Quantifying Hair Color Fading

The essence of the cosmetic industry involves an ability to transform one’s appearance—and there are few changes in hair care as dramatic as hair color. An old adage suggests that brunettes would rather be blonde, and blondes would rather be brunette; however, fashion concerns may give way to other anxieties with age, such as the ability to mask undesirable grey hairs. By far, the largest segment of the hair color market is permanent color products, which produce long-lasting changes that persist throughout everyday habits and practices. The efficacy of these products begins with small, water soluble dye-precursors diffusing into hair fibers, wherein they are coaxed into reactions that produce larger pigmented molecules.1 In other words, hair serves as a reaction vessel in which this chemistry takes place; and the larger size of these dye molecules inhibits their diffusion from the fibers to yield an enduring effect.

However, as Figure 1 shows, some color bleeding does occur with subsequent washing, allowing the original dyed color to progressively fade with time. Of course, new hair growth necessitates color treatments every 4-6 weeks, which refreshes the original treatment, but there can be general dissatisfaction with the true “permanence” of the desired end color. Around 10-12 years ago, shampoos and conditioners began to appear on the market with claims specifically addressing this issue. They bore statements of being “color-safe” or “helping to preserve color.” This positioning strongly resonated with consumers and led to a deluge of similarly positioned products. In relation, the present article discusses the various contributors to the color fade process, with a specific emphasis on the use of measurements to quantify and understand the problem.

Natural Hair Color Changes

The natural color of hair is produced by the presence of melanin pigment granules in the inner cortical structure. This native color is a function of both the type and concentration of melanin present. Increasing levels of eumelanin produce blonde, brown or black hair, respectively; while pheomelanin contributes red coloration. The inability for hair follicles to continue manufacturing melanin leads to unpigmented, i.e., grey hair. Therefore, efforts aimed at darkening hair involve the use of dyes to overwhelm the innate contribution of melanin.

Conversely, the reverse scenario involving hair lightening must first destroy the natural pigment. This “bleaching” procedure is commonly performed by treating hair with oxidizing agents such as hydrogen peroxide, although melanin also can be broken down by photo-chemical oxidation as a result of prolonged exposure to sunlight. The more-energetic nature of UV irradiation is often presumed to be the culprit of this sun-lightening, although an overwhelming number of scientific publications point the finger at the visible portion of the spectrum as the major contributor to this phenomenon.2

UV irradiation is known to break down hair proteins in a variety of ways,3 one of which involves degradation of tryptophan to form kynurenines.4 This process is thought to be the origin of undesirable photo-yellowing in blonde and grey hair.

Measuring Changes in Hair Color

Perhaps the most common means of quantifying color involves the CIELAB system, which is used by many commercially available colorimeters. In this approach, color is measured in three dimensional space in terms of L*, a* and b* coordinates (see Figure 2), where L* provides a measure of lightness on a scale of 0 to 100, a* denotes the red-green color range (positive value denotes higher red) and b* represents the yellow-blue color range (positive value denotes higher yellow). As such, changes in color from an initial state can be represented in terms of ΔL, Δa and Δb. It is also common to see an overall color change (ΔE) reported, as calculated by Equation 1:

ΔE = √ (ΔL2 + Δa2 + Δb2) Eq. 1

Sometimes one may also encounter a change in Chroma (ΔC), which ignores the contribution of the lightness/darkness. This can be calculated using Equation 2:

ΔC = √ (Δa2 + Δb2) Eq. 2

Experimental Details

A recurring theme in this series of articles on hair testing5-8 has been how product treatment likely represents the greatest source of error/variability in any laboratory test. Such error may occur two-fold in experiments to evaluate washout color fading, where it is first necessary to reproducibly dye a suitable number of hair tresses and then perform repeated washing cycles under equally controlled  conditions.

Commercial permanent color products contain two solutions, the dye and a developer, which are mixed to initiate the aforementioned reaction process. After this step, the composition of the resulting blend will constantly change as small dye precursors progressively react to become larger dye molecules. The dyeing performance can therefore vary considerably depending on the duration between product mixing and actual application to hair. For this reason, it is imperative to apply the freshly mixed dye/developer blend to hair tresses as quickly as possible to obtain consistent dyeing results.

This can be accomplished, in part, by dyeing hair in small, manageable sub-batches, i.e., 10-15 tresses each, using multiple samples of the desired product/formulae as opposed to trying to treat all the tresses in a single large batch. At the same time, multiple pairs of hands will give rise to a faster, more controlled dyeing process. For example, a three-person, tag-team approach is recommended, wherein the first person uses a syringe to quickly measure and apply a consistent amount of dye to the hair; the second uses a stylist brush to rapidly distribute product in a relatively even manner; and the third then hangs the tresses and provides additional periodic massaging to ensure uniform dispersal and optimal results.

Color Washout Protocol

After the appropriate period of development time, color must be removed from hair—also using a suitably controlled and reproducible rinsing protocol. The quality of the entire dyeing process will be reflected in the standard deviation of the L*, a* and b* values of freshly prepared tresses. A quality protocol should yield standard deviations in the range of 1-2 color units. Accordingly, a suitable number of replicate tresses, i.e., 8-10, are necessary for each color fade test cell to provide appropriate statistical rigor.

A highly controlled washing and rinsing protocol involves wetting tresses in an exact way, followed by syringe application of a precise amount of product. The hair is massaged for a given time in a manner that approximates real-life usage conditions. The product is then removed by rinsing tresses for an exact duration using water having a controlled temperature and flow rate.

Color Fade Culprits

It is widely believed by consumers that shampoos are capable of stripping color from permanently dyed hair. However, technical consideration of this presumption leads to some inconsistencies. The cleansing mechanism of shampoos relies on the use of surfactants to form micelles that are able to solubilize hydrophobic soils from the hair surface. Yet, the permanent dye molecules reside within the hair and are not exposed to this same environment. It therefore seems unlikely that surfactant molecules penetrate this deeply into hair, and certainly not in a manner that allows micelles to be formed.

In relation, Figure 3 shows color fade results comparing hair washed with a full strength shampoo, a half-strength shampoo and a water control. Results indicate that all three treatments gave rise to comparable levels of color fade. It therefore can be concluded that water is the cause of this color fade, and that shampoos themselves do not induce additional detrimental effects. Sharing this observation with colleagues throughout the industry has produced many echoings of this conclusion.

However, there are other external factors that change the color of permanently dyed hair. Perhaps the most obvious offender is sunlight, which can photochemically degrade dye molecules as well as natural melanin. Figure 4 shows the color change results (ΔE) for permanently dyed hair as a function of irradiation time within a commercial accelerated weathering chamber. Similarly, Figure 5 shows how dye molecules can be thermally degraded by the extreme temperatures encountered during heat styling by straightening and curling irons.

A further major contributing factor to color fade is the condition of the hair itself. The degree to which hair swells in water is well-recognized to significantly increase as the internal structure of hair becomes damaged. Accordingly, diffusion both into and out of hair is facilitated. Figure 6 shows washout color fading results in which the only variable is the initial state of the hair. As can be seen, color fade is relatively slow for hair in the virgin state prior to dyeing, but effects are considerably magnified in damaged hair.


Unquestionably, color fading of permanently dyed hair is an issue for consumers, yet there are different mechanisms by which this process can occur. Dye molecules can leach from the hair during washing, be photo-degraded by sunlight, and/or thermally degraded by heat styling devices. The relative contributions of these factors will depend on the habits and practices of the individual but it seems likely that the consumer’s hair condition also plays a role. As has been shown, wash fading of dyes is significantly exacerbated in damaged hair, presumably due to increased fiber swelling, which leads to enhanced diffusion.

It also seems likely that the relative contribution of these factors will vary with time. For example, washout leaching occurs more readily in freshly dyed hair but declines with increasing wash cycles (see Figure 1). Therefore, any color change that occurs 1-2 weeks after a dye treatment is more likely the result of contributions from the effects of sun and heat than other sources. It also is not unreasonable to suspect that the relative contribution of these factors may be shade-dependent. For example, there is a well-entrenched belief that red coloration is especially prone to fading. Clearly, dye molecules of different sizes will diffuse differently. Their differing adsorption spectra will dictate variations in interactions with sunlight, while different decomposition temperatures will yield dissimilar responses to heat.

The color fading of hair is clearly a convoluted process but with the origins of the problem now well-understood, cosmetic chemists possesses the knowledge to develop the appropriate strategies to address this issue.


  1. JM Marsh, Hair Coloring, in Practical Modern Hair Science, T Evans and RR Wickett, eds, Allured Books, Carol Stream, IL USA (2012)
  2. E Hoting, M Zimmermann and H Hocker, Photochemical alterations in human hair. II. Analysis of melanin,J Cosmet Sci 46 181-190 (1995)
  3. KR Millington, Photoyellowing of wool. Part 2: Photoyellowing mechanisms and methods of prevention, Color Technol 122, 301-316 (2006)
  4. J Jackowicz and RL McMullen, Tryptophan fluorescence in hair–Examination of contributing factors, J Cosmet Sci 62 291-304 (2011)
  5. TA Evans, Measuring hair strength—Part 1. Stress-strain curves, Cosm & Toil 128(8) 590–594 (Aug 2014)
  6. TA Evans, Measuring hair strength—Part 2. Fiber breakage, Cosm & Toil 128(12) 854–859 (Dec 2013)
  7. TA Evans, Evaluating hair conditioning with instrumental combing, Cosm & Toil 126(8) 558–563 (Aug 2011)
  8. TA Evans, Measuring the water content of hair, Cosm & Toil 129(2) 64-69 (Mar 2014)
More in Efficacy