This work was carried out at the Center for Studies and Research on Renewable Energy in Hann Bel Air, Dakar, Senegal. The experimental setup for superheated steam roasting of the coffee-Guinea pepper blend comprises a roasting chamber, a steam generator, a superheater heat exchanger and a burner. The burner is powered by biogas from a geomembrane biodigester with a capacity of 4 m³. The cylindrical glass roasting chamber measures 0.45 m in height and 0.03 m in diameter (see Figure 1 ). Saturated steam is produced by a generator equipped with a 0.33 L tank and 1300 W of power. This steam is then superheated using a 2 m-long copper coil, which is heated by a biogas-fueled burner. The Arabica coffee beans had an initial moisture content of 11.06%, while the Guinea peppercorns had an initial moisture content of 8.75%. Five coffee beans were introduced into the roasting chamber, where they were in direct contact with the superheated steam to ensure uniform roasting. This small batch was selected as part of a laboratory-scale experimental approach to obtain accurate temperature change and roasting time measurements using integrated thermocouples. Using five beans provides a better understanding of heat transfer mechanisms and thermochemical effects on a small scale. This is a preliminary step to further large-scale tests aimed at validating the reproducibility of the process under industrial conditions. Once roasted, the coffee beans are immediately mixed with the Guinea peppercorns at a ratio of 10%. The residual heat from the coffee beans is then used to roast the peppercorns through simple agitation until the roasting process is complete.

The quality of roasted coffee is highly dependent on roasting temperature and duration. These parameters are meticulously measured using K-type thermocouples installed inside each coffee bean, at the inlet of the superheated steam, and at the outlets of both the roaster and the saturated steam generator. To achieve this, holes measuring 5 mm in depth and 1 mm in diameter were drilled into the green coffee beans, which were then fixed and suspended in the roasting chamber. The data collected by the thermocouples is recorded in real time using a data logger connected to a computer ( Figure 2 ). To minimize uncertainties, this experimental procedure was repeated three times. The roasted beans’ properties, including color, temperature evolution, moisture content, and mass loss, are analyzed and correlated with the experimental parameters.

The moisture content is analyzed in the laboratory using the oven-drying method based on ASTM D 3172 [28] . This method involves drying the coffee in an oven at 105˚C for 24 hours. After cooling in a desiccator, the mass loss is calculated as a percentage using Equation (1):

$H\left(\text{\%}\right)=\frac{m_1-m_2}{m_1-m_0}\times 100$ (1)

m₀: Mass of the empty capsule.

m₁: Mass of the empty capsule plus the sample before drying.

m₂: Mass of the capsule plus the sample after drying.

Figure 3 illustrates the equipment used.

Three steam flow rates were investigated. These flow rates were experimentally measured based on the known volume of the steam generator tank (0.330 L) and the time required for complete vaporization under three different control knob settings: minimum, intermediate, and maximum.

For a measurement uncertainty on mass U(m) and an uncertainty on time U(t), the uncertainty on the flow rate U(D) is determined using Equation (2).

$\frac{U\left(D\right)}{D}=\sqrt{{\left(\frac{U\left(m\right)}{m}\right)}^2+{\left(\frac{U\left(t\right)}{t}\right)}^2}$ (2)

For a water mass of 330 g with U(m) = ±1 g, U(t) = ±1 s, D steam flow rate and a duration of 825 s:

$\frac{U\left(m\right)}{m}=\frac{1}{330}=0.00303$ (3)

$\frac{U\left(t\right)}{t}=\frac{0.0167}{13.75}=0.00121$ (4)

$\frac{U\left(D\right)}{D}=\sqrt{{0.00303}^2+{0.00121}^2}=0.00327$ (5)

$U\left(D\right)=24\times 0.00327$ (6)

$D=24\pm 0.08\text{g}/\text{min}$ (7)

The same approach was applied to the intermediate and maximum settings, yielding flow rates of 27 ± 0.09 g/min and 30 ± 0.10 g/min, respectively.

Cow dung was used as the raw material for the biodigester. It was collected in 50 kg bags and 25 L containers after the slaughter of cows. The feeding of the biodigester was carried out in two stages. On the first day, eight 25 L containers of liquid substrate were introduced into the biodigester. On the second day, 507.48 kg of solid substrate were mixed with 216.5 L of water and added. After filling and hermetic sealing, the temperature, pH and chemical composition of the biogas were regularly monitored. These characteristics were analyzed using a G5000 biogas analyzer connected to the biodigester ( Figure 4 ). The indoor and outdoor temperatures were measured using a Fourier transform infrared spectroscopy (FTIR) T420 thermal imaging camera ( Figure 5 ). To ensure a continuous and stable supply, the biogas was stored in balloons. A water trap was connected to the biodigester to minimize moisture content during combustion. The hydrogen sulfide (H₂S) content was reduced using an H₂S purifier ( Figure 6 ).

Monitoring biogas production helps evaluate the efficiency of the biodigester based on its temperature and the composition of the biogas produced. The internal temperature ranges from 38˚C to 72.9˚C, indicating mesophilic and thermophilic phases conducive to anaerobic digestion. The temperature evolution ( Figure 7 ) shows a rapid initial increase, followed by fluctuations influenced by external conditions.

Analysis of biogas composition ( Figure 8 ) reveals a gradual increase in CH₄ concentration from 29% to 59.1% by the thirteenth day, indicating optimal conditions for methanization. The transformation of volatile acids reduces CO₂ concentration, while CO and H₂S levels remain low, necessitating appropriate management.

The statistical analysis of methane content, as summarized in Table 1 , indicates a mean value of 52% (±9%) with a relatively narrow 95% confidence interval (±3.9%). These results provide evidence of the biodigester’s stable and efficient performance.

Based on the total 740 L of substrate introduced and the experimental conditions, the hydraulic retention time (HRT) ranges between 5 and 7 days. This duration corresponds to a batch feeding test phase, which is typical of laboratory-scale experiments. Mudhoo et al. [29] reported that under mesophilic/thermophilic conditions, one kilogram of fresh manure produces an average of 0.22 m³ of biogas. However, for an introduced mass of 507.48 kg, the total volume of biogas produced is estimated at 112 m³ over an active period of approximately 20 days. This results in an average daily yield of 5.6 m³/day, of which about 3.3 m³/day is methane (at 59.1% CH₄). These performances confirm the biodigester’s ability to supply the thermal energy required for superheated steam production. Increasing the substrate volume or adopting continuous feeding would meet the energy needs of a medium-scale artisanal roasting unit.

The biogas produced is a mixture of combustible gases with a methane (CH₄) content of 59.1%, exceeding the minimum threshold (50% - 55%) required for stable combustion. A flammability test was conducted using an atmospheric burner (see Figure 9 ). Combustion is characterized by the emission of a blue flame, which confirms the presence of methane in a significant proportion greater than 50%.

Figure 10 illustrates the evolution of saturated steam and superheated steam temperatures measured upstream of the beans over time. The average measured temperature of the superheated steam is 284˚C. The temperature of the saturated steam is relatively stable, oscillating slightly around an average of 107˚C. This reflects a controlled process in the generator where the steam reaches its saturated state at a constant temperature. The steam produced circulates through a coil exchanger where it is re-superheated by the burner, resulting in a significant increase in steam temperature.

The thermal profile analysis during roasting reveals a progressive increase in coffee bean temperature, directly correlated with their heat absorption capacity. Initial fluctuations, attributed to factors such as bean position within the roasting chamber, particle size, and residual moisture content, momentarily affect this progression. However, the slight temperature differences observed between the beans gradually decrease, leading to a satisfactory thermal uniformity by the end of the process. At the target temperature of 250˚C, the beans undergo full aromatic development, ensuring a uniform and consistent roast. Figure 11 illustrates the temporal evolution of superheated steam temperatures at the inlet and outlet of the roasting chamber, as well as the corresponding coffee bean temperature profile.

The calculation of the mean, standard deviation, and confidence interval allows for the statistical characterization of coffee bean temperatures during roasting. This statistical analysis makes it possible to assess the stability, thermal uniformity, and reproducibility of the roasting process. To avoid dispersion caused by the initial temperature rise, only temperatures ≥ 200˚C were considered, as presented in Table 2 .

The results indicate that coffee bean temperatures remained both homogeneous and stable, exhibiting variations of less than 10˚C between individual beans and moderate confidence intervals (±12 to ±17˚C). These findings provide strong evidence of the thermal stability and reproducibility of the roasting process.

Figure 12 illustrates the uniformly roasted coffee beans.

A comparative study of three steam flow rates (24 g/min, 27 g/min, and 30 g/min) highlights the critical impact of this parameter on process efficiency. Figure 13 illustrates the temperature evolution of the coffee beans for each steam flow rate.

The moisture loss curve of coffee beans, measured during roasting, is divided into three distinct phases, each reflecting the product’s physicochemical transformations. During the initial drying phase (0 to 150 seconds), the moisture content of the beans decreases from 11.06% to 7.08%, corresponding to the removal of surface free water. It is followed by the actual roasting phase (150 to 310 seconds), during which the moisture content further decreases to 4.67%, accompanied by the first crack, caused by the expansion of the beans under heat. The final phase (310 to 610 seconds), marked by a second crack, sees the residual moisture reach

1.5%. Figure 14 illustrates the moisture loss of the coffee beans throughout the roasting process.

The statistical analysis of moisture loss in the beans, as shown in Table 3 , indicates a consistent decrease in moisture content from 11.06% to 1.5%, with a mean value of 5.65% ± 2.8%, confirming its consistency and reproducibility.

The most commonly used measure to determine the degree of roasting is the color of the bean. Roasting occurs in five stages: drying, yellowing, the first crack, roast development, and the second crack. Figure 15 illustrates the development of bean color during the different roasting stages.

The blending of coffee and Guinea pepper is carried out immediately after roasting the coffee beans. The residual heat from the freshly roasted coffee beans is used to roast the Guinea peppercorns by simply stirring the roasted coffee and pepper mixture until it reaches the desired roasting level. The results showed a moisture loss in Guinea peppercorns from 8.75% to 3.23%. This significant decrease in moisture, corresponding to a loss of 5.52%, results from heat transfer between the freshly roasted coffee beans and the Guinea peppercorns during blending. The thermal energy causes rapid evaporation of water from the peppercorns, facilitated by their structure, which promotes moisture release in the form of steam. Figure 16 illustrates the roasted coffee-Guinea pepper mixture.

The roasted coffee-Guinea pepper blend is prepared using coffee beans roasted to the second crack, an advanced stage of roasting where aromatic compounds are highly developed. The infrared spectra shown in Figure 17 reveal heightened carbonyl and aromatic bands, indicating a strong presence of pyrazines, furfural, and aldehydes. This aromatic richness from the coffee is complemented by compounds in the roasted pepper, such as phenols and terpenes.

Table 4 illustrates the influence of roasting phases on the aromatic profile.

Control experiments using traditional roasting methods were indeed conducted for comparative purposes, as illustrated in Figure 18 . This figure displays the infrared (IR) spectra of the coffee-Guinea pepper blend roasted using the traditional method and superheated steam roasting.

The spectra obtained by Fourier transform infrared spectroscopy (FTIR) reveal that the roasting of the coffee-Guinea pepper blend induces significant changes in its chemical composition. The attenuation of bands associated with volatile groups C=O around 1740 cm⁻¹, O-H around 3295 cm⁻¹, and C-H around 2924 cm⁻¹ indicates the thermal degradation of fatty acids, alcohols, and terpenes initially present in green coffee and Guinea pepper. Simultaneously, new bands appear around 1620 cm⁻¹, corresponding to condensed aromatic structures, and in the 1000 - 1300 cm⁻¹ range, attributed to secondary volatile compounds resulting from thermal transformation processes. Figure 19 illustrates the changes in chemical composition before and after roasting.

The sensory characteristics of the roasted coffee-Guinea pepper blend produced using superheated steam were compared to two samples roasted using the conventional method. For each sample, 20 g of ground coffee was filtered with 258.7 mL of hot water, followed by the addition of 46 g of sugar. Figure 20 illustrates the filtered beverage samples.

For each sample, five panelists provided their evaluation of color and aroma intensity. Figure 21 illustrates the tasting process.

Analysis of the evaluations from the five panelists reveals a clear superiority of the superheated steam roasting method compared to conventional methods. As shown in Table 5 , superheated steam roasting received significantly higher average scores across all sensory criteria.

The sensory analysis, conducted alongside physicochemical measurements, established a correlation between the molecular characteristics revealed by FTIR spectra and the sensory perception of the roasted coffee.

Superheated steam roasting was distinguished by a visibly darker color and a significantly higher aromatic intensity. These observations were confirmed by FTIR results, which showed better preservation of key functional groups such as C=O, C-H, and C-O—essential to the aromatic and flavor profile of coffee.

From a sensory standpoint, the panelists assigned an overall score of 7.8 to the superheated steam-roasted coffee, compared to 5.8 for the conventionally roasted sample. The correlation between the darker color (+33%) and enhanced aromatic intensity (+36%) with overall satisfaction highlights a clear qualitative advantage of this method. These findings support the hypothesis that the gentle and uniform roasting induced by superheated steam allows for a gradual release of aroma compounds without excessive thermal degradation, which is in perfect agreement with the FTIR spectroscopic signature observed.

Our experimental work on superheated steam roasting of coffee was compared with the results of Bottazzi et al. [30] . Figure 22 and Figure 23 illustrate the evolution of temperature and moisture losses over time, as well as the analysis of the linearity of the result.

The five roasted coffee beans correspond to a mass of 25 g. For a daily production of 5.6 m³ of biogas, including 3.3 m³/day of CH₄, roasting with superheated steam at 250˚C for 450 seconds was carried out. This roasting duration and temperature were achieved using 0.03 m³ of biogas characterized by a CH₄ content of 59.1%, corresponding to an average lower heating value (LHV) of 21 MJ/m³. Under these conditions, the energy consumed is:

$E=\frac{0.03\times 21}{0.025}$ (8)

$E=25.2\text{MJ}/\text{kg}$ (9)

$E=\frac{25.2}{3.6}$ (10)

$E=7\text{kWh}/\text{kg}$ (11)

In terms of CO₂ emissions, the combustion of methane (CH₄) produces CO₂ according to the stoichiometric reaction:

${\text{CH}}_4+2{\text{O}}_2\to {\text{CO}}_2+2{\text{H}}_2\text{O}$ (12)

The CO₂ emitted from the combustion of 0.03 m³ of biogas (containing 59.1% CH₄) is estimated as follows:

$V\left({\text{CH}}_4\right)=0.03\times 0.591=0.01773{\text{m}}^3$ (13)

${\text{CO}}_2\text{emis}=0.01773\times 1.9$ (14)

${\text{CO}}_2\text{emis}=0.0337\text{kg}{\text{CO}}_2$ (15)

For 1 kg of coffee, the corresponding emission is calculated as:

${\text{CO}}_2\text{emis}=\frac{0.0337}{0.025}$ (16)

${\text{CO}}_2\text{emis}=1.35\text{kg}{\text{CO}}_2/\text{kg}\text{coffee}$ (17)

Furthermore, according to Rahman et al. [31] , the net CO₂-eq emissions from biogas combustion are significantly lower, approximately 0.043 kg CO₂-eq/m³, since biogenic CO₂ is not accounted for in the carbon footprint.

Thus, the emissions associated with 0.03 m³ of biogas are:

$0.03\times 0.0043=0.00129\text{kg}{\text{CO}}_2\text{-eq}$ (18)

${\text{CO}}_2\text{emis}=\frac{0.00129}{0.025}$ (19)

${\text{CO}}_2\text{emis}=0.0516\text{kg}{\text{CO}}_2\text{-eq}/\text{kg}\text{coffee}$ (20)

Compared to the findings of Tsai et al. [32] (0.1576 kg CO₂/kg), our roasting technology demonstrated an 11.22% reduction in energy consumption and a 67% reduction in CO₂ emissions. These results indicate that the proposed technology is a clean, sustainable energy alternative well-suited to the context of sustainable development.

The observed temperature ( Figure 7 ) variations highlight the biodigester’s ability to reach and maintain optimal thermal conditions for bacterial activity. The rapid initial temperature increase indicates effective heat retention, while the fluctuations reflect the influence of climatic conditions. The system’s insulation ensures stable thermal conditions, optimizing bacterial activity and improving biogas production. This thermal stability is essential for sustaining the activity of mesophilic and thermophilic bacteria, which is critical to the overall efficiency of biogas production. The results in Figure 8 show an increase in methane content. This evolution of methane is due to thermal conditions that are favorable to methanogenic bacteria. The decrease in CO₂ and the consistently low levels of CO and H₂S also indicate good substrate conversion and a limited presence of toxic byproducts. The effective management of these components, which characterize the biogas, ensures the safety and durability of the biodigester. These results confirm the proper functioning of the system and the production of high-quality biogas.

Biogas combustion test ( Figure 9 ) demonstrates the ability of the biogas and burner to superheat the saturated steam required for roasting the coffee-Guinea pepper blend. The resulting blue flame indicates efficient combustion, confirming that the biogas is of high quality and suitable for thermal applications. The methane content exceeding the 50% threshold further confirms the efficiency of the digestion process and the quality of the biogas, highlighting the potential of the biodigester for energy recovery in agri-food processes.

The slight variations observed ( Figure 10 ) are the result of manual handling of the saturated steam generator. The produced steam circulates through a coil heat exchanger, where it is reheated by the biogas burner. This reheating leads to a significant increase in steam temperature, demonstrating the efficiency of the biogas burner. The fluctuations in the superheated steam temperature, although less pronounced after the initial peak, indicate active regulation of the burner to maintain the temperature required for roasting the coffee and Guinea pepper blend. These thermal performances highlight that the biogas burner is well-suited for the roasting process.

Superheated steam plays a central role in this mechanism by transferring thermal energy primarily through convection and conduction. A portion of this energy is used to vaporize the residual moisture within the beans, leading to a slight decrease in steam temperature at the roasting chamber outlet. However, this phenomenon is effectively managed through the stabilization of the heat flux, ensuring precise control over the roasting conditions. The system’s energy efficiency is further enhanced by high-performance chamber insulation, which minimizes heat losses and enables rigorous thermal regulation. Although residual temperature variations among individual beans persist, they remain negligible and do not compromise the final product quality, as demonstrated by the uniformly roasted beans in the experimental results.

A flow rate of 24 g/min ( Figure 13 ), while economical in steam consumption, significantly restricts heat transfer to the beans, extending the roasting duration and increasing the risk of incomplete roasting, which hinders aroma development. At 27 g/min, the temperature rise accelerates; however, persistent thermal inertia delays the achievement of optimal conditions. Conversely, a 30 g/min flow rate facilitates a rapid and uniform temperature increase, reaching the required 250˚C in a shorter time. This flow rate not only enhances roasting uniformity but also preserves the coffee’s organoleptic properties by minimizing prolonged exposure to intermediate temperatures. These findings emphasize the necessity of sufficient energy input to ensure efficient and reproducible roasting while improving the overall system efficiency.

These three phases illustrate the moisture loss in coffee beans during the roasting process ( Figure 14 ). The initial drying phase is characterized by uniform heat penetration, which helps prevent irregularities in bean coloration and ensures the development of the characteristic coffee aroma. During the cracking phase, the moisture content in the beans gradually decreases, while sugar caramelization and Maillard reactions take place, defining the distinctive sensory qualities of roasted coffee. The observed transformations, closely linked to temperature and roasting duration, highlight the balance between moisture removal and the preservation of aromatic compounds, which is essential for producing a high-quality final product.

Coffee roasting follows a complex process divided into several essential steps to achieve the desired flavor profile ( Figure 15 ). It begins with the drying phase, where the green beans lose their moisture and gradually change color from light green to dark brown. During roasting, the beans slightly increase in volume, causing a cracking sound characterized by a silvery skin and the release of small balls. At this point, the first crack occurs due to the heat, causing an increase in internal gas pressure, which leads to the cracking of the beans. This phenomenon is accompanied by a distinct noise that signals the beginning of the development of coffee aromas. During this period, roasting can be adjusted or stopped according to the desired profile, as the temperature of the beans rises more slowly despite constant heat. After the first crack, the development of the beans continues, influencing their balance between acidity and bitterness. A longer roast reduces acidity while increasing bitterness. Continuing the process leads to a second crack beyond 200˚C. At this stage, the oils from the beans appear on the surface, and an intense flavor due to carbonization is characterized by the release of smoke. This step results in the loss of subtle flavors, giving the coffee a robust but less refined taste.

This moisture loss in Guinea pepper contributes to their roasting, enhancing their characteristic taste and aroma. The residual moisture content of 3.23% in the roasted Guinea pepper supports preservation by limiting the risk of mold or degradation. This process optimizes the use of the heat generated by the coffee beans without requiring additional energy input, making it both efficient and well-suited for sustainable production.

The analysis of Figure 17 confirms that roasting coffee up to the second crack significantly enhances its aromatic complexity, primarily due to the development of pyrazines and furfural, as evidenced by the pronounced carbonyl and aromatic bands in the IR spectra of Figure 17 . The addition of roasted Guinea pepper further enriches this profile with terpenes and oxidized phenols, contributing subtle peppery and smoky notes. The comparison of IR spectra highlights key evidence validating the efficacy of superheated steam roasting. Both curves exhibit bands around 2923 cm⁻¹ (C-H bonds) and 1027 - 1043 cm⁻¹ (C-O bonds), indicating the preservation of aromatic functional groups and volatile components in the blend. This demonstrates that superheated steam induces chemical transformations comparable to those of traditional roasting. The superheated steam spectrum shows enhanced retention of intensity in specific characteristic bands, suggesting more effective preservation of aromatic compounds. This reflects a more uniform and less destructive thermal treatment. Superheated steam roasting is significantly faster, reducing energy consumption while ensuring sufficient product transformation, a key indicator of efficiency. The absence of bands indicating excessive decomposition (e.g., in the 1600 - 1700 cm⁻¹ range or below 800 cm⁻¹) for superheated steam points to reduced carbonization or degradation of organic structures.

The effectiveness of superheated steam roasting is confirmed by:

While achieving results comparable to traditional methods, as demonstrated by the IR spectra. These findings support the viability of this approach for roasting the coffee-Guinea pepper blend, combining efficiency with minimal structural compromise. In addition to this chemical analysis, the energy consumption of biogas-powered superheated steam roasting proves to be significantly lower and more efficient overall compared to conventional roasting processes using wood or butane gas. This method delivers superior overall energy efficiency, attributed to its enhanced heat transfer and reliance on renewable energy sources.

Bottazzi et al. [30] used a hot air roaster with temperatures ranging from 220˚C to 260˚C. The results show similar trends in terms of temperature development. During the drying phase, steam reaches 150˚C in 100 seconds, compared to 130˚C for hot air, due to more efficient heat transfer through conduction and convection. Hot air, by contrast, relies solely on convection. During roasting, temperatures converge between 180˚C and 200˚C between 150 and 200 seconds. However, steam attains the critical temperature threshold faster, reflecting superior thermal penetration. Beyond 200 seconds, steam temperature exceeds 250˚C more rapidly and stabilizes at a higher level. The linear regression curve, with an R² value of 0.9213 ( Figure 20 ), indicates a strong correlation between measured data, confirming the method’s reliability. A maximum temperature difference of 5˚C, corresponding to a relative error of 2.04%, was observed between this experimental study (250˚C) and the reference (245˚C) [29] . This discrepancy is attributed to manual steam flow control and the inherent limitations of thermocouples (±1.5˚C). Superheated steam, with a heat capacity of 2 kJ/kg·K, reaches 250˚C in 450 seconds, compared to 600 seconds for hot air (1.005 kJ/kg·K), while maintaining a maximum temperature deviation of 5˚C. This method reduces residual moisture to 1.5% (vs. 2.5% for hot air) and operates in a low-oxygen environment, minimizing aromatic compound losses. Superheated steam roasting emerges as a reliable, sustainable solution for large-scale production, delivering performance consistent with prior research. Its efficiency in heat transfer, temperature control, and reduced energy consumption positions it as a superior alternative to conventional hot-air methods.