Abstract
Sublingual sprays occupy a commercially attractive space in nutraceutical and pharmaceutical delivery: they bypass hepatic first-pass metabolism, exploit the highly vascularised sublingual mucosa, and offer needle-free rapid absorption. The conventional solution to formulating complex botanical and amino acid blends has been to include ethanol at concentrations of 15–40%, where it acts simultaneously as solvent, wetting agent, and antimicrobial preservative. As consumer demand, regulatory guidance, and paediatric or alcohol-sensitive indications push formulators toward alcohol-free aqueous platforms, a cascade of stability failures emerges. This article examines those failure modes in physicochemical depth — crystallisation of amino acids, phase separation of lipophilic botanical fractions, and nozzle obstruction — and then surveys the engineered architectures that can circumvent them.
1. The Appeal and the Problem
Sublingually delivered liquids reach systemic circulation within minutes. The sublingual mucosa presents a non-keratinised epithelium with a mean thickness of only 100–200 µm and dense capillary perfusion, making it among the most permeable mucosal surfaces accessible without invasive devices. [^1] In a simple ethanolic solution, lipophilic botanical actives and polar amino acids alike remain solubilised: ethanol disrupts the hydrogen-bonding network of water, depresses the dielectric constant of the medium, and creates a miscible organic continuum in which both hydrophilic and hydrophobic solutes can coexist. Remove the ethanol and replace it with water, glycerin, or aqueous glycerin blends, and thermodynamic reality reasserts itself with considerable force.
Three principal failure mechanisms dominate in practice:
- Crystallisation and salting-out of amino acids at high concentrations or low temperatures
- Phase separation and agglomeration of lipophilic botanical fractions
- Nozzle clogging as the downstream mechanical consequence of both
Each has a distinct physicochemical origin and demands a tailored engineering response.
2. Amino Acid Crystallisation in Aqueous Solutions
2.1 Solubility Thermodynamics
Amino acids dissolved at the concentrations typical of functional nutraceutical sprays — taurine at 50–200 mM, glycine at 100–500 mM, L-theanine at 10–50 mM — exist as supersaturated or near-saturated solutions in water, particularly when chilled during storage or shipment. Their crystallisation behaviour is far from simple.
Glycine, the most extensively characterised example, exists in three polymorphic forms (α, β, γ). Recent nucleation studies demonstrate that polymorph outcome is exquisitely sensitive to environmental conditions. Cotting et al. showed in 2025 that sodium chloride — a near-universal excipient in liquid formulations — stabilises the metastable β-glycine polymorph for hours and dramatically alters the classical nucleation pathway: γ-glycine ultimately nucleates on the surface of β-glycine crystals rather than directly from solution, a mechanism that runs counter to the previously accepted model. [^5] Wang and Tiwary independently confirmed in 2025 that elevated ionic strength generically enhances polymorph metastability, accelerating nucleation of thermodynamically disfavoured forms. From a formulation standpoint, this matters enormously: a spray containing even physiologically relevant electrolyte levels can initiate an unanticipated crystallisation pathway, producing crystals with a different shape, density, and dissolution rate than the formulator anticipated.
For taurine, recent crystallisation studies reveal that process conditions determine crystal morphology with precision. Wu et al. demonstrated in 2020 that sodium sulfate (a common ionic excipient) modifies taurine crystal morphology from needle-shaped to columnar by selectively adsorbing on the (011) and (11-1) crystal faces and inhibiting their growth. Needle-shaped taurine crystals are particularly hazardous from a device standpoint: they interlock upon settling and form dense, intractable plugs. A 2025 study using differential scanning calorimetry to map taurine crystal defects found that gradient cooling from 80°C to 15°C substantially changes internal defect structure, with larger crystals containing approximately 15.6 times more internal moisture than smaller equivalents — defects that release water upon storage, locally increasing solute concentration and triggering secondary nucleation events.
2.2 Salting-Out Interactions
The simultaneous presence of multiple amino acids and ionic excipients creates competition for the water of solvation. Naderi et al., studying aqueous ternary systems of amino acids and quaternary ammonium salts, found systematic salting-out behaviour driven by unfavourable solute–solute interactions, with the strength of the effect following the order serine > glycine > alanine > proline. [^2] In a spray formulation containing taurine, glycine, and L-theanine together with potassium sorbate or sodium benzoate as preservatives, the ionic environment generated by the preservative salt can cross the threshold that initiates salting-out of the amino acids — even when each individual component remains below its nominal saturation concentration in pure water.
Guin et al. further demonstrated concentration- and temperature-dependent switching between salting-in and salting-out for alanine and threonine in ammonium sulphate media, with salting-out dominating at higher electrolyte concentrations. This behaviour implies that cooling a correctly formulated spray (which may be salted-in at room temperature) can shift the equilibrium to the salting-out regime, initiating crystallisation during cold-chain storage or in an unheated warehouse in winter.
2.3 The Role of Mechanical Agitation
Vesga et al. established that stirring promotes the metastable α-polymorph of glycine, while γ-glycine (the stable form) nucleates preferentially under quiescent conditions. [^4] A sublingual spray bottle undergoes repeated mechanical agitation during transport and use. Each actuation generates shear through the pump mechanism, and this repeated perturbation may selectively promote metastable polymorph nucleation — forms that subsequently transform to more stable, less soluble polymorphs upon rest, producing a progressively worsening precipitation problem over the product's shelf life.
3. Botanical Extract Phase Separation in Aqueous Matrices
3.1 The Compositional Complexity Problem
Botanical extracts are not single-compound entities. A liquid extract of valerian, ashwagandha, passionflower, or Centella asiatica contains simultaneously: flavonoids and other polar polyphenols (log P typically −1 to +2), condensed tannins (high molecular weight, amphiphilic), resinous terpenoid fractions (log P +3 to +6), and trace essential oil components (log P +4 to +8). These coexist in ethanolic solution because ethanol expands the miscibility window. In an aqueous-glycerin matrix, the system is thermodynamically unstable with respect to the lipophilic fractions.
Sepperer and Tondi's fractionation work on industrial tannin extracts demonstrated that industrial tannin powders contain 20–25% hydrocolloids alongside their polyphenolic content, and that selective solubility behaviour differs sharply between these fractions depending on solvent polarity. [^6] When transferred to a predominantly aqueous medium, the hydrophobic tannin oligomers and resins — which readily dissolved in the acetone/ethanol extraction medium — aggregate via hydrophobic stacking interactions and eventually phase-separate.
3.2 Mechanisms of Destabilisation
- Ostwald ripening of fine droplets formed upon dilution from an ethanolic concentrate: small lipophilic droplets dissolve preferentially and redeposit on larger ones, driving progressive coarsening until macroscopic phase separation occurs.
- Tannin–protein interactions, when protein-based excipients (gelatine, casein hydrolysates) are present, produce precipitates at low ionic strength that can occlude pump channels.
- Essential oil component autoxidation: monoterpene alcohols and sesquiterpenes undergo autoxidative polymerisation in the absence of the antioxidant environment provided by ethanolic solutions, producing resinous precipitates.
Ueoka and Moraes found that liquid crystal formation in emulsified botanical formulations using cetearyl alcohol significantly enhanced stability, and that formulations containing glycolic extracts from Centella asiatica and Hamamelis virginiana remained homogeneous over 90 days under thermal cycling only when a structured liquid-crystal phase was deliberately induced. Absent such structuring, botanical-containing emulsions showed progressive phase separation driven by extract-induced disruption of the emulsifier film.
4. Nozzle Clogging: The Engineering Consequence
4.1 Mechanisms of Obstruction
Nozzle clogging in sublingual and nasal spray devices occurs through two principal routes that often operate in concert:
- Evaporative crystallisation at the nozzle tip: between actuations, the small liquid volume retained in the nozzle orifice (typically 2–10 µL) loses water to evaporation. As the water activity drops, supersaturation is quickly achieved for any solute present above 50 mM. Taurine and glycine, at typical nutraceutical spray concentrations of 100–300 mM, will crystallise at the nozzle tip within hours of last use, forming a microcrystalline seal that must be mechanically disrupted by the next actuation. Repeated crystallisation–dissolution cycles damage the orifice geometry, enlarging the orifice irregularly and changing spray angle and droplet size distribution.
- Particle agglomeration in the delivery channel: botanical resin droplets and tannin aggregates in the sub-micron to micron size range undergo Brownian collision and progressive aggregation. Unlike reversible flocculation, resin-mediated aggregation is often irreversible — the viscoelastic resin film at the droplet surface confers an energy barrier against redispersion. This aggregated material accumulates at the valve seat and the nozzle insert, the points of maximum local pressure differential and minimum internal diameter.
Device studies confirm how sensitive spray performance is to even modest changes in nozzle geometry. Tong et al. showed that 10 µm particles are optimal for sublingual/nasal delivery, and that the spray cone angle and nozzle insertion depth together determine deposition with high sensitivity.[^8] A partially obstructed nozzle that increases effective orifice diameter by even 20% dramatically shifts the droplet size distribution upward, moving particles out of the optimal deposition range and reducing mucosal contact.
Seifelnasr et al. found that the nozzle retraction distance during actuation — nominally around 5.5 mm in standard multi-dose pumps — is a critical determinant of initial deposition pattern and drug loss to the pharynx.[^7] Partial obstruction changes the effective retraction dynamics, further compromising reproducibility.
4.2 Detection and Prediction
Nozzle clogging in alcohol-free formulations is notoriously difficult to predict from accelerated stability data alone, because the evaporative concentrating mechanism operates primarily at ambient humidity and room temperature — conditions that accelerated stability protocols at 40°C/75% RH do not replicate faithfully. The most predictive test is a repeated use/rest cycling study at the anticipated worst-case in-use temperature and humidity.
5. Engineering Solutions: Advanced Solubilisation Architectures
The engineering response to these failure modes has converged on four principal technology platforms, each addressing a distinct thermodynamic root cause.
5.1 Nanoemulsions
Oil-in-water nanoemulsions with droplet radii below 100 nm represent the most direct solution to the phase separation problem for lipophilic botanical fractions. At this scale, the kinetics of Ostwald ripening slow dramatically (ripening rate scales with droplet radius cubed), and the formulation remains optically transparent — a significant consumer acceptance advantage for sublingual sprays.
Choi and McClements' comprehensive review of nanoemulsion delivery systems for nutraceuticals identifies the key design parameters: lipid phase composition, emulsifier type and concentration, and processing energy input. For botanical extracts, medium-chain triglycerides (MCT) are preferred as the lipid phase because they solubilise a broad range of terpenoid and phenolic lipophilics and are generally recognised as safe for oral mucosal application. Polysorbate 80 and lecithin are the most commonly employed emulsifiers; at concentrations above the critical micelle concentration but below levels that cause mucosal irritation, they form stable interfacial films that resist coalescence.
Aboalnaja et al. characterised the two strategic uses of nanoemulsions in delivery: as a delivery vehicle (nanoemulsion delivery systems, NDS, where the bioactive is dissolved in the lipid phase) and as an excipient system (NES, co-administered with the primary product to improve bioaccessibility). For sublingual sprays, the NDS architecture is most relevant: it simultaneously solubilises the lipophilic fractions and presents them at the mucosa as nanoscale lipid droplets that merge readily with the mucosal lipid film.
5.2 Polymeric Micelles and Self-Micellising Systems
Polymeric micelles formed from amphiphilic block copolymers (poloxamers, PEG-phospholipid conjugates) or natural amphiphiles (saponins, glycyrrhizin) provide a thermodynamically stable solubilisation environment for molecules of intermediate log P. Their critical micelle concentration is typically orders of magnitude lower than that of small-molecule surfactants, meaning micellar solubilisation is maintained even after the significant dilution that occurs when a sublingual spray contacts the pool of saliva under the tongue.
Nanomicelle delivery for nutraceuticals has shown particular promise for curcumin, coenzyme Q10, and lipophilic vitamins — all of which share log P and molecular weight characteristics similar to terpenoid botanical actives. The additional advantage of polymeric micelles for spray applications is that their core is essentially anhydrous, meaning lipophilic actives loaded within the core do not interact with water molecules and are protected from hydrolytic degradation — a failure mode for some terpene esters and resinous glycosides.
5.3 Cyclodextrin Inclusion Complexation
For compounds of defined molecular geometry — many flavonoids, individual terpenoids, and some amino acid derivatives — cyclodextrin inclusion complexation provides precision solubilisation through host–guest chemistry. β-Cyclodextrin and its hydroxypropyl derivative (HPβCD) are the most widely used, offering cavity dimensions suited to molecules of molecular weight 200–500 Da.
Singh and colleagues' broad review of phytochemical–cyclodextrin complexes documents solubility improvements of 5- to 50-fold for compounds ranging from curcumin and quercetin to artemisinins and dihydromyricetin. The complexation simultaneously addresses solubility, chemical stability (the host cavity shields the guest from oxidation and hydrolysis), and taste masking — relevant for sublingual formulations where the drug is in prolonged contact with taste receptors.
The recent patent review by Costa et al. on propolis–cyclodextrin systems highlights how this approach can be extended to complex botanical resin matrices: propolis, whose activity derives from a broad spectrum of lipophilic flavonoids and terpenoids, becomes both water-soluble and shelf-stable upon HPβCD complexation, with demonstrated applications in sublingual and buccal pharmaceutical products. Critically for the alcohol-free challenge, CD complexation replaces the solvating function of ethanol with a supramolecular mechanism that does not require organic solvents.
5.4 Nanostructured Lipid Carriers and Solid Lipid Nanoparticles
Nanostructured lipid carriers (NLC) combine a solid lipid matrix with a liquid lipid internal phase, creating an imperfect crystal lattice that can accommodate a higher drug load than pure solid lipid nanoparticles (SLN) with reduced expulsion upon storage. For sublingual delivery, particles in the 50–200 nm range produced by high-shear homogenisation or ultrasonication provide the necessary fineness to pass through the pump orifice without obstruction. Suryawijaya et al.'s NLC work with green tea extract found that a 50:50 solid/liquid lipid ratio gave the best stability and smallest particle size (approximately 360 nm), while higher solid lipid ratios drove phase separation upon thermal cycling — a clear design constraint for alcohol-free botanical spray formulations.
5.5 Two-Component Device Architectures
When physicochemical engineering of the liquid phase alone cannot achieve the required stability, device engineering offers a parallel solution. Rautiola and Siegel demonstrated a pneumatic nasal spray device capable of mixing a solid and liquid component during actuation, thereby keeping the drug in its most stable (solid or lyophilised) state until the moment of delivery. This approach is conceptually applicable to sublingual sprays: amino acids stored as a dry powder and botanical nanoemulsion stored as a separate liquid are mixed only at the point of actuation, eliminating the stability challenge entirely at the cost of device complexity.
