Discoloration of liquid foundation across the shade color diversity

Abstract

Due to the increasing demand for diversity and inclusivity, beauty companies have expanded their liquid foundation ranges to cover a broad spectrum of skin tones and brightness levels. However, discoloration over time remains a concern. This study investigates discoloration trends across 32 foundation shades from four global brands. Foundations were applied to an opacity chart at a thickness of $200~\upmu$m, and color changes were measured over time using a spectrophotometer, reported in the CIE1976L*C*h color space. Our results indicate that discoloration trends depend on foundation lightness, with darker shades exhibiting less $L^*$ value reduction than lighter shades within each brand. Additionally, one brand demonstrated superior color retention. Scanning Electron Microscopy (SEM) images-acquired using secondary electron mode-offered qualitative comparisons of surface morphology across formulations. These findings highlight the importance of considering discoloration patterns when formulating foundations tailored to specific market demographics.

Introduction

Facial foundations are pigmented cosmetic products applied to the entire face and typically worn for eight hours or longer before removal¹. They are widely used to create an even skin tone, conceal blemishes, and serve as a base for additional makeup. Among various forms-liquid, cream, and powder-liquid foundations are most preferred by consumers² due to their hydration properties, natural appearance, and buildable coverage. They also support even pigment distribution across the skin’s surface.

Pigments impart color to foundations and are typically in powder form, insoluble in water or organic solvents³. Commonly used pigments include red, black, and yellow iron oxides, whose proportions determine the specific shade of each foundation⁴. The extent to which these pigment particles are evenly dispersed in the liquid carrier is referred to as pigment dispersion⁵. Proper dispersion is essential to prevent flocculation or agglomeration⁶, which can negatively impact color stability in cosmetics⁷. However, due to the hygroscopic nature of pigment powders and their tendency to form aggregates, processing techniques such as wetting, stabilization, and coating are crucial for achieving optimal dispersion.

Nevertheless, consumer concerns persist regarding foundation darkening during wear. While this phenomenon is often colloquially described as “oxidation,” the underlying mechanisms remain insufficiently explained in cosmetic science. Since the pigments themselves (e.g., iron oxides and titanium dioxide) are already in stable oxidized forms, other factors such as ingredient evaporation, surface changes, or formulation variables may be involved. Despite wide anecdotal discussion, including in beauty forums and brand guidance, the specific causes and trends of discoloration are not well understood.

Recent studies have examined the issue of foundation discoloration. For instance, Yan et al. developed techniques to quantify and describe color changes in liquid foundations, applying commercially available products to both human skin and an opacity chart before measuring color with a spectrophotometer⁸ . Their findings revealed that discoloration manifests as a decrease in $L^*$ and $h^\circ$ values within the CIE1976 $L^*C^*h$ (hereafter, CIELCh) color space. Two key parameters were introduced: $D_t$, representing the extent of discoloration over time, and $T_{\Delta E}$, indicating the time required for discoloration to reach a critical level. Furthermore, Huang et al. investigated the effects of exposure to sebum, sweat, and light on foundation darkening, concluding that sebum had the largest impact, whereas sweat and light exerted comparatively minor effects⁹. Additionally, Chen et al. explored the intrinsic mechanisms of foundation darkening, demonstrating that factors such as product volatilization rate and pigment coating methods influence discoloration¹⁰. Moreover, Lam discussed how pigment dispersion and surface treatment affect the overall performance of long-wear cosmetics, including their color presentation on skin over time⁶.

However, existing empirical studies have primarily focused on lighter shades^8,10, leaving a gap in understanding how discoloration affects deeper shades. This gap is particularly relevant given the differences in pigment composition across varying shade brightness levels, which may influence discoloration trends. Additionally, as Lam⁶ explains, while pigment dispersion is primarily crucial for achieving uniform initial color and formulation consistency, its role in post-application discoloration remains uncertain and underexplored.

The beauty industry has undergone a transformation in recent years, driven by the growing demand for inclusive foundation shades. Many companies now adopt a glocalization approach-addressing global consumer preferences while respecting local trends and ethnic characteristics. This has resulted in the development of foundation shades ranging from the lightest to the deepest, accommodating a diverse spectrum of skin tones.

To address these gaps, this study investigates foundation discoloration across a broad range of shades, from deep to pale, and examines how discoloration patterns vary by shade. By analyzing widely available commercial products, our research aims to reflect and communicate real-life consumer experiences regarding foundation discoloration, offering valuable insights to both consumers and manufacturers. Furthermore, if consistent discoloration trends emerge across brands with different formulations, production processes, and environments, it would enhance the robustness and applicability of our findings.

Accordingly, we measured the color of liquid foundations applied to an opacity chart at five specific time points: immediately after application and at two-hour intervals over an eight-hour period. The results of these measurements are presented alongside Scanning Electron Microscopy (SEM) analysis. We then discuss these findings, emphasizing how discoloration trends relate to pigment dispersion. Finally, we summarize our conclusions and provide insights into foundation formulation improvements.

Results and analyses

This section presents the results of discoloration observed over time in 32 liquid foundation products. Discoloration is quantified through changes in overall color difference ($\Delta E$), as well as shifts in lightness ($L^*$), chroma ($C^*$), and hue angle ($h^\circ$), based on the CIE 1976 $L^*a^*b^*$ color space (hereafter, CIELab). Each parameter is analyzed at 2-hour intervals over an 8-hour period following application.

Observing the discoloration over time

The color measurements of all 32 foundation products at the initial time point-immediately after application to the opacity chart-are presented in Table 1. Each brand includes eight different shades, with shade 1 being the lightest and shade 8 the darkest. The products are labeled using the format (Brand)-(Shade number), such as A-1, A-2, A-3, and so on. The lightness ($L^*$) values of the liquid foundations range from 34.05 to 86.50, with a mean of 60.10 and a standard deviation of 16.13.

To analyze discoloration trends, we employed the CIELCh color space, which describes the brightness, vividness, and hue characteristics of a color. CIELCh is an alternative representation of the CIELab color space, where $a^*$ and $b^*$ indicate redness (or greenness) and yellowness (or blueness), respectively. Instead of using $a^*$ and $b^*$ directly, $C^*$ in the CIELCh space represents the chroma, or vividness, of a color, which is calculated as the distance from the neutral axis (where $a^* = 0$ and $b^* = 0$). The $h^\circ$ value refers to the hue angle, derived from the $a^*$ and $b^*$ coordinates.

Since discoloration results from simultaneous changes in $a^*$ and $b^*$, $h^\circ$ provides a more straightforward way to illustrate hue shifts. Finally, $L^*$ represents the lightness of a color, ranging from 0 (black) to 100 (white). In the following subsections, discoloration is described in terms of overall color change ($\Delta E$), followed by individual analyses of $L^*$, $C^*$, and $h^\circ$.

$$\begin{aligned} & C^* = \sqrt{(a^*)^2 + (b^*)^2} \end{aligned}$$

(1)

$$\begin{aligned} & h^\circ \ = \tan ^{-1} \left( \frac{b^*}{a^*} \right) \quad \text {[degrees]} \end{aligned}$$

(2)

$$\begin{aligned} & \Delta E^*_{ab} = \sqrt{(L^*_2 - L^*_1)^2 + (a^*_2 - a^*_1)^2 + (b^*_2 - b^*_1)^2} \end{aligned}$$

(3)

Table 1 The color measurements for products from brands A, B, C, and D used in the experiments. For each brand, shade 1 indicates the lightest shade and shade 8 the darkest.

Overall discoloration: changes of $\Delta E^*$

The change of $\Delta E^*$ after two, four, six, and eight hours for all foundations was calculated using Eq. 3 and plotted in Fig. 1. The results show that the change of $\Delta E^*$ intensifies over time across all products. Brand B exhibits the largest $\Delta E^*$, followed by Brand C, Brand D, and then Brand A. In general, the most pronounced discoloration is observed in the intermediate shades, specifically shades 5 and 6 of each brand. Figure 1A shows that Brand A demonstrates minimal discoloration, with all products exhibiting a $\Delta E^*$ of less than 2.2. In particular, the darkest shade, A-8 experienced the smallest discoloration, less than 1.0 in $\Delta E^*$. Color difference of 1.0 is commonly accepted as the threshold for a just noticeable difference (JND) under standard viewing conditions¹¹. Also, $\Delta E^*$ of all other shades of Brand A were considered as perceptible on close inspection. However, most of the rest showed $\Delta E^*$ bigger than 3.0, interpreted as the easily noticeable difference. In the following sections, the discoloration trends were identified in terms of lightness, redness, and yellowness.

Fig. 1

$\Delta E^*$ of foundations calculated and plotted at 0, 2, 4, 6, and 8 h-time points. $\Delta E^*$ is most severe in brand B, followed by brand C, D, and A. $\Delta E^*$ is largest in the intermediate shades 5 and 6 of each brand. (A) Brand A. All products only experienced minimal $\Delta E^*$ compared with other brands. (B) Brand B. B-5 and B-6 show the greatest $\Delta E^*$. (C) Brand C. C-5 and C-6 show the greatest $\Delta E^*$. (D) Brand D. D-5 and D-6 show the greatest $\Delta E^*$.

Discoloration by $L^*$ changes

The $L^*$ values of liquid foundations Brand A, B, C, and D that were recorded right after application, and after two, four, six, and eight hours are plotted and presented in Fig. 2. Dark shades exhibited the least variation when compared to their lighter counterparts within the same brand. Specifically, the darkest shades, shades 7 and 8 of each brand, demonstrated relatively minimal changes in $L^*$ values when compared to the other shades within their respective brands. This trend underscores the consistency of minimal $L^*$ change among the dark shades within each brand, which shows that the lightness of dark shade foundations is relatively unchanged. Although this pattern is observed across all brands, the degree of $L^*$ change varies among them. Notably, as shown in Fig. 3A, Brand A displayed minimal changes across all shades. Furthermore, the extent of darkening did not strictly align with the reduction in lightness. In other words, our data showed that darkening was not less pronounced as the shade color deepened.

Fig. 2

$L^*$ value of foundations plotted at 0, 2, 4, 6, and 8 h-time points. Each brand’s two darkest shades (shades 7 and 8) show less $L^*$ change than other shades. (A) Brand A. All products only experienced minimal $L^*$ change, with A-7 and A-8 showing the least change. (B) Brand B. B-7 and B-8 show the least $L^*$ change. (C) Brand C. C-7 and C-8 show the least $L^*$ change. (D) Brand D. D-7 and D-8 show the least $L^*$ change.

Discoloration by $C^*$ changes

The chroma ($C^*$) values of liquid foundations from Brands A, B, C, and D were calculated using Eq. 1 and are plotted in Fig. 3.

The results indicate that $C^*$ values decreased across all shades in Brands A and D. In contrast, for Brands B and C, shades 1 and 2 (the lightest shades) exhibited an increase in $C^*$, while the remaining shades (3 to 8) showed a decrease.

For Brands A, C, and D, the most substantial chroma decrease occurred in shades 5 and 6. However, in Brand B, B-7 demonstrated the greatest reduction in $C^*$. As illustrated in Fig. 3A, Brand A maintained its chroma relatively well, exhibiting minimal overall changes.

Fig. 3

$C^*$ values of foundations calculated and plotted at 0, 2, 4, 6, and 8 hours. For Brands B and C, an increase in $C^*$ values is observed for shades 1 and 2, while shades 3 to 8 exhibit a decrease. In contrast, all shades from Brands A and D show a decrease in $C^*$. (A) Brand A: Minimal chroma changes, with all shades (A-1 to A-8) decreasing in $C^*$. (B) Brand B: $C^*$ values of B-1 and B-2 increase, while B-3 to B-8 decrease. (C) Brand C: $C^*$ values of C-1 and C-2 increase, while C-3 to C-8 decrease. (D) Brand D: $C^*$ values of all shades (D-1 to D-8) decrease.

Discoloration by $h^\circ$ changes

The hue ($h^\circ$) values of all foundation products were calculated using Equation 2 and plotted in Fig. 4.

As time progressed, changes in $h^\circ$ became more evident across all products. In Brands B and C, a consistent decrease in $h^\circ$ was observed across all shades after 8 hours. In contrast, for Brands A and D, a decrease in $h^\circ$ was observed in shades 1 to 6, while shades 7 and 8 exhibited an increase.

A general trend emerged, indicating that darker foundation shades exhibited smaller changes in $h^\circ$ compared to lighter shades within the same brand. As shown in Fig. 4A, Brand A demonstrated minimal $h^\circ$ variation compared to other brands.

Fig. 4

Time-series changes in hue angle ($h^\circ$) for foundation shades from Brands A to D, measured at 0, 2, 4, 6, and 8 hours. A decrease in $h^\circ$ corresponds to a shift toward a redder hue, while an increase indicates a shift toward a yellower hue. To assist with intuitive interpretation, a vertical color stripe is shown along the y-axis, representing hue angles ranging from $45^\circ$ to $80^\circ$ at a fixed lightness ($L^* \approx 70$) and chroma ($C^* \approx 40$) level. (A) Brand A: Shades A-1 to A-6 exhibit a decrease in $h^\circ$, while A-7 and A-8 show a slight increase. (B) Brand B: All shades (B-1 to B-8) show a consistent decrease in $h^\circ$. (C) Brand C: All shades (C-1 to C-8) decrease in $h^\circ$ over time. (D) Brand D: Shades D-1 to D-6 decrease, whereas D-7 and D-8 slightly increase in $h^\circ$ during the 8-hour period.

Scanning electron microscopy (SEM) analysis

For the SEM analysis, four foundation shades were selected (Table 2): A-1, A-8, B-1, and B-8. These shades represent the lightest and darkest categories from Brand A, which exhibited the least discoloration, and Brand B, which showed the most severe discoloration (see Fig. 1). By comparing brands with minimal and severe discoloration, as well as the lightest and darkest shades within each, this analysis provides a focused contrast to identify surface characteristics potentially linked to discoloration.

Table 2 Initial $L^*$ values, $L^*$ values after 8 hours, and $L^*$ changes of selected foundations.

Figure 5 shows secondary electron (SE) images captured using a Hitachi S-4800 SEM. SE imaging provides detailed topographical contrast and was used here to qualitatively examine surface textures across selected samples. Brand A (Fig. 5A and B) displays smoother and more homogeneous surfaces, while Brand B (Fig. 5C and D) exhibits rougher textures and more apparent surface irregularities.

While SE imaging does not provide compositional information or confirm pigment dispersion mechanisms, the differences in surface morphology may offer visual context to the optical behaviors observed in discoloration. For example, surfaces with more visible topographical variation and texture (as seen in Brand B) may lead to increased light scattering or shadowing, potentially contributing to the perceptual darkening over time.

We emphasize that these observations are qualitative and exploratory. No causal inferences are drawn regarding pigment mobility or chemical changes. As described in the Method section, the SEM analysis was limited to secondary electron imaging without conductive coating. Therefore, our interpretation focuses on qualitative topographical features rather than quantitative pigment characterization.

Fig. 5

SEM images (secondary electron mode) at scale bars of $500~\upmu$m, $50~\upmu$m, and $10~\upmu$m for foundations A-1 (A), A-8 (B), B-1 (C), and B-8 (D). (A) and (B) show smoother and more uniform surface textures, while (C) and (D) display greater topographical irregularities. These qualitative differences may correspond to the observed discoloration trends but should not be interpreted as direct evidence of pigment dispersion or composition s.

Trend summary of discoloration

The results indicate that discoloration trends vary depending on the brightness levels of the foundation shades. Based on these observations, the foundation shades were categorized into four groups (L1–L4) according to their lightness ($L^*$) values. This categorization is summarized in Table 3, which highlights the major directional changes in $L^*$, $a^*$, $b^*$, $C^*$, and $h^\circ$ for each group.

Overall, most foundation samples exhibited noticeable darkening over time, with the exception of those in the dark shade category (L4), which showed relatively stable lightness. Specifically, the $L^*$ values in the L4 group decreased markedly less than in the L1, L2, and L3 categories across all brands. Given the minimal decrease, we conclude that the lightness of dark shades (L4) remains comparatively stable.

The $a^*$ values increased over time for the light and medium-light categories (L1 and L2), whereas they decreased in the medium and dark categories (L3 and L4). The $b^*$ and $C^*$ values declined in the medium-light, medium, and dark categories (L2, L3, and L4), while in the lightest category (L1), the changes in $b^*$ and $C^*$ varied by brand. The $h^\circ$ value decreased in the L1, L2, and L3 groups, whereas in L4, trends varied across brands.

Summarizing the chromatic discoloration trends: foundations in the L2 and L3 groups became less vivid and saturated over time. Tone changes were also evident-light and medium-light shades (L1 and L2) shifted toward a redder tone, with $a^*$ increasing by approximately 1.26 on average. In contrast, medium and dark shades (L3 and L4) became paler, with $a^*$ and $b^*$ decreasing by 2.16 and 4.44, respectively, indicating reduced chromatic intensity. Among the tested brands, Brand A was the only one that maintained relatively stable color across all shade categories.

We acknowledge that the boundary values used for categorizing lightness were based on the four brands tested in this study. These thresholds may not generalize across all liquid foundation formulations.

To offer a more comprehensive overview of discoloration behavior, representative numeric values for parameters-$L^*$, $a^*$, $b^*$, and $\Delta E^*$-at 0 h, 2 h, and 4 h are consolidated in Table 4. This summary enables clearer comparison across brands and shade categories and supports the interpretation of key trends observed in the dataset.

Table 3 Categorization of foundation shades based on lightness levels. Discoloration is characterized by an increase or decrease in $L^*$ (lightness), $a^*$ (redness), $b^*$ (yellowness), $C^*$ (chroma), and $h^\circ$ (hue angle). The $L^*$ values of dark shade foundations remained stable. The decrease in hue angle suggests that the foundation appeared redder than its initial color. The lightness category thresholds were derived from the current dataset and may not be universally applicable to all liquid foundation products.

Table 4 Representative colorimetric values ($L^*$, $a^*$, $b^*$, $\Delta E^*$) measured at 0 h, 2 h, and 4 h across all tested brands and shade categories. The summarized data illustrate the direction and magnitude of discoloration in early time intervals, where most color changes occurred. This consolidated view supports visual comparison and complements the charts presented as Figs. 1, 2, 3, 4.

Discussion

The identification of distinct darkening patterns based on lightness level underscores the need for brands to allocate resources toward addressing discoloration issues most relevant to their consumer base. Brands that actively minimize discoloration not only enhance product reliability but also strengthen their competitive advantage by appealing to consumers seeking stable, long-lasting foundation shades.

In this study, spectrophotometer measurements showed that the $L^*$ values of dark shades exhibited negligible change over time. SEM images, used here as qualitative topographic references, showed that samples from darker shades tended to exhibit rougher and more irregular surface textures compared to lighter shades. While this may be visually interpreted as surface clustering, we refrain from attributing these features directly to pigment agglomeration or redistribution after application. It is plausible that higher pigment concentration contributes to surface complexity; however, further analysis using complementary techniques (e.g., EDX, Cryo-SEM) would be needed to support such conclusions. Additionally, dark pigments absorb more light and reflect less, making discoloration less noticeable. In contrast, lighter shades reflect more light, making even slight darkening more visible. Thus, despite surface irregularities, the high pigment content and optical absorption properties of dark shades may contribute to lower perceptual discoloration.

While discoloration trends were observed across brands, the severity of the discoloration varied. In our results, shades from Brand A generally remained below this value, suggesting imperceptible discoloration to most users. In contrast, many shades from Brands B, C, and D exceeded 3.0 in $\Delta E^*$, which is considered a clearly visible difference. These benchmarks provide a more intuitive understanding of the severity of discoloration in practical terms. This may reflect differences in formulation strategies beyond pigment composition, including dispersion stability and surface morphology. A more stable formulation can contribute to improved visual consistency over time, which is critical for product performance.

To better contextualize these findings, we also considered perceptual thresholds of color differences. A change in $\Delta E^*$ of approximately 1.0 is widely accepted as the threshold for a JND under standard viewing conditions¹¹. More sophisticated analysis methods such as $\Delta E_{94}$, $\Delta E_{00}$ and $\Delta E_{\text {CMC}}$ will be employed to better align the reported color differences with human perceptual sensitivity. These approaches will be supported by repeated measurements and improved statistical power to refine the evaluation of subtle color changes. In addition, while this study primarily analyzed discoloration based on $L^*$, $a^*$, and $b^*$, we acknowledge that color shifts often follow nonlinear trajectories that may not be fully captured by single-axis plots. The perceptual changes observed across multiple dimensions (Figs. 2, 3, 4) suggest that future analyses may benefit from trajectory-based visualization methods to better interpret the underlying colorimetric dynamics.

Furthermore, discoloration trends may be more accurately assessed when foundation shades are applied to human skin rather than inert substrates. As demonstrated in a previous study⁸, evaluating color changes on diverse skin types offers a more realistic understanding of perceptual discoloration and its cosmetic implications. In particular, testing directly on human faces may yield different thresholds of color tolerance compared to standardized substrates, due to the complexity of facial skin conditions and structural features. Future research incorporating in vivo testing across a range of skin tones is needed to validate and extend the current findings.

Discoloration is a complex phenomenon influenced by multiple factors, including formulation ingredients and manufacturing processes. Foundations consist of solvents, pigments, emollients, oils, waxes, thickeners, and humectants¹². The interactions among these components may contribute to discoloration, demanding further investigation to develop more stable formulations. Beyond pigment properties, formulation finish type-particularly oil composition-may also influence discoloration. Dewy foundations, which typically contain higher concentrations of emollients and natural oils, tend to exhibit more visible discoloration over time. In contrast, matte formulations often include oil-absorbing agents that may reduce such changes. Although this study did not analyze oil composition directly, the observed brand-level differences suggest that finish type may be a contributing factor. Studies are expected to clarify the relationship between oil content, finish type, and long-term color stability.

To enable such progress, further investigation using controlled formulation variables will be essential to establish causal relationships and deepen understanding of the mechanisms driving foundation discoloration. While this study proposed plausible explanations-such as differences in surface morphology and finish type-the underlying causes remain speculative due to the use of commercially available products with undisclosed formulations. Together, these enhancements will allow for a deeper and more realistic understanding of foundation discoloration in everyday use.

Materials

Selection of liquid foundations

Liquid foundations from four global brands (referred to as Brand A, B, C, and D) were selected for this study based on their popularity, as indicated by sales rankings on major e-commerce platforms such as Sephora and Amazon. The selection criteria included products ranked among the top 10 best-selling foundation items in June 2023, which marketed claims of long-lasting wear and resistance to color change, medium to full coverage, a matte or semi-matte finish, and a diverse shade range to accommodate various skin tones. All products were purchased in South Korea between June and July 2023. Each brand was represented by eight foundation shades, covering a range of brightness levels and undertones. These foundations were applied to an opacity chart, and their colors were measured at multiple time points.

By testing commercially available foundations, this study aims to determine whether discoloration trends remain consistent across different manufacturers, each employing distinct formulations, production techniques, and processing environments. While the detailed ingredient lists are proprietary, all four foundations are comparable in terms of general formulation characteristics. Each is marketed as a long-wear, oil-free, matte-finish foundation and contains commonly used base ingredients such as water, dimethicone, glycerin, and iron oxides. Some products mention additional features (e.g., climate-adaptive technology or fruit extracts), but no extreme differences in base formulation are evident from public documentation.

The findings will provide insights into the foundation discoloration phenomena encountered by consumers, offering valuable information for both manufacturers and end-users.

Liquid foundation preparation

Liquid foundation is typically applied to the face, but direct color measurement on the skin presents challenges. First, it is difficult to control application thickness due to the uneven surface of the skin. Second, foundation color can be affected by individual sebum and sweat production, which varies due to hormonal cycles, skin type, and environmental conditions¹³. Since this study focuses on foundation discoloration independent of skin interactions, color measurements were performed using an opacity chart instead of human skin.

To ensure consistency, liquid foundation was applied to the opacity chart using a stainless steel, four-sided applicator⁸. Environmental conditions were controlled at 27 °C ± 1 °C and 46% ± 2% humidity, simulating standard indoor conditions in summer.

Approximately 4–5 drops of foundation were dispensed onto the opacity chart. For three of the four products (Brands A, C, and D), which use pump-type packaging, each drop was approximately 0.15 mL, leading to an estimated total volume of 0.60–0.75 mL. Brand B was in a tube format and dispensed manually to a comparable amount. The foundation was then evenly spread to a thickness of $200~\upmu \text {m}$ using the applicator. This thickness was selected to ensure complete opacity-preventing the underlying chart from influencing measurements-and to allow multiple measurements on a single spread area. It also optimized material usage while maintaining measurement accuracy. Pilot testing confirmed that thinner layers (e.g., $< 200~\upmu \text {m}$) resulted in partial visibility of the chart’s black/white boundary, compromising color fidelity.

This setup ensured that foundation discoloration was evaluated under standardized and repeatable conditions, eliminating inconsistencies that could arise from skin-based testing.

Methods

Color measurement using a spectrophotometer

All color measurements were performed using a Konica Minolta CM-2600d spectrophotometer, a device widely used in color science for its high reproducibility and compliance with international standards. The device uses a pulsed xenon arc lamp as its light source, providing broad-spectrum illumination that closely matches CIE standard illuminant D65. Measurements were conducted in SCI mode with a viewing geometry of $8^\circ$:diffuse (d/$8^\circ$) and a measurement aperture of 3 mm. Ultraviolet (UV) components are filtered by default in the standard setting, minimizing fluorescence effects.

$L^*$, $a^*$, and $b^*$ values were calculated according to the CIE1976 $L^*a^*b^*$ color space, as defined in ISO 11664-4:2019¹⁴. Chroma ($C^*$) and hue angle (${\hbox {h}}^\circ$) were derived from a and b as introduced in Eqs. 1 and 2. Reflectance and geometric measurement conditions were consistent with ISO 7724-1 to -3 guidelines. All measurements were performed under controlled laboratory conditions to ensure repeatability and minimize environmental variation.

Figure 6 illustrates the color measurement process on an opacity chart. Measurements for each product and trial were conducted within the same foundation spread to minimize variation. Four to five drops of foundation were dispensed onto the opacity chart and uniformly spread to a thickness of $200~\upmu \text {m}$ using an applicator (Fig. 6A). The first measurement was taken within approximately 10 seconds after spreading the product surface, in the upper section of the spread, as shown in Fig. 6B. The second measurement, performed after a two-hour interval, was taken slightly lower than the first measurement (Fig. 6C). This process was repeated for subsequent measurements, ensuring that the final spread included four distinct locations corresponding to the four measurement time points.

This method was implemented to reduce potential factors affecting color consistency and to optimize both time and resources. Since each measurement was taken from an untouched section of the spread, the foundation remained uncontaminated, ensuring accurate color assessment.

Fig. 6

Process of foundation color measurement using an opacity chart and spectrophotometer. (A) Liquid foundation (4–5 drops) was dispensed onto an opacity chart and uniformly spread. (B) Spectrophotometer measurement at the initial time point. (C) Spectrophotometer measurement at the second time point (2 hours).

Collection of SEM images

Scanning Electron Microscopy (SEM) analysis was conducted to investigate potential causes of discoloration by examining the surface morphology and pigment dispersion of selected foundation shades. A Hitachi S-4800 SEM was used to capture high-resolution images of pigment distribution, providing insights into differences between brands and shades exhibiting varying degrees of discoloration.

SEM is widely utilized in the paint and coatings industry for assessing pigment behavior^15,16,17, particularly in conjunction with techniques such as Cryo-SEM, AFM, TEM, or EDX to evaluate particle composition and distribution at the microscale. In contrast, the present study employed standard secondary electron (SE) imaging, which primarily offers qualitative insights into surface topography-such as smoothness, roughness, and apparent clustering of pigment particles. We therefore refrain from inferring detailed pigment dispersion mechanisms or chemical composition based solely on the SE images. Rather, the SEM images are intended to support visual comparisons of surface characteristics across brands and shade categories.

In this study, foundations were dried for 72 hours at 25 °C before imaging, without any additional coating or conductive treatment. While this preparation allowed the foundation to stabilize and immobilize pigment particles, we acknowledge that the absence of sputter-coating may have introduced charging artefacts, particularly given the non-conductive nature of the samples. Furthermore, the drying process may have created skin layers with trapped moisture, which could lead to surface disruption or crater-like features under vacuum conditions. These limitations constrain the interpretability of the observed structures. Future studies may benefit from conductive coatings, EDX analysis, or cryo-SEM techniques to minimize artefacts and better characterize true surface morphology and pigment distribution.