Paul Wan Sia Heng

FDA advocates building quality into a product through PAT.

  One of the ways this can be achieved is by in-, at- and on-line measurements of critical product and process attributes with various PAT-enabled equipment. In-line PAT involves mounting the instrument directly onto the process line without any sample extraction. For at-line PAT, the instrument is physically separate from the process line, but is located within the production floor, and periodic samples are taken and tested. On-line measurements require an eductor system to extract a portion of the product from the bulk flow for testing, after which it is returned to the main line. Both at-line and on-line applications allow for additional sample manipulation such as dispersion, whereas in-line application samples are evaluated in the path of product flow. 

The focus on the concept of Quality by Design instead of relying on end-product testing and release, has resulted in manufacturing industries investing considerable resources on systems that enable continuous monitoring of critical processes and product parameters. Concurrently, attention is on the design of analytical tools to meet PAT requirements. 

The function of PAT is to monitor critical quality attributes of a product or intermediate material. For a particulate system, particle size and particle size distribution are major considerations. From the production of toner and cement to raw drugs and final dosage forms, the particle size distribution of the product is an important, defining characteristic and has a huge impact on quality. PAT allows the manufacturer to achieve and maintain a high level of process understanding and control. While particle sizing by PAT has been reported for the measurement of primary particles undergoing milling and crystallization processes,

  little information is available on utilizing PAT for sizing products that have been created with processes such as spray drying and microencapsulation.  

Spray drying is common in the pharmaceutical and food industries, and can be used for the microencapsulation of a solid or an oily liquid. Spray-dried microspheres can control release and protect core materials against oxidation and degradation, as their walls form a physical and permeability barrier against the effects of the external environment. Other useful applications include taste masking and solubility enhancement. 

Controlling microsphere size is a must as it is an important process variable that can affect product performance. The conventional method of sizing involves periodic sampling and subsequent analysis using off-line techniques, but these have limitations associated with late feedback response times and sampling errors. More importantly, the information lacks the sensitivity required for it to be used in the detection of fluctuations that indicate changing process trends. If left unchecked, these could affect overall product quality.  

Utilizing PAT as an in-process monitor during spray drying could offer better process control, improve product quality and, ultimately, result in products of greater value, thus increasing cost-effectiveness.  

This article discusses the use of an in-line particle sizer, Insitec (Malvern Instruments; UK), with a pilot-scale spray dryer for the production of microspheres. Both the in-line and at-line use of the instrument are explored and evaluated.  

The Insitec system comprises several parts: the optical head, an interface box, and computer and data analysis software. The optical head functions as a standard laser diffraction system. For the in-line set-up, the optical head is directly connected to the process stream (Figure 1). Unlike the on-line systems more commonly used in the industry (which are designed for bigger product throughput), there is no eductor that extracts part of the product flow for measurement before returning it to the main product line. In the smaller spray dryer, the entire product flow is monitored. This is only possible if the product flow stream is not too large. In the at-line set up, the laser module is not connected to the product flow stream and is positioned instead adjacent to the spray dryer, where it works as a separate system. Sampling is conducted when necessary. 

Figure 1: Layout of spray dryer with in-line and at-line Insitec.			

An important requirement for the optimal performance of Insitec is the purge air flow, which is set at an air velocity greater than that of the primary gas flow. Purging is essential to prevent the adherence of particles to the surface of the optical lenses, which would cause constant background counts and affect the accuracy of the results. For the at-line set up, there is usually an additional purge air supply 

 the venturi that aids in particle dispersion. In this case, the air flow rate must be high enough to sufficiently break down agglomerates, but not too high so as to cause particle attrition. 

A spray dryer was used to produce oil-encapsulated microspheres in the 10–35 μm range. The oily nature of the product meant that conventional off-line sizing using a benchtop laser diffraction system was not possible, as the presence of oil deposits not only caused considerable agglomeration problems, but constantly contaminated the sample conduit, necessitating extensive cleaning after each run. Thus, particle sizing had to be performed using light microscopy interfaced with an image analysis system — a tedious and time-consuming method that relied heavily on operator skill and experience because the small sample used made it prone to sampling errors. Microscopic measurements are difficult to use as a real-time quality control system for large-scale production, and are best used during the initial R&D phase for the elucidation of detailed particle morphology and shape information. They are also useful as a visual verification system for data obtained using laser diffraction. 

As with most process equipment, spray dryer operation involves a period of start-up and shut-down, during which product quality is most prone to fluctuation. In most cases with spray drying, microsphere size is a concern. With any off-line sizing methods, it is virtually impossible to gather time-sensitive information on product quality and the exact duration needed for the product to fall within required specifications.  

In this particular spray drying operation, an estimation of 10 min was used. During this time, the microspheres produced were discarded to ensure stable product quality. The in-line sizing set-up continuously monitored microsphere size throughout the operation, providing information on the exact duration required for the process to stabilize from start-up, which allowed a real-time decision to be made regarding when a steady state was achieved. In this study, approximately 2 min was found to be an adequate waiting time for a stabilized product to be attained (Figure 2). As the duration of start-up was now accurately determined, precise switch-over to product collection could be achieved, avoiding unnecessary product waste and establishing greater confidence in product quality. 

Figure 2: Time history graph of particle size during spray dryer start-up.			

Despite the ability of the in-line system to provide real-time information about the process, it had its limitations. The most apparent was the inability of the in-line sizers to accurately size particles that tended to agglomerate. It was a significant problem for this particular application because of the propensity of the microspheres to aggregate, as a result of their small size and presence of surface oil. Agglomeration was most pronounced for microspheres produced at high-oil loadings. This resulted in grossly elevated particle size results with bimodal particle size distributions skewed towards the right. Discrete microspheres and agglomerates were undifferentiated during measurement, and mean size values were not representative of true microsphere sizes. 

The inaccurate size measurement was not caused by the accuracy of the sizer, but was related to its set-up. As mentioned previously, the in-line configuration measured particles collectively as they passed through the sizing zone, without the aid of an additional eductor or deaggregative purge air for the particle stream. Agglomerates were measured as single particles as they moved in transit through the sizer.  

This problem could be solved by incorporating an additional sample dispersion system directly into the process line, such as the addition of an in-line purge air system at the point just before the microspheres reach the particle sizer (Figure 3). In this way, agglomerates could be broken up into discrete particles for analysis. However, its effects on the primary product flow need to be studied to ensure that minimal disruption occurs. Care should also be taken for friable products. One way to reduce this would be to initiate the purge air flows at periodic intervals when measurements were taken, as opposed to leaving it on throughout the run. As such, a totally enclosed system could be achieved with effects similar to at-line systems. 

Figure 3: (a) In-line and (b) on-line purge air system.			

Appropriate real-time analysis of the captured particle size distributions could also be performed. Data analysis could be conducted to mathematically separate the bimodal curves into two single curves, or to report particle size as two separate values: one for primary particle size and the other for agglomerates (Figure 4). This could be a useful qualitative indicator of the extent of agglomeration of a particular product. If it is certain that particles greater than a specific size are agglomerates and not of interest, the range of data acquisition could be modified to disregard those particles. 

Figure 4: Separation of bimodal distribution into two separate modes. 			

Besides the in-line use of Insitec, it was also used at-line by positioning it adjacent to the spray dryer. Samples were withdrawn at periodic intervals and sized.  

For the same spray dried formulation, particle size distributions were narrower and the second modes were absent, implying that most agglomerates were broken up by the eductor air flow and that discrete microspheres were sized. Although this set-up effectively sized discrete microspheres, it was not as useful as an in-process tool for the detection of subtle changes in product quality because of the practical limitations in sampling frequency — although response times were relatively quicker than conventional benchtop sizing equipment. The need for manual sampling and feeding required trained operators, and would likely be prone to human errors and could be biased as opposed to in- and on-line systems. It was also a destructive analysis with no sample recovery: a factor which would have an impact on the production of costly materials. Additionally, unlike the in-line set up, it was not a totally enclosed system. This gives rise to safety and environmental concerns for potent or toxic materials. It would also be less suitable in cases where aseptic conditions were required. 

In this article, the experience of using a PAT instrument (Insitec) with a pilot-scale spray dryer for microsphere sizing was described. The instrument was used both in-line and at-line, and the objectives of process monitoring and microsphere sizing were achieved. Modifications were proposed to enable the particle sizer to meet the specific requirements of the spray drying operation more effectively. Nevertheless, the system was found to be a useful tool for rapid and convenient sizing of oil-loaded microspheres compared with the conventional method of light microscopy. It was also able to provide real-time information about the process and product size. 

 is a doctorate student of the National University of Singapore. 

			 is Associate Professor with the Department of Pharmacy, National University of Singapore. 

1. FDA — Process Analytical Technology (PAT) Initiative, February 2005. 

2. S. Brenek, D. am Ende and P. Rose, "High shear wet milling to achieve particle size with 

Paul Wan Sia Heng

Articles