1.Fluidized Bed Granulation Technology

1.1 Structure and Working Principle of Fluidized Bed

A fluidized bed granulator consists of a product container, distribution plate, spray nozzles, filter bags, air inlet & outlet ports, and material inlet & outlet ports.

Purified process air is heated and delivered into the container through the distribution plate, fluidizing and heating the powder feedstock. Meanwhile, binder liquid is atomized and sprayed into the fluidized powder bed. Fine powders agglomerate to form seed nuclei, which further grow into wet granules. Granulation and drying proceed simultaneously within the equipment, yielding soft, porous granules with a large specific surface area.

1.2 Advantages of Fluidized Bed Granulation

(1) Mixing, granulation and drying are completed in a single unit operation, featuring simplified production procedures and high automation level.

(2) The resulting granules possess regular spherical shape, uniform particle size and excellent solubility.

(3) The granules exhibit favorable flowability and compressibility. During tableting, the weight variation of tablets is minor; the obtained tablets have superior disintegration performance and excellent appearance quality.

(4) Minimal or negligible migration of soluble components occurs between granules, which reduces the risk of tablet content non-uniformity induced by such migration.

(5) The whole process is operated in a sealed vessel without dust emission, fully complying with GMP specifications.

(6) Fluidized bed granulation is suitable for the preparation of traditional Chinese medicine

(TCM) granules, especially those with high extract content and relatively low excipient dosage.

It is also applicable to granulation of heat and moisture-sensitive pharmaceutical substances.

Despite the numerous merits of fluidized bed technology, multiple process parameters including inlet air temperature, air volume, liquid spray rate and atomization pressure generally need to be adjusted during fluidized bed granulation to sustain stable fluidization of powder materials. Such adjustments rely on manual visual inspection or particle property data obtained via sampling analysis. The ultimate goal is to make critical quality attributes of granules such as moisture content and particle size distribution meet pre-defined specifications. Therefore, these influencing factors cannot be overlooked.

2.Factors Affecting Fluidized Bed Granulation

2.1 Granulation Raw Materials

First, the hydrophilicity of raw materials determines the selection and dosage of binders. Second, both hydrophilic and hydrophobic powders shall pass through an 80-mesh sieve; otherwise, the finished granules will have mottled color or excessive particle size. In addition, hygroscopic raw materials feature strong cohesiveness, poor flowability and high moisture absorption capacity, which shall be dried in advance before granulation. 

2.2 Inlet Air Temperature

High inlet air temperature produces granules with small particle size, high friability, low bulk density and poor flowability. Partial binder droplets evaporate completely before contacting powder materials, resulting in excessive fine powder in finished granules. Excessively high temperature leads to rapid solvent evaporation on granule surfaces, generating a large number of oversized dark-colored granules with dry exterior yet wet interior. Besides, certain powders tend to soften under high temperature accompanied by increased viscosity and deteriorated flowability, which gradually form massive agglomerates. 

2.3 Inlet Air Humidity

High inlet air humidity prevents timely drying of wet granules and causes adhesion between powder particles. When hygroscopic traditional Chinese medicine extract powder serves as the base material, high inlet air humidity often triggers massive agglomeration during material preheating and further leads to bed collapse. 

2.4 Binder Viscosity

High-viscosity binders generate large atomized droplets, yielding granules with large particle size, low friability, high hardness and severe agglomeration as well as abundant fine powder. Low-viscosity binders form tiny droplets that hinder granulation; the resulting granules contain excessive loose fine powder. 

2.5 Binder Feed Flow Rate

A high binder feed flow rate produces large, low-friability granules. With fixed atomization air pressure, increased binder flow rate raises the bulk density of granules. Excessively high flow rate will induce bed collapse, while low flow rate generates small-sized granules. 

2.6 Atomization Air Pressure

Raised atomization pressure atomizes binders into fine droplets with weakened powder wetting capacity, forming small-sized, high-friability granules. Excessively high pressure disrupts stable fluidization and causes turbulent air flow, leading to localized powder agglomeration. Conversely, low atomization pressure generates large binder droplets and large-size granules. [5]

2.7 Fluidizing Air Volume

Fluidizing air volume directly governs the fluidization status of raw materials. Excessively high air volume pushes the fluidized powder bed too close to the spray gun, resulting in uneven granule size distribution. Insufficient air volume leads to poor fluidization performance and potential bed collapse. 

2.8 Nozzle Position

The nozzle is mounted at the top of the fluidization chamber, and its installation height influences granule particle size distribution. To achieve the narrowest possible particle size distribution, the spray coverage area shall be adjusted to match the surface area of the wet powder bed. The closer the nozzle is to fluidized powder, the larger the granule particle size and the lower the friability. However, an overly short distance will cause the same adverse effects induced by excessive fluidizing air volume. 

2.9 Static Bed Depth

Static bed depth refers to the filling depth of raw materials inside the vessel before fluidization, which is determined by the designed production capacity of equipment and inherent properties of raw materials. For conical fluidized beds, low static bed depth facilitates obvious granule growth. Moderately shallow static bed delivers superior fluidization, heat transfer and mass transfer efficiency. If the bed depth is too shallow, air flow will penetrate directly through the powder layer without forming stable fluidization, thereby impairing granule quality. 

2.10 Other Influencing Factors

Additional factors including spray gun type, filter bag quality and filter bag shaking frequency also exert certain impacts on final granule quality.

3.Application of Process Analytical Technology in Fluidized Bed Granulation

Although operators have accumulated abundant practical experience in fluidized bed granulation, conventional control methods are plagued by prominent drawbacks including strong subjectivity and delayed parameter regulation, which fail to meet the stringent quality control requirements for fluidized bed granulation processes.

In recent years, researchers have developed a variety of process analytical technologies (PAT) based on near-infrared spectroscopy, Raman spectroscopy, microwave resonance, machine vision inspection and other analytical principles. These PAT tools are applied for real-time online monitoring of critical granule quality attributes during fluidized bed granulation, such as granule moisture content and particle size distribution, as well as for automatic endpoint determination of the granulation process.

3.Application of Process Analytical Technology in Fluidized Bed Granulation

Although abundant practical experience has been accumulated for fluidized bed granulation, conventional control methods suffer from prominent drawbacks including strong subjectivity and delayed parameter regulation, which fail to satisfy stringent quality control requirements for the granulation process. In recent years, researchers have developed various Process Analytical Technology (PAT) tools based on near-infrared spectroscopy, Raman spectroscopy, microwave resonance, machine vision inspection and other analytical principles. These PAT tools are adopted for real-time in-line monitoring of critical granule quality attributes during fluidized bed granulation, such as granule moisture content and particle size distribution, as well as for automatic judgment of granulation endpoint.

3.1 Spatial Filter Velocimetry (SFV)

Spatial Filter Velocimetry (SFV) is a particle velocity measurement method that relies on observation via a front-end receiving device implementing spatial filtering for moving objects. When particles pass through a laser beam, shadows are projected onto a linear fiber-optic array; the generated optical signal is proportional to particle velocity, and a single optical fiber generates a secondary pulse signal. With known pulse duration and particle velocity, the chord length of particles can be calculated. Chord Length Distribution (CLD) is affected by multiple variables such as particle morphology and particle orientation during movement.

To better reflect the underlying particle size distribution, Fischer et al. proposed a mathematical method to convert SFV-derived CLD data into particle size distribution, and verified the reliability of the conversion algorithm via periodic offline reference sampling and analysis. The measuring unit of the SFV probe is equipped with a sapphire window, which is kept clean by an internal compressed air supply system to prevent contamination. The internal air flow also guarantees effective dispersion of high-concentration particles and optimizes particle movement within the measuring zone. This technology supports particle size measurement ranging from 50 μm to 6000 μm, with a measurable particle velocity range of 0.01~50 m·s⁻¹ and a data acquisition rate up to 20,000 particles per second. Measurement results can be reported in multiple formats, including sieve distribution (mass fraction, passing fraction), volume-based distribution, number-based distribution, and velocity distribution.

Meanwhile, Hudovornik et al. compared measurement results obtained by SFV with offline sieve analysis and static image analysis. They found that SFV probe data exhibits favorable correlation with offline particle sizing methods. SFV can serve as an effective tool for coating thickness determination and real-time monitoring of attrition of coated pellets, yet it fails to detect agglomeration of coated particles.

In-line SFV technology is highly sensitive to variations in particle size during granulation and features an ultra-short measurement cycle, which helps deepen understanding of fluidized bed granulation and pellet coating processes. Furthermore, continuous rapid measurement of particle size distribution throughout fluidization improves process consistency, boosts production efficiency and strengthens overall product quality control.

3.2 Focused Beam Reflectance Measurement (FBRM)

Focused Beam Reflectance Measurement (FBRM) is an analytical technique for tracking particle growth dynamics. A low-power laser beam rotates at a linear speed of 2~8 m·s⁻¹ through a sapphire window. When the laser beam hits particles, light is backscattered. The backscattered beam is captured by an optical sensor, and electronic circuitry calculates the travel time of the laser beam across the full width of a particle. Combined with the known rotational speed of the laser beam, the chord length of individual particles can be computed. Thanks to high-speed laser rotation, chord lengths of thousands of particles can be measured per second and represented as chord length bins in CLD profiles.

Alshihabi et al. investigated the influence of FBRM probe installation position on analytical performance. Horizontal probe mounting was found to introduce noise, high signal scattering and unidentifiable curves in CLD profiles, which arises from powder deposition covering the probe window during liquid spraying. Tilting the probe at a 45° angle enables process air flow to sweep away any powder adhering to the window, keeping the window dry and clear throughout granulation. This configuration maximizes the number of sampled particles contacted by the laser beam, thus guaranteeing measurement accuracy of FBRM. The authors also proposed that more precise reference sizing techniques such as laser diffraction can further improve the accuracy of FBRM particle size prediction.

3.3 Near-Infrared Spectroscopy (NIRS)

The near-infrared (NIR) electromagnetic spectrum covers wavelengths from 780 nm to 2526 nm, corresponding to wavenumbers ranging from 12820 cm⁻¹ to 3959 cm⁻¹. Primary spectral information in the NIR region originates from overtone and combination band absorption of hydrogen-containing functional groups, including –CH, –NH, –OH and –SH. Such spectral signals are highly dependent on the chemical composition, physical components and inherent properties of tested samples, rendering NIRS widely adopted for real-time monitoring of granulation processes. Quantitative analysis via NIRS relies on chemometric software: calibration models are established by correlating reference chemical assay data with corresponding NIR spectral data, followed by determination of model parameters. The validated model is then applied for quantitative real-time monitoring of critical material attributes during production.

For instance, Frake et al. installed an NIR probe within the product bed of a top-spray fluidized bed granulator to continuously acquire spectral data of granules, enabling simultaneous monitoring of moisture content evolution and particle size variation throughout granulation.

In recent years, PAT has gained growing recognition from global regulatory authorities including the U.S. Food and Drug Administration (FDA) and European Medicines Agency (EMA), and has become an effective technical tool for process optimization and pharmaceutical quality research. PAT ensures robust process performance and consistent product quality. Deploying PAT monitoring during fluidized bed manufacturing delivers intuitive, detailed insight into dynamic particle evolution, securing process reliability and uniform product quality.

PAT also creates favorable conditions for transforming batch-mode fluidized bed granulation into continuous manufacturing processes, and provides critical reference value for the future implementation of continuous fluidized bed granulation technology across the pharmaceutical industry.

3.4 Acoustic Emission Technology

In addition to the aforementioned PAT tools, relevant research indicates that implementation of these analytical technologies requires partial modification of production equipment. Such intrusive sampling methods demand direct contact between measurement probes and process materials, which elevates the risk of pharmaceutical product contamination. Accordingly, acoustic emission technology is proposed as a non-invasive strategy for in-line process monitoring. Data acquisition and conversion devices are mounted externally on the granulator vessel wall, and process detection is completed via subsequent computer signal analysis.

Acoustic emission (AE) technology is a passive acoustic wave detection technique. Acceleration sensors installed on the outer wall of the granulator capture vibration signals and convert them into electrical signals. The electrical signals are amplified by a signal amplifier, collected by a data acquisition card, and processed by computer algorithms to characterize real-time process statusWiley.

During fluidized bed granulation, particle-particle collisions and particle-wall collisions and friction generate characteristic acoustic signals. Existing studies have verified that these acoustic signals contain abundant information correlated with particle physical properties including particle size, density and elastic modulus. Therefore, by continuously collecting acoustic signals during production and applying appropriate signal analysis algorithms, acoustic emission technology can extract comprehensive state information of the fluidized bed granulation process. Featuring convenient installation, non-invasive measurement, high sensitivity and low operational cost, acoustic emission technology is suitable for in-line monitoring of pharmaceutical manufacturing processes. Multiple published studies have validated its application for real-time measurement of critical quality attributes such as granule moisture content and particle size distribution.

Furthermore, research has applied acoustic emission technology to in-line monitoring of fluidized bed granulation for traditional Chinese medicine (TCM) granules. Taking the fluidized bed granulation process of 10 batches of Yangwei Granules as the research object, two core quality attributes (granule moisture content and particle size distribution) were selected as research indicators. The Partial Least-Squares Regression (PLSR) method was adopted to establish a quantitative correlation model between acoustic emission spectral signals and the physical properties of TCM granules. This model enables real-time reflection of dynamic granulation process status, and provides technical support for real-time feedback control of fluidized bed granulation processes for TCM granular formulations.

Conclusion

Process Analytical Technology (PAT) has been widely applied in pharmaceutical manufacturing. It enables real-time in-process monitoring of hazardous chemical reactions, solvent recovery and moisture levels during drying. In pharmaceutical formulation production, PAT can be utilized to monitor fluidized bed granulation, blend homogeneity, film coating and other core unit operations. Moreover, PAT also boasts extensive application prospects in traditional Chinese medicine (TCM) and biopharmaceutical industries.

Replacing traditional off-line laboratory quality control testing with PAT delivers multiple prominent benefits. It not only elevates overall production efficiency, cuts laboratory testing costs and reduces waste liquid generated during chemical testing to realize eco-friendly green manufacturing, but also improves batch-to-batch consistency of finished pharmaceutical products. At present, pharmaceutical regulatory authorities worldwide encourage pharmaceutical manufacturers to deploy PAT throughout their production workflows. It is believed that with joint efforts from regulatory authorities and domestic pharmaceutical enterprises, an increasing number of PAT projects will be implemented for real-time process monitoring in China’s pharmaceutical manufacturing lines. Ultimately, intelligent manufacturing and continuous manufacturing will be fully realized across the whole domestic pharmaceutical industry.