an industrial aqueous stream,” Chem. Eng. Res. Des. 88(9A), 1259-1263. DOI: 10.1016/j.cherd.2009.09.012

Venkatesh, V. (1992). “Lime reburning: The rotary lime kiln,” in: Chemical Recovery in the Alkaline Pulping Processes, R. P. Green and G. Hough (eds.), 3rd Ed., TAPPI Press, Atlanta, Ch. 8, pp. 153-179.

Venkatesh, V., and Nguyen, X. N. (1992). “Evaporation and concentration of black liquor,” in: Chemical Recovery in the Alkaline Pulping Processes, R. P. Green and G. Hough (eds.), 3rdEd., TAPPI Press, Atlanta, Ch. 2, pp. 5-35.

Wachinski, A. M. (2017). Environmental Ion Exchange: Principles and Design, CRC Press, Taylor & Francis Group, Boca Raton, FL, USA.

Wallmo, H., Richards, T., and Theliander, H. (2009). “An investigation of process parameters during lignin precipitation from kraft black liquors: A step towards an optimised precipitation operation,” Nordic Pulp Paper Res. J. 24(2), 158-164. DOI: 10.3183/NPPRJ-2009-24-02-p158-164

Wartena, R. C., Winnick, J., and Pfromm, P. H. (2000). “Electrolytic recovery of molten pulping chemicals: An alternative to the kraft causticizing process,” in: Molten Salts XII, Proceedings, Book Ser.: Electrochem. Soc. Ser. 99(41), 685-693.

Wartena, R., Winnick, J., and Pfromm, P. H. (2002a). “Recycling wood pulping chemicals by molten salt electrolysis: Cyclic voltammetry of mixtures containing Na2CO3 and Na2SO4,” J. Appl. Electrochem. 32(4), 415-424. DOI: 10.1023/A:1016334307899

Wartena, R., Winnick, J., and Pfromm, P. H. (2002b). “Recycling kraft pulping chemicals: Cyclic voltammetry of molten salt mixtures containing Na2CO3, Na2SO4, Na2S/Na2Sx and Na2O/Na2O2,” J. Appl. Electrochem. 32(7), 725-733. DOI: 10.1023/A:1020182322704

Wiegandt, H. F., and von Berg, R. L. (1980). “Myths about freeze desalting,” Desalination 33, 287-297. DOI: 10.1016/S0011-9164(00)88571-9

Wienhaus, O., Bernaczyk, Z., and Pecina, H. (1990). “Research into the precipitation of lignin from sulfate black liquors, particularly by carbon-dioxide,” Papier 44(11), 563-569.

Young, R. A. (1994). “Comparison of the properties of chemical cellulose pulps,” Cellulose 1(2), 107-130. DOI: 10.1007/BF00819662

Yu, L., and Ni, Y. H. (2006). “Decreasing calcium oxalate scaling by partial substitution of Mg(OH)2 for NaOH in the peroxide bleaching of mechanical pulps,” TAPPI J. 5(2), 9-12.


The work of Martin A. Hubbe was supported by and endowment from the Buckman Foundation. The authors would like to express their great thanks to two people who read the text and indicated many items in need of correction or clarification: Marcelo Seckler, Univ. Sao Paulo, Dept. Chem. Engn., Polytech. Sch., Av Prof Luciano Gualberto, Trav 3,380, BR-05508010 Sao Paulo, SP, Brazil; and Marjatta Louhi-Kultanen, Aalto Univ., Sch. Chem. Engn., POB 16100, FI-00076 Espoo, Finland. Thanks are also extended to Ramo Alén, University of Jyväskylä, Finland, who provided input regarding the inorganic content of black liquor.


Assumed Stoichiometry for Calculated Values in Table 2

Formation of sodium sulfate (anhydrous basis)

Na2S [78 g/mole] + 2O2 Na2SO4 [142 g/mole]

Ratio 142/78 = 1.82 (Na2SO4 formed per Na2S oxidized)

Na2SO3 [126 g/mole] + ½ O2 Na2SO4 [142 g/mole]

Ratio 142/126 = 1.13 (Na2SO4 formed per sulfite oxidized)

Na2S2O3 [158 g/mole] + 2 O2 + 2NaOH [40 x 2 = 80 g/2 moles]

2Na2SO4 [142 g/mole x 2 = 284 g/2 moles] + H2O

Ratio1 = 284/158 = 1.80 (Na2SO4 formed per thiosulfate oxidized)

Ratio2 = 80/284 = 0.281 (NaOH used up per Na2SO4 formed)

Formation of sodium carbonate (anhydrous basis)

2 NaOH [2 x 40 g/mole = 80 g/mole] + CO2 Na2CO3 [106 g/mole] + H2O

Ratio = 106/80 = 1.32 (Na2CO3 formed per NaOH consumed)