Typical Crystallizer Design

Obviously, there are many different configurations of batch and continuous crystallizers. The design used for a particular chemical system must be selected in light of the nature of the material being crystallized and the desired properties of the product, such as purity, habit, and CSD. For simplicity, the discussion 

For any process, the lack of accurate process measurements limits the successful implementation of model identification and control. Crystallizers are dispersed-phase systems, and the shortcomings of on-line measurement techniques for these types of systems are particularly evident. 

An accurate measurement of the concentration of the solute in the continuous phase can be obtained by sampling and evaporating to dryness, but this is, of course, prohibitively time-consuming for control purposes. There are several methods for the rapid measurement of continuous-phase concentration. The measurement of refractive index has been shown to provide quick and accurate concentration determination (Kelt and Larson 1977 Mullin and Leci 1972 Sikdar and Randolph 1976). Garside and Mullin (1966) suggested the use of an on-line densitometer to determine solution concentration, and Witkowski (1990) successfully used densitometry for concentration measurements to estimate the kinetic parameters of an isothermal crystallizer. It should 

---

Figure 9.1 Typical crystallization process systems design procedure...

The principal variables that must be controlled in crystallization are the solution supersaturation, the crystal surface area available for growth, and the nucleation rate. These are affected by multiple interacting secondary variables, which may be divided into two categories—equipment design variables and dynamic variables affecting the crystallization. It is the secondary dynamic variables, such as those listed in Table 9.1, to which automatic process control is applied in a typical crystallizer for these vari-... 

CP forms yellow monoclinic crystals with crystal density 1.97 g cm [11]. It is compatible with most metallic and ceramic materials used in typical detonator design and also epoxies cured by amines or anhydrides. This is interesting since many organic explosives and amine cured epoxy materials have compatibility problems. It is also compatible with PETN [12]. 

Xylene Isomerization. After separation of the preferred xylenes, ie, PX or OX, using the adsorption or crystallization processes discussed herein, the remaining raffinate stream, which tends to be rich in MX, is typically fed to a xylenes isomerization unit in order to further produce the preferred xylenes. Isomerization units are fixed-bed catalytic processes that are used to produce a close-to-equiUbrium mixture of the xylenes. To prevent the buildup of EB in the recycle loop, the catalysts are also designed to convert EB to either xylenes, benzene and lights, or benzene and diethylbenzene. 

A typical heat treatment cycle, as illustrated in Figure 1, comprises both nucleation and crystallization temperature holds, but some glass-ceramics are designed to nucleate and/or crystallize during the ramp itself, eliminating the need for multiple holds. 

Crystallization batches range from 30,000 to 60,000 Hters for each pan. Continuous centrifugals are typically used for second, third, and affination steps continuous vacuum pans are less common but are used in the U.S. for intermediate strikes. Most horizontal batch crystallizers have been replaced by continuous units, and all are designed for controlled cooling of the massecuite to maintain supersaturation. 

Crystallizers with Fines Removal In Example 3, the product was from a forced-circulation crystallizer of the MSMPR type. In many cases, the product produced by such machines is too small for commercial use therefore, a separation baffle is added within the crystallizer to permit the removal of unwanted fine crystalline material from the magma, thereby controlling the population density in the machine so as to produce a coarser ciystal product. When this is done, the product sample plots on a graph of In n versus L as shown in hne P, Fig. 18-62. The line of steepest ope, line F, represents the particle-size distribution of the fine material, and samples which show this distribution can be taken from the liquid leaving the fines-separation baffle. The product crystals have a slope of lower value, and typically there should be little or no material present smaller than Lj, the size which the baffle is designed to separate. The effective nucleation rate for the product material is the intersection of the extension of line P to zero size. 

---