eISSN: 1984-3909

Coffee Science, e192184, 2024

Doi: https://doi.org/10.25186/.v19i.2184

Effect of extraction processing parameters using a ribbon blender on

the physicochemical properties of coffee
Paula Andrea Mayorga Barriga1 , Ruth Yolanda Ruiz Pardo2 , Fabian Leonardo Moreno Moreno2

1
Universidad de La Sabana, Facultad de Ingeniería, Maestría en Diseño y Gestión de Procesos, Chía, Cundinamarca, Colombia
²Universidad de La Sabana, Facultad de Ingeniería, Grupo de Investigación en Procesos Agroindustriales, Chía, Cundinamarca, Colombia
Contact authors: [email protected]; [email protected]; [email protected]
Received in October 25, 2023 and approved in April 2, 2024

ABSTRACT
In this study, we investigated the impact of water-coffee ratio, time, and stirring speed on percolation in a horizontal ribbon blender. We analyzed their
influence on total soluble solids, extraction rate, titratable acidity, and extraction yield. The coffee extract was obtained in a pilot unit at a constant temper-
ature of 85°C and varying the water-coffee ratio (w/w) from 4:1 to 10:1; the stirring speed between 30 and 95 RPM and the extraction time from 10 to 60
min. It was determined that the water-coffee ratio was the factor that had a significant influence on all the response variables, while time and stirring were
significant for the extraction rate and titratable acidity. The optimal conditions of soluble solids, extraction yield, extraction rate, titratable acidity, and
chlorogenic content were a water temperature of 85°C, a water-coffee ratio (w/w) of 4:1, and a stirring speed of 66 rpm for 10 min. At these conditions,
an extract of 5.85% Total Dissolved Solids, 14.54% as yield, an extraction rate of 654.8 g/h, and a content of 5.62 mg of CGA/mL was obtained. Hence, this
study presents an alternative process to obtain coffee extract in producing soluble coffee at a low industrial scale.
Key words: Chlorogenic acid; Extraction yield; Foods; Optimization; Soluble coffee.

1 INTRODUCTION on the quality of the bean, the roasting process, particle size
after grinding, and the extraction method, where conditions,
Coffee is one of the most important food commodities parameters, and variables in the process can affect the sensory
worldwide. Its economic significance is mainly due to the and chemical characteristics of coffee. (Cordoba et al., 2021a).
beverage or extract obtained from roasted and ground coffee The extraction time is a key factor, as soluble compounds
beans, used to produce soluble coffee (Ocampo; Alvarez, 2017) such as organic acids, sugars, and caffeine are extracted
Instant coffee, a growing preference, recorded consumption quickly, while less soluble compounds require more time to
surpassing 1.6 million tonnes in 2017. (Sulewska et al., be removed from the food matrix (Mestdagh; Glabasnia;
2021), whose popularity is attributed to the convenience of its Giuliano, 2017). Likewise, the water-to-coffee ratio used for
preparation and the shelf life of the product (Capek et al., 2014). extraction greatly affects the extract, as an unsuitable selection
The industrial production of this coffee is carried out using will result in underdeveloped flavors and reduced extraction
roasted and ground beans. Initially, the compounds present in yield of the coffee extract (Angeloni et al., 2019). Temperature
the coffee are extracted, followed by the concentration stage of favors the solubility of many compounds in the coffee bean.
the extract through evaporation or technologies such as freeze However, very high temperatures can lead to the extraction
concentration. Finally, the extract is sent to a spray dryer or of undesired compounds in the extract, affecting its sensory
a freeze dryer before packaging the product (Sulewska et al., perception (Mestdagh; Glabasnia; Giuliano, 2017). Finally,
2021). agitation and particle size of the coffee beans determine the
In general, at industrial scale, the extraction of coffee extraction rate of the process and consequently the total solid
extract is carried out in large percolators, where roasted and content achieved in a specific time. Those variables promote
ground coffee beans are in contact with water at 200 °C and the transfer of compounds to the extract by increasing the
1500 kPa to extract volatile and non-volatile compounds extraction surface area and the interface between water and
present in them (Benincá et al., 2016). At this stage, three coffee (Cordoba et al., 2020).
main processes are identified: the solubilization of solutes In recent years, interest in coffee extraction has
from the food matrix, the diffusion of these solutes through increased. Nevertheless, due to the complexity of the process
the coffee bean pores, and finally, the solubilization of solutes to obtain a high-quality product, the studies conducted so
in the extract (Fuller; Rao, 2017). The objective of extraction far have been small-scale and focused on obtaining coffee
to produce soluble coffee is to remove the highest amount of extract for immediate consumption, where the concentration
soluble solids (TDS) since they influence the operational yield of soluble solids in the extract reaches only between 1.3%
and represent the proportion of dissolved material in the extract and 3.2% (Angeloni et al., 2019); while at an industrial scale,
over the total mass of coffee. This result can vary depending fixed-bed percolation systems are employed at high pressure,
2024 | Lavras | Editora UFLA | www.coffeescience.ufla.br
 BARRIGA, P. A. M. et al.

which can yield around 30% (Benincá et al., 2016). However,   x  b

2

in our best knowledge, there are no reported studies of non- y  ae (1)

2c 2
pressured pilot-scale equipment. An alternative for industrial
coffee extract production is the ribbon blender solid mixer, The average particle size was calculated using Equation 2.
whose purpose is promoting coffee ground-water mixing and
enhance convective mass transfer. The study of extraction in  a 
Ln  
pilot-scale units allows identifying alternatives to industrial  Tm   b
 X c (2)
production with simpler and more economical equipment 0.5
than traditional percolation equipment. The term “pilot-
scale” refers to a scaled-down version of the equipment Where a, b, and c are the model parameters, and Tm is
used in preliminary testing before potential upscale to larger the sample size used.
production. The use of a pilot-scale setup is crucial at this
stage as it allows for controlled experimentation in a smaller,
2.2.2 Experimental design
more manageable setting. This approach enables meticulous A d-optimal response surface model to obtain 21
parameter adjustments, facilitating a detailed understanding experimental combinations was obtained by the software
of their impact on the extraction process. The insights gained Design Expert® Version 11.0 (Stat-Ease Inc, Minneapolis,
from this pilot-scale exploration lay the foundation for MN). The water temperature was set at 85 °C, and 3 L of
optimizing conditions before considering large-scale industrial water per trial, medium-high roast of coffee. The design
production. The aim of this study the effect of different water- factors corresponding to agitation speed (between 30 and 95
to-coffee ratios, extraction times, and stirring in a pilot-scale RPM), time (between 10 and 60 min), and water-to-coffee
ribbon blender on the percentage of TDS, yield, extraction ratio (between 4:1 to 10:1 w/w) that would maximize the total
rate, titratable acidity, and pH. soluble solids, extraction rate, and acidity.
Statistical techniques, such as Analysis of Variance
(ANOVA) and regression analysis, were applied using the same
2 MATERIAL AND METHODS software to examine the relationships between independent
factors and dependent response variables. In addition, the
2.1 Materials software facilitated the determination of the most favorable
conditions, conducting a desirable analysis to identify those
Medium-high roast and ground non-rated coffee
that enhance the response variables.
(Coffea arabica) (Aro, Colombia), purchased from a local
store in Bogotá, Colombia.. NaOH CAS 1310-73-2 (Sigma
2.2.3 Extraction process
Aldrich, USA) and 0.15% potassium bitartrate CAS 868-14-4
For coffee extraction, the ribbon blender equipment
(Sigma-Aldrich) were used for titratable acidity.
depicted in Figure 1, designed by University of La Sabana,
2.2 Methods was employed. The unit operates in batch mode. The unit
consists of a chamber with a ribbon blender type agitator that
promotes contact between water and coffee. The chamber
2.2.1 Particle size distribution
is cylindrical with dimensions 18 cm in diameter by 80 cm
The average particle size of the coffee was calculated in length. It has a 20 L capacity water storage tank, with
according to the NTC 2441 standard for particle size electrical resistances that heat the water used in extraction.
distribution (Instituto Colombiano de Normas Técnicas - After mixing the coffee and water, there is a filter with a
ICONTEC, 2011). The sieves were stacked on top of each pressure frame that allows the coffee to be pressed and
other in decreasing mesh order in the AS 200-Retsch analytical the extract separated, which is collected in a final tank.
sieve shaker, and 65 g of roasted and ground coffee was placed Before connecting the equipment to the electrical source,
on the top sieve, covered with a lid. The equipment was run the water tank (2) was filled, and then it was connected,
for 5 mins, and the accumulated fractions of coffee in each the temperature was established at 85°C on the equipment
sieve were collected and weighed, and the percentage of each controller (6). While the water in the tank reached the
fraction relative to the initial sample was calculated, analyzing required temperature, roasted and ground coffee was placed
the results in CurveExpert Pro Version 2.2.3 software in the coffee hopper (1), ensuring it is closed to prevent its
(Hyams Development). The information of sieve diameter passage to the extraction tank (3). When the water reached
and cumulative percentage was analyzed in CurveExpert 85°C, the agitation for the extraction tank was turned on at
Pro Version 2.2.3 software (Hyams Development) using a the set speed using the equipment controller, and the passage
Gaussian model fit, as shown in Equation 1 and 2. valve (7) to the extraction tank was opened. Subsequently,

Coffee Science, 19:e192184, 2024

 Effect of extraction processing parameters using a ribbon blender on the physicochemical properties of coffee

the coffee hopper was opened to let the coffee fall to the Where TDS represents the extracted soluble solids,
extraction tank. The mixing operated at the set speed until Wextr is the weight of coffee extract (g), and Wcoffee is the weight
the operational time. After the mixing time was achieved, the of roasted and ground coffee used in the trial (g).
extract pass to a press filter (4) where the remaining solids
of the process were collected. The filter is a 20 cm squared 2.2.7 Extraction Rate
frame with a filter clothe and a press system. Ultimately, the The extraction rate was reported in g/h and calculated
coffee extract was stored in the product tank (5). using Equation 4 (Moroney et al., 2016).

2.2.4 Statistical Analysis   TDS  

A D-optimal response surface design was used. The  Wextr *  100  
 
Extration rate   (4)
obtained results were analyzed using Design-Expert V11  t 
software, and the significance of the factors in the extraction  
 60 
process was determined using a p-value of 0.05. Likewise,
a desirability analysis was performed to obtain the factor Where Wextr is the weight of coffee extract (g), TDS
configuration that would fit a replicable prediction model at an represents the extracted soluble solids, and t represents the
industrial level, maximizing the total dissolved solids (TDS), stirring time in minutes.
extraction rate, operation yield, and titratable acidity of the
coffee extract. 2.2.8 pH and Titratable Acidity
The pH of the coffee extract was measured using
2.2.5 Total dissolved solids
a Mettler Toledo FiveEasyTM F20 benchtop pH meter.
The total dissolved solids (TDS) were determined using The determination of titratable acidity in the sample was
the PAL-COFFEE BX-SST refractometer (Atago, Japan). carried out using the A.O.A.C 920.92 reference method. A
0.1 M NaOH solution was prepared and then standardized
2.2.6 Yield using a 0.15% potassium bitartrate solution to determine
The extraction yield is determined as the ratio of the the equivalence point. The coffee extract samples were then
extracted solids to the total amount of roasted ground coffee, it titrated potentiometrically with the standardized NaOH
was determined using Equation 3 (Zhang et al., 2022). solution until they reached pH 7. The results are expressed
in mg of CGA/mL coffee (mg CGA/mL coffee) (Vezzulli et
TDS *Wextr
Yield  (3) al., 2021).
Wcoffee

Figure 1: Diagram of the pilot coffee extractor.

Coffee Science, 19:e192184, 2024

 BARRIGA, P. A. M. et al.

3 RESULTS Table 2: Parameters of the Gaussian model for the

particle size distribution.

3.1 Particle Size Distribution a 102.54

The average particle size of the roasted and ground b 63.42

coffee beans is presented in Table 1. c 332.76
EE 0.0026
Table 1: Coffee recovered in the sieving process. R 2
0.99

Sieve N° Sieve Diameter (µm) % m/m

16 1180 1.2 Using the above information and Equation 2. the
18 1000 2.5 average particle size for the roasted and ground coffee used
20 850 4.3 was calculated to be 383.67 µm.
35 500 34.9
45 355 25.9
3.2 Characterization of the obtained extracts
60 250 16.2 The relationship between the evaluated
70 212 12.1 parameters (water-to-coffee ratio, stirring speed, and
extraction time) as well as the response variables
Bottom 0 2.9
are presented in Table 3. Each treatment 1 to 21
Total 100.0
correspond to the combination of the studied factors
(water to coffee ration, stirring speed and extraction
Table 2 presents the obtained parameters for a, time) shown on each line. The effect of operational
b, and c, as well as the correlation coefficient and the parameters on each response variable is explained in
standard error of the model of Equation 1. the following sections.

Table 3: Experimental Design, Extraction Yield, and Characterization of the Obtained Extracts.
Water to coffee Stirring speed Extraction Extraction Titratable acidity (mg Extraction rate
TDS (%) pH
ratio (rpm) time (min) yield (%) CGA/mL coffee) (g/h)
1 4.0 57.6 60.0 6.57 10.42 5.66 10.38 78.17
2 7.0 62.5 35.0 3.65 11.87 5.63 4.63 87.24
3 6.9 77.8 60.0 4.37 14.91 5.78 5.42 64.72
4 10.0 95.0 60.0 2.80 20.03 5.84 3.08 60.11
5 10.0 30.0 10.0 1.87 17.74 5.89 1.29 321.48
6 4.0 95.0 10.0 5.12 12.61 5.8 4.97 567.26
7 6.6 30.0 60.0 2.74 15.00 5.53 3.40 68.69
8 10.0 95.0 10.0 2.52 14.35 5.94 2.37 258.26
9 10.0 95.0 60.0 2.29 22.22 5.63 2.75 66.67
10 10.0 95.0 10.0 2.05 18.03 5.9 2.64 324.92
11 5.3 93.1 35.0 3.66 15.87 5.68 3.48 153.54
12 4.0 30.0 10.0 5.90 7.79 5.98 5.19 350.23
13 10.0 30.0 60.0 2.16 19.10 5.66 2.85 57.82
14 8.6 95.0 35.0 2.72 17.32 5.66 3.08 103.35
15 4.0 95.0 60.0 5.92 12.50 5.67 7.28 93.74
16 4.0 95.0 10.0 4.19 13.82 5.96 2.35 621.87
17 10.0 62.5 22.5 2.85 14.82 5.69 4.57 118.71
18 10.0 30.0 60.0 2.26 16.92 5.61 3.16 50.76
19 10.0 30.0 10.0 1.92 15.48 5.9 1.32 278.75
20 5.5 62.5 10.0 3.89 14.50 5.64 4.44 474.02
21 4.0 30.0 38.8 4.23 13.60 5.69 3.80 157.95

Coffee Science, 19:e192184, 2024

 Effect of extraction processing parameters using a ribbon blender on the physicochemical properties of coffee

3.2.1 Total Soluble Solids

The results of TDS obtained in the experimentation
follow a normal distribution and range from 1.87 to 6.57 %. The
collected data was analyzed in Design-Expert V11 software
using a quadratic model. Table 4 presents the significance of
the factors.

Table 4: ANOVA of the Quadratic Model of TDS.

p-Valor
Model < 0.0001 *
A (Water-to-coffee ratio) < 0.0001 *
B (Stirring speed) 0.4432
AB 0.5053
A 2
0.0249 *
B 2
0.0168 *
Lack of Fit 0.1008
*Significant influence, p-value < 0.05.
Figure 2: Response surface of extracted TDS. Water-to-coffee
ratio vs. Stirring speed.
According to the Table 4, it can be observed that the
obtained model for the process is significant. Similarly, it can be
Table 5: ANOVA of the linear model of yield extracts.
observed that the only significant factor on its own is the water-to-
coffee ratio while stirring speed is significant in the quadratic factor. p-Valor
Furthermore, the lack of fit result is non-significant, Model 0.0002 *
indicating that the proposed model will fit well and, therefore, A (Water-to-coffee ratio) < 0.0001 *
its replicates will be satisfactory. Additionally, a determination B (Stirring speed) 0.0892
coefficient of 0.87amd and adjusted R² value of 0.83 were C (Extraction time) 0.0780
obtained, indicating that despite the variations that may arise Lack of Fit 0.3247
from the non-significant factors excluded from the model, it
*Significant influence, p-value < 0.05.
mostly fits the experimental values.
These results reinforce the observed impact of the water-
to-coffee ratio, as detailed in Figure 2, where it is confirmed that  Yield 5.15206  0933063( A) (6)
this parameter significantly influences Total Soluble Solids (TDS)
A determination coefficient of 0.68 and an adjusted
attainment, with a 4:1 ratio proving optimal for maximizing solids
coefficient of 0.62 were obtained.
extraction during the coffee extraction process.
As observed in Figure 3, when lower water-to-coffee
Based on the above, Equation 5 proposes a formula
ratios are used, extraction yields are lower.
in terms of the significant factors present in the model. This
equation can be used to make predictions about the TDS 3.2.3 Extraction Rate
response obtained from different operating conditions.
In the 21 samples, extraction rates ranged from 50.76
TDS  2
8.73708  1.95285( A)  0.107263( B)  0.100286( A )  0.000893( 2
to
(5) B ) 621.87 g/h. Table 6 clearly shows that the extraction rate
100286( A2 )  0.000893( B 2 ) is significantly affected by the water-to-coffee ratio, stirring
speed, extraction time, and their interaction.
3.2.2 Extraction Yield The determination coefficient was 0.96 and the
Concerning the coffee extracts, extraction yields ranged adjusted correlation coefficient was 0.93, indicating a
from 7.79 to 22.22%. As seen in Table 5, the water-to-coffee good fit of the model (Equation 7) and suggesting that
ratio significantly influences the extraction yield, while stirring it can be used for accurate predictions of the extraction
speed and extraction time does not have a significant effect on rate.
this variable. Additionally, Equation 6 describes the extraction
yield as a linear model, with a coefficient of determination R² Extration rate 675.5696  27.9562( A)  2.4523( B)  22.3734(C )  (7)
of 61.97%, which is influenced by the factors B and C. 0.59598( AC )  0.18617(C )2

Coffee Science, 19:e192184, 2024

 BARRIGA, P. A. M. et al.

Figure 3: Response surface of extraction yield. Water-

to-coffee ratio vs. Stirring speed.

Table 6: ANOVA of the quadratic model of extraction rate.

p-Valor
Model < 0.0001 *
A (Water-to-coffee ratio) 0.0005 *
B (Stirring speed) 0.0466 *
C (Extraction time) < 0.0001 *
AB 0.0618
AC 0.0065 *
BC 0.2120
A² 0.9191
B² 0.7495
C² 0.0015 *
Lack of Fit 0.3247 Figure 4: Response surface of extraction rate (A) Water-to-Coffee
*Significant influence, p-value < 0.05. Ratio vs. Time (min) (B) Stirring Speed (RPM) vs. Time (min).

Figures 4A and 4B demonstrate that a 4:1 water-to- Table 7: ANOVA of the Quadratic Model for TA.
coffee ratio and a shorter 10-minute extraction time lead to p-Valor
elevated extraction rates. Furthermore, higher stirring speeds
Model < 0.0001 *
are linked to increased extraction rates.
A (Water-to-coffee ratio) < 0.0001 *
3.2.4 Titratable Acidity (TA) and pH C (Extraction time) 0.0005 *
The obtained extracts exhibited titratable acidity B² 0.0003 *
values ranging from 1.29 to 10.38 mg CGA/mL coffee. As
Lack of Fit 0.4118
shown in Table 7, the water-to-coffee ratio and extraction
*Significant influence, p-value < 0.05.
time have a significant effect on the model. Similarly
to the total soluble solids model, the quadratic term for
stirring speed shows a significant influence on titratable Equation 8 presents the non-hierarchical model for
acidity. determining titratable acidity (mg CGA/mL coffee) in the

Coffee Science, 19:e192184, 2024

 Effect of extraction processing parameters using a ribbon blender on the physicochemical properties of coffee

obtained extracts. The determination coefficient was 0.84 and ph 5.81542  0.003946(C ) (9)
the adjusted coefficient for the model was 0.79, indicating that
it can be used for accurate predictions. Figure 6 shows a clear linear correlation between
titratable acidity and the content of soluble solids present in
TA  4.91665  2.82898( A)  0.02864(C )  0.175409( A2 )  0.002575( B 2 ) (8) the extract, while no relationship is evident between pH and
the extracted solids.
Figure 5 illustrates that extracting at lower water-
to-coffee ratios and operating for one-hour results in higher
acidity values.

(A)

Figure 5: Response Surface of Titratable Acidity. Water-to-

Coffee Ratio vs. Extraction Time (min).

(B)
The coffee extracts had a pH range of 5.53 to 5.98 Figure 6: (A) Linear Regression of Titratable Acidity vs TDS.
consistent with findings reported by Muzykiewicz-Szymanska (B) Linear Regression of pH vs. TDS.
et al. (2021) for Colombian coffee extracts obtained under
similar time and temperature conditions. Their study indicated
pH values ranging from 5.03 to 6.28. Furthermore, an analysis Therefore, the relationship between titratable acidity
of variance was performed to assess the factors influencing pH and pH of the extracts was studied. Similar to the results
in relation to the studied variables. Table 8 reveals that the water- obtained by Gloess et al. (2013), Figure 7 shows no correlation
to-coffee ratio and stirring speed have no significant impact on between pH and titratable acidity, as the Pearson coefficient (r)
the pH of the extracts, while extraction time stands out as the was less than 0.5, and the correlation between them was not
sole factor with a notable influence on pH, demonstrating an statistically significant (Rao; Fuller, 2018).
inversely proportional relationship (Equation 9).
3.3 Operating Conditions and Desirability
Table 8: ANOVA of the Linear pH Model. Analysis
Based on the obtained results, a desirability analysis
p-Valor
was conducted to determine the operating conditions that
Model 0.0126 *
would maximize the percentage of soluble solids, extraction
A (Water-to-coffee ratio) 0.6165
yield, extraction rate, and titratable acidity in the coffee
B (Stirring speed) 0.5093 extracts with desirability of 74%, which were found to be a
C (Extraction time) 0.0017 * ratio of 4:1, 66 RPM, and 10 minutes. The expected values for
Lack of Fit 0.2743 the response variables as well as the experimental values are
*Significant influence, p-value < 0.05. presented in Table 9.

Coffee Science, 19:e192184, 2024

 BARRIGA, P. A. M. et al.

amount of water, the amount of solids to be extracted will

be greater and therefore the TDS will be higher. The higher
the solute gradient, the faster the mass transfer (Cordoba et
al., 2020). This ratio not only promotes increased diffusion
but also elevates the concentration of soluble solids, playing
a significant role in flavor development (Lingle, 2011).
A high TDS content and high extraction yield promotes a
strong bitter coffee. The balance between the water-to-coffee
ratio and the concentration of soluble solids is critical for
achieving the desired taste profile of coffee. In practice, a
Figure 7: Correlation between pH and Titratable Acidity. 4:1 ratio enhances the extraction of flavors and aromatic
compounds while ensuring a substantial concentration of
these elements in the final extract. The highest water -coffee
Table 9: Predicted and experimental operational conditions at
ratio, such as 10:1, may result in low brew strength. The
the optimal point.
highest TDS is desirable for an instant coffee production
Predicted values Experimental values process. The influence of the water-to-coffee ratio is further
TDS (%) 6.02 ±1.06 5.85 ± 0.05 evident in the extraction yield, as lower ratios lead to reduced
Yield (%) 11.15 ± 3.08 14.53 ± 1.48 yields. Despite a low ratio can increase a solvent retention
Titratable acidity within the food matrix, it results in higher levels of soluble
6.57 ± 2.65 5.62 ±0.19
(mgCGA/mL coffee) solids (Wankat, 2022; Guinard et al., 2023). Figure 3. Shows
Extraction rate (g/h) 499.99 ± 138.68 654.8 ± 66.8 that the highest extraction yield was obtained at the highest
water to coffee ration and the highest stirring speeds. This
result is consistent with the mass transfer theory. The higher
With the most favorable extraction conditions the concentration gradient and missing increase the mass
determined under the studied variables, three confirmation transfer rate. However. Only, the water to coffee ratio had a
trials were conducted. The average results obtained from significant difference. Extraction rates values were between
these trials were 5.85% TDS content, an extraction yield of 15 and 20%, which correspond to under-developed and ideal
14.53%, an extraction rate of 654.8 g/h, and a TA value of extraction yield respectively reported to coffee beverage
5.617 mg CGA/mL coffee. The TDS value is above the 1.5% (Batali; Ristenpart; Guinard, 2020).
value obtained in an extraction for coffee drinks. It is suitable Stirring is a common practice in coffee extraction,
for an instant coffee production process for subsequent known for its role in enhancing compound diffusion and
concentration or freeze-drying operations. The acidity value maintain concentration gradients to increase the mass
of 5.7 is comparable to those reported by Cordoba et al. (2020) transfer rate. However, our study suggests that stirring has
for Colombian coffee. The extraction yields are slightly lower a low influence in the solid’s extraction within the studied
than those reported for coffee drinks in drip brewers (Batali; range of agitation speeds, as reported by Dueñas-Rivadeneira
Ristenpart; Guinard, 2020). These results confirm that the et al. (2016).
chosen operating conditions, based on the desirability analysis Furthermore, it is evident in Table 5 that the water-to-
and optimization, reached the desired outcomes for the coffee ratio, stirring speed, and extraction time significantly
response variables. influence the extraction rate, as well as the interaction
between the water-to-coffee ratio and extraction time. This
4 DISCUSSION can be explained by the fact that the extraction technique in
a belt-based system involves complete immersion, ensuring
The results obtained in this study provide valuable thorough mixing of the roasted and ground coffee with water.
insights of new coffee extraction process made in a ribbon Additionally, the operating temperature helps facilitate the
blender unit. By examining various factors and their extraction of a greater amount of compounds from the food
interactions, we can better understand the nuances involved in matrix in a shorter time compared to extraction techniques
optimizing coffee process. at lower temperatures agreeing with the data obtained by
The water-to-coffee ratio had significance for all the Cordoba et al. (2021b).
response variables. A ratio of 4:1 is shown to be particularly A closer look at Figures 4A and 4B demonstrates that
effective in maximizing solids extraction, because of a 4:1 water-to-coffee ratio and a short extraction time, like 10
the higher amount of coffee solid that can be extracted. minutes, lead to higher extraction rates. This result is explained
Although the concentration gradient is lower with a smaller because the shorter the time the higher the rate at a constant

Coffee Science, 19:e192184, 2024

 Effect of extraction processing parameters using a ribbon blender on the physicochemical properties of coffee

TDS. Furthermore, the data corroborates with Zhang et al. consumption extracts by at least 81%, and although its
(2022) in which higher stirring speeds contribute to increased performance is approximately half of what is achieved with
extraction rates by improving mass transfer between coffee industrial equipment using high pressures and temperatures,
grounds and water. it remains a valuable alternative for efficient coffee extract
Figure 5 shows the total acidity of the extract. Values production for the obtainment of soluble coffee.
ranged from 4 to 10 mg CGA/mL coffee. Values are similar to
those reported by Cordoba et al. (2020) for Colombian coffee 5 CONCLUSIONS
and greater than those reported by Rao and Fuller (2018) for
Colombian coffee cold brewed. This result shows that the The investigation delved into the effect of operational
extraction process in a ribbon blender allows to obtain high parameters of the extraction with a ribbon blender on
acidity from the coffee. In the scope of Titratable Acidity (TA) the physicochemical properties of coffee. By varying the
and pH, our findings in Figure 5 reveal a direct correlation water-to-coffee ratio, stirring speed, and extraction time,
between lower water-to-coffee ratios and higher acidity values. we scrutinized their influence on key response variables.
This relationship can be explained by the preference for the The outcomes revealed a substantial effect of the water-
elution of acidic coffee components due to their high solubility to-coffee ratio on all response variables, except pH, while
in water, particularly at higher temperatures (Cordoba et al., stirring speed impacted the extraction rate and titratable
2020; Schwarzmann; Washington; Rao, 2022). Furthermore, acidity. Additionally, extraction time influenced both pH and
Figure 6 sheds light on a discernible linear correlation between extraction rate. Notably, the analysis of pH and titratable
titratable acidity and the content of soluble solids within the acidity unveiled their lack of correlation, emphasizing the
extract, reaffirming the viability of extracted soluble solids as distinct nature of the measurements. Employing a desirability
an effective metric for approximating titratable acidity in the analysis based on the obtained results, we identified optimal
context of coffee (Batali et al., 2021). conditions—85°C water, a 4:1 water-to-coffee ratio for 10
Figure 7 shows no correlation between pH and TA minutes, and a stirring speed of 66 rpm—yielding a coffee
as reported by Cordoba et al. (2020). Analyzing the pH extract with 5.85% soluble solids, an extraction yield of
range of coffee extracts, we find consistency with the 14.53%, an extraction rate of 654.8 g/h, and titratable acidity
results reported by Muzykiewicz-Szymanska et al. (2021). of 5.617 mg chlorogenic acid/mL coffee. In conclusion, the
The pH values of the extracts fall within a range of 5.53 operating conditions selected, optimized through desirability
to 5.98, corresponding to similar conditions. Our study analysis, successfully achieve the desired outcomes for
also investigates the factors influencing pH, revealing TDS, extraction yield, extraction rate, and titratable acidity.
the absence of a significant relationship between pH and The ribbon blender equipment studied can be used for
extracted solids. While both acidity measures offer insights coffee extraction on small scales, with a possible use for
into coffee’s acidity, titratable acidity proves to be a more the production of soluble coffee. It provides extraction
precise measure, accounting for acidic protons even when levels suitable for a further concentration or lyophilization
they aren’t fully dissociated. In contrast, pH measurement is operation. This simple and economical equipment that
related to dissociated hydrogen ions, offering a more limited operates in batch mode can be an alternative to traditional
perspective on the coffee’s acidity profile, as discussed by percolation systems, useful at high production scales.
Muzykiewicz-Szymanska et al. (2021) and Schwarzmann, These findings offer valuable insights into optimal
Washington and Rao (2022). operating conditions for coffee extraction at a in a ribbon
The investigation of the relationship between pH and blender unit, aligning with the specified objectives.
titratable acidity, reaffirms the lack of a meaningful correlation, Furthermore, they underscore the potential of ribbon mixer
consistent with earlier research findings. This outcome can be equipment for producing concentrated coffee extracts,
attributed to the likelihood that many acids in coffee extracts holding promise for small-scale industrial production of
may not be completely deprotonated, minimally influencing soluble coffees. A further comparison with a standard
pH measurements. commercial unit is recommended.
Nevertheless, a more comprehensive assessment of
their impact can be achieved through the measurement of 6 AUTHOR CONTRIBUTIONS
titratable acidity using an alkaline compound, as suggested by
Gloess et al. (2013). Conceptual Idea: F.L. Moreno; RY Ruiz; Methodology
The utilization of a pressure-free ribbon blender, while design: F.L. Moreno, Data collection: P.A. Mayorga, Data
not matching the performance of industrial high-pressure and analysis and interpretation: F.L. Moreno; RY Ruiz; P.A.
high-temperature equipment, presents an efficient alternative Mayorga, and Writing and editing: P.A. Mayorga, F.L. Moreno;
for coffee extract production, outperforming immediate RY Ruiz.

Coffee Science, 19:e192184, 2024

 BARRIGA, P. A. M. et al.

7 REFERENCES INSTITUTO COLOMBIANO DE NORMAS TÉCNICAS -

ICONTEC. NTC 2441:2011. Café torrado e moído. Método para
ANGELONI, G. et al. What kind of coffee do you drink? An determinação do tamanho médio de partícula por distribuição
investigation on effects of eight different extraction methods. granulométrica. 2011. Bogotá D.C., Colombia. 2011.
Food Research International, 116:1327-1335, 2019.
INSTITUTO COLOMBIANO DE NORMAS TÉCNICAS
BATALI, M. E. et al. Titratable acidity, perceived sourness, - ICONTEC. Café tostado y molido. Método para la
and liking of acidity in drip brewed coffee. ACS Food determinación del tamaño de partícula por distribución
Science & Technology, 1(4):559-569, 2021. granulométrica (NTC 2441), pp. 1-9, Bogotá D.C.,
Colombia. 2011.
BATALI, M. E.; RISTENPART, W. D.; GUINARD, J. X.
Brew temperature, at fixed brew strength and extraction, LINGLE, T. R. The coffee brewing handbook: A systematic
has little impact on the sensory profile of drip brew guide to coffee preparation. Specialty Coffee Association
coffee. Scientific Reports, 10:16450, 2020. of America. 2011

BENINCÁ, C. et al. Pressure cycling extraction as an MESTDAG, F.; GLABASNIA, A.; GIULIANO, P. The brew.
alternative to percolation for production of instant coffee. Extracting for excellence. In: FOLMER, B. (Ed.), The craft
Separation and Purification Technology, 164:163-169, and Science of Coffee, Academic Press, p. 355-380, 2017.
2016.
MORONEY, K. M. et al. Coffee extraction kinetics in a well-mixed
CAPEK, P. et al. Coffea arabica instant coffee-chemical system. Journal of Mathematics in Industry, 7:3, 2016.
view and immunomodulating properties. Carbohydrate
Polymers, 103:418-426, 2014. MUZYKIEWICZ-SZYMAŃSKA, A. et al. The effect of
brewing process parameters on antioxidant activity and
CORDOBA, N. et al. Coffee extraction: A review of caffeine content in infusions of roasted and unroasted
parameters and their influence on the physicochemical Arabica coffee beans originated from different countries.
characteristics and flavour of coffee brews. Trends in Molecules, 26(12):3681, 2021.
Food Science & Technology, 96:45-60, 2020.
OCAMPO LOPEZ, O. L.; ALVAREZ-HERRERA, L. M.
CORDOBA, N. et al. Chemical and sensory evaluation Tendencia de la producción y el consumo del café en
of cold brew coffees using different roasting profiles Colombia. Apuntes del Cenes, 36(64):139-165, 2017.
and brewing methods. Food Research International,
141:110141, 2021a. RAO, N. Z.; FULLER, M. Acidity and antioxidant activity of
cold brew coffee. Scientific Reports, 8:16-30, 2018.
CORDOBA, N. et al. Specialty and regular coffee bean
quality for cold and hot brewing: Evaluation of sensory SCHWARZMANN, E. T.; WASHINGTON, M. P.; RAO, N.
profile and physicochemical characteristics. LWT, Z. Physicochemical analysis of cold brew and hot brew
145:111363, 2021b. peaberry coffee. Processes, 10(10):1989, 2022.

DUEÑAS-RIVADENEIRA, A. et al. Determinación de las SULEWSKA, A et al. Advanced instrumental

condiciones de extracción de compuestos fenólicos a partir de characterization of the coffee extracts produced by pilot
chuquiraga jussieuijf gmel usando la lixiviacion de muestras scale instant coffee process. European Food Research
sólidas. Tecnología Química, 36(2):166-175, 2016. and Technology, 247:1379-1387, 2021.

FULLER, M.; RAO, N. Z. The effect of time, roasting VEZZULLI, F. et al. Sensory profile of Italian espresso
temperature, and grind size on caffeine and CGA brewed arabica specialty coffee under three roasting
concentrations in cold brew coffee. Scientific Reports, profiles with chemical and safety insight on roasted beans.
7(1):1-9, 2017. International Journal of Food Science & Technology,
56(12):6765-6776, 2021.
GLOESS, A. N. et al. Comparison of nine common coffee extraction
methods: instrumental and sensory analysis. European Food WANKAT, P. Separation process engineering: Includes
Research and Technology, 236(4):607-627, 2013. mass transfer analysis. 5th. Pearson. 2022. 1167p.

GUINARD, J.-X. et al. A new coffee brewing control chart ZHANG, L. et al. Extraction and physicochemical
relating sensory properties and consumer liking to brew characteristics of high pressure-assisted cold brew coffee.
strength, extraction yield, and brew ratio. Journal of Future Foods: A Dedicated Journal for Sustainability
Food Science, 88(5):2168-2177, 2023. in Food Science, 5:100-113, 2022.

Coffee Science, 19:e192184, 2024