MDI & DPI particle physics — from actuation to alveoli. Inertial impaction, gravitational sedimentation, and Brownian diffusion determine where inhaled drug actually lands.
Every inhaled therapy begins with a device converting a metered drug dose into an aerosol cloud of respirable particles. Pressurized metered-dose inhalers (pMDIs) and dry powder inhalers (DPIs) achieve this through fundamentally different physics — one driven by propellant expansion, the other by the patient's own inspiratory airflow — and each imposes its own constraints on technique and resulting particle-size distribution.
Pressurized MDI (pMDI): • A metering valve doses 25–100µL of drug suspended or dissolved in liquefied hydrofluoroalkane (HFA) propellant (HFA-134a or HFA-227). • On actuation, the propellant flash-evaporates as it exits the nozzle at high velocity (peak ~20–30 m/s at the orifice, decelerating rapidly within the first 10cm). • Initial droplets are large (30–45µm) but shrink rapidly as residual propellant evaporates, reaching a near-final MMAD of roughly 2–5µm within a few centimeters of the mouth. • Requires actuation-inhalation coordination: the patient must trigger the device within about the first second of a slow, deep inhalation. Poor timing is the dominant source of MDI failure.
Dry Powder Inhaler (DPI): • Drug is pre-metered as micronized powder, typically blended with coarse lactose carrier particles (agglomerates) or formulated as engineered low-density particles. • No propellant — the device is entirely passive. The patient's own inspiratory effort creates turbulent airflow inside the device (through a grid, cyclone, or spiral channel) that shears the powder agglomerates into respirable-sized fragments. • Because de-aggregation energy scales with flow rate, DPIs require a minimum inspiratory flow — typically 30 L/min for low-resistance devices and up to 60 L/min for high-resistance devices — to achieve their labeled fine-particle fraction. • Removes the coordination problem entirely (breath-actuated) but introduces a flow-rate dependence problem instead: children, the elderly, and patients in acute bronchospasm may be physically unable to generate sufficient flow.
A landmark inhaler-technique audit (Melani et al., Respiratory Medicine 2011, >1,600 patients) found critical errors in the majority of both MDI and DPI users, and technique errors were directly associated with worse asthma/COPD control and more emergency visits — underscoring that device physics only delivers its theoretical performance when technique is correct.
The mouth and pharynx form a narrow, sharply angled conduit — roughly a 90° turn at the back of the throat — that any inhaled particle must negotiate before reaching the trachea. Particles with enough momentum cannot follow the air as it curves; they travel in a straight line into the mucosal wall. This single mechanism, inertial impaction, accounts for the largest and most avoidable loss of inhaled dose.
Whether a particle impacts or follows the airflow streamline around a bend is governed by the Stokes number:
St = (ρp · d² · U) / (18 · µ · Dh)
where ρp is particle density, d is particle (aerodynamic) diameter, U is air velocity, µ is air viscosity, and Dh is the hydraulic diameter of the airway at the bend. Physically, St compares a particle's stopping distance to the characteristic size of the flow curvature.
• St ≪ 1: the particle has negligible momentum relative to the airway geometry — it follows the streamline smoothly around the bend and continues downstream. This favors small, slow-moving particles. • St ≳ 1: the particle's inertia dominates — it cannot turn with the air and impacts the wall at the bend. Because St scales with d² and with U, impaction risk rises sharply for larger particles and for higher plume velocities.
This is precisely why pMDI plume velocity matters so much: the propellant-driven jet exits at up to ~30 m/s, and a patient who fires the device directly into a closed or poorly-positioned mouth delivers particles at high U straight into the 90° oropharyngeal bend. Coarse fractions above roughly 5µm MMAD are impaction-dominated almost regardless of technique; particles above 10µm rarely escape the mouth at all.
Spacers and valved holding chambers (VHCs) intervene on both variables at once: they add distance and volume between the device and the mouth, allowing the plume to decelerate (lowering U) and allowing residual propellant to evaporate (shrinking d toward the true drug-particle size) before inhalation begins. Classic chamber studies (Newman et al., beclomethasone MDI) showed oropharyngeal deposition falling from roughly 80% with direct actuation to around 17% with a chamber — while lung deposition simultaneously improved.
Because oropharyngeal deposition is swallowed and absorbed systemically from the GI tract, it is not merely wasted dose — for inhaled corticosteroids it is a direct driver of local side effects (oral candidiasis, dysphonia) and contributes to systemic corticosteroid exposure. Spacer use and post-dose mouth rinsing are standard clinical mitigations for exactly this reason.
Particles that survive the oropharyngeal bend and continue into the trachea and bronchial tree encounter progressively narrower, lower-velocity airways across roughly 16 generations of branching (Weibel model). In this zone, both inertial impaction and gravitational sedimentation contribute to deposition, and the 1–5µm size window — the "respirable" or "fine particle" fraction — is specifically engineered to land here, on the airway smooth muscle and epithelium that carry the drug's pharmacological targets.
As the airway tree branches (trachea → bronchi → bronchioles), two deposition mechanisms trade off dominance:
Impaction remains significant in the larger, higher-velocity central airways (generations 0–6), particularly at each bifurcation where the same Stokes-number logic from the oropharynx applies at smaller scale. Particles at the upper end of the respirable range (3–5µm) deposit preferentially here.
Sedimentation becomes dominant as airway diameter shrinks and residence time increases in the smaller bronchi and bronchioles (generations ~7–16). Under gravity, a particle reaches terminal (Stokes) settling velocity when drag balances weight:
Vs = ρp · d² · g / (18 · µ)
For a unit-density (1 g/cm³) spherical particle, this gives roughly 0.03 mm/s at 1µm, 0.27 mm/s at 3µm, and 0.75 mm/s at 5µm — small absolute speeds, but airway residence time in the bronchioles (low velocity, small cross-section) is long enough for meaningful settling, especially with an end-inspiratory breath-hold.
Why 1–5µm is the design target: below ~1µm, particles have too little mass to impact or settle efficiently in this timeframe and mostly continue toward the alveoli or are exhaled; above ~5µm, impaction losses in the mouth and central airways dominate before the particle ever reaches the bronchioles. The 1–5µm window is the practical intersection where enough particles survive the oropharynx yet still deposit before reaching the alveolar region — landing directly on the β2-adrenergic receptors of airway smooth muscle (bronchodilators) or the glucocorticoid receptors of the bronchial epithelium (inhaled corticosteroids).
Published lung deposition for optimized inhaler/technique combinations: budesonide via Turbuhaler DPI achieves approximately 32% lung deposition (Borgström et al.), versus roughly 10–20% for a poorly-used conventional MDI — nearly a two-fold difference driven almost entirely by particle size control and technique.
Particles smaller than roughly 1–2µm are light enough that inertial impaction and even simple sedimentation become inefficient within the timescale of a single breath. In the vast, low-velocity alveolar region these particles are instead governed by a competition between slow gravitational settling and random Brownian diffusion — a regime that is exploited deliberately for systemic drug delivery, and that also explains why the smallest aerosol particles are the most likely to simply be exhaled again.
In the alveolar region, airflow velocity drops close to zero (the enormous ~70m² cross-sectional area of ~300 million alveoli spreads the same tidal volume across a huge area), so neither impaction nor bulk convection meaningfully deposits particles. Two slower mechanisms take over:
Gravitational sedimentation: still governed by Vs = ρp·d²·g/(18µ), but at 1µm this settling speed (~0.03mm/s) only deposits a meaningful fraction of particles if the patient performs an end-inspiratory breath-hold of several seconds — allowing time for particles to fall the short distance to the alveolar wall before exhalation.
Brownian diffusion: below roughly 0.5µm, thermal collisions with gas molecules randomize particle trajectories enough that diffusion — not gravity — becomes the dominant deposition mechanism, described by the Stokes–Einstein diffusion coefficient D = kT/(3πµd). Diffusion is more effective as particles get smaller, since D scales inversely with d.
Because sedimentation favors larger sub-micron particles and diffusion favors smaller ones, classical aerosol theory predicts a "most penetrating particle size" around 0.3–0.5µm where total deposition efficiency in the respiratory tract is at a minimum — particles in this narrow window are the most likely of any size to be exhaled entirely unretained, sometimes described in filtration/inhalation toxicology literature as the most penetrating particle size (MPPS).
This same physics is deliberately harnessed for systemic (not just local) drug delivery. Because the alveolar epithelium is only one cell layer thick with an enormous absorptive surface area, drug deposited here enters the bloodstream almost as fast as an intravenous injection. Inhaled insulin exploited exactly this: Exubera (Pfizer, withdrawn 2007) used a bulky device and coarser aerosol that limited practical adoption, while Afrezza (Technosphere insulin, approved 2014) uses engineered fumaryl diketopiperazine microparticles with an MMAD of roughly 1–3.5µm that dissolve rapidly at alveolar pH, producing an insulin absorption profile with onset around 12–15 minutes — faster than rapid-acting subcutaneous insulin analogs.
The same deep-lung deposition pathway that enables ultra-fast systemic drug absorption also means sub-micron combustion particles, viral aerosols, and environmental nanoparticles reach the alveoli efficiently — the identical physics (Stokes number, sedimentation, Brownian diffusion) underlies both therapeutic inhaler design and inhalation toxicology / respiratory infection transmission modeling.
Every deposition loss described so far — oropharyngeal impaction, flow-rate-dependent DPI de-aggregation, and sub-micron exhalation — has a corresponding engineering or behavioral countermeasure. Modern inhaler design combines device physics (spacers, breath-actuated valves) with particle engineering (low-density porous particles) to push a larger fraction of the labeled dose into the 1–5µm therapeutic window and out of the mouth or exhaled breath entirely.
Spacers and valved holding chambers (VHCs): by adding a rigid air volume between the MDI actuator and the mouth, spacers accomplish two things simultaneously — they let the high-velocity propellant plume decelerate before inhalation (lowering the Stokes number at the oropharyngeal bend), and they let residual HFA propellant evaporate so particles shrink toward their true drug-particle aerodynamic diameter before being inhaled. A one-way valve on true VHCs also removes the actuation-inhalation coordination requirement entirely, since the aerosol is held in the chamber until the patient inhales at their own pace. Anti-static chamber materials further reduce electrostatic particle loss to the chamber walls, a problem that plagued early plastic spacers.
DPI flow-rate dependence and its mitigation: because DPI de-aggregation energy comes from the patient's inspiratory effort, fine particle fraction is inherently flow-dependent — too little flow leaves drug agglomerated with the lactose carrier (which then impacts in the throat as an oversized particle); paradoxically, too much flow can also increase inertial impaction losses in the mouth and throat. Device engineering (cyclone geometries, optimized internal resistance) narrows the flow range over which the device performs consistently, and patient training targets a steady, forceful inhalation rather than a slow one (the opposite technique from an MDI).
Particle engineering — porous, low-density particles: aerodynamic diameter dae relates to the physical (geometric) diameter dgeo by dae = dgeo·√ρ. By manufacturing large, geometrically bulky but highly porous, low-density particles (density ~0.1 g/cm³ instead of ~1 g/cm³ for a solid crystal), formulators can produce a particle with a small aerodynamic diameter — landing in the 1–5µm respirable window — while its large physical size makes it far less prone to the agglomeration and electrostatic clumping that plague fine milled powders, and largely removes the flow-rate dependence problem. Novartis's PulmoSphere technology (spray-dried porous particles) is used in products such as TOBI Podhaler (inhaled tobramycin for cystic fibrosis) and combination COPD therapies, achieving fine particle fractions well above those of conventional jet-milled DPI powders even at low inspiratory effort.
Combining a well-fitted VHC with a corticosteroid/bronchodilator MDI and correct technique, or a porous-particle DPI at adequate flow, can push total lung deposition from the 10–20% typical of unaided, poorly-coordinated MDI use up into the 50–60% range reported in optimized studies — roughly a three- to five-fold improvement achieved entirely through device and formulation engineering, without changing the drug molecule at all.