Three-Component Blend Control

In the above method, an autotitrator repeatable to +0.5% relative error (or 0.01 wt%) was employed to measure H2O2 wt%, independent of the relative amounts of DI water and slurry. A densitometer with an accuracy of + 0.0005 g/cc was used to estimate the percentage volume of slurry in the mixture (repeatability = +3% slurry by volume), for a specific concentration of H2O2. In the above blend, the measured H2O2 wt% concentration was used to control the speed (or strokes/batch) of the W2000 slurry pump, whereas the density, obtained using a coriolis mass flow meter, controlled the speed of the DI water pump. On the basis of the property curves obtained from bench 

---

Monitoring and control of CMP slurry properties is essential for effective and uniform CMP processes. Bench-top blend sensitivity analysis helps in identification of the most sensitive blend monitoring and control parameter. Two- and three-component blends of CMP slurries can be created and monitored based on the measurements of density, wt% solids, refractive index, pH, and oxidizer level. Typical silica oxide slurry blend ratio is controlled using density as a control parameter, whereas tungsten and copper CMP slurries usually need an autotitrator for periodic monitoring of the oxidizer level. 

A further degree of coupling is found between variables that are similar in nature. Imagine a three-component, blend in which it is desired to control both density p and viscosity p. A problem arises because a change in either of two components may affect both density and viscosity in the same direction. This differs from the cases studied earlier, in that while Ml and m2 affected one variable in the same direction, they affected the other in opposite directions. Let the mathematical model of the three-... 

Blending may be accomplished on a batch or continuous basis. Batch blending processes consist of three sequential steps weighing and loading the components, blending, and discharging. Unlike a continuous blender, the retention time in a batch blender is rigidly defined and controlled, and is the same for all of the particles. 

Three examples of simple multivariable control problems are shown in Fig. 8-40. The in-line blending system blends pure components A and B to produce a product stream with flow rate w and mass fraction of A, x. Adjusting either inlet flow rate or Wg affects both of the controlled variables andi. For the pH neutrahzation process in Figure 8-40(Z ), liquid level h and the pH of the exit stream are to be controlled by adjusting the acid and base flow rates and w>b. Each of the manipulated variables affects both of the controlled variables. Thus, both the blending system and the pH neutralization process are said to exhibit strong process interacHons. In contrast, the process interactions for the gas-liquid separator in Fig. 8-40(c) are not as strong because one manipulated variable, liquid flow rate L, has only a small and indirec t effect on one controlled variable, pressure P. 

In the catalyst preparation area where the fire occurred, aluminum alkyl and isopentane are mixed in a batch blending operation in three 8000-gallon kettles. The flow rates of components are regulated by an operator at the control room. Temperature, pressure, and liquid level within the kettles are monitored by the control room operator. The formulated catalyst is stored in four 12,000-gallon vertical storage tanks within this process unit. Aluminum alkyl is a pyrophoric material and isopentane is extremely flammable. Each vessel was insulated and equipped with a relief valve sized for external fire. 

In essence, the test battery should include XRPD to characterize crystallinity of excipients, moisture analysis to confirm crystallinity and hydration state of excipients, bulk density to ensure reproducibility in the blending process, and particle size distribution to ensure consistent mixing and compaction of powder blends. Often three-point PSD limits are needed for excipients. Also, morphic forms of excipients should be clearly specified and controlled as changes may impact powder flow and compactibility of blends. XRPD, DSC, SEM, and FTIR spectroscopy techniques may often be applied to characterize and control polymorphic and hydrate composition critical to the function of the excipients. Additionally, moisture sorption studies, Raman mapping, surface area analysis, particle size analysis, and KF analysis may show whether excipients possess the desired polymorphic state and whether significant amounts of amorphous components are present. Together, these studies will ensure lotto-lot consistency in the physical properties that assure flow, compaction, minimal segregation, and compunction ability of excipients used in low-dose formulations. 

A necessary restriction in the linear programming input is the requirement that the volume fractions of the final blend components add up to 1.0. Other helpful restrictions are that minimum amounts of given solvents be present in the formulation, or conversely, that upper limits be set on the concentrations of given solvents. A Rule 66-type formulation could be handled in this manner, but a more satisfactory manner is to use three equations placing maximum restrictions on the three classes as specified by the Rule. For example, Class I solvents must be less than 0.05 volume fraction, Class II solvents must be less than 0.08 volume fraction, and the sum of the three classes must be less than 0.20 volume fraction. As Nelson (I) discussed, the viscosity may be controlled although we have not found this to be a significant factor, and we often omit the specification. In this case the logarithms of the viscosities of the... 

Three examples of simple multivariable control systems are shown in Fig. 8-39. The in-line blending system blends pure components A and B to produce a product stream with flow rate w and mass fraction of A, X. Adjusting either inlet flow rate or Wg affects both of the controlled variables w and x. For the pH neutralization process in Fig. 

---