Evolution and Stability of Post-Fermentative Copigmentation in Listán Negro Red Wine Using Caffeic Acid and Glucose

Abstract

The Listán Negro cultivar, a red grape variety endemic to the Canary Islands, has been traditionally characterized by limited color stability, often leading to significant pigment loss during early storage. The present study investigates the efficacy of caffeic acid as a hydroxycinnamic cofactor for copigmentation when it is introduced during the post-fermentative stage in young, lightly colored red wines. Ninety young red wine samples were prepared using a factorial design. Caffeic acid was added at four concentrations, both independently and in combination with glucose, and was monitored over 158 days. Initial spectrophotometric analysis revealed a dose-dependent hyperchromic effect, with color intensity increasing by up to 12.8% for caffeic acid alone and 15.9% when combined with glucose. Accompanying bathochromic shifts (1–3 nm) were consistent with copigmentation interactions. Although the effect decreased over time at lower concentrations, doses ≥ 480 mg/L maintained improved color retention after storage. The addition of glucose alone did not improve color stability and showed a transient chromatic response. These results indicate that post-fermentative addition of caffeic acid provides a short-term enhancement of color intensity in low-anthocyanin red wines; however, this effect is strongly concentration-dependent and decreases over time, thereby identifying a threshold for persistent post-fermentative copigmentation.

1. Introduction

Wine color is one of the most decisive sensory attributes influencing consumer perception and market acceptance, acting as an immediate indicator of varietal identity, technological management, and aging status [1]. In red wines, chromatic expression is primarily governed by anthocyanins, flavonoid pigments extracted from grape skins during maceration [2]. However, despite their central role, anthocyanins are intrinsically unstable molecules whose visible form depends on a dynamic equilibrium influenced by pH, ethanol, temperature, sulfite concentration, and the presence of cofactors that may enhance or reduce color intensity [3].

At wine pH (3.2–4.0), anthocyanins coexist in multiple interconvertible forms: the red flavylium cation (A⁺), the quinoidal base (A⁰), the colorless hemiketal (carbinol pseudobase), and the chalcone form [4]. Under these conditions, hydration reactions strongly favor the formation of colorless species, resulting in a significant loss of chromatic expression despite the presence of measurable pigment concentrations [5].

Two principal stabilization mechanisms contribute to preserving red color during vinification and storage: covalent condensation reactions leading to polymeric pigments, and non-covalent copigmentation interactions [6]. Copigmentation is defined as the molecular association between anthocyanins and colorless cofactors—such as hydroxycinnamic acids, flavonols, or other phenolics—through π–π stacking interactions [7]. These interactions can partially stabilize the flavylium cation by reducing the accessibility of water to reactive positions, thereby shifting the equilibrium toward colored species rather than completely preventing hydration. Spectrophotometrically, this phenomenon is characterized by a hyperchromic effect (increase in absorbance) and a bathochromic shift (displacement of the maximum absorbance toward different wavelengths). These changes arise from modifications in the electronic environment of the chromophore due to intermolecular interactions affecting π-electron energy levels [7].

Hydroxycinnamic acids, particularly caffeic acid, are among the most studied cofactors due to their planar structure and high electron density, which favor stacking interactions [8]. Numerous studies have demonstrated that copigmentation may account for up to 30–50% of the visible color in young red wines under optimal conditions [9,10,11]. However, the magnitude of this contribution is strongly matrix-dependent and highly sensitive to ethanol concentration, which weakens hydrophobic and Van der Waals interactions responsible for complex stability [12]. In addition, sugars such as glucose have been proposed to influence copigmentation through hydrogen-bond networks in aqueous systems, although their actual stabilizing role in finished wines remains unclear [13].

The relevance of these mechanisms becomes evident in grape cultivars characterized by relatively low anthocyanin levels and limited pigment stability. The Listán Negro variety, a native Vitis vinifera red grape endemic to the Canary Islands and widely cultivated in most of the Archipelago, is typically associated with modest anthocyanin concentrations, typically ranging between 250 and 400 mg/L [14], and rapid chromatic decline during early storage [15]. Previous studies conducted on Listán Negro red wines have demonstrated that copigmentation can represent a substantial proportion of the initial color expression, with marked visual consequences measurable in CIELAB space, with ΔE*_ab values well above the perceptibility threshold [16]. Furthermore, it has been reported that copigmentation in Listán Negro significantly influences not only color intensity but also hue mutation, contributing to a darker and more purplish visual profile in young wines [16], which is generally associated with higher consumer acceptance.

This intrinsic chromatic limitation is further exacerbated under certain winemaking conditions, particularly when extraction during maceration is moderate or when early oxidative processes occur. These characteristics make this variety particularly suitable for evaluating post-fermentative strategies aimed at improving color intensity and stability, as even moderate chromatic enhancements may have a measurable visual impact.

Although this study focuses on Listán Negro, the approach may be relevant to other red grape varieties characterized by moderate anthocyanin content and limited color stability, such as Pinot Noir, Grenache or Sangiovese [17]. Previous studies have shown that copigmentation mechanisms operate similarly across different varietal matrices, although their magnitude depends on phenolic composition and ethanol content [18]. Therefore, the present work may provide a framework for evaluating post-fermentative color enhancement strategies in other low-color red wines.

From a commercial perspective, insufficient color intensity in red wines is often perceived as a quality limitation, as consumers tend to associate deeper and more saturated hues with higher quality and aging potential [19]. Therefore, strategies aimed at enhancing and stabilizing color in such wines are of practical relevance, particularly when they can be applied after fermentation without modifying the vinification process. Accordingly, the present study evaluates whether post-fermentative interventions may partially compensate for this intrinsic chromatic limitation, acting as a potential corrective approach for wines that have already been produced.

Most reported research has focused on prefermentative or co-maceration strategies, where cofactors are present in an aqueous must environment and may later participate in condensation reactions forming more stable polymeric pigments [20,21,22,23]. Under these conditions, increases in color intensity between 25% and 60% have been observed [24]. However, the efficiency of copigmentation once alcoholic fermentation has concluded remains insufficiently explored. In finished wines containing between 13 and 14% ethanol, non-covalent complexes are expected to be less stable, and their contribution to long-term color preservation may differ substantially from that observed in must [25]. As a consequence, the magnitude, persistence, and technological relevance of post-fermentative copigmentation cannot be directly extrapolated from model systems or must-based studies.

Therefore, there is a need to evaluate how post-fermentative conditions influence copigmentation behavior under realistic wine conditions, particularly in terms of concentration dependence and temporal stability. The present study aims to evaluate the chromatic evolution and stability of post-fermentative copigmentation induced by caffeic acid in a young Listán Negro red wine. By combining spectrophotometric analysis with CIELAB colorimetry and ΔE*_ab perceptibility thresholds, this work seeks to (i) quantify the magnitude of the hyper-chromic and bathochromic effects, (ii) determine the persistence of these effects over a 158-day storage period, and (iii) identify potential concentration thresholds required to achieve perceptible and sustained chromatic effects in this type of young red wine. The key novelty of this study lies in the experimental evaluation of post-fermentative copigmentation under realistic hydroalcoholic wine conditions, combined with the identification of a minimum concentration threshold required to achieve perceptible and persistent chromatic effects.

From an enological perspective, post-fermentative interventions aimed at correcting or improving specific attributes of wine are well established. Additions of compounds such as tannins, mannoproteins, or even oxygen are routinely applied after fermentation to adjust structure, stability, or sensory properties [26]. In this context, the addition of caffeic acid can be conceptually framed as a targeted modulation of phenolic interactions within the existing matrix. In the case of hydroxycinnamic acids and glucose, these compounds are intrinsically present in grapes and wines, with caffeic acid typically occurring at concentrations of a few tens of mg/L depending on the variety and processing conditions [27]. It should also be noted that the concentrations evaluated in this study were selected to explore system behavior across a wide range, rather than to reproduce typical endogenous levels. However, the direct addition of pure hydroxycinnamic acids such as caffeic acid is not currently authorized under existing regulatory frameworks from the International Organisation of Vine and Wine (OIV). Nevertheless, structurally related phenolic compounds, including hydroxycinnamic acid derivatives, are already introduced into wines through the use of approved oenological products such as commercial tannins, which are plant-derived extracts with complex phenolic composition. Therefore, the present approach should be interpreted as research conducted under controlled experimental conditions, while also providing insight into the behavior of phenolic systems relevant to existing enological practices.

2. Materials and Methods

2.1. Preparation of Red Wine Samples

A lightly colored young red wine from the Listán Negro cultivar produced in the Tacoronte-Acentejo region (Tenerife, Spain) was used for this study. This red wine was fermented in stainless steel vats and supplied by Bodegas Monje S.L (Tenerife, Canary Islands, Spain). At the time of the study, the wine had been aged for two months.

Ten liters of wine were supplemented with defined concentrations of caffeic acid (120, 240, 480, and 960 mg/L) and, where applicable, glucose (10 g/L) and homogenized in a 16 L glass container. All additions were performed gravimetrically using analytical-grade reagents (≥99% purity Sigma-Aldrich, Merck KGaA, Darmstadt, Germany). After addition, samples were homogenized for 10 min under controlled stirring (300 rpm) and protected from light to minimize oxidation. Exposure to dissolved oxygen was minimized during all handling steps.

A total of 90 experimental samples were prepared using a factorial design with 10 different treatments and 9 replicates each. The treatments included:

• Control group: Original red wine without additions (n = 9).
• Glucose control: Wine supplemented with 10 g/L of glucose (n = 9).
• Caffeic acid series: Wine with four increasing concentrations of caffeic acid (120, 240, 480, and 960 mg/L; n = 9 each, total n = 36).
• Caffeic acid plus glucose series: wine containing each of the four caffeic acid concentrations combined with 10 g/L glucose (n = 9 each, total n = 36).

Red wine samples were transferred into hermetically sealed containers and stored under refrigeration. Samples were stored at 9 ± 1 °C under dark conditions to minimize oxidation and preserve their initial characteristics throughout the experimental period.

The selected concentrations of caffeic acid (120–960 mg/L) were chosen to cover a range from levels comparable to naturally occurring hydroxycinnamic acids in wines [28] to supra-physiological concentrations allowing the exploration of saturation effects in copigmentation systems [11]. This range was selected to evaluate concentration-dependent behavior and identify potential threshold effects under post-fermentative conditions. All experimental additions were performed for research purposes under controlled laboratory conditions and do not reflect authorized oenological practices for commercial wine production.

The glucose concentration (10 g/L) was selected to simulate residual sugar conditions found in off-dry commercial wines [29] and to evaluate its potential role as a modulator of copigmentation through hydrogen-bonding interactions [13].

2.2. Physico-Chemical Characterization

Initial characterization of the base wine was conducted using official analytical methods from the OIV [29]. All parameters were measured in triplicate. Density and alcoholic strength (% vol) were determined using densitometric methods; total dry extract was determined by gravimetry; total, fixed and volatile acidity were measured by titration; pH was determined using a calibrated pH meter; sulfur dioxide (free and total) was determined by iodometric titration; and malic acid was quantified using enzymatic analysis.

Additionally, total polyphenol index (TPI), tannins and total anthocyanin content were measured according to professional oenological procedures using absorbance at 280 nm, 320 nm and 520 nm [26].

These parameters were considered relevant due to their potential influence on anthocyanin stability and copigmentation equilibria. Among them, pH, ethanol content, and sulfur dioxide levels are known to affect color expression and pigment evolution in red wines [7]. Furthermore, several of these variables are interrelated, such as free and total SO₂, ethanol and density, and acid content and pH.

2.3. Spectrophotometric Measurements and Chromatic Parameters

The evolution of color was monitored at three intervals: immediately after addition, 87 days, and 158 days. All measurements were performed using a Perkin Elmer Lambda 25 spectrophotometer (PerkinElmer, Shelton, CT, USA) with a 1 cm light path. Each measurement was performed in triplicate for each replicate sample, and instrument calibration (autozero) was performed daily using distilled water as a blank.

Visible absorption spectra were recorded, specifically measuring absorbance at 420, 520, and 620 nm. Ultraviolet absorption at 280 and 320 nm was measured after dilution (1:100).

The following standard chromatic parameters were calculated according to Glories [30] and the OIV [29]:

• Color Intensity (CI) [28]:

CI = A₄₂₀ + A₅₂₀.

(1)

• Total Color Intensity (TCI) [29]:

TCI = A₄₂₀ + A₅₂₀ + A₆₂₀.

(2)

• Tonality (Hue):

Hue = A₄₂₀/A₅₂₀

(3)

Detailed chromatic characterization was performed in the CIELAB color space [29,31] using PECSS software v. 3.0 (Perkin-Elmer). The tri-stimulus values (X, Y, Z) were used to determine the following coordinates:

• Lightness (L*): from black (0) to white (100).
• a* and b*: representing the red-green and yellow-blue axes, respectively.
• Chroma (C*): indicating color saturation calculated as follows:

C* = (a² + b²)^0.5

(4)

• Hue angle (H*): representing the color tone obtained as follows:

H* = arctan(b*/a*)

(5)

To evaluate the copigmentation effect, the relative color increase (hyperchromic effect) was calculated as:

Relative Color Increase = %Color = [(A_max − A_control)/A_control] × 100

(6)

where A_max is the maximum absorbance of the treated sample and A_control is the maximum absorbance of the control.

The wavelength shift (∆λ) was recorded to identify bathochromic (red-to-blue) or hypsochromic (red-to-yellow) shifts.

2.4. Total Color Difference (∆E*_ab) and Visual Perception Analysis

To evaluate the overall visual impact of copigmentation and the perceptible color changes over time, the total color difference (∆E*_ab) was calculated using the CIE 1976 formula [32]. This parameter represents the Euclidean distance between two points in the CIELAB color space (L*, a*, b*) and was determined as follows:

∆E*_ab = (∆L*² + ∆a*² + ∆b*²)^0.5

(7)

where ∆L*, ∆a* and ∆b* represent the differences in lightness, redness, and yellowness, respectively, between the treated wine samples (with varying doses of caffeic acid and glucose) and the corresponding control at each time point. Additionally, ∆E*_ab was used to monitor the temporal color stability by comparing the coordinate values at 158 days with initial measurements.

To further investigate the nature of visual changes, a partitioning analysis of the individual coordinate contributions was performed to determine whether perceived differences were mainly due to lightness (L*) or chromatic shifts (a*, b*). The percentage contributions were calculated using the following equations:

• Contribution of Lightness (%Cont.∆L*):

%Cont.∆L* = [(∆L*)²/(∆E*_ab)²] × 100

(8)

• Contribution of Redness (%Cont.∆a*):

%Cont.∆a* = [(∆a*)²/(∆E*_ab)²] × 100

(9)

• Contribution of Yellowness (%Cont.∆b*):

%Cont.∆b* = [(∆b*)²/(∆E*_ab)²] × 100

(10)

This analysis was applied to all treatments across the three sampling times.

2.5. Statistical Analysis

Data are expressed as mean ± standard deviation (n = 9). Statistical analyses were performed using SPSS software (version 17.0, IBM Corp., Armonk, NY, USA).

Normality and homoscedasticity of the data were assessed using the Shapiro–Wilk and Levene tests, respectively. One-way analysis of variance (ANOVA) was used to evaluate treatment effects at each time point. When significant differences were observed (p < 0.05), Duncan’s multiple range test was applied.

Additionally, two-way ANOVA was performed to evaluate the combined effects of caffeic acid and glucose addition over time. Statistical significance was set at p < 0.05.

3. Results

3.1. Initial Physico-Chemical Characterization of the Base Wine

The oenological characterization of the young Listán Negro red wine (

Table 1. Oenological characterization of Listán Negro red wine used in experimental trials.

Analytical Parameter	Mean Value (±SD)	Reference Method
Physicochemical Parameters
Density (g/cm3)	0.9930 ± 0.0002	[29] OIV-MA-AS2-01A
Specific Gravity (20°/20°)	0.9948 ± 0.0003	[29] OIV-MA-AS2-01A
Alcoholic strength (% vol)	13.41 ± 0.12	[29] OIV-MA-AS312-01
Total Dry Extract (g/L)	31.6 ± 0.5	[29] OIV-MA-AS2-03A
pH (pH units)	3.54 ± 0.02	[29] OIV-MA-AS313-15
Sulfur Dioxide
Total SO2 (mg/L)	42 ± 10	[29] OIV-MA-AS323-04A2
Free SO2 (mg/L)	13 ± 5	[29] OIV-MA-AS323-04A
Acidity profile
Total Acidity (g. tartaric acid/L)	5.20 ± 0.15	[29] OIV-MA-AS313-01
Volatile Acidity, VA (g. acetic acid/L)	0.50 ± 0.05	[29] OIV-MA-AS313-02
Fixed Acidity (g. tartaric acid/L)	4.60 ± 0.17	[29] OIV-MA-AS313-03
Malic Acid (g/L)	0.15 ± 0.04	[29] OIV-MA-AS313-11
Phenolic composition
TPI (280 nm Absorbance Units)	47.0 ± 1.2	[26]
Tannins (g/L)	3.10 ± 0.25	[26]
Anthocyanins (mg/L)	319 ± 15	[26]

) revealed a physicochemical profile consistent with the varietal typology and regional standards [14]. The wine exhibited an alcoholic strength of 13.41% vol and a pH of 3.54, values within the typical range for red wines and compatible with anthocyanin stability. Total acidity (5.20 g/L as tartaric acid) and low volatile acidity (0.50 g/L as acetic acid) confirmed the good sanitary state of the starting material. Notably, the residual concentration of L-malic acid (0.15 g/L) indicates the successful completion of malolactic fermentation, ensuring biological stability prior to the experimental trials.

Regarding the phenolic profile, the Total Polyphenol Index (TPI) of 47.0 is typical for young red wines intended for early consumption or short aging periods [26]. In addition, the anthocyanin content (319 mg/L) was significantly lower than that found in high-color-intensity cultivars such as Cabernet Sauvignon or Syrah (typically 700–800 mg/L) [33,34]. This relatively low pigment concentration, characteristic of the Listán Negro variety, provides a suitable baseline for evaluating the effects of post-fermentative treatments on color evolution.

3.2. Spectrophotometric Data and Color Intensity

The evolution of spectrophotometric parameters from the beginning of the experiment (t₀) to the end of storage (t₂ = 158 days) is summarized in

Table 2. Evolution of spectrophotometric and CIELAB parameters during storage.

Assay	Trial	A620	A520	A420	IC	TCI	Hue	L*	a*	b*	C*	H*
Initial color parameters (t0 = 0 days)	C0	0.69 c (0.19)	3.469 d (0.20)	2.28 d (0.26)	5.75 e (0.46)	6.44 e (0.65)	0.66 ab (0.03)	23.16 a (1.55)	53.19 a (1.73)	33.22 a (0.64)	62.72 a (1.81)	31.99 a (0.36)
	C1	1.04 ab (0.12)	3.87 bc (0.06)	2.60 bc (0.06)	6.46 d (0.12)	7.50 cd (0.25)	0.67 a (0.01)	13.83 b (3.12)	40.35 bc (4.75)	20.64 bc (5.51)	45.32 b (6.84)	27.09 b (2.86)
	C2	1.06 ab (0.11)	4.04 ab (0.01)	2.76 ab (0.13)	6.79 bc (0.14)	7.86 abc (0.25)	0.68 a (0.03)	13.28 b (2.89)	40.19 bc (4.13)	20.35 bc (4.75)	45.05 bc (5.90)	26.86 bc (2.55)
	C3	1.19 a (0.22)	4.19 a (0.14)	2.82 a (0.22)	7.00 ab (0.37)	8.19 a (0.59)	0.67 a (0.03)	10.52 c (5.29)	36.44 d (7.33)	16.24 d (8.20)	39.89 d (10.30)	24.02 d (4.74)
	C4	1.07 ab (0.12)	4.06 ab (0.03)	2.62 bc (0.01)	6.68 bcd (0.04)	7.76 bc (0.08)	0.65 bc (0.01)	12.82 bc (3.05)	39.57 bc (4.48)	19.44 bc (5.30)	44.09 bc (6.46)	26.17 bc (2.99)
	C0G	1.06 ab (0.11)	3.82 c (0.04)	2.63 bc (0.03)	6.46 d (0.07)	7.52 cd (0.18)	0.69 a (0.01)	13.34 b (2.59)	39.42 bc (3.99)	19.97 bc (4.60)	44.19 bc (5.71)	26.86 bc (3.25)
	C1G	1.06 ab (0.15)	3.97 bc (0.06)	2.62 bc (0.03)	6.59 cd (0.02)	7.65 bc (0.17)	0.66 ab (0.02)	13.37 b (3.81)	40.08 bc (5.36)	20.18 bc (6.22)	44.87 bc (7.71)	26.72 bc (3.25)
	C2G	1.06 ab (0.16)	3.98 bc (0.05)	2.72 abc (0.07)	6.70 bcd (0.13)	7.76 bc (0.28)	0.68 a (0.01)	13.37 b (3.99)	40.18 bc (5.65)	20.23 bc (6.43)	44.99 bc (8.06)	26.72 bc (3.26)
	C3G	1.02 b (0.11)	4.06 ab (0.02)	2.57 c (0.02)	6.63 bcd (0.01)	7.65 bc (0.12)	0.63 c (0.01)	14.15 b (2.96)	41.50 bc (4.33)	21.42 bc (5.06)	46.70 bc (6.25)	27.30 b (2.60)
	C4G	1.10 b (0.13)	4.18 a (0.02)	2.74 ab (0.01)	6.93 a (0.01)	8.02 ab (0.12)	0.66 bc (0.01)	12.37 bc (3.16)	39.24 c (4.63)	19.01 c (5.32)	43.61 bc (6.59)	25.80 bc (3.05)
Intermediate color parameters (t1 = 87 days)	C0	1.02 ab (0.07)	4.63 bc (0.20)	3.09 ab (0.06)	7.73 cd (0.26)	8.75 cd (0.32)	0.67 a (0.02)	14.83 a (1.44)	43.87 ab (1.85)	24.53 ab (2.48)	50.33 ab (2.87)	29.20 a (1.39)
	C1	1.04 ab (0.02)	4.79 ab (0.15)	3.13 ab (0.03)	7.92 bc (0.17)	8.96 abc (0.19)	0.65 ab (0.02)	14.37 ab (0.45)	43.57 ab (0.46)	23.93 bc (0.65)	49.73 ab (0.68)	26.07 b (5.08)
	C2	1.02 ab (0.07)	4.79 ab (0.13)	3.09 ab (0.09)	7.88 bc (0.21)	8.90 bc (0.28)	0.65 ab (0.01)	14.70 a (1.66)	44.03 ab (2.25)	24.43 ab (2.71)	50.37 ab (3.39)	28.93 a (1.55)
	C3	1.00 b (0.10)	4.80 ab (0.28)	3.09 ab (0.16)	7.89 bc (0.44)	8.89 bc (0.54)	0.64 bc (0.01)	15.17 a (1.39)	44.70 a (1.95)	25.13 a (2.76)	51.30 a (3.32)	29.27 a (1.15)
	C4	1.08 a (0.07)	4.97 a (0.16)	3.16 a (0.03)	8.13 a (0.19)	9.21 a (0.26)	0.63 bc (0.02)	16.15 a (0.12)	41.65 bc (0.56)	27.17 a (0.68)	46.80 bc (1.02)	27.00 ab (1.06)
	C0G	1.00 b (0.07)	4.59 c (0.23)	3.01 b (0.06)	7.60 d (0.30)	8.60 d (0.37)	0.65 bc (0.02)	10.67 c (1.35)	30.20 d (1.70)	17.53 d (1.25)	34.90 d (1.55)	20.07 c (1.30)
	C1G	0.99 b (0.01)	4.63 bc (0.09)	3.03 b (0.14)	7.66 cd (0.02)	7.65 bc (0.17)	0.65 ab (0.02)	15.37 a (0.42)	44.73 a (0.55)	25.43 a (0.91)	51.43 a (1.05)	29.60 a (0.45)
	C2G	1.00 b (0.08)	4.64 bc (0.30)	3.01 b (0.09)	7.66 cd (0.38)	7.76 bc (0.28)	0.65 ab (0.03)	15.20 a (1.91)	44.37 a (2.43)	25.17 a (3.10)	51.00 a (3.93)	29.47 a (1.45)
	C3G	1.06 ab (0.07)	4.92 a (0.10)	3.16 a (0.03)	8.07 a (0.15)	7.65 bc (0.12)	0.64 bc (0.01)	13.90 b (1.70)	43.07 ab (2.35)	23.20 bc (2.85)	48.93 b (3.65)	28.23 b (1.55)
	C4G	1.02 ab (0.05)	4.96 a (0.10)	3.16 a (0.05)	8.12 a (0.15)	8.02 ab (0.12)	0.64 bc (0.01)	14.80 a (1.39)	44.40 a (1.74)	24.77 ab (2.30)	50.83 a (2.62)	29.10 a (1.25)
Final color parameters (t2 = 158 days)	C0	0.97 ab (0.03)	4.67 bc (0.06)	3.14 a (0.05)	7.81 bc (0.11)	8.78 bc (0.13)	0.67 a (0.00)	16.33 b (0.73)	46.83 b (1.04)	27.38 ab (1.13)	54.23 b (1.50)	30.33 a (0.52)
	C1	0.96 b (0.06)	4.63 c (0.11)	3.01 b (0.12)	7.64 c (0.22)	8.59 c (0.28)	0.65 bc (0.01)	16.50 b (1.31)	46.83 b (1.74)	27.43 ab (2.06)	54.30 b (2.65)	30.55 a (0.90)
	C2	0.94 b (0.08)	4.59 c (0.21)	2.96 b (0.11)	7.54 c (0.31)	8.49 c (0.39)	0.64 c (0.01)	16.83 b (2.05)	47.10 ab (2.63)	27.80 a (2.98)	54.70 b (3.96)	30.48 a (1.34)
	C3	1.06 a (0.09)	4.89 ab (0.06)	3.10 (0.09)	7.99 ab (0.14)	9.05 ab (0.23)	0.63 c (0.02)	14.08 c (1.88)	43.50 c (3.12)	23.50 c (3.52)	49.48 c (4.67)	28.28 b (1.93)
	C4	1.02 ab (0.03)	4.94 a (0.04)	3.09 a (0.01)	8.03 a (0.03)	9.05 a (0.06)	0.63 c (0.00)	15.25 bc (1.03)	45.70 bc (1.49)	25.60 bc (1.63)	52.35 bc (2.19)	29.25 ab (0.81)
	C0G	0.97 ab (0.06)	4.63 bc (0.13)	3.09 a (0.08)	7.72 bc (0.19)	8.69 bc (0.25)	0.67 a (0.01)	16.25 b (1.37)	46.65 b (1.75)	27.08 ab (2.15)	53.85 b (2.75)	30.10 a (1.05)
	C1G	0.92 b (0.04)	4.54 c (0.13)	2.93 b (0.05)	7.47 c (0.15)	8.39 c (0.20)	0.65 bc (0.01)	17.48 a (0.79)	48.20 a (1.40)	28.93 a (1.69)	56.23 a (2.15)	30.93 a (0.59)
	C2G	0.91 b (0.07)	4.66 bc (0.07)	2.96 b (0.11)	7.63 c (0.14)	8.54 c (0.21)	0.64 c (0.02)	17.60 a (2.06)	48.48 a (2.54)	29.25 a (3.32)	56.68 a (4.15)	31.05 a (1.46)
	C3G	0.91 b (0.10)	4.74 abc (0.17)	2.95 b (0.12)	7.69 bc (0.28)	8.60 bc (0.38)	0.62 c (0.01)	17.55 a (2.69)	48.38 a (3.50)	29.13 a (4.24)	56.50 a (5.49)	30.90 a (1.73)
	C4G	0.97 ab (0.10)	4.98 a (0.13)	3.12 a (0.07)	8.10 a (0.17)	9.07 a (0.25)	0.63 c (0.01)	16.15 b (2.45)	46.55 b (3.16)	27.18 ab (3.39)	53.93 b (4.63)	30.15 a (1.78)

Values are expressed as the mean with standard deviation in brackets (n = 9). Different lowercase letters within the same column indicate statistically significant differences (p < 0.05, Duncan’s test). A620, A520, A420, CI, and TCI are expressed in Absorbance Units (A.U.); L*, a*, b*, C*, H* are expressed in CIELAB Units (C.U.).

. Values represent the mean of nine replicates (n = 9), with the standard deviation shown in parentheses. Significant differences within the same column and time point are indicated by different lowercase letters, according to Duncan’s post hoc test (p < 0.05).

The experimental conditions were defined as follows: C0, Red wine control; C1, C2, C3, and C4, Red wine with 120, 240, 480, and 960 mg/L of caffeic acid, respectively; C0G, Control + 10 g/L glucose; C1G, C2G, C3G, and C4G, Red wine with 120, 240, 480, and 960 mg/L of caffeic acid, respectively, each supplemented with 10 g/L glucose.

Overall, a dose-dependent increase in color intensity was observed, with higher caffeic acid concentrations associated with greater chromatic enhancement. This effect was more pronounced at initial time points and tended to decrease over the storage period.

At t₀, the addition of caffeic acid resulted in a significant increase in color intensity in a dose-dependent manner, with C3 reaching the highest initial value (7.002 A.U.) compared to the control, C0 (5.747 A.U.). Throughout the study, all treatments showed an increase in color intensity over time. By the end of the experiment (t₂), the highest color intensity values were recorded in trials with the highest caffeic acid concentrations (C4 and C4G), with C4G reaching an IC of 8.099 A.U., compared to 7.807 A.U. in the control (C0).

The chromatic coordinates (L*, a*, b*, C*, and H*) were significantly influenced by both the dose of caffeic acid and the presence of glucose. At t₀, a marked decrease in lightness (L*) was observed after the addition of caffeic acid, with C3 showing the lowest value (10.52 C.U.) compared to C0 (23.16 C.U.). By the end of the storage period (t₂), treatments supplemented with both glucose and caffeic acid (C1G, C2G, and C3G) showed higher luminosity (L*) and higher values of a* (redness) and b* (yellowness) than the control and the non-glucose trials. The H* coordinate followed a similar trend to the spectrophotometric Hue, with C1G, C2G, and C3G maintaining the highest values at the end of the study (p < 0.05).

The spatial evolution of selected wine samples within the CIELab color sphere is illustrated in Figure 1. The untreated control (C0) exhibited a progressive shift toward higher b* and L* values over time. In contrast, C3 (480 mg/L caffeic acid) showed limited displacement within the color space, remaining in a darker and more saturated region between intermediate and final time points. A different pattern was observed for the glucose-only treatment (C0G), which showed notable variability over time, with an intermediate shift toward lower saturation followed by a return to values similar to the control at the end of the experiment.

3.3. Evidence of Copigmentation

Immediately after the addition of caffeic acid, a dose-dependent hyperchromic effect was observed in all samples (

Table 3. Evolution of copigmentation parameters during wine storage.

Trial	Copigmentation (t0 = 0 Days)				Copigmentation (t1 = 87 Days)				Copigmentation (t2 = 158 Days)
	λmax	∆λ	AbsMax.	%Color	λmax	∆λ	AbsMax.	%Color	λmax	∆λ	AbsMax.	%Color
C0	527.0	0.0	3.707	0.0%	527.3	0.0	4.689	0.0%	525.6	0.0	4.689	0.0%
C1	529.0	2.0	3.825	3.2%	528.0	0.7	4.84	3.2%	526.0	0.4	4.653	−0.8%
C2	529.0	2.0	3.977	7.3%	527.3	0.0	4.793	2.2%	526.5	0.9	4.679	−0.2%
C3	530.0	3.0	4.063	9.6%	528.0	0.7	4.856	3.6%	526.8	1.2	4.933	5.2%
C4	528.0	1.0	4.181	12.8%	528.0	0.7	5.026	7.2%	526.0	0.4	4.975	6.1%
C0G	529.0	2.0	3.770	1.7%	526.7	−0.6	4.648	−0.9%	526.0	0.4	4.665	−0.5%
C1G	529.5	2.5	3.946	6.5%	527.5	0.2	4.71	0.4%	526.0	0.4	4.57	−2.5%
C2G	530.0	3.0	4.036	8.9%	527.3	0.0	4.695	0.1%	526.0	0.4	4.689	0.0%
C3G	530.0	3.0	4.256	14.8%	528.0	0.7	4.968	5.9%	526.8	1.2	4.769	1.7%
C4G	530.0	3.0	4.296	15.9%	528.5	1.2	4.988	6.4%	526.5	0.9	5.004	6.7%

Note: λ_max (maximum absorbance wavelength) and ∆λ (bathochromic shift) are expressed in nm. Abs_Max indicates the maximum absorbance value measured, and %Color corresponds to the color increase defined in Equation (6).

). The magnitude of this effect depended on cofactor concentration and decreased over time.

At t = 0, several effects were observed following caffeic acid addition:

• Hyperchromic Effect: The increase in maximum absorbance (Abs_Max) relative to the control ranged from 3.18% (120 mg/L, C1) to 12.79% (960 mg/L, C4).
• Bathochromic Shift: All treated samples showed a shift in the maximum absorption wavelength (λ_max) toward longer wavelengths (red-to-blue shift). This displacement reached up to 3 nm and was observed across the concentration range.
• Effect of Glucose: The addition of 10 g/L of glucose increased the initial response, with maximum color increases reaching 15.9% in the C4G treatment.

Monitoring at 87 days and 158 days showed a progressive decrease in the copigmentation effect, particularly:

• 87 days: All treatments maintained higher color values than the control, although the magnitude of the increase was reduced. For example, the C4 treatment decreased from 12.8% to 7.2%. At this stage, an increase in color intensity was also observed in the control (Table 2).
• 158 days: Low concentrations (120 and 240 mg/L) showed values similar to or slightly below the control (−0.77% and −0.21%, respectively).
• High concentrations: Treatments with 480 and 960 mg/L maintained higher color values than the control, with increases of 5.20% and 6.10%, respectively.
• Glucose Effect: The initial effect of glucose was not maintained at later stages, with differences becoming negligible at the end of the experiment.

3.4. Evolution of Visual Perception and Total Color Difference (∆E*_ab)

The chromatic impact of copigmentation was evaluated through the total color difference (ΔE*_ab) over the 158-day study period (

Table 4. Evolution of Total Color Difference (ΔE*_ab) throughout trials and storage.

Trial	Comparison	Initial ΔE*ab (t = 0)	Intermediate ΔE*ab (t = 87 d)	Final ΔE*ab (t = 158 d)
C1	vs. C0	20.25 ± 10.00 a	0.81 ± 0.15 c	0.18 ± 0.05 d
C2	vs. C0	20.79 ± 9.50 a	0.23 ± 0.10 c	0.71 ± 0.20 cd
C3	vs. C0	26.99 ± 12.10 a	1.08 ± 0.45 c	5.59 ± 1.12 a
C4	vs. C0	21.96 ± 8.80 a	3.70 ± 0.95 b	2.37 ± 0.65 bc
C0G	vs. C0	21.48 ± 7.20 a	15.91 ± 3.40 a	0.36 ± 0.10 d
C1G	vs. C0	20.92 ± 9.00 a	0.58 ± 0.20 c	2.37 ± 0.55 bc
	vs. C0G	0.69 ± 0.12 b	17.19 ± 4.10 a	2.71 ± 0.85 b
C2G	vs. C0	20.83 ± 8.50 a	1.22 ± 0.35 c	3.01 ± 0.70 b
	vs. C0G	0.65 ± 0.10 b	16.66 ± 3.80 a	3.18 ± 0.90 b
C3G	vs. C0	18.90 ± 6.40 a	1.32 ± 0.40 c	3.08 ± 0.75 b
	vs. C0G	2.66 ± 0.55 b	14.28 ± 3.20 a	3.17 ± 0.80 b
C4G	vs. C0	22.65 ± 7.90 a	0.58 ± 0.15 c	0.39 ± 0.12 d
	vs. C0G	1.38 ± 0.25 b	16.47 ± 4.00 a	0.17 ± 0.05 d

Note: Values in bold indicate a visual difference detectable by the human eye (ΔE*_ab > 2.7).

). Values above 2.7 were considered perceptible as the ΔE*_ab threshold [32].

At t₀, all treatments showed ΔE*_ab values between 20.25 and 26.99, indicating strong visual differences relative to the control. This initial effect was mainly associated with a decrease in lightness (L*). Over time, a reduction in ΔE*_ab values was observed for all treatments, particularly at lower concentrations.

By the end of the experiment, treatments with 120 and 240 mg/L showed ΔE*_ab values below 1.0, indicating no perceptible difference relative to the control. In contrast, the 480 mg/L treatment (C3) maintained a ΔE*_ab value of 5.59, remaining above the visual perception threshold.

For treatments including glucose, ΔE*_ab values remained higher at intermediate time points (approximately 14–17 at 87 days), but decreased substantially by the end of the experiment, approaching values similar to the control.

The temporal evolution of ΔE*_ab for selected treatments is shown in Figure 2. All treatments initially exceeded the visual perception threshold, followed by a general decrease over time. By 158 days, only moderate and high caffeic acid concentrations maintained perceptible differences.

The contribution of CIELAB coordinates to ΔE*_ab is presented in

Table 5. Evolution of Color Difference (ΔE*_ab) and percentage distribution throughout trials.

Period	Trial	Comparison Ref.	ΔEab*	%ContΔL*	%ContΔa*	%ContΔb*
Initial (t = 0)	C1	vs. C0	20.3 ± 10.0 a	21.2 ± 2.4 b	40.2 ± 3.8 a	38.6 ± 3.5 a
	C2	vs. C0	20.8 ± 9.5 a	22.6 ± 2.1 b	39.1 ± 3.2 a	38.3 ± 3.1 a
	C3	vs. C0	27.0 ± 12.1 a	21.9 ± 1.9 b	38.5 ± 2.8 a	39.6 ± 3.4 a
	C4	vs. C0	22.0 ± 8.8 a	22.1 ± 2.0 b	38.5 ± 3.0 a	39.4 ± 3.2 a
	C0G	vs. C0	21.5 ± 7.2 a	20.9 ± 1.8 b	41.1 ± 2.5 a	38.0 ± 2.9 a
	C1G	vs. C0G	0.7 ± 0.1 b	2.1 ± 0.5 b	1.1 ± 0.3 b	96.8 ± 4.2 a
	C2G	vs. C0G	0.6 ± 0.1 b	5.3 ± 1.2 b	2.5 ± 0.6 b	92.2 ± 5.0 a
	C3G	vs. C0G	2.7 ± 0.6 b	13.9 ± 3.5 c	31.5 ± 4.1 b	54.6 ± 5.8 a
	C4G	vs. C0G	1.4 ± 0.2 b	61.2 ± 7.4 a	3.8 ± 0.9 c	35.0 ± 4.6 b
Intermediate (t = 87 d)	C1	vs. C0	0.8 ± 0.1 c	32.2 ± 4.5 b	13.7 ± 5.1 c	54.1 ± 6.2 a
	C2	vs. C0	0.2 ± 0.1 c	32.1 ± 3.8 b	12.2 ± 4.3 c	55.7 ± 5.9 a
	C3	vs. C0	1.1 ± 0.4 c	34.6 ± 5.2 a	35.2 ± 4.8 a	30.2 ± 3.9 b
	C4	vs. C0	3.7 ± 0.9 b	12.7 ± 2.8 c	35.9 ± 4.1 b	51.4 ± 5.5 a
	C0G	vs. C0	15.9 ± 3.4 a	6.8 ± 1.2 d	73.5 ± 4.8 a	19.7 ± 2.3 c
	C1G	vs. C0G	17.2 ± 4.1 a	7.5 ± 1.4 d	74.3 ± 5.2 a	18.2 ± 2.1 c
	C2G	vs. C0G	16.7 ± 3.8 a	7.7 ± 1.6 d	74.0 ± 4.9 a	18.3 ± 2.2 c
	C3G	vs. C0G	14.3 ± 3.2 a	6.3 ± 1.1 d	74.4 ± 4.5 a	19.3 ± 2.4 c
	C4G	vs. C0G	16.5 ± 4.0 a	6.3 ± 1.3 d	74.4 ± 5.0 a	19.3 ± 2.5 c
Final (t = 158 d)	C1	vs. C0	0.2 ± 0.1 d	70.0 ± 8.2 a	0.0 ± 0.0 c	30.0 ± 4.5 b
	C2	vs. C0	0.7 ± 0.2 cd	22.3 ± 3.5 b	27.6 ± 4.1 b	50.1 ± 6.4 a
	C3	vs. C0	5.6 ± 1.1 a	16.3 ± 3.4 c	35.5 ± 4.2 b	48.2 ± 5.5 a
	C4	vs. C0	2.4 ± 0.6 bc	20.7 ± 4.1 b	22.7 ± 3.9 b	56.6 ± 6.9 a
	C0G	vs. C0	0.4 ± 0.1 d	6.2 ± 1.1 b	3.1 ± 0.8 b	90.7 ± 4.8 a
	C1G	vs. C0G	2.7 ± 0.5 b	20.6 ± 3.2 b	35.5 ± 4.5 a	43.9 ± 5.1 a
	C2G	vs. C0G	3.2 ± 0.9 b	17.0 ± 2.8 c	31.8 ± 3.9 b	51.2 ± 6.2 a
	C3G	vs. C0G	3.2 ± 0.8 b	15.6 ± 2.5 b	27.8 ± 3.6 b	56.6 ± 7.1 a
	C4G	vs. C0G	0.2 ± 0.1 d	34.6 ± 5.5 a	34.6 ± 5.1 a	30.8 ± 4.3 a

Note: Values in bold indicate a visual difference detectable by the human eye (ΔE*_ab > 2.7).

. At t₀, all caffeic acid treatments showed a similar distribution, with approximately 21–22% contributions from lightness (%Cont.∆L*) and around 39% each from chromatic coordinates (redness %Cont.∆a* and yellowness %Cont.∆b*).

At intermediate stages, particularly in glucose-only cotreatments, the contribution of redness (∆a*) increased. By the end of the experiment, the contribution of ∆a* decreased at low concentrations, while ∆b* became more relevant in higher-dose treatments.

The temporal evolution of coordinate contributions is illustrated in Figure 3. Different patterns were observed depending on treatment, with higher caffeic acid concentrations showing more stable chromatic distributions over time compared to lower doses and glucose-only treatments.

4. Discussion

The results of this study provide a detailed evaluation of the potential and limitations of post-fermentative copigmentation under realistic wine conditions as a technological tool for color enhancement in red wines, particularly in those produced from grape cultivars characterized by moderate anthocyanin concentration and limited pigment stability, such as Listán Negro [14,15]. Unlike previous studies conducted in model or pre-fermentative systems, the results indicate that the effectiveness of this strategy is constrained by ethanol content and depends on achieving sufficient cofactor concentrations to obtain perceptible chromatic effects.

Although the primary objective of this study was to evaluate color enhancement through post-fermentative copigmentation, caffeic acid was selected as a representative hydroxycinnamic cofactor due to its broader role in wine phenolic chemistry. In addition to its involvement in copigmentation equilibria, such compounds may participate in redox reactions and phenolic evolution processes, although these aspects were not specifically addressed in the present work [35].

From a regulatory perspective, caffeic acid is a naturally occurring phenolic compound present in grapes and wines. However, the direct exogenous addition of phenolic compounds is subject to regulatory restrictions. Therefore, the concentrations evaluated in this study should be interpreted from a mechanistic standpoint, and any practical application would require compliance with existing regulations. Accordingly, the post-fermentative addition of pure caffeic acid should not be considered a standard or authorized winemaking practice, and its evaluation in this study is therefore limited to a controlled experimental context.

While the addition of pure hydroxycinnamic acids is not currently included among the oenological practices authorized for commercial winemaking, it is important to note that structurally related phenolic compounds are already introduced into wine through the use of approved oenological products. In particular, commercial oenological tannins, widely used for color stabilization and mouthfeel enhancement, include flavanols and hydroxycinnamic acid derivatives. In this context, the present results may inform the optimization of such products rather than for the direct use of pure caffeic acid.

From a toxicological perspective, caffeic acid and its derivatives are widely present in the human diet. Although no specific acceptable daily intake (ADI) has been established, dietary exposure has been documented. For instance, esters of caffeic acid are consumed in amounts between 500 and 1000 mg per day in regular coffee drinkers [36]. However, these values should not be directly extrapolated to justify the addition of caffeic acid to wine at the concentrations evaluated in this study. Any practical application would require a comprehensive safety assessment, including intake estimation and regulatory evaluation.

In relation to labeling, the addition of non-authorized compounds would require regulatory approval and appropriate declaration to ensure consumer transparency. Therefore, labeling and safety considerations should be regarded as prerequisites for any potential application of this approach.

The progressive attenuation of the copigmentation effect during storage may be explained by the instability of both anthocyanins and cofactors in hydroalcoholic media. Anthocyanins undergo degradation through oxidation, nucleophilic attack, and structural rearrangements, leading to less intensely colored or polymeric derivatives. Similarly, hydroxycinnamic acids may participate in oxidative or condensation reactions, reducing their availability as copigments over time. In addition, ethanol weakens hydrophobic interactions, contributing to the dissociation of non-covalent π–π stacking complexes. These combined effects are consistent with previous reports describing a decline in monomeric anthocyanins during wine aging [18].

4.1. Mechanism of Initial Color Enhancement

The immediate hyperchromic effect (3% to 16% increase) and the bathochromic shift (1 to 3 nm) observed after caffeic acid addition constitute spectrophotometric evidence of copigmentation complex formation. Such spectral changes are consistent with π–π stacking interactions between anthocyanins and planar cofactors [5,7].

At the experimental pH (3.54), anthocyanins are distributed among several interconvertible forms, with a substantial fraction existing as colorless hemiketal species [37]. The planar structure of caffeic acid facilitates π–π stacking interactions with the flavylium chromophore, reducing water accessibility to the C-2 position and shifting the equilibrium toward the colored form. Importantly, this supramolecular interaction increases visible intensity without increasing the relatively low anthocyanin concentration of the red wine used in this study (319 mg/L), suggesting that the chromatic enhancement is primarily structural in nature.

From an analytical perspective, this enhancement is captured by the increase in maximum absorbance (Abs_Max) and derived parameters such as %Color, as well as by global indices like CI and TCI, which reflect the hyperchromic effect. In contrast, these parameters do not provide information about the qualitative nature of the color change, which must be interpreted through CIELAB coordinates. Similar copigmentation effects have been reported in model systems and pre-fermentative conditions, where stronger and more persistent chromatic enhancements are typically observed due to the lower ethanol content and simpler matrix composition [38,39]. In comparison, the present results indicate a reduced magnitude and persistence of these effects under post-fermentative hydroalcoholic conditions.

The dose-dependent response observed here is consistent with association models previously described for anthocyanin–cofactor systems, where the magnitude of the hyperchromic effect increases with cofactor concentration until equilibrium saturation is approached [40]. However, the present results suggest that this relationship is not strictly linear across the entire concentration range, as the increase in color intensity tends to level off at higher doses, indicating the approach to a saturation regime in the copigmentation system.

4.2. Visual Impact and CIELAB Interpretation

From a perceptual standpoint, the initial ΔE_ab* values (20.25–26.99) indicate a marked visual change, exceeding established human perceptibility thresholds [32]. The decrease in Lightness (L*) indicates that copigmentation reduces transparency, producing a darker visual appearance. This visual response contrasts with the more moderate increases observed in spectrophotometric parameters, indicating that relatively small changes in absorbance can translate into larger perceptual differences in wine color.

The initial partitioning analysis revealed a balanced contribution of ΔL*, Δa*, and Δb* coordinates (≈22%, ≈39%, and ≈39%, respectively), indicating that post-fermentative copigmentation simultaneously modifies brightness and chromatic tone. This distribution suggests that the effect is not limited to darkening but also involves changes in chromatic coordinates. Similar patterns have been reported in copigmented wines, where stacking interactions are associated with darker and more purplish tonalities [16,39].

Accordingly, the initial color modification appears to result from combined changes in lightness and chromatic components (Δa* and Δb*), rather than from a single dominant parameter.

The maintenance of lower hue angle (H*) values in high-dose treatments may indicate a delayed shift toward yellow/orange tones typically associated with oxidative evolution. This behavior is consistent with previous observations that hydroxycinnamic cofactors can transiently stabilize red tonalities in young wines [9], although this effect appears to decrease over time as copigmentation interactions weaken.

4.3. Glucose Effect and Matrix Influence

The initial enhancement observed in glucose-supplemented red wine samples suggests a short-term synergistic interaction. In model systems, sugars may reinforce anthocyanin–cofactor complexes through extensive hydrogen bonding involving polyhydroxyl groups and by decreasing water activity, thereby limiting nucleophilic water attack on the flavylium cation [41,42,43]. The observed increase in relative color from 12.79% to 15.89% at the highest caffeic acid dose is consistent with this interpretation.

However, when glucose is added alone (C0G), the results reveal that it does not provide a stable improvement in color, but rather induces a variable and transient chromatic response. This behavior is evidenced by high initial ΔE*_ab values followed by a marked decrease during storage, indicating that glucose alone does not stabilize the anthocyanin system in hydroalcoholic media.

The progressive loss of the glucose effect over time suggests that such interactions are unstable in wine conditions. Ethanol is known to weaken hydrophobic interactions and reduce copigmentation constants by interfering with π–π stacking and decreasing solvent polarity [25]. In a red wine containing 13.4% (v/v) ethanol, the stability of weak sugar-mediated networks is likely reduced, which may explain the disappearance of the glucose-related effect by day 158. Thus, glucose appears to act as a transient facilitator of copigmentation rather than a stabilizing agent.

At the intermediate stage (t = 87 days), ΔE*_ab values remained relatively high despite a decline in absorbance. The partitioning analysis showed that up to 74% of the perceptual difference was driven by the Δa* (redness) coordinate. This observation suggests a partial decoupling between spectrophotometric intensity and perceptual color coordinates, suggesting that residual stabilization of the flavylium form persists even as overall absorbance decreases. Similar behavior has been reported in other copigmented systems [39].

Additionally, the trajectory observed for the glucose-only treatment suggests that glucose may influence the redox and colloidal balance of the system in the absence of effective copigments. This may contribute to the evolution toward less stable chromatic states during storage, although this effect cannot be directly confirmed from the present data.

4.4. Concentration Threshold and Negative Low-Dose Effect

A key result of this study is the identification of a concentration-dependent stability threshold. Although all treatments exhibited initial increases in color, only doses ≥ 480 mg of caffeic acid/L maintained perceptible chromatic differences after 158 days (ΔE*_ab = 5.59 for C3).

These results indicate that the relationship between caffeic acid concentration and color stabilization is not linear, but instead involves a minimum effective threshold, below which the copigmentation effect is not sustained over time.

At lower concentrations (120–240 mg/L), color values at the end of the experiment were similar to or lower than those of the control. This inversion suggests that sub-threshold concentrations do not provide sustained stabilization and may be associated with decreased color retention during storage.

This behavior may be explained by the dissociation of weak copigmentation complexes over time, leaving anthocyanins more exposed to hydration and oxidative degradation [2,44,45]. In addition, caffeic acid and its derivatives are readily oxidizable phenolics that may undergo autoxidation or redox cycling, particularly under low free SO₂ (13 mg/L at the beginning of the present study), potentially generating reactive intermediates that contribute to pigment degradation [45,46]. Under these conditions, caffeic acid at low concentrations may not effectively stabilize the system and could contribute indirectly to pigment degradation. However, this interpretation should be considered cautiously, as the present study does not include direct measurements of oxidative pathways.

These observations highlight the importance of achieving sufficient cofactor concentration to maintain copigmentation equilibria. From an applied perspective, insufficient dosing may result in limited or no improvement in color stability under post-fermentative conditions.

4.5. Evolution Toward More Stable Pigments

The progressive increase in the contribution of the Δb* coordinate in high-dose treatments (480 mg/L) may suggest a chromatic evolution toward more orange tonalities, which has been associated with the formation of more stable anthocyanin-derived pigments, such as pyranoanthocyanins [47,48,49]. In these treatments, the contribution of Δb* (yellowness) becomes dominant, exceeding 50% of the total color change in several samples. This behavior may reflect the gradual transformation of initially copigmented anthocyanins into more stable covalent derivatives, such as pyranoanthocyanins or tannin–anthocyanin condensation products, which accumulate during wine aging and are known to impart more orange or tawny hues as monomeric anthocyanins decline [47,48,49]. In parallel, the temporal evolution of color may also reflect the degradation of native anthocyanins through hydration, oxidation, and condensation reactions, leading to both loss of color intensity and formation of new pigment structures.

Pyranoanthocyanins exhibit enhanced resistance to hydration and SO₂ bleaching, as well as distinct spectral signatures (shifted λ_max and altered molar absorptivity) compared to native flavylium forms [50,51]. Accordingly, the late-stage color persistence observed in high-dose treatments may involve a partial transition from purely non-covalent copigmentation toward more stable covalently bound pigmented structures such as pyranoanthocyanins [52,53]. However, this interpretation should be considered cautiously, as the present study did not include direct chemical profiling (e.g., HPLC analysis) to confirm the formation or accumulation of these compounds. Therefore, this hypothesis is based on the observed spectrophotometric and CIELAB trends rather than on direct compositional evidence.

In this context, copigmentation may act as an initial stabilizing mechanism that maintains anthocyanins in their colored form, potentially increasing their participation in subsequent reactions, leading to more stable pigments. This could partly explain the greater persistence observed at 480 and 960 mg/L, although this transition cannot be confirmed from the present data alone. A limitation of this study is the absence of chromatographic characterization of individual phenolic compounds, which would be required to verify the formation and evolution of specific pigment species. Future work should include HPLC-based profiling to further elucidate the molecular mechanisms underlying the observed chromatic evolution.

4.6. Post- vs. Pre-Fermentative Efficiency

In contrast to the extensive body of literature focused on prefermentative or fermentative copigmentation strategies, fewer studies have addressed the modulation of wine color through post-fermentative interventions. The present results contribute to this emerging area by demonstrating that the addition of caffeic acid to a fully fermented wine matrix can still induce measurable, albeit moderate, improvements in color intensity and visual perception. This suggests that copigmentation remains operative in hydroalcoholic systems, albeit under less favorable conditions.

The maximum color increase observed in this study (~16%) is markedly lower than the 25–60% increases reported when caffeic acid is added during pre-fermentative stages [8,54]. This difference is likely related to the destabilizing effect of ethanol on non-covalent copigmentation equilibria [25], as well as to the reduced opportunity for copigmented anthocyanins to participate in subsequent polymerization reactions under hydroalcoholic conditions [55,56].

Additionally, the absence of a continuous extraction phase in post-fermentative systems limits the availability of both anthocyanins and complementary phenolics, further constraining the formation of stable pigment networks. In aqueous must systems, copigmentation may act as a precursor stage facilitating the formation of stable pigmented polymers during fermentation [2,57]. In contrast, post-fermentative addition occurs after this stage, limiting the integration of added cofactors into long-term pigment structures [1].

Accordingly, post-fermentative copigmentation may be better understood as a corrective strategy, rather than a structural modification of the wine matrix. In the case of Listán Negro wines, characterized by moderate anthocyanin concentration and limited chromatic stability, caffeic acid addition provides a measurable short-term enhancement of color. However, long-term stabilization is concentration-dependent and remains less effective than prefermentative strategies.

Despite these limitations, post-fermentative interventions offer practical flexibility, as they allow modification of wine color after fermentation without altering the vinification process. This may be particularly relevant for wines already produced with suboptimal color. Nevertheless, it should be noted that the experimental conditions of this study were controlled and may not fully represent the variability of commercial winemaking environments.

4.7. Comparison with Recent Post-Fermentative Approaches

Recent studies applying post-fermentative strategies based on phenolic enrichment provide a useful framework to contextualize these findings. For instance, post-fermentative maceration with overripe grape seeds has been shown to enhance chromatic stability, leading to wines with lower lightness, higher chroma, and visually perceptible differences compared to untreated controls [58]. Similarly, the combined post-fermentative addition of ripe and overripe seeds further improved color stability and intensity, particularly when overripe seeds were used, an effect associated with increased formation of polymeric pigments [59].

These approaches differ mechanistically from the direct addition of a pure copigment such as caffeic acid. While seed-derived treatments introduce a complex mixture of flavan-3-ols, phenolic acids, and tannins capable of promoting both copigmentation and subsequent polymerization reactions, the present study isolates the effect of a single hydroxycinnamic acid. This difference may account for the lower magnitude and persistence of the chromatic changes observed here. The initial increase in color suggests that copigmentation interactions between caffeic acid and anthocyanins are rapidly established, but their persistence over time appears limited in the absence of additional pathways leading to more stable pigment formation.

Accordingly, the present results suggest that copigmentation alone, when applied post-fermentatively, may not be sufficient to ensure long-term chromatic stability. This interpretation is consistent with model wine studies showing that caffeic acid can promote short-term copigmentation, whereas long-term color evolution depends on the formation of more stable derived pigments, such as pyranoanthocyanins, which are influenced by the overall phenolic composition [9]. In finished wine matrices, where anthocyanin concentration is already reduced and the system is closer to equilibrium, the capacity of a single copigment to induce sustained chromatic changes may therefore be limited.

Recent studies have also highlighted that the timing of phenolic addition plays a critical role in determining its effectiveness [10,23]. Interventions applied before or during fermentation benefit from combined effects on anthocyanin extraction, copigmentation, and early polymerization processes, whereas post-fermentative additions act on an already established matrix and are therefore primarily corrective rather than structurally transformative.

Within this context, the moderate increases in color intensity observed in this study (up to ~16%) are consistent with the expected response of a system in which copigmentation is not coupled to subsequent pigment stabilization pathways.

Overall, the present findings are consistent with previous reports indicating that post-fermentative strategies can be used to adjust wine color, but their effectiveness is generally lower than that of earlier interventions. Nevertheless, their practical relevance lies in their flexibility, as they allow modification of chromatic properties after fermentation without altering the vinification process. In this framework, the identification of a dose-dependent response and a minimum effective concentration for caffeic acid represents a relevant contribution to the understanding of post-fermentative copigmentation behavior.

5. Conclusions

Post-fermentative addition of caffeic acid to young, lightly colored red wines resulted in an immediate, dose-dependent hyperchromic effect and a bathochromic shift, consistent with the rapid formation of anthocyanin–cofactor stacking complexes in hydroalcoholic media. The visual impact was substantial, with pronounced decreases in lightness (L*) and CIELAB shifts exceeding commonly accepted perceptibility thresholds.

This chromatic enhancement was primarily reflected by spectrophotometric parameters (Abs_Max, %Color, CI, TCI), while CIELAB coordinates revealed that the effect involved both an increase in color intensity and modifications in hue and saturation.

Glucose produced a short-term enhancement of the copigmentation response, but this effect was transient and diminished during storage, indicating that hydrogen-bond-mediated stabilization is not sustained in ethanol-containing wine matrices. When added alone, glucose did not improve long-term color stability and showed a variable and unstable chromatic evolution.

Although all doses initially increased color intensity, a minimum effective concentration (≥480 mg/L) was required to preserve perceptible chromatic differences after 158 days. Lower doses did not provide long-term stabilization and, in some cases, resulted in final color intensities below the untreated control, suggesting that sub-threshold cofactor levels may be associated with increased pigment degradation under low SO₂ conditions.

These results indicate that post-fermentative copigmentation acts as a short-term mechanism, whose persistence depends on achieving a sufficient cofactor concentration and may be linked to the subsequent formation of more stable pigment structures.

Post-fermentative copigmentation with caffeic acid provides a short-term corrective enhancement of visual appearance and may delay hue evolution toward yellowish tones. However, its technological efficiency is lower than that of prefermentative applications, where copigments are introduced prior to the establishment of the alcoholic matrix and more effectively promote the incorporation of anthocyanins into stable polymeric structures.

Therefore, post-fermentative caffeic acid addition may be regarded as a practical approach for improving the initial appearance of low-anthocyanin wines such as Listán Negro. However, long-term chromatic stabilization is more likely to require early-stage interventions that combine copigmentation with subsequent polymer formation. The identification of a minimum effective concentration represents a relevant outcome of this study for understanding post-fermentative color modulation.

Future work should address the combined influence of cofactor concentration, oxygen exposure, and SO₂ levels to better understand their role in post-fermentative color evolution in lightly pigmented red wines.