Heat and mass transfer of scraped surface heat exchanger used for suspension freeze concentration

2021-05-07 10:55:06 Ftherm Scraped Surface heat exchanger 2870


  • The heat and mass transfer modeling and experimental study for multi-pass freeze concentration were presented.

  • The ice production rate and energy efficiency of the system were correlated with the heat and mass transfer.

  • Heat transfer and ice production were correlated with concentration ratio, partition coefficient and recovery yield.


Freeze concentration (FC) by suspension crystallization is a complex process due to the combination of a scraped-surface heat exchanger, crystallizer, and wash column. Following the study of a three-in-one structure of a multi-pass FC published earlier this year, modeling and experiments of the heat and mass transfer of this freeze concentrator are presented in this paper. The experimental assessment of the system performance, including the measured values of the heat transfer coefficient, ice production rate, and energy efficiency as well as their correlation with the concentration ratio and partition coefficient are presented in this article.


Multi-pass freeze concentration
Apple juice
Suspension crystallization
Heat and mass transfer

1. Introduction

Freeze concentration (FC) is a nonthermal processing technology of liquid food, in which a portion of the water in the aqueous solution is frozen and converted into relatively pure ice crystals, after which it is removed from the liquid phase to concentrate the remaining solution. It can be used to concentrate or pre-concentrate heat-sensitive aqueous solutions, such as milk (Habib and Farid, 2007Sanchez et al., 2011), fresh fruit juices (Orellana-Palma et al., 2017Petzold et al., 2013Bayindirli et al., 1993), other liquid foods (Moreno et al., 2014a), and biological solutions (Moreno et al., 2014b). Studies have shown that compared with evaporation concentration, FC has the advantage of producing less thermal denaturation of the solution, and thus, it can better maintain the original flavor, nutrition, and color of liquid foods (Benedetti et al., 2015Moreno et al., 2014cMiyawaki et al., 2016a). Moreover, the latent heat of water freezing is almost one seventh of the latent heat of water evaporation (Lide and Haynes, 2010), which offers potential for energy savings for de-watering of aqueous solutions.

Ice crystals can be formed from aqueous solutions in two ways: progressive crystallization (Miyawaki et al., 20052016bZambrano et al., 2018) and suspension crystallization (Huige and Thijssen, 1972Qin et al., 2007). In the former, the water freezes on the cooling surface, forming an ice layer progressively and a concentrated liquid phase. Thus, it is also known as layer crystallization. This method of FC is called progressive FC. When the ice layer extends to the entire vessel to form an ice block, it is known as block FC (Moreno et al., 2014cZambrano et al., 2018). The advantage of this technology is the low equipment cost and simple operation management. However, the ice layer has a poor heat transfer coefficient of less than 0.1 kW m−2 K−1 (Qin et al., 2003aPronk et al., 2010Hasan et al., 2017), and a huge cooling surface area is required for practical applications. In addition, the ice layer tends to entrain liquid sacs and causes severe solute loss (Miyawaki et al., 2016a2016bSamsuri et al., 2015).

In the latter method, a scraped-surface heat exchanger (SSHE) is used to clean ice scaling from the cooling surface and improve the heat transfer coefficient, the value of which is correlated to the speed of the scraper and the solution concentration and varies between 200 and 1000 W m−2 K−1, corresponding to a specific ice production rate up to 40 kg h−1 m−2 (Qin et al., 2003b2016Abichandani et al., 1987Rao and Hartel, 2006). However, the newly formed ice crystals are very small and must undergo Ostwald ripening for efficient solid-liquid separation (Pronk et al., 2005Thijssen and Spicer, 1974Qin et al., 2008). The ice crystals will be compressed into a close-packed ice bed and then separated with the mother liquor in a wash column (wash tower) (Schwartzberg et al., 1990). This process and its equipment are complex, and the costs of investment and maintenance are high, which hinders its application. Therefore, it is believed that future research and development in the field of freeze concentrators should focus on novel crystallizer design, to reduce equipment costs and increase the process efficiency, as well as efficient cooling methods, with an emphasis on a single unit operation that incorporates both crystallization and separation in one piece of equipment (Randall and Nathoo, 2015). This requires a comprehensive understanding and modeling of the process of suspension FC, including the heat and mass transfer of the SSHE, the correlation of the ice production rate with the system cooling, and the recovery yield of the soluble solid with ice crystallization.

Moreover, a common concentration scenario, such as to concentrate a dilute solution from 10 to 20°Bx (or higher), requires a dehydration of 50% (or even more). When such an amount of water is converted into ice crystal particles, ice actually fully fills the whole crystallizer and leaves concentrated liquid between the ice particles (Qin et al., 2009a), which deteriorates the heat and mass transfer quickly. The ice must be removed from the concentrated mother liquor to achieve a higher concentration. The de-watering requirements for concentrating a 10°Bx aqueous solution to a higher concentration are shown in Table 1, which indicates that when a final concentration of more than 20°Bx is required, more than one batch operation is required unless a continuous crystallization and separation method is used.

To address the above problems, a three-in-one FC structure was proposed, in which the SSHE, crystallizer, and wash-column were united in a single piece of equipment, and multi-pass (batch) suspension crystallization was applied to achieve a higher product concentration (Ding et al., 2019Qin et al., 2019). A screw scraper was further used in the SSHE. Most of the previous studies on SSHEs were based on straight-blade scrapers. Information on screw-scraper SSHEs is lacking, especially for ice production. This study focused on the modeling and experiments of the heat and mass transfer of a FC system using screw-scraper SSHE.

The juice (raw material) used in this study was “Huiyuan Youth 100% Apple Juice,” which was available in 1-L packs in the local market, with a concentration of 10.5°Bx.

Scraped surface heat exchanger-Ftherm-Votator-Contherm-Terlotherm-SSHEs

2. Experimental apparatus and procedure

The experimental system was basically the same as that described in our previous paper with minor modifications. It mainly consisted of a refrigeration unit, a scraper heat exchanger, a crystallizer, a measurement and control unit, and two storage tanks (Ding et al., 2019). The SSHE, an important part of the FC system that is associated with the heat and mass transfer, used a screw scraper, as shown in Fig. 1. The scraper not only provided scraping action on the cooling surface but also helped to drive ice and juice to float upward into the suspension crystallizer located above. When the juice was cooled to or below its freezing point, dendritic nuclei of the ice precipitated on the cooling surface first (Qin et al., 2009b), which was then scraped off to mix with the solution to form an ice slurry. The ascending ice crystals were intercepted by the perforated plate below the top cover of the crystallizer, but the juice passed through the perforated plate and flowed out of the top cover and returned to the bottom of the SSHE to form external circulation. The process flowchart is shown in Fig. 2, where the pre-treatment steps were completed by the juice manufacturer, and the freeze concentration steps were carried out using the experimental setup described above.

During the ice production period, i.e., the freezing stage, ice Ostwald ripening occurred in the crystallizer simultaneously, which required at least for 3 h for dilute aqueous solutions, such as the original apple juice at a concentration of approximately 10°Bx (a higher juice concentration requires longer ripening times). During this time, newborn micron-size dendritic ice nuclei grew to several hundreds of microns to make the final washing operation possible. When the crystallizer was fully filled by ice particles and formed a tightly packed ice bed, freezing of the pass-1 FC ended and was followed by a wash process, in which 0 °C water was introduced into the crystallizer from the top cover to displace the concentrated juice of the ice bed in a form of plug flow. The concentrated juice was discharged at the bottom of the SSHE simultaneously. The same volume of water as that of the concentrated juice collected was added. The flow rate was controlled under the critical velocity to maintain plug-flow in the ice bed, which produced a sharp wash front, or interface, between the washed and un-washed ice beds. This prevented the mixing of water and the concentrated juice. The concentration of the latter was approximately 1.8 times that of the original juice. If targeting a higher concentration, the collected pass-1 FC juice would sent through a second FC process, which is referred to as the pass-2 FC, and so on. For more details of the FC system and experimental procedure, please refer to our previous published articles (Ding et al., 2019Qin et al., 2019).