Before diving into technical details, here's a simple roadmap of our hardware project:
What it is: We built a Dry Powder Inhaler (DPI) to deliver our freeze-dried engineered probiotics into the lungs.
For our therapy to work, the bacteria must reach the deep lung regions. Standard inhalers are not optimized for live microbial powders, so we designed one that balances efficiency, safety, and ease of use.
How it works:
1. The probiotic powder is loaded in a capsule.
2. When the patient inhales, air enters the device and creates a swirling airflow inside the chamber.
3. This turbulence breaks up clumps of powder and carries fine particles through the mouthpiece.
4. The particles, with an optimized size around 3 micrometers, travel to the lower airways where they are most effective.
What we achieved:
Low inhalation resistance (0.0222 KPa^(1/2)/L/min), making it suitable even for patients with reduced breathing capacity.
Consistent fine particle fraction, targeting alveolar deposition.
A simple, low-cost design that can be 3D printed and tested.
In short, this inhaler is the bridge between our biology and the patient: it turns our engineered probiotic into a therapy that can actually be delivered to the lungs.
The next sections explain in detail how we designed, simulated, built, and tested the device.
Delivering live probiotics to the deep lung requires an inhaler that produces enough turbulence to break up powder agglomerates and form aerosol in the 1–5 µm aerodynamic range; particles around 3 µm maximize bronchiolar and alveolar deposition. (1,2) Conventional DPI designs are built around small-molecule APIs and may not meet the dispersal, resistance, or payload-integrity needs of microbial powders; therefore we designed a purpose-built DPI that balances dispersion, low resistance, and manufacturability. (2,3)
To iterate the design efficiently we combined SOLIDWORKS 2022 SP1.0 CAD with steady-state CFD (SOLIDWORKS Flow Simulation) and focused experimental validation (bench emission tests and pressure-drop measurements) to converge on the final geometry. CFD is a well-established tool in inhaler design and was used here to predict flow, vorticity, particle trajectories and deposition before committing to multiple physical prototypes. (4,5)
Design targets (driven by the biology, patient usability, and engineering constraints):
1. Low device resistance: target ≈ 0.0222 kPa^0.5·L⁻¹·min⁻¹ (when loaded) — within the lower half of commercial capsule DPIs. (3)
2. High fine particle fraction (FPF): > 60% for particles 5 µm. (2)
3. MMAD ≈ 3.0 µm (±0.4 µm) to maximize deep lung deposition. (1)
4. High emitted dose (ED): > 85% (minimize first-dose loss). (2,4)
5. Robust across flow rates: consistent performance from weak to strong inhalations (60–100 L/min). (3,6)
6. Manufacturability & usability: 3D-printable prototype geometry, snap-fit assembly, low actuation force, and design transitions to injection molding. (7)
These targets guided each geometry and mechanism choice described below.
Below we list the principal parts with the engineering rationale and the design constraints that informed them.
Mouthpiece length: 65 mm (tuned by simulation across 30–80 mm).
Internal diameter: 10 mm — selected to deliver appropriate exit velocity for 3 µm particles.
Protective grid: 1.5 mm circular apertures arranged hexagonally (~68% porosity) to prevent large fragments while straightening flow and reducing local jetting and throat impaction. (See CFD evidence in Section 5.) (8)
Mouthpiece component showing the protective grid at the inlet, the optimized length for flow development, and the precisely dimensioned outlet. The transparent rendering reveals the internal geometry that guides airflow and particles toward the patient's respiratory system.
Tangential inlets: two × 3.5 mm (180° apart) to induce cyclonic flow and capsule rotation.
At 100 L/min total inspiratory flow the inlet velocity is ~40 m/s, producing a coherent vortex that encourages deagglomeration with minimal pressure penalty. Chamber wall curvature and radial clearances were optimized to reduce stagnation. (4)
Capsule chamber shows the two tangential inlets that create cyclonic flow, the optimized internal volume for capsule accommodation, and the lightweighting features that reduce material usage. The cutaway view reveals the smooth internal surfaces and strategic material removal patterns.
Capsule: size 3 HPMC (≈ 15.9 × 5.8 mm) chosen for structural stability and favorable moisture behaviour compared with gelatin. (9)
Piercing: two opposing 0.5 mm holes (steel pins); fewer larger apertures produce coherent jets that assist deagglomeration and lower MMAD vs many small holes. (10)
Pin geometry: conical pins with 30° tip, stroke 3.0 mm — sufficient to fully perforate a ~100 µm capsule wall.
Actuation: multi-spring return for low actuation force and durability (>10,000 cycles); tactile/audible "click" indicates full puncture. FEM confirmed component longevity and elastic return. (7)
Piercing button components (×2) with integrated spring mechanisms. Each button features a precision pin for creating controlled apertures and four supporting springs for reliable return action. The ergonomic button surface provides comfortable finger contact during operation.
Snap-fit protective cover with dust-tight lip; polypropylene base cap chosen for low moisture uptake and sterilization compatibility. Transparent base enables visual alignment checks. (material selection and sealing validated in prototyping tests).
Base cap component featuring optimized ribbing structure for strength and lightweighting. The design incorporates snap-fit features for tool-free assembly and strategic material removal to reduce manufacturing costs and environmental impact.
We used SOLIDWORKS Flow Simulation with Lagrangian particle tracking for steady-state analyses. The workflow follows inhaler CFD best practice and the approaches described in inhaler literature. (4,5)
Simplified fluid domain by removing non-fluid solids; mesh reduced from ~12 million → ~4.2 million cells to focus compute on the flow region.
Local mesh refinement targeted capsule perforations and the protective grid.
Mesh independence was established: pressure-drop variation was <2% with further refinement (Grid Convergence Index / Richardson extrapolation principles applied). (7)
Boundary conditions: inlet at ambient pressure, temperature ≈ 20 °C; outlet specified as volumetric flow = 100 L/min (0.001667 m³/s).
Air model: incompressible; walls treated as smooth and adiabatic.
Turbulence: k - ε model used (suitable for average turbulent statistics in this geometry); selected turbulence closure tuned by comparison to literature and sensitivity checks. (4)
Computational mesh of the DPI showing refined regions near walls and in areas of high velocity gradient. The mesh contains approximately 4.2 million cells with 8-12 cells spanning the smallest flow passages. Mesh refinement is particularly dense around the capsule apertures, at the tangential inlets, and within the mouthpiece grid where accurate resolution of flow gradients is critical for predicting particle behavior.
Case 1 — empty device: baseline flow and resistance (no capsule).
Case 2 — loaded device: size 3 HPMC capsule pierced by two 0.5 mm holes.
Tracked outputs: velocity fields, pressure distribution, turbulent kinetic energy / vorticity, streamlines, discrete particle trajectories and predicted deposition fractions. Where possible, CFD outputs were cross-checked with bench measurements (pressure-drop, emitted dose).
Outlet boundary condition specifying the volumetric flow rate of 100 L/min (0.001667 m³/s). This flow rate represents the peak inspiratory flow rate for an adult patient with severe uncontrolled asthma (Type 2), based on clinical data showing mean PIFR of 106 L/min with standard deviation of 16 L/min.
Inlet velocity: ~40.2 m/s.
Pressure drop: ≈ 4.824 kPa.
Computed resistance: 0.02196 kPa^0.5·L⁻¹·min⁻¹.
Reynolds number: ~9,250–11,700 (turbulent flow).
Showing the air trajectories from the inlet through the DPI device to the outlet. The variety of the velocity magnitude is presented as a variety of colors using contour palette.
Showing an XZ plane cut plot from the internal base of the capsule chamber through the DPI device’s flow domain to the outlet. The variety of the velocity magnitude is presented as a variety of colors using contour palette.
Showing the air trajectories from the inlet through the DPI device to the outlet. The variety of the pressure magnitude is presented as a variety of colors using contour palette.
Showing an XZ plane cut plot from the internal base of the capsule chamber through the DPI device’s flow domain to the outlet. The variety of the pressure magnitude is presented as a variety of colors using contour palette.
Pressure drop: ≈ 4.909 kPa.
Resistance: 0.02216 kPa^0.5·L⁻¹·min⁻¹ (+0.9% vs empty).
Local jet speeds: >60 m/s at capsule perforations (coherent jets).
CFD fields show strong cyclonic vortices and recirculation zones that facilitate deagglomeration and entrainment into the mouthpiece flow (4,5).
Showing the air trajectories from the inlet through the DPI device to the outlet. The variety of the velocity magnitude is presented as a variety of colors using contour palette.
Showing an XZ plane cut plot from the internal base of the capsule chamber through the DPI device’s flow domain to the outlet. The variety of the velocity magnitude is presented as a variety of colors using contour palette.
Showing an XShowing the air trajectories from the inlet through the DPI device to the outlet. The variety of the pressure magnitude is presented as a variety of colors using contour palette.Z plane cut plot from the internal base of the capsule chamber through the DPI device’s flow domain to the outlet. The variety of the pressure magnitude is presented as a variety of colors using contour palette.
Showing an XZ plane cut plot from the internal base of the capsule chamber through the DPI device’s flow domain to the outlet. The variety of the pressure magnitude is presented as a variety of colors using contour palette.
Predicted MMAD: ≈ 3.0 µm (±0.4 µm). (1,2)
FPF (<5 µm): >60%. (2)
Emitted dose (ED): >85%; bench tests measured residual deposition <5% , consistent with CFD trends.
Particle tracking shows the majority of generated fines are transported through the mouthpiece with minimal wall losses. (5,6)
Three-dimensional velocity magnitude flow trajectories showing the complex three-dimensional flow structure created by the capsule. Airflow enters tangentially, accelerates around the capsule, forms jets through the apertures, and eventually converges toward the mouthpiece. The color coding reveals velocity stratification from high-velocity inlet jets (red) to moderate-velocity core flow (green) to slower wall boundary layers (blue).
Surface pressure distribution on the capsule itself showing high pressure (blue) on windward surfaces facing the incoming flow and low pressure (yellow-red) on leeward surfaces in the wake region. The pressure difference creates aerodynamic forces that can rotate and translate the capsule. The apertures appear as localized low-pressure spots (red) where flow accelerates dramatically to ~60 m/s, creating a pressure drop via Bernoulli's principle.
Collectively, these results place the prototype among high-efficiency capsule DPIs while maintaining low inhalation resistance.
The measured device resistance (0.0222 kPa^0.5·L⁻¹·min⁻¹) is in the low–mid commercial range, supporting ease of inhalation by patients with impaired lung function. (3)
The +0.9% resistance increase with a loaded capsule indicates good aerodynamic robustness and low sensitivity to the capsule state.
MMAD ≈ 3.0 µm, FPF >60%, ED >85% are consistent with design criteria for deep lung targeting and efficient therapeutic delivery. (1,2,4)
The device was designed for manufacturability from the start: snap-fits, ribbing, internal lightweighting and a 3D-print prototype route that converts to injection molding for scale-up. (7)
CFD-guided development reduced physical prototyping cycles and is consistent with the literature recommendation to combine simulation with targeted experiments. (4,5)
Vorticity magnitude variation around outlet perimeter with capsule (Case 2). The distribution is similar to Case 1, with values ranging from 200-1,400 s^-1 and higher vorticity near walls (positions 0-90° and 270-360°). The capsule slightly increases average vorticity by ~5% due to additional turbulence generation in the chamber, which may enhance particle dispersion.
Air density variation around outlet perimeter with capsule (1.202-1.206 kg/m³), essentially identical to Case 1. The uniform density confirms incompressible flow conditions are maintained regardless of capsule presence.
Our results demonstrated the inhaler can reproducibly disperse particles with minimal dose loss while still standing usability for the patient, having resistance in individual ranges. The application of numerical simulation in tandem with experimental validation shortened development cycles and served as a solid scientific foundation for design changes.
This hardware provides not only a delivery system for our engineered probiotic but also offers a modular scaffold that future teams could exploit for other therapeutic microorgs. Future directions to help the realization of the clinical relevance of this engineered prototype include cascade impactor studies, microbial viability assays, and usability testing.
1. Islam N, Cleary MJ. Developing an efficient and reliable dry powder inhaler for pulmonary drug delivery — a review. Med Eng Phys. 2012;34(4):409–425.
2. Dhoble S, Merwin J, et al. Design, development, and technical considerations for dry powder inhalers. Drug Discov Today. 2024; in press.
3. Shur J, Lee S, Adams W, et al. Effect of device design on the in vitro performance and comparability for capsule-based dry powder inhalers. AAPS J. 2012;14(3):667–676.
4. Dal Negro RW. Dry powder inhalers and the right things to remember: a concept review. Multidiscip Respir Med. 2015;10:13.
5. Ruzycki CA, Javaheri E, Finlay WH. The use of computational fluid dynamics in inhaler design. Expert Opin Drug Deliv. 2013;10(3):307–323.
6. Wong W, Fletcher DF, Traini D, Chan HK, Young PM. The use of computational approaches in inhaler development. Adv Drug Deliv Rev. 2012;64(4):312–322.
7. Sulaiman M, Liu X, Sundaresan S. A CFD-DEM investigation of powder transport and aerosolization in ELLIPTA® dry powder inhaler. Powder Technol. 2022;409:117817.
8. Krüger P, et al. Inspiratory flow resistance of marketed dry powder inhalers (DPI). European Respiratory Journal. Supplement. 2014;44:4635.
9. Nadarassan DK, et al. Aerodynamic characteristics of a dry powder inhaler at low flow rates. J Aerosol Sci. 2010;41(11):983–991.
10. Lexmond AJ, Kruizinga TJ, Hagedoorn P, et al. Effect of inhaler design variables on paediatric use of dry powder inhalers. PLoS ONE. 2014;9(6):e99304.
11. Bock S, et al. Additive manufacturing in respiratory sciences: 3D printing for inhaler prototyping and particle carriers. (Review) 2022.
12. Fletcher DF, et al. Effects of grid and exit geometry on inhaler flow in CFD studies. [Conference or technical article].
13. Dhoble S, et al. (cited above) – chamber and inlet design discussion.
14. Magramane S, et al. Comparative evaluation of HPMC and gelatin capsules for inhalation.
15. Tong Z, Zheng B, Yang R, Yu A, Chan HK. CFD-DEM investigation of the dispersion mechanisms in commercial dry powder inhalers. Powder Technol. 2013;240:19–24.
