Sensory Investigation of Toothpaste Foam Using Rheology and Dynamic Foam Analysis

ABSTRACT

Consumer acceptance of toothpaste formulations is heavily influenced by oral tactile sensations, particularly foaming behavior. While Home Use Tests (HUTs) provide definitive consumer insights, they are resource‐intensive and time‐consuming. This study establishes an accelerated laboratory methodology to predict consumer perception of toothpaste foaming using instrumental rheology and dynamic foam analysis (DFA). Fifteen custom made toothpaste formulations varying in surfactant systems, abrasives, and sensates were evaluated. Laboratory evaluation included foam generation via a customized blender‐rheometer geometry to monitor the viscoelastic properties (tan δ) of a paste‐water slurry, paste yield stress analysis, and structural profiling via DFA. Multivariable linear regression revealed that the initial foam loss factor (tan δ) correlates strongly with consumer panel ratings for both Foam Degree (R ² = 0.7) and Foam Liking (R ² = 0.6). Integrating paste yield stress (σ _y) into a multivariate model markedly enhanced the prediction (R ² = 0.8). Alternative multivariable models using tan δ and initial DFA parameters—specifically bubble count and average bubble radius—similarly yielded highly accurate predictions of mouthfeel (R ² = 0.8). The findings demonstrate that formulations with low paste yield stress and high foam visco‐dominance (higher tan δ) facilitate bubble formation and produce rich, consumer‐preferred “liquidish” textures. Alternative multivariable models using tan δ and DFA parameters were also presented. This dual rheological framework provides an instrumental proxy to screen and optimize toothpastes prior to panel studies.

This study bridges classical foam physics and consumer perception in oral‐care systems. By integrating parent paste yield stress with foam viscoelastic properties and microstructural metrics from a Dynamic Foam Analyzer (DFA), robust models (R ² = 0.8) were established to directly predict consumer‐perceived foam degree.

1. Introduction

Foams are prototypical soft materials whose macroscopic behavior is governed by interfacial forces, geometry, and time‐dependent microstructural changes. Foundational studies on foams and concentrated emulsions have shown that densely packed bubbles form elastic networks with a finite yield stress and characteristic deformation mechanisms controlled by surface tension, capillarity, and bubble morphology (Princen 1983). Under small‐amplitude shear, aqueous foams are commonly characterized by an elastic modulus G′, a viscous modulus G″, and a loss factor tan δ = G″/G′, all of which depend on gas volume fraction, film thickness, continuous‐phase rheology, and surfactant chemistry (Hohler and Cohen‐Addad 2005; Dollet and Raufaste 2014). Parallel work on foam drainage, coarsening, and film rupture has further clarified how gravity‐driven liquid flow and diffusive gas transfer transform initially fine, homogeneous foams into coarser, heterogeneous networks with larger bubbles, thinner films, and altered rheological signatures over time (Requier et al. 2024). Moreover, recent advances in soft matter physics have emphasized that measuring the yield stress (σ _y) of structured fluids requires careful methodological consideration, as linear steady‐shear extrapolations (e.g., Herschel‐Bulkley fits) and strain‐sweep crossover points (G′ = G″) capture distinct static versus dynamic structural yielding behaviors (Dinkgreve et al. 2016). In complex fluid foams containing particulate suspensions, solids or insoluble films further modify this bulk‐interfacial rheological coupling, altering interstitial liquid drainage, film elasticity, and structural resistance to shear (Pitois and Rouyer 2019). Furthermore, advanced dynamic rheometry has demonstrated that aging foams undergo time‐dependent microstructural transformations, which directly govern the temporal dissipation of viscoelastic moduli and flow curves (Carraretto et al. 2022).

While this body of work has established robust physical principles for model systems, similar research on foam microstructure, stability, and mouthfeel is highly prominent across various food matrices but remains noticeably absent for oral‐care applications like toothpaste foam. In the food and beverage industry, extensive structural and sensory frameworks exist for low‐, intermediate‐, and high‐viscosity matrices. For instance, the biochemical composition and state of fat and proteins are known to critically regulate the foamability, overrun, and sensory creaminess of dairy systems and skim milk foams (Mondal and Niranjan 2019). Similarly, in intermediate‐viscosity matrices such as aerated chocolates, parameters like gas hold‐up and uniform bubble size distribution directly dictate consumer preferences for crispness or velvety mouthfeel (Mondal and Niranjan 2019). This coupling between physical traits and human perception is also vital in carbonated drinks like beer, sparkling water, and sparkling wine, where bubble dynamics, drainage rates, and foam collar stability directly influence aroma release, mouthfeel, and overall quality scores (Gonzalez Viejo et al. 2019). Furthermore, structural building blocks like polymeric saccharides have been successfully deployed to optimize overrun, decrease friction coefficients, and replace small molecular sugars to concurrently preserve physical foam frameworks and elevate consumer creaminess rankings in non‐fat whipped cream analogues (Han et al. 2025).

Yet, despite these detailed advancements in food matrices, virtually no parallel research translates these structural and tribo‐rheological parameters into the consumer experience of toothpaste foams. During brushing, a structured paste is diluted with saliva and aerated, generating a dynamic foam that consumers experience as “rich,” “creamy,” “liquid‐like,” or conversely as “thin,” “airy,” or rapidly collapsing. These tactile impressions play a decisive role in judgments of product quality, cleaning power, and overall liking. In practice, oral‐care formulations are therefore optimized using empirical Home Use Tests (HUTs), even though food science paradigms suggest that many aspects of mouthfeel should be predictable from foam rheology and microstructure, analogous to the relationships established in conventional foam studies (Princen 1983; Dollet and Raufaste 2014; Dinkgreve et al. 2016).

From a sensory perspective, key questions remain insufficiently addressed. First, how do viscoelastic properties of the in‐mouth foam—especially the loss factor tan δ, which encodes the balance between liquid‐like and solid‐like behavior—map onto consumer descriptors such as “creamy,” “velvety,” or “airy”? Classical foam rheology links higher tan δ to wetter, more dissipative foams with thicker liquid films, and lower tan δ to drier, more elastic networks (Princen 1983). However, systematic evidence connecting these regimes to consumer ratings of foam abundance and liking in toothpaste is scarce. Second, the microstructural evolution processes described in the foam physics literature—drainage, coarsening, and bubble coalescence (Weaire 2008; Pitois and Rouyer 2019) are expected to influence how stable, dense, and “full” the foam feels over the brushing period. Yet quantitative links between measurable structural metrics (average bubble size, bubble count) and perceived foam texture in oral‐care applications are limited. Third, the initial yield behavior of the parent paste, a central concept in foam rheology (Princen 1983; Dollet and Raufaste 2014), is rarely considered from a sensory standpoint, even though the paste must first yield and incorporate air before any foam‐based mouthfeel can emerge.

The present study addresses these gaps by explicitly integrating conventional foam rheology concepts with consumer sensory outcomes for toothpaste formulations. We investigate 15 experimental toothpastes with varied surfactant systems, abrasives, and sensates, and characterize: (i) the yield stress σ _y of the bulk paste, which governs the ease of structural breakdown and air incorporation; (ii) the initial viscoelastic properties of the resulting foam, with emphasis on the loss factor tan δ as a marker of “liquid‐like” versus “solid‐like” behavior; and (iii) the initial average bubble size R _avg and bubble count BC using Dynamic Foam Analyzer (DFA) to capture structural attributes implicated in foam stability and fullness, in line with prior work on foam drainage and coarsening (Weaire 2008). These instrumental measurements are then directly compared to HUT panel ratings of Foam Degree and Foam Liking.

By demonstrating that combinations of paste yield stress and foam tan δ, or of foam tan δ with microstructural metrics obtained by DFA, can robustly predict consumer evaluations (up to R ² = 0.8), this work translates the theoretical and experimental insights from classical foam rheology into a practical sensory design tool for oral‐care products. In doing so, it links the physical language of moduli, loss factors, and bubble statistics to the perceptual language of foam “degree” and “liking,” and provides a quantitative pathway to accelerate formulation development while reducing reliance on resource‐intensive sensory testing.

2. Materials and Methods

2.1. Materials

A total of 15 experimental toothpaste formulations (DOE1‐DOE15) were evaluated in this study. All prototypes were prepared using the same base formula, keeping the baseline solvent system and polymer/thickener matrix constant. Modifications were strictly limited to the surfactant systems, abrasive types/blends, and sensates to generate variations in foaming behavior while keeping the underlying paste structure identical.

While these formulations varied significantly in their compositions, detailed information regarding the specific chemical identities and concentrations remains proprietary. This study focuses exclusively on the relationship between instrumental measurements and the sensory perception of foaming. The physical properties of the formulations are summarized in Tables 1 and 2, and are described below.

TABLE 1.

Summary of the formulas, their consumer ratings, and rheological properties.

#	Formulas			Consumer ratings		Paste rheology			Foam rheology (1:3 paste: water)
	Surfactant	Abrasive	Sensate	Foam degree	Foam liking	σ y (Pa)	K (Pa s n )	n	G′ (Pa)	G″ (Pa)	tan δ
DOE1	Surfactant A + B	Abrasive A	Sensate 1	6.11 ± 0.53	6.54 ± 0.58	224	34	0.56	5.73 ± 0.53	2.57 ± 0.11	0.45 ± 0.06
DOE2	Surfactant A + C	Abrasive A	Sensate 1	6.10 ± 0.59	6.81 ± 0.58	246	53	0.50	5.66 ± 0.42	2.83 ± 0.11	0.50 ± 0.02
DOE3	Surfactant A + B	Abrasive B + C	Sensate 1	6.64 ± 0.60	6.65 ± 0.67	167	38	0.57	6.11 ± 0.86	3.09 ± 0.37	0.51 ± 0.06
DOE4	Surfactant A + C	Abrasive B + C	Sensate 1	6.32 ± 0.65	6.35 ± 0.70	165	32	0.58	5.71 ± 0.17	2.47 ± 0.20	0.43 ± 0.03
DOE5	Surfactant A + C	Abrasive B	N/A	5.96 ± 0.58	6.65 ± 0.62	223	26	0.69	7.54 ± 2.11	2.48 ± 0.36	0.34 ± 0.04
DOE6	Surfactant A + C	Abrasive A	Sensate 2	6.66 ± 0.57	7.66 ± 0.64	262	43	0.54	4.92 ± 0.40	2.72 ± 0.31	0.56 ± 0.08
DOE7	Surfactant A + C	Abrasive A	N/A	6.70 ± 0.53	7.34 ± 0.64	240	45	0.52	5.34 ± 0.53	3.01 ± 0.32	0.57 ± 0.08
DOE8	Surfactant A + C	Abrasive B	Sensate 1	5.70 ± 0.62	6.24 ± 0.71	244	11	0.90	6.00 ± 0.90	2.30 ± 0.16	0.39 ± 0.03
DOE9	Surfactant A + C	Abrasive B + C	Sensate 2	5.89 ± 0.58	6.21 ± 0.54	193	40	0.54	6.41 ± 1.20	2.70 ± 0.24	0.43 ± 0.06
DOE10	Surfactant A + C	Abrasive B + C	N/A	6.02 ± 0.60	6.65 ± 0.68	192	44	0.53	5.37 ± 0.28	2.35 ± 0.16	0.44 ± 0.02
DOE11	Surfactant A + B	Abrasive B	Sensate 2	6.17 ± 0.63	6.52 ± 0.65	214	61	0.49	5.38 ± 0.44	2.59 ± 0.24	0.48 ± 0.07
DOE12	Surfactant A + C	Abrasive B	Sensate 2	5.20 ± 0.63	5.93 ± 0.65	309	13	0.93	7.66 ± 1.04	2.50 ± 0.32	0.33 ± 0.00
DOE13	Surfactant A + B	Abrasive B	N/A	5.76 ± 0.73	7.00 ± 0.65	268	29	0.66	6.96 ± 1.18	3.13 ± 0.46	0.46 ± 0.12
DOE14	Surfactant A + B	Abrasive A	Sensate 2	5.89 ± 0.72	6.47 ± 0.71	286	48	0.50	9.61 ± 1.47	4.36 ± 0.46	0.46 ± 0.12
DOE15	Surfactant A + B	Abrasive B + C	N/A	6.24 ± 0.54	7.04 ± 0.51	275	45	0.56	9.46 ± 3.39	4.28 ± 0.43	0.48 ± 0.16

TABLE 2.

DFA results.

#	Dynamic foam analyzer (DFA)
	Average bubble radius, R avg (initial)	Average bubble radius, R avg (final)	Bubble count, BC (initial)	Bubble count, BC (final)
DOE1	53.3 ± 6.2	158.0 ± 13.1	68.1 ± 32.4	5.9 ± 1.2
DOE2	52.7 ± 2.4	157.0 ± 11.9	52.3 ± 20.9	4.9 ± 2.3
DOE3	61.7 ± 8.6	253.7 ± 21.2	44.9 ± 16.7	2.8 ± 0.4
DOE4	56.7 ± 1.3	228.0 ± 30.1	52.6 ± 11.9	3.4 ± 1.6
DOE5	67.3 ± 8.8	189.3 ± 12.9	34.8 ± 10.1	4.3 ± 0.3
DOE6	64.7 ± 6.8	182.7 ± 33.6	34.7 ± 11.2	4.9 ± 2.3
DOE7	60.3 ± 5.3	196.7 ± 19.7	50.5 ± 10.0	4.8 ± 0.5
DOE8	59.0 ± 4.5	214.7 ± 35.6	44.6 ± 8.0	3.1 ± 0.6
DOE9	73.7 ± 11.9	240.7 ± 18.4	37.6 ± 11.6	3.5 ± 0.5
DOE10	44.7 ± 2.8	182.0 ± 21.5	88.5 ± 6.1	4.8 ± 0.8
DOE11	45.5 ± 1.0	261.0 ± 2.4	67.3 ± 6.2	3.4 ± 0.6
DOE12	49.0 ± 6.3	214.0 ± 19.6	72.1 ± 28.0	3.4 ± 0.3
DOE13	48.0 ± 3.2	267.0 ± 3.9	73.9 ± 14.8	2.7 ± 0.1
DOE14	55.0 ± 28.5	176.3 ± 28.4	86.7 ± 62.6	5.4 ± 0.5
DOE15	47.3 ± 4.3	241.0 ± 32.3	77.5 ± 13.9	3.2 ± 0.6

2.2. Methods

2.2.1. Consumer Panel Study

A 15‐day sensory study was conducted to evaluate the perception of foaming across all 15 experimental toothpaste formulations. The study employed a sequential monadic, blind evaluation design with a balanced incomplete block format. A total of 165 regular toothpaste users, aged 18–54, were recruited for the panel. Over the course of the evaluation, each participant tested a subset of five distinct toothpaste samples, utilizing each for a three‐day period. To ensure procedural consistency, all participants were provided with a standardized mid‐tier whitening toothbrush. Participants selected their preferred bristle stiffness (soft or medium) at the start of the study and used that same stiffness for all tested samples. Subjects were instructed to brush thoroughly twice per day in place of their usual oral care products. Following each three‐day usage cycle, participants completed an online questionnaire focusing on two primary dimensions of particle perception:

1. “Foam Degree”: Participants were asked, “How does the toothpaste feel in your mouth?” on a scale ranging from 0 (not foamy at all) to 10 (extremely foamy).

2. “Foam Liking”: Participants were asked, “How much do you like the foaminess of the toothpaste?” on a scale ranging from 0 (hate it) to 10 (love it).

2.2.2. Rheological Tests

2.2.2.1. Rheology of the Paste

Rheological measurements were conducted on a DHR‐20 rheometer (TA Instruments/Waters, New Castle, DE) with a 4‐paddled vane spindle. This spindle is the most appropriate one for rheological measurements of fluids that possess yield stress, since it minimizes the fluid disturbance (Barnes and Carnali 1990; Estellé et al. 2008; Teoman et al. 2021). Shear rate sweep experiments performed, increasing the shear rate from 1/s to 30/s, 10 s per point, to obtain the viscosity profiles from rest. Collected apparent shear rate and stress data were parameterized in terms of the Herschel‐Bulkley equation (Herschel and Bulkley 1926):

$$\sigma ={\sigma }_{\mathrm{y}}+K{\dot{\gamma }}^n$$ (1)

where σ is the stress (Pa), and $\dot{\gamma }$ is the shear rate (1/s), while the yield stress, σ _y (Pa), consistency index, K (Pa s ⁿ ), and the flow index, n, are the fitting parameters.

2.2.2.2. Rheology of the Resulting Foam

To emulate the intensive shear of manual toothbrushing, foams were generated using a blender cell integrated inside a rotational rheometer. A paste‐to‐water slurry was prepared at a 1:3 mass ratio. This slurry was mixed inside the cell for 30 s to produce a uniform foam. Immediately following foam generation (t = 0 s), time‐sweep experiments were initiated for a duration of 1 min. The testing was conducted under strict controlled‐strain conditions at 0.1% strain to ensure the foam micro‐structure remained within the linear viscoelastic region (LVR). The elastic modulus (G′), viscous modulus (G″), and the loss factor (tan δ = G″/G′) were monitored continuously, and their initial (t = 0 s) values were reported here.

2.2.3. DFA Tests

A DFA (Krüss GmbH) was used to capture microstructural metrics (bubble size and count) over time. Because the 1:3 slurry utilized in rheology exceeded viscosity thresholds suitable for sparging optics, a 1:10 paste‐to‐water slurry was deployed for DFA testing. 50 mL of the 1:10 slurry was introduced into the glass column, and air was sparged through a porous glass filter with pore size ranging from 16 to 40 μm at a constant flow rate of 0.3 L/min until the foam height reached 190 mm. Since the air blowing stopped, the camera positioned at 190 mm height recorded the evolution of the bubbles in the course of foam decaying over 5 min. The integrated visual imaging software monitored the foam structure, extracting the average bubble radius (R _avg) and the total bubble count per unit area (count/mm²), BC, at the initial generation state (t = 0 min) and tracking structural changes up to t = 5 min.

3. Results

3.1. Panel Study Results

The consumer sensory panel provided a distinct baseline for evaluating how variations in toothpaste formulation impact human perception of foam and its texture. Figure 1 presents the mean panel ratings for “Foam Degree” and “Foam Liking” across all 15 evaluated prototypes (DOE1‐DOE15), revealing a wide sensory distribution driven by the modified ingredient systems. Perceived volume ranged from a low Foam Degree of 5.20 for DOE12 to a high of 6.70 for DOE7, while consumer preference ranged from a low Foam Liking score of 5.93 (DOE12) up to a peak of 7.66 (DOE6).

FIGURE 1.

Foam degree and liking panel ratings for all 15 formulations.

A key finding from this data is that Foam Degree and Foam Liking scores follow similar trends across the sample set, indicating that perceived foam volume is a primary driver of overall satisfaction. This relationship is highlighted by top‐performing prototypes like DOE6 and DOE7, which combined high perceived volume with peak liking scores. Conversely, the single poorly performing formulation, DOE12, was systematically penalized by the panel, scoring lowest in both metrics. These tightly matched sensory evaluations confirm that generating an abundant, robust foam is essential for satisfying consumer expectations, providing clear performance benchmarks for the subsequent physical and microstructural analysis.

3.2. Rheological Results

3.2.1. Rheological Results for the Pastes

The rheological results for five selected pastes are shown in Figure 2. The steady‐state flow behavior and structural resistance of the bulk parent matrices were evaluated through shear rate sweep experiments. Figure 2 presents the measured shear stress as a function of shear rate from a resting state for a selected subset of the most physically distinct formulations (DOE1, DOE4, DOE6, DOE10, and DOE14). All formulations exhibit clear non‐Newtonian, shear‐thinning characteristics with a distinct yield stress. The experimental data points are parameterized and fitted using the Herschel‐Bulkley model (Equation 1), as indicated by the solid regression lines. The resulting yield stress values reveal significant structural differences among these five representative pastes. Formulations DOE14 (σ _y = 286 Pa) and DOE6 (σ _y = 262 Pa) exhibit the highest curves on the plot, reflecting a highly rigid, robust internal network structure. This high resistance to initial deformation means that substantial physical effort or manual shear is required to break down the paste skeleton. In contrast, formulation DOE4 exhibits the lowest curve and the lowest yield stress (σ _y = 165 Pa), indicating a highly compliant and easily collapsible bulk matrix. Formulations DOE1 (σ _y = 224 Pa) and DOE10 (σ _y = 192 Pa) fall in the intermediate range, illustrating the broad physical spectrum captured by the experimental design. These varying baselines paste yield stresses serve as the initial physical gateway during brushing; lower‐curve formulas like DOE4 streamline the breakdown process to facilitate early bubble formation, whereas higher‐curve formulas like DOE14 act as a physical barrier to rapid phase inversion and initial air incorporation.

FIGURE 2.

Rheological results of selected pastes. All other rheological results are presented in Table 1.

3.2.2. Rheological Results for the Resulting Foam

The transient viscoelastic properties of the generated toothpaste foams were monitored via controlled‐strain time‐sweep experiments to capture structural evolution under resting conditions. Figure 3 displays the dynamic evolution of tan δ over a 1‐min period for a selected subset of formulations (DOE2, DOE5, DOE7, DOE8, and DOE10). Immediately following high‐shear generation (t = 0 min), the foams exhibit their peak loss factor values, ranging from a highly visco‐dominant tan δ of 0.57 for DOE7 down to an elastically dominant value of 0.34 for DOE5. This initial variation reflects the dramatic baseline shifts in microstructural layout and fluid film thickness separating the newly incorporated air bubbles. As time progresses, all formulations display a continuous, monotonic decrease in tan δ, indicating a progressive transition toward a more structured, elastic solid‐like behavior (G′ dominance) as the foam network ages. However, the kinetics of this decay differ significantly between the variants. High‐performing, consumer‐preferred formulations like DOE7 and DOE2 maintain higher overall tan δ trajectories throughout the window, preserving their “liquidish,” compliant nature longer to support a creamy oral sensation. In contrast, elastically dominant formulas like DOE5 and DOE8 remain consistently low, dropping below a tan δ of 0.30 within the first 30–45 s. This rapid consolidation into a rigid, highly elastic structure is associated with the accelerated interstitial liquid drainage, driving the formation of the thin, brittle, and hollow foam geometries that consumers penalize.

FIGURE 3.

Loss factor (tan δ) progression in time for selected samples.

3.3. Results of DFA Tests

The microstructural evolution and destabilization kinetics of the generated foam heads were captured quantitatively via DFA (Table 2) and confirmed visually through optical progression panels (Figure 4). At the initial state (t = 0 min), all 15 formulations generated relatively fine, uniform, and dense bubble structures (yet still marginally different), with initial average bubble radii (R _avg) tightly bounded between 44.7 and 73.7 μm and initial bubble counts (BC) spanning from 34.7/mm² to 88.5/mm². This initial uniformity is visually evident across the t = 0 min optical cells, which display a homogeneous matrix of small, tightly packed green micro‐bubbles. However, tracking the structural progression up to t = 2 min reveals a stark divergence in foam stability across the sample set. Stable, high‐performing prototypes—such as DOE3, DOE4, DOE6, and DOE7—effectively preserved their uniform, small‐bubble infrastructure over time, demonstrating excellent spatial retention with minimal bubble growth or room for large voids. Conversely, unstable networks underwent severe microstructural decay driven by rapid liquid drainage and bubble coalescence. In cells corresponding to DOE1, DOE5, and DOE12, individual bubble walls ruptured, causing a rapid shift toward maximum bubble radii and a catastrophic drop in bubble density (such as DOE12 collapsing to a final count of just 3.4/mm²). This visual and quantitative degradation explicitly tracks the transition of a uniform dispersion into a coarse, heterogeneous network dominated by oversized, irregular white and gray air pockets. Notably, the data also reveals that certain formulations, like DOE10, can nucleate an extraordinarily dense and fine initial dispersion (BC = 88.5/mm², R _avg = 44.7 μm), yet still suffer a severe structural collapse by t = 2 min due to insufficient film stabilization.

FIGURE 4.

Visual results from DFA during the testing. Note: Only the first 2 min are shown for visual comparison purposes.

3.4. Statistical Analysis

3.4.1. Correlations Using Rheology

To quantify the relationship between laboratory parameters and human mouthfeel, single‐variable and multivariate linear regression models were constructed. As shown in the single‐variable plots in Figure 5, tan δ correlates with consumer Foam Degree (R ² = 0.7) and Foam Liking (R ² = 0.6) panel ratings while having low p‐values (p < 0.05).

FIGURE 5.

Correlations of tan δ with Foam Degree and Liking ratings.

The predictive accuracy improves significantly when accounting for the structural breakdown of the bulk parent matrix prior to air incorporation by integrating the bulk paste yield stress σ _y alongside tan δ into a multivariate linear regression model. This multivariable linear regression modeling was performed using JMP Pro software. Model optimization was conducted via backward stepwise elimination, starting from an initial model including all paste rheological parameters (σ _y, K, n), initial foam viscoelastic metrics (G′, G″, tan δ), and initial structural metrics (BC, R _avg). Variables were iteratively removed until only statistically significant terms (p < 0.05) remained in the final predictive equations.

This rheological model increases the predictive power to an R ² of 0.8. As seen in the first actual‐by‐predicted plot in Figure 6, the data points tightly align along the line of unity by factoring in its moderate yield stress threshold.

FIGURE 6.

Predicted foam degree (using rheological parameters shown on the plot) vs. actual foam degree ratings.

3.4.2. Correlations Using Rheology and DFA

Alternative multivariate frameworks were established by pairing tan δ with initial structural metrics captured via DFA. These models evaluate how fluid rheology operates in tandem with immediate bubble architecture to dictate the perceived mouthfeel. Combining tan δ with the initial bubble count (BC) or initial average bubble radius (R _avg) yields a highly predictive proxy for consumer foam perception, achieving an R ² of 0.8. Both of the corresponding parity plots in Figure 7 demonstrate that the combination of foam rheology and bubble characteristics effectively models the consumer response without requiring paste yield parameters. While tan δ remains the statistically dominant term across all configurations, the inclusion of initial DFA metrics provides a robust secondary benchtop pathway to accurately screen prospective candidate matrices prior to large‐scale panels.

FIGURE 7.

Predicted foam degree (using rheological and DFA parameters shown on the plots) vs. actual foam degree ratings.

4. Discussion

The findings of this study establish a dual‐parameter instrumental framework that intends to bridge the gap between laboratory benchtop measurements and subjective consumer mouthfeel during toothbrushing. A primary insight from this research is the critical role of the foam loss factor, tan δ, in driving both the perceived abundance and hedonic appreciation of toothpaste foam. Formulations characterized by a higher tan δ exhibit a higher viscous compliance over elasticity, which microstructurally corresponds to tighter packing of air bubbles insulated by a thicker interstitial liquid layer. Perceptually, this arrangement translates directly to a velvety, creamy, and thick mouthfeel that consumers instinctively associate with premium products.

This is prominently exemplified by formulations like DOE6 and DOE7, which yielded the highest tan δ values (0.56 and 0.57) and subsequently secured the highest consumer ratings for Foam Liking (7.66 and 7.34). Conversely, systems with dominant elasticity feature a lower loss factor, resulting in an “airy” foam network with larger bubble diameters and exceptionally thin interstitial liquid layers that drain rapidly. Consumers register these rigid, thin networks as structurally hollow, leading to lower hedonic satisfaction. For instance, DOE12 featured a low tan δ of 0.33 directly mapping to the lowest reported consumer Foam Degree (5.20) and Foam Liking (5.93) across the entire sample set. This linear correlation is robustly demonstrated by the single‐variable regression models, where tan δ alone accounts for 70% of the variance in Foam Degree (R ² = 0.7) and 60% of the variance in Foam Liking (R ² = 0.6). However, the data emphasizes that foam rheology alone is an insufficient predictor of overall consumer acceptance. Before a consumer can experience the velvety texture of a visco‐dominant matrix, the parent toothpaste paste must first undergo structural breakdown in the initial seconds of brushing to permit air entrainment and slurry homogenization.

This phase of foam evolution is governed by the paste yield stress (σ _y), which represents the physical threshold required to deform the bulk paste's internal structural skeleton. Formulations possessing a lower paste yield stress streamline the initial breakdown of the paste skeleton, thereby facilitating immediate bubble formation and continuous air incorporation under the manual shear of toothbrushing. High yield stress values act as a physical barrier to rapid phase inversion, delaying or compressing the foam generation window. This phenomenon explains why incorporating σ _y alongside tan δ into a multivariate linear regression framework enhances the predictive capacity of the model remarkably to R ² = 0.8. The negative coefficient for σ _y mathematically confirms that lower paste yield stresses are structurally advantageous for optimizing consumer‐perceived foam abundance.

This physical mechanism is independently validated by the DFA metrics, where preferred formulations maintain a controlled bubble morphology, whereas low tan δ variants show rapid shifting toward maximum bubble radii accompanied by a steep decline in overall bubble count, documenting severe film drainage and bubble coalescence over time.

In fact, high predictive accuracy is achieved by combining foam rheology with visual metrics from DFA at the initial generation state (t = 0 min). Statistical modeling reveals that pairing tan δ with initial bubble count BC yields a strong multi‐variable proxy for foam degree (R ² = 0.8), where tan δ acts as the primary driver (p < 0.001) and bubble count provides a significant baseline check (p = 0.03). Similarly, evaluating the tan δ alongside initial bubble size R _avg accounts for 80% of the variance in consumer response (R ² = 0.8), with tan δ continuing to dictate the model's predictive weight (p < 0.001).

From an industrial standpoint, these intersecting multivariate frameworks provide formulators with multiple highly effective diagnostic pathways to sweep and optimize prospective toothpaste candidate matrices. R&D teams can successfully substitute resource‐intensive human sensory panels by leveraging either paste yield thresholds (σ _y) or dynamic fluid structures (BC, R _avg) in tandem with time‐sweep foam rheology (tan δ) to programmatically deliver rich, velvety, consumer‐preferred oral textures.

It is important to note that the predictive correlation (R ² = 0.8) reported here was obtained with prototypes sharing an identical baseline solvent, binder polymer, and thickener matrix, with changes limited to the surfactants, abrasives, and sensates. Because major changes in binder type or continuous phase composition alter bulk rheology and mouthfeel, this model serves as a promising starting point for similar toothpastes rather than a universal rule. Future studies testing different base formulas and alternative polymer networks will be needed to validate and expand these findings.

Finally, a few technical limitations of this baseline study should be highlighted. First, deionized water was used instead of human or artificial saliva, omitting natural salivary proteins and mucins that can alter surface tension and film elasticity. Second, instrument operating limits mandated different slurry dilution ratios between rheology (1:3) and DFA optical sparging (1:10). Lastly, consumer ratings reflected 3‐day aggregated averages across both morning and evening brushing occasions, which did not isolate potential diurnal variations in salivary flow. Despite these limitations, the strong correlations demonstrate that this methodology successfully provides an effective laboratory screening tool to predict consumer perception and narrow down candidate samples prior to costly panel studies. Future studies addressing alternative base formulas, salivary matrices, and time‐resolved sensory logging will help validate and broaden the generalizability of these instrumental methodologies.

5. Conclusion

This work presents a preliminary, laboratory‐scale framework to evaluate consumer toothpaste foaming preferences using physical instrumentation. A promising predictive correlation (R ² = 0.8) was observed by pairing the viscoelastic loss factor of the diluted foam (tan δ) with the yield stress of the bulk paste (σ _y). Alternatively, pairing this initial foam loss factor with DFA metrics at the initial generation state provides a useful diagnostic proxy, with multivariate models utilizing initial bubble count (tan δ + BC) and initial bubble radius (R _avg) accounting for 80% (R ² = 0.8) of the variance in consumer response within this study. The data suggest that consumers favor “liquidish” over “airy” textures. Within this baseline formula architecture, prototypes achieved higher consumer ratings when combining a low yield stress paste structure to facilitate air introduction with a visco‐dominant fluid matrix to support dense, velvety micro‐bubbles. Utilizing this rapid benchtop method provides formulators with an initial screening tool to evaluate candidate samples prior to large‐scale consumer testing.

Author Contributions

Monica Morelli: conceptualization, investigation, data curation. Paulo Kolle: data curation, resources, investigation. Jonghun Lee: conceptualization, investigation, methodology. Aileen Cabelly: investigation, conceptualization, data curation. Baran Teoman: conceptualization, writing – original draft, investigation. Mark Vandeven: conceptualization, investigation. Jimmy Yu: data curation. Andrei Potanin: writing – review and editing, conceptualization, investigation, supervision. Phil Mai: data curation.

Funding

This work was supported by Colgate‐Palmolive Company.

Ethics Statement

This study was approved by Colgate‐Palmolive's safety and regulatory committee.

Consent

Written informed consent was obtained from all study participants.

Conflicts of Interest

The authors declare no conflicts of interest. All authors are employees of the Colgate‐Palmolive Company.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.