From coffee industry waste materials to skin-friendly products with improved skin fat levels

Abstract

Spent coffee grounds (SCG), which are the residue obtained from the treatment of coffee with hot water or steam, can be used for industrial applications, due to the high content in lipids. The cosmetic products might be a suitable application for these types of residues because the barrier properties of the stratum corneum (SC) are largely dependent on the intactness of the lipid lamellae that surrounds the corneocytes. The purpose of this work was to assess the feasibility of using the lipid fraction of SCG extracted with supercritical carbon dioxide in the development of new cosmetic formulations with improved skin lipids (sebum) and hydration. The use of spent coffee lipid extract in cosmetic industry seems to be a suitable approach to recycle the wastes from coffee industry. Emulsion containing 10% of the lipid fraction of SCG (SpentCofOil cream) presented promising characteristics in the improvement of sebum skin levels with a good acceptance by consumers when compared to an emulsion containing 10% w/w of green coffee oil (GreenCofOil cream) and a placebo without coffee oil (NoCofOil cream).

Practical applications: In this work, the authors develop and characterize a cream containing 10% of the lipid fraction of SCG extracted with supercritical carbon dioxide with improved skin lipids (sebum) and hydration.

Abbreviations:

GC, gas chromatography; O/W, oil-in-water; SC, stratum corneum; SC-CO₂, supercritical carbon dioxide; SCG, Spent coffee grounds; TEWL, transepidermal water loss

Introduction

In a country like Portugal, where coffee consumption is so deeply rooted in the cultural habits of the population, the total import of coffee is increasing within time and it reached almost 55 thousand tons in 2009 1. Spent coffee grounds (SCG), which are the residue obtained from the treatment of coffee with hot water or steam for extracting flavor substances therefrom, can be used for industrial applications such as to produce high quality biodiesel 2, 3. Due to their high content in carbohydrates, lipids, and proteins 4, the cosmetic products might be a suitable application for these types of residues.

Brazil is the largest coffee producer in the world 5. On average, a fifth of the Brazilian coffee production consists of defective beans which after roasting process decrease the final beverage quality 6. Several studies have been developed in order to find an alternative use for these defective coffee beans. One of the alternatives being considered is the cosmetic application of the oil extracted from the defective beans.

The human stratum corneum (SC) consists of several layers of keratinized corneocytes embedded in a lipid matrix of ordered lamellar structure. SC intercellular lipids particular composition allows a highly ordered arrangement of lipids playing an essential role in keeping an optimal skin barrier and in regulating the skin hydration. Deficiency of ceramides, cholesterol, essential fatty acids, and triglycerides leads to enhanced transepidermal water transport in addition to dryness of the skin, i.e., xerosis 7. The etiology of dry skin is variable but is often related to skin disorders and changes in lipid composition, particularly ceramides, as they are the major lipid constituent of lamellar sheets, followed by cholesterol, fatty acids, and triglycerides. A correct lipid ratio in SC is necessary to maintain the lipidic lamellae 8, 9. Ceramides play an important role and with cholesterol, free fatty acids, and cholesterol sulfate, which are ionized at physiological pH, they form ordered structures.

Traditionally, dry skin treatments were based on rehydration of the epidermis, through the use of emollients and/or an occlusive emulsion. Emollients or moisturizers are often used in the treatment of xerosis with the aim of improving skin hydration and sebum levels 8, 10, 11.

The purpose of this work was to assess the feasibility of using the lipid fraction of SCG in the development of cosmetic formulations (oil-in-water (O/W) creams) with improved skin hydration and sebum capacity. Their physicochemical characterization, stability, biological effects, and sensory acceptability were evaluated and compared with formulations containing green coffee oil and no coffee oil.

Materials and methods

Materials

Tego®Care 450, Abil® 350, Cutina® GMS, and MultiEx Naturotics® and propylenoglycol were obtained from Evonik Industries AG, Cognis Group and DS, Produtos Químicos and Sigma (Sigma, St. Louis, MO, USA).

Green coffee oil was supplied by COOXUPÉ – Cooperativa de Cafeicultores de Gauxupé, (in Guaxupé, Minas Gerais), Brazil. The green coffee beans were pressed and come back to be pressed again for a further three times and then the oil is bottled with nitrogen.

SCG were supplied by NovaDelta – Comércio e Indústria de Cafés, S.A. (Campo Maior, Portugal).

Methods

Spent coffee lipid extraction

The SCG were dried in an oven at 378 K for 12 h with air circulation to remove moisture. The residual moisture content of dried SCG was measured by Karl Fischer titration and had an average value of 1% w/w.

The lipid fraction of SCG was extracted with supercritical carbon dioxide (SC-CO₂). Conventional oil extraction from food crops and agro-industrial residues involves hazardous organic solvents, like n-hexane. SC-CO₂ technology allows an environmentally friendly process whereby extraction/separate recovery of oil and bioactive substances from biomass can be done without degradation 12, 13. Through manipulation of temperature and pressure, the density of SC-CO₂ is adjusted to allow complete separation of oil and bioactive solutes that the matrix may contain. Recently, it has been demonstrated the feasibility of SGC oil extraction by SC-CO₂ 2.

The spent coffee oil was extracted in a high pressure extraction plant shown schematically in Fig. 1. Gaseous carbon dioxide (Air Liquide, Lisboa, Portugal) is first compressed to the desired extraction pressure by means of a gas compressor (Burton Corblin, A0C 400, 0963) and then heated to the desired temperature by passing through a tube-in-tube coil heat exchanger. Supercritical carbon dioxide flows at the desired pressure and temperature conditions upwards through a packed bed of SCG held in a basket placed previously inside the extraction vessel (316SS; internal diameter 6.4 cm; total length 59.6 cm). The extraction pressure is controlled by an electro-pneumatic control valve (Von rohr armaturen AG, VEGP700 F59, Muttenz, Switzerland) where depressurization of CO₂ flow stream exiting the extraction vessel takes place. The extracted oil is separated from the gas stream in two cyclones (Separex 4140/CY01 AS2, Champigneulles, France) connected in series and the regenerated CO₂ is recycled back to the compressor.

Figure 1. High pressure plant used for spent coffee oil extraction.

SC-CO₂ extractions were performed using 0.5 kg of dry SCG per batch at 55°C and 250 bar with an average CO₂ flowrate of 15 kg/h for ca. 1 h of extraction. The collected oil was stored in an autoclaved and light protected flask at −20°C in a freezer for subsequent analysis.

Coffee oil analyses

Free fatty acids

The free fatty acid content of the coffee oils was determined by titration. One gram of oil was added to 50 mL of solvent mixture (ethanol/diethyl ether 1:1, v/v, Panreac, 99.5%: José M. Vaz Pereira S.A., Lisboa, Portugal) previously neutralized. Potassium hydroxide 0.1M (PRONOLAB, José M. Vaz Pereira S.A., Lisboa, Portugal) in ethanol was added until the solution turned from yellow to pink. Phenolphthalein was used as a pH indicator.

Unsaponifiable lipids

Unsaponifiable lipids were determined following the AOCS Official Method 14. Their content was expressed as weight percent.

Fatty acid analysis

The fatty acid composition of the coffee oils was determined by direct transesterification of the lipids to the corresponding methyl esters and then quantitatively analyzed by gas chromatography (GC), according with the method of Lepage and Roy 15 with modifications. Ten micrograms of oil extract was transmethylated with 2 mL of methanol:acetyl chloride (95:5 v/v). The mixture was sealed in a Teflon-lined vial under nitrogen atmosphere and heated at 80°C for 1 h. The vial contents were then cooled, diluted with 1 mL water, and extracted with 2 mL of n-hexane. The organic layer was dried over Na₂SO4.

Methyl esters were quantitatively analyzed by GC in a Thermo Trace GC ULTRA (Thermo Unicam, Lisboa, Portugal), equipped with a flame ionization detector (FID) and a split/splitless injector. The injector and detector temperatures were set at 280°C. The split flow was set at 134 mL/min. The analytical column was a TR-Biodiesel (F) from Thermo Unicam (Lisboa, Portugal). 0.5 µL sample volume was injected by means of an automatic injector TriPlus. Helium was used as carrier at a constant flow of 2 mL/min. The oven temperature program was 230°C for 13 min. Peak identification was carried out using known standards (FAME mixture C4–C24, Sigma–Aldrich, Sintra, Portugal). Methyl-heptadecanoate (Nu-Check-Prep, Elysian, USA) was used as the internal standard (IS). The data were processed with the Chrom-Card software.

pH determination

The pH was measured using a pH meter, Crison, model micropH 2002 (Barcelona, Spain) with a glass electrode (Eutech, USA), specially used for food oils.

Preparation of O/W emulsions containing coffee lipid extract

Non-ionic O/W creams containing 10% w/w of coffee lipids were prepared: SpentCofOil cream with 10% w/w of spent coffee lipid extract, GreenCofOil cream with 10% w/w of green coffee oil and NoCofOil cream without coffee lipids.

To prepare the creams, the oily (Tego®Care 450, Abil® 350, Cutina® GMS, and MultiEx Naturotics®) and aqueous phases (purified water and propyleneglycol) were heated separately until reach 75°C, then the oily phase was added to the aqueous phase and the system was mixed (Helipath® 130 rpm) with constant agitation until the temperature reached 30°C.

Physical characterization of the emulsions

Macroscopic organoleptic characteristics were analyzed. The pH was controlled using a pH meter Metrohm® pH Meter 744, glass electrode. The apparent viscosity and rheological profile were evaluated using the Brookfield® RV DV-II, SSA, spindleSC4-27.

The size distribution was measured by light scattering using a Malvern Mastersizer 2000 coupled with a Hydro S accessory. For a correct turbidity, about 0.5 g of each formulation, corresponding to an obscuration between 25 and 28%, was added in the sample chamber containing 150 mL of water and analyzed using a stirrer of 700 rpm.

Biological effects

The epidermal capacitance, transepidermal water loss (TEWL) and skin surface lipids of the SC for all creams were evaluated with a Corneometer CM 820, a Tewameter TM210, and a Sebometer SM 810 (C + K Electronics GmbH) on a panel of uniform volunteers for 28 days. The volunteers had given their written and informed consent before the evaluation (n = 10, young healthy females – 18–25 y.a., same professional activity) and included in the study after written and informed consent. The formulations were applied in the forearm and the results were compared with a defined control area (anatomically equivalent) on the same forearm. Measurements were performed under standardized conditions, at room temperature 16, 17.

Sensorial analysis

To evaluate the acceptability of all formulations, a questionnaire was answered by each volunteer, by the assessment of sensory attributes such as texture, odor, after feel, greasiness, hydration, spreadability, and tackiness using a hedonic scale from 1 to 5.

Statistical analysis

Data were compared using a two-way ANOVA (95% confidence level). Results are expressed as mean ± SD. Statistical significance was considered at a probability level of p<0.05.

Results

Spent coffee lipid extraction

Supercritical carbon dioxide extraction of SCG obtained a yellow to yellow/brown viscous oil extract, mainly collected in the first cyclone, C1, of the high pressure extraction plant (see Fig. 1). The extraction yield of spent coffee oil at 250 bar and 50°C was 12.1% on a dry weight basis (i.e., % g-oil/100 g-dry spent coffee). Comparatively, Arya and Rao 4 have reported an oil yield of 13.4% from SCG by soxhlet extraction with n-hexane as the solvent.

Coffee oils analyses

The analyses to the coffee oils used in this work are given in Tables 1 and 2. The results are comparable with previous results published in the literature 6, 18. Palmitic (C16:0) and linoleic (C18:2) were the major fatty acids detected in both oil samples comprising almost 80% of all fatty acids. Spent coffee lipid extract showed a pH of 3 while the green coffee oil showed a pH of 6.

Table 1. Free fatty acids and unsaponifiable lipids of spent coffee oil and green coffee oil

Compound	Spent coffee oil	Green coffee oil
Free fatty acids, % w/w	1.6	1.4
Unsaponifiable matter, % w/w	5.5	11.3

Table 2. Fatty acid composition of spent coffee oil and green coffee oil

Fatty acid, % w/w	Spent coffee oil	Green coffee oil
Fatty acids are designated by the number of carbon atoms: number of double bonds.
Palmitic acid (16:0)	33.1	37.1
Stearic acid (18:0)	9.1	7.9
Oleic acid (18:1)	12.2	9.2
Linoleic acid (18:2)	44.7	44.5
Linolenic acid (18:3)	0.8	1.2

The acidity of the lipidic fraction collected from green coffee beans or from their consequent residues – SCG – is due not only to free fatty acids but also to phenolic acids, nonvolatile aliphatic acids (e.g., malic and quinic acids), and volatile carboxylic acids (e.g., acetic, propanoic, butanoic acids, etc.). Coffee origins and species, growth conditions, processing method, roasting degree, and beverage extraction type influence the brew acidity, affecting the coffee's aroma and flavor. In our case, two different raw materials were employed – green coffee beans and SCG – from different origins and growth conditions as well as two different extraction processes – pressing and supercritical CO₂ extraction, explaining the different pH of the oils obtained.

Physico-chemical characterization of obtained creams

Concerning the macroscopic properties, both creams appeared yellow, glossy, and as semi-mobile emulsions. The results (pH values, apparent viscosity, and particle size) are shown on Table 3.

Table 3. pH, apparent viscosity, and particle size values for the emulsions

	pH	Apparent viscosity (Pa s)	Particle size (µm)
Mean ± SD (n = 3).
NoCofOil cream	6.37 ± 0.01	30.1 × 10^3 ± 5.25	27.54 µm ± 1.78
GreenCofOil cream	6.63 ± 0.01	40.2 × 10^3 ± 4.20	36.98 µm ± 2.53
SpentCofOil cream	4.89 ± 0.02	48.2 × 10^3 ± 4.56	22.99 µm ± 1.96

The pH data showed that the incorporation of the spent coffee oil decreased the pH value to 4.89 ± 0.02 (SpentCofOil cream). The pH of the other creams were 6.37 ± 0.01 (NoCofOil cream) and 6.63 ± 0.01 to GreenCofOil cream. These creams needed an adjustment in pH with orthophosphoric acid to achieve pH values suitable for skin application (between 4.5 and 5.5). This is due to the different pH of the lipids.

Continuous shear experiments measure the ability of each system to resist structural breakdown during the standardized shearing procedure.

Since most fluids are non-Newtonian, nonlinear models are needed to describe the change in viscosity with shear rate or shear stress.

Flow curves are shown in Fig. 2 for all creams. When the apparent viscosity decreases with the increase in shear rate, the fluid is considered pseudoplastic or shear thinning. The samples also showed a loop of anti-thixotropic or rheopetic.

Figure 2. Rheograms of NoCofOil cream (♦), GreenCofOil cream (x), and SpentCofOil cream (▴) at 25°C.

The apparent viscosity calculated in the apex of the curve is 48.2 × 10³ ± 4.56, 40.2 × 10³ ± 4.20, and 30.1 × 10³ ± 5.25 Pa s, respectively. Thus the inclusion of coffee oil contributes to obtain a product with improved stability.

Droplet size distribution showed a monomodal population for NoCofOil cream and SpentCofOilcream (Fig. 3). The mean particle size was: 27.54 µm ± 1.78 to NoCofOil cream, 22.99 µm ± 1.96 to SpentCofOil cream, and 36.98 µm ± 2.53 to GreenCofOil cream.

Figure 3. Droplet size distribution of NoCofOil cream (gray line) GreenCofOil cream (dotted line), and SpentCofOil cream (black).

Biological effects

The lipid components were chosen taking into consideration the lipid composition of SC. Spent coffee lipid extract and green coffee oil were added because they are mainly composed by triglycerides, essentially made of linoleic acid, palmitic acid and oleic acid, and residual antioxidants. These substances naturally occur in the skin lipid matrix.

The effect of the formulations on the skin hydration and on the maintenance of the intactness of the lipid lamellae was studied in ten selected volunteers. After a 28-day application, it was shown that these creams were nonirritating to the skin, as it was expected due to the selected components. In order to avoid the inter-individual variations in the SC hydration levels, the baseline value (control area) was always registered as an internal control of the experiment.

The results obtained showed that SpentCofOil and GreenCofOil creams were significantly different (p<0.01) from the control area, indicating that there was an effect on epidermal capacitance, TEWL and on the skin sebometry after 28 days (Figs. 4–6). The NoCofOil cream showed the lowest epidermal capacitance and sebum and the highest TEWL when compared to the formulations containing coffee oils.

Figure 4. Values of epidermal capacitance during 28 days. (mean ± SD, n = 10).

Figure 5. Values of TEWL during 28 days (mean ± SD, n = 10).

Figure 6. Comparison of skin surface lipids between control, NoCofOil cream GreenCofOil cream and spentCofOil cream.

Sensorial analysis

The results obtained from the sensory evaluation (Fig. 7) done by volunteers showed that both creams met consumer appeal and acceptance requirements. However, in future work, the odor should be improved as volunteers gave low scores (±50%) to products tested with coffee oils.

Figure 7. Sensorial profiles of NoCofOil cream, GreenCofOil cream, and SpentCofOil cream.

The SpentCofOil cream presents the highest score for the texture, greasiness, and skin feel on application, with a low score for the odor.

Discussion

The vehicle used to deliver the topical lipids was carefully selected, not only because it played an important role in the product efficacy and stability, but also because topical cream formulations must provide easy application on the skin surface without leaving any oil residue. The rationale behind this study was the physicochemical characterization of creams containing different lipids and simultaneously assessing their biological effects. In addition to the traditional emollients and occlusive constituents we intended to provide spent coffee oil, which is the residue obtained from the treatment of coffee.

The lipid fraction of SCG was extracted with supercritical carbon dioxide, an environmentally friendly solvent that allows the extraction and recovery of the oil at such conditions that no degradation of the lipid composition of the oil may occur.

All the formulations developed within this work were suitable to skin application: acidic pH and pseudoplastic or shear thinning behavior. This behavior occurs because entities in the fluid referred to as “flocs” tend to disassemble or assemble when stress is applied.

Apparent viscosity values provide a comparison of the resistance to structural breakdown between the emulsions and the loop areas compare the amount of structure that fractures in the standardized cycle. The inclusion of spent and green coffee oil seems to slightly increase the resistance to structural breakdown when compared with placebo (NoCofOil cream).

The emulsion maintained the mean particle size after oil coffee inclusion indicating that the oil did not destabilize the emulsion structure. Moreover, with the inclusion of the oil the population becomes thinner, i.e., with a more homogeneous population. However the inclusion of green coffee oil seems to slightly destabilize the original emulsion structure since we obtained a bi modal population with higher particle size. As known the emulsion stability is related to its particle size and distribution. Smaller the particle size, higher the stability. The results obtained show that the SpentCofOil cream present lower values when compared to the other formulations, especially with the GreenCofOil cream suggesting a better stability.

In general, there is a correlation between SC hydration and TEWL values, as lower TEWL (intact epidermal barrier function) corresponds to normal hydration state of the horny skin layer 19.

Estimation of the dynamics in SC hydration is used in efficacy claim studies on topically applied potentially hydrating (moisturizing) agents 20, 21. The increase in SC hydration is correlated with the improvement of the skin barrier function (lowering TEWL) 22. In this study it was observed that the SpentCofOil and GreenCofOil creams increased the epidermal capacitance and lowered the TEWL suggesting skin hydration.

Being a major component of this superficial layer, sebum lipids take part in the non-specific protective mechanisms of the skin barrier. Sebum is produced by the sebaceous glands (with higher density on the forehead, chest, and back) and consists predominantly of triglycerides, wax esters, and squalene 23. The role of sebum for the epidermal barrier was demonstrated in asebia mice (with profound sebaceous gland hypoplasia) 24. Despite the unaffected permeability barrier, asebia mice displayed epidermal hyperplasia, inflammation, and decreased (>50%) SC hydration, associated with a reduction in sebaceous gland lipids. The barrier abnormalities were attributed to the insufficient glycerol levels, derived from the triglyceride hydrolysis.

Elias and coworkers 25 showed that application of exogenous selected lipid mixtures optimized barrier repair in murine skin. In our study we used creams containing sebum-like lipids (SpentCofOil and GreenCofOil creams). When applied in human volunteers they had a restored effect as well as an increase in the sebum levels when compared with the cream with no coffee oil (NoCofOil cream) and with the control area as shown on Fig. 6.

Therefore, both creams (SpentCofOil and GreenCofOil creams) significantly increased skin hydration and sebum levels suggesting that the barrier properties of the skin were increased, probably by supplementation of skin lipids.

It is clearly shown that greasiness and skin feel on application were the attributes with the most significant difference whereas hydration and tackiness were the ones with the least difference among them. These observations are in accordance with the results obtained so far in epidermal capacitance, TEWL, and sebometry.

Conclusion

The use of spent coffee lipid extract in cosmetic industry seems to be a suitable approach to recycle the wastes from coffee industry. In this sense, a cream containing 10% by weight of the lipid fraction of SCG extracted with supercritical carbon dioxide was developed and characterized in this work, showing improved skin lipids (sebum) and hydration qualities. This may suggest that the barrier properties of the skin were increased, probably by supplementation of skin lipids.

Moreover the coffee oil presented promising characteristics in the improvement of fat skin levels with a good acceptance by consumers.

Acknowledgements

We gratefully acknowledge the support by NovaDelta – Comércio e Indústria de Cafés, S.A. (Campo Maior, Portugal) and COOXUPÉ. A. Paiva would like to acknowledge Fundacão para a Ciência e a Tecnologia (FCT, Portugal) through the contract SFRH/BPD/44946/2008 and Dr. Inês Casais for the drawing support.

The authors have declared no conflict of interest.