Effect of climate, growing region, country of origin, and post-harvest processing on the of content chlorogenic acids (CGAs) and aromatic compounds in roasted coffee beans

Abstract

The paper presents an analysis of the content of caffeine, tocopherols, and phenolic compounds as well as the volatile compound profile and volatile compound emission intensity in relation to the cultivation parameters of the Typica variety of Arabica coffee from Peru, Costa Rica, Guatemala, and Ethiopia. The study provides a detailed description of the cultivation and post-harvest parameters of the coffee types selected for the analyses. Special emphasis was placed on the analysis of the plantation altitude effect on instrumentally determined aromatic parameters and bioactive properties. The analyses were performed with the use of high-performance liquid chromatography, gas chromatography, and an electronic nose. The investigation results indicate a significant effect of the altitude of coffee cultivation on the most important biological and chemical properties of coffee beans, e.g. caffeine content, phenolic content, intensity of volatile compound emission, and coffee aroma. Other factors, such as shading and post-harvest processing, were also found to interact with the cultivation altitude and influence the content of these coffee attributes.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-16126-x.

Introduction

Coffee beans are one of the most widespread products of agricultural origin, and coffee beverages are an element of everyday culinary culture in most countries around the world. Coffee is the world’s second most traded commodity by volume, after crude oil, and its popularity is increasing. According to the USDA Foreign Agriculture Service (USDA Foreign Agriculture Service (16 December 2024)), the total world coffee production achieved over 169.18 million 60-kg Bags in the 2023/2024 season^1,2.

Coffee drinkers may soon see their morning treat get more expensive, as the price of coffee on international commodity markets has hit its record high level. In December, the price for Arabica beans, which account for the majority of global production, topped $3.44 a pound, having jumped more than 80% this year, while the cost of Robusta beans hit a fresh high in September. Coffee traders expect crops to shrink after the world’s two largest producers, Brazil and Vietnam, were hit by bad weather in addition to the popularity of the drink that continues to grow. This indicates that the value of the coffee market will continue to grow. Additionally, coffee consumption in China has more than doubled in the last decade.

Differences in sensory properties can affect consumer preferences and emotions or attitudes toward coffee drinking. Using chemometric methods and the GC-MS technique, Bressanello et al. (2017) have indicated that the coffee aroma is one of most significant quality attributes. However, most people drink low-quality coffee based on Arabica grown in a flat area at a low altitude in the Santos region in Brazil, to which Vietnamese Canephora (Robusta) with poor properties and taste is added. Therefore, most coffee brands found in retail chains probably contain fewer minerals and antioxidants than “specialty” coffee roasted excellently in craft roasters. Tieghi et al. (2024) studied variable factors and parameters in the coffee production chain that influenced its quality; however, a question should be asked about the most influential factor on the neuroprotective properties of coffee. The new methods of processing (natural, washed, honey, fermentation, and maceration) influence the volatile and neuroprotective compounds contained in coffee beans and infusions, which depend on bean quality, roasting methods, coffee aroma, climate and geographical location of the plantation, altitude, and variety^3–5. However, the impact of post-harvest processing, roasting level, grinding and brewing procedure, or potential synergy of all factors on health-promoting effects of coffee should be investigated. Kolb et al. (2020)⁶ studied the health effects of coffee consumption, while Wołosiak et al. (2023) and Farah and Lima (2019) analyzed the effects of consumption of chlorogenic acids with health benefits and risks. They found that the chemical composition of coffee is intricately linked to such factors as geographical origin, processing methods, and climatic conditions. High-quality coffee, particularly specialty varieties, is known for its superior flavor profiles. However, it is worth checking whether these coffee types have high levels of bioactive compounds, such as chlorogenic acids (CGAs) and antioxidants. CGAs are pivotal antioxidants, offering neuroprotective benefits and reducing the risks of chronic diseases, including Alzheimer’s and Parkinson’s⁹.

Coffee has long been considered a stimulant and an element of a not very healthy lifestyle. There is still a discussion about the effect of coffee on the risk of neurological diseases, cardiovascular diseases, type 2 diabetes, and even sudden vascular events, but the research results dispel the myth about the negative impact of coffee when drunk in recommended amounts on human health. A study conducted by Lee et al. (2020)¹⁰ indicated the high health-promoting importance of polyphenolic compounds and their neuroprotective effects on the brain. Despite many studies^11,12it is difficult to determine which of the nearly a thousand ingredients of coffee is mainly responsible for its beneficial effects. Possibly, a number of ingredients interact with each other and result in such positive effects. The aim of the study was to determine the impact of physicochemical changes occurring during the processing of coffee beans on the physical and sensory parameters of coffee as well as the content of antioxidant compounds and neuroprotective properties. The analyses were performed to elucidate these complex relationships and the impact of physicochemical processes occurring during all stages of coffee processing on the health-promoting properties of coffee beans.

As in the study carried out by Knysak 2017¹³, the goal of the first stage of the present research was to identify factors that influence the quality of coffee beans from different regions, the content of volatile compounds after roasting, and the taste of the resulting infusions. Based on previous results¹⁴the selection of coffee varieties gives an opportunity for chromatographic and mass-spectrometry verification and identification of the chemical composition of coffee infusions in order to determine the impact of cultivation conditions, location of plantations, harvesting methods, and coffee bean processing on their flavor and quality. An important task is to assess the content of phenolic compounds and the impact of the roasting process on the degradation or increase in the amount of these compounds, the health-enhancing properties of coffee beans, and the content of aromatic compounds in roasted beans.

Marek et al. (2020) and Rusinek et al. (2024a) selected varieties of coffee from diverse regions to assess the impact of all important cultivation and climatic factors on the quality of coffee beans used in further infusion processes, which facilitated confirming the hypotheses regarding the quality of coffee beans. This material did facilitate determination of the impact of fruit processing, fermentation, and drying methods, which affect not only the quality of coffee beans but also their health-promoting properties (e.g. the content of flavonoids and other polyphenols), or their potential harmful effects. Therefore, the main element of this study was the identification of the synergistic effect of all physicochemical and biological mechanisms influencing the aroma and health-promoting properties of coffee.

Results and discussion

Content of phenolic compounds, tocopherol, and caffeine

The analysis revealed significant differences in the total phenolic content in the coffee samples originating from the different countries (See Supplementary Materials). The highest content of total phenolic compounds expressed in gallic acid equivalents was detected in the coffee samples from Ethiopia (30.4 mg/g) (Fig. 1). The lowest content was found in the samples from Guatemala (24.1 mg/g). The Peruvian and Costa Rican coffee samples contained 28.9 mg/g and 28.2 mg/g of total phenolic compounds, respectively. These values ​​did not differ statistically from the Ethiopian samples. The analysis of the total phenolic compounds performed using the Folin-Ciocialteu reagent has certain limitations resulting from the specificity of this approach. This method does not distinguish phenolic compounds from other non-phenolic reducing agents, which may lead to overestimation of the total phenolic content¹⁶. Hence, in the next step, the content of individual phenolic compounds (CQAs) was determined using chromatographic techniques. The analyzed extracts contained 8 chlorogenic acid derivatives, i.e. 3-caffeoylquinic acid, 4-caffeoylquinic acid, 5-caffeoylquinic acid, 4-feruloylquinic acid, 5-feruloylquinic acid, 3,4-dicaffeoylquinic acid, 3,5-dicaffeoylquinic acid, and 4,5-dicaffeoylquinic acid. The content of individual CQAs and their total content are shown in Figs. 2 and 3.

Fig. 1.

Tocochromanol content in coffee beans from Ethiopia, Peru, Costa Rica, and Guatemala.

Fig. 2.

Heat map of volatile compounds and coffee beans from Ethiopia, Peru, Costa Rica, and Guatemala.

Fig. 3.

Projection of variables (volatile compounds, caffeine, bioactive compounds, tocopherols, and VOC emission levels) on planes PC1 and PC2.

Regardless of their origin, all the coffee samples had the highest content of 5-CQA (4.2 mg/g − 6.2 mg/g), followed by 4-CQA (2.1 mg/g − 3.3 mg/g) and 3-CQA (1.7 mg/g − 2.4 mg/g) (Fig. 2). In terms of the country of origin, the highest (statistically significant) content of these substances was determined in the samples from Ethiopia and Peru, followed by the Costa Rican coffee, whereas the lowest content was detected in the samples from Guatemala. The analysis of the content of FQA derivatives did not show any major significant differences between the samples. Only the content of 5-FQA differed significantly in the case of coffee from Peru. The content of di-CQA derivatives in the analyzed coffee samples was in the range of 0.19–0.32 mg/g (3,4-diCQA), 0.15–0.27 mg/g (3,5-diCQA), and 0.27 mg/g − 0.44 mg/g (4,5-diCQA) (Fig. 2). The samples from Peru exhibited the significantly highest content of these compounds, whereas their content in the Guatemalan coffee had the statistically significant lowest values. Figure 3 shows the total content of CQAs in the analyzed coffee samples. The highest content was found in the samples from Peru (13.8 mg/g), followed by the coffee from Ethiopia (13.7 mg/g), and these samples did not differ statistically significantly. The other samples had significantly lower amounts of these compounds. In total, 10.9 mg/g and 9.6 mg/g of CQA were determined in the Costa Rican and Guatemalan coffee, respectively.

The main phenolic compounds present in coffee, i.e. chlorogenic acids (CGAs), play a key role in determining its aroma and taste at various stages of production. CGAs are present in large amounts in green coffee beans, but their content is reduced during fermentation and roasting processes^17,18. During fermentation, phenolic compounds are metabolized by microorganisms, resulting in a decrease in the CGA concentration and an increase in the levels of volatile phenolic compounds, such as phenol, guaiacol, and vinylguaiacol¹⁷. This process exerts a significant effect on the sensory profile of coffee, i.e. it reduces coffee acidity and contributes to greater aroma complexity. During the roasting process, CGA is further degraded to quinic and caffeic acids, which are responsible for coffee astringency and bitterness. Additionally, free phenolic acids, such as caffeic acid or ferulic acid, undergo decarboxylation accompanied by generation of intense volatile compounds, e.g. guaiacol, p-vinylguaiacol, and phenols, which are involved in the characteristic smoky, clove-like, and spicy notes of roasted coffee¹⁹. The intensity of these notes depends on the roasting degree; dark-roast beans are characterized by increased amounts of phenylindanes, which contribute to the distinct bitterness of the beverage²⁰. In technological processes, some phenolic compounds are lost while others react with Maillard products, leading to further changes in coffee flavor and aroma²¹. Phenolic compounds contained in coffee beans are responsible for its multidimensional sensory profile and its different flavors and aromas; their quantity and quality depend on the processing methods and roasting degree.

The presence of two tocopherol homologs α-T and β-T was detected in the analyzed coffee samples; their content is shown in Fig. 4.

Fig. 4.

Projection of cases (coffee beans from Ethiopia, Peru, Costa Rica, and Guatemala) on planes PC1 and PC2.

*Values marked with different letters indicate statistically significant differences (p < 0.05) based on Tukey’s HSD test.

The analyzed coffee samples exhibited the highest content of homolog β-T, whose amount ranged from 5.7 mg/100 g to 8.0 mg/100 g. The content of homolog a-T was in the range of 1.8 mg/100 g to 2.5 mg/100 g. In terms of the country of origin, the data on the content of tocochromanols differed from the results of the CQA content. The highest tocopherol content was determined in the samples from Guatemala (in total 10.5 mg/100 g), whereas the lowest amounts of these compounds were found in the Ethiopian coffee (in total 7.5 mg/100 g) (Fig. 4). The samples from Peru and Costa Rica did not differ significantly in the content of homolog α-T, but there were differences in the level of homolog β-T. The lipid fraction containing triacylglycerols, terpenes, tocopherols, and sterols prevents aroma evaporation and loss during the roasting process. Furthermore, emulsified lipids present in espresso coffee infusions play a role of aroma carriers and texture promoters^22,23.

The caffeine content in all the analyzed samples was similar and ​​did not differ statistically significantly (Fig. 5). This is probably related to the characteristic traits of Arabica coffee varieties²⁴. During the roasting process, caffeine is very stable. This odorless substance carries bitterness and may contribute to this sensory characteristic in coffee beverages²⁵.

Fig. 5.

Projection of variables (caffeine, bioactive compounds, tocopherols, and altitude above sea level) on planes PC1 and PC2.

The investigated compounds were subjected to a correlation analysis, and the matrix obtained is ​​presented as a heatmap in Fig. 6. The analysis revealed a negative correlation especially between the content of total phenolic compounds, 3-CQA, 4-CQA, and 5-CQA and the content of the tocopherol homologs (α-T and β-T). The coefficients of correlation of the content of total phenolic compounds with tocopherols α-T and β-T were − 0.69 and − 0.76, respectively. In the case of 3-CQA, the correlation coefficient values were − 0.669 (α-T) and − 0.79 (β-T). These values were − 0.70 and − 0.83 for 4-CQA as well as -0.67 and − 0.79 for 5-CQA (with α-T and β-T, respectively).

Fig. 6.

Projection of cases (coffee beans from Ethiopia, Peru, Costa Rica, and Guatemala) on planes PC1 and PC2.

Analysis of selected volatile compounds

Table 1 ; Fig 7 show the percentages of volatile compounds in each coffee type. Noteworthy is the high content of 2-oxopropanal in the coffee sample from Peru. Its content in the other samples ranged from 2.90% to approximately 8% This compound, which has antibacterial properties, is also found in substantial amounts in some honeys²⁷. Methyl-D3 1-diderterio-2-propenyl ether is a characteristic compound present in Arabica coffee but does not occur in Robusta coffee. The lowest concentration of this compound was determined in the Ethiopian coffee samples.

Table 1.

Environmental and process conditions of coffee production

Country	Peru	Costa Rica	Guatemala	Ethiopia
Species	Arabica	Arabica	Arabica	Arabica
Variety	Typica	Typica	Typica	Typica
Bean Body	Strictly Hard Bean (SHB)	Strictly Hard Bean (SHB)	Strictly Hard Bean (SHB) Full, Round	Premium Cherry Selection
Meters a.s.l.	1600	1540	1650	2065
Region	Amazonas San Nicolas Rodriguez de Mendoza	La Pastora Estate, San Marcos de Tarrazu, Tarrazu Canton	Huehuetenango San Pedro Necta,	Sidamo/West Arsi, Oromia
Soil	Clay minerals soil	Soil richly volcanic	Clay minerals soil	Rich, fertile red soil
Wether and growing condition	Coffee growing In an agroforestry with slight water stress	Air often humid	coffee is shade-grown harvested at dry season	Semi-forest
Coast	Pacific	Pacific Central	Pacific Central	Central Rift Valley
Harvest period	March-September	March-September	January - March	January
Ripness	Gentle squeeze, if they feel a bit soft, they’re just right	Color change of coffee cherries from green to intense red.	coffee is harvested during the dry season when the coffee cherries are bright red, glossy, and firm.	Harvested at peak ripeness.
Harvest method	Coffee beans are harvested by hand, in which they can be strip picked or selectively picked	Beans are harvested by hand by coffee pickers. A very efficient coffee picker can fill 20 cajuelas (an official unit of measurement established by the Costa Rican government) each day	Harvest is primarily done by hand, where skilled workers selectively pick ripe coffee cherries from the trees.	Heirloom coffee is harvested by hand. Farmers select only the ripest cherries. This careful selection ensures quality. After harvesting, the beans are processed
Processing after harvest	Anaerobic fermentem for 48 to 60 hours	Traditionally Fully washed fermentaion	Fully washed the coffee is submerged in water and fermented for 48 to 72 hours	Fully Washed, hand sorted, then floated to further remove defects. Depulped. Wet fermented for 48-72 hours. Soaked in clean water for five hours.
Drying metod	Cherry is laid In thin layers on Bed, turned frequently up to 30 days for cherry to dry	With patios and elevated beds for varied kinds of drying	Sun-dried. Drying can take up to four weeks, and it is very tricky to ensure that no mouldy or off flavors get into the beans.	Dried on raised beds until moisture content reaches 10.5%.
Roasted been characteristics	balanced acidity, good body, and classic notes of sweet chocolate and nuts.	notes of cocoa, fruit, and an inherent sweetness. Often include chocolate, stone fruits, and berries	acidic, smokey and full-bodied characteristics with hints of chocolate and caramel	deep sweetness and dark fruit characteristics in addition to citrus, peach and florals

Fig. 7.

Heat map of volatile compounds and coffee beans from Ethiopia, Peru, Costa Rica, and Guatemala

A large amount of pyridine present in roasted coffee is produced via the pyrolysis of the alkaloid trigonelline^28,29. Pyridine is responsible for the bitter notes and vegetal aroma of coffee, which is advantageous for some consumers but undesirable for others. Its excess is harmful to the human organism. A darker coffee roast degree is associated with generation of higher levels of pyridine. Dominant or elevated amounts of pyridine are a drawback of the coffee roasting process³⁰. However, the pyridine content depends not only on the roasting degree but also on the variety and the cultivation region, as shown in Table 2 ³¹. Another mechanism of the generation of volatile chemical compounds in coffee is the degradation of chlorogenic acid (CGA) yielding such compounds as catechol, 4-ethylcatechol, phenol, benzoic acid, 2,5-dimethylfuran, 2-methylcyclohexanone, 1,2-cyclohexadione, furfuryl alcohol, pyridine, 2-cyclohexen-1-one, 3-ethylcatechol, and others. Many different and complex reactions are involved in the generation of volatile chemical compounds from chlorogenic acids. These mechanisms include dehydration, decarboxylation, acid deformation, keto-enol tautomerization, homolytic bond cleavage, and formation of free radicals³². As shown by the present results, the highest content of pyridine was determined in the coffee samples from Peru, whereas its lowest amounts were detected in the Guatemalan samples. The presence of butan-2-one was determined in the Ethiopian and Costa Rican coffee samples. This compound is responsible for the sharp hay-like or grassy aroma. The content of 2-methylpirimidine, representing the pyrazine group, was comparable in all the coffee samples, except for its slightly lower amounts in the Peruvian coffee. Other pyrazines (e.g. 4,6-dimethylpyrimidine, 2-pethylpyrazine), which give coffee roasted, nutty, cereal, cracker-like, or toast-like flavors, were detected in all the coffee samples. The highest content of 2-furancarboxaldehyde, a compound from the furan group responsible for the grassy and hay-like aroma of coffee, was detected in the Ethiopian coffee, whereas the Peruvian samples contained the lowest amounts of this compound. Another compound from the furan group is 2-furanmethanol, whose content was estimated at approximately 8%. It is responsible for the bitter flavor and generates a faint burnt aroma³³. The aldehyde 5-methylfuran-2-carbaldehyde generating a spice-caramel aroma reminiscent of maple was detected in all the coffee types. Its content was low and ranged from 2.62% in the Peruvian and Costa Rican coffee samples to 3.73 in the coffee from Guatemala. 2-furylmethyl acetate, i.e. furfuryl alcohol, may be responsible for coffee browning during the roasting process³⁴. Its lowest content was determined in the coffee from Ethiopia (2.22%), while it ranged from 3.43 to 3.75% in the other samples. Piperidine (1-methyl-2-cyano-2-piperidine) with an unpleasant odor typical of amines was present in all the coffee samples in a low range between 1.53% and 2.83%. N, N-dimethylpyridyn-4-amine is a pyridine derivative. The content of this compound ranged from 1.55% in the Ethiopian coffee samples to 2.82% in the samples from Guatemala.

Electronic nose results

Table 2. The results of the analyses of the intensity of volatile compound emission are presented in Table 3. The VOC emission intensity varied depending on the country of coffee origin and cultivation conditions, including the altitude a.s.l¹⁵. The maximum response ΔR/Rmax ​​was exhibited by the coffee samples from Ethiopia. This indicates high aromatic potential and diversity of the aroma bouquet of this coffee. Sensors detecting organic volatile substances and food aromas, i.e. TGS 2602, TGS 2603, and AMS-MLV-P2, exhibited the lowest response values in the case of the Guatemalan coffee. This may indicate the lowest aromatic potential of this variety. The results of the determination of the VOC emission intensity presented in Table 5 rank the coffee samples in terms of aroma strength from the most intense exhibited by the Ethiopian coffee (the highest cultivation altitude) through the coffee from Peru and Costa Rica to the Guatemalan samples³⁵.

Table 2.

Mean values and standard deviation (±) of chemical compounds detected in all Arabica coffees beans from Brazil, Ethiopia, Guatemala, Costa Rica, and Peru

No,	Name of compounds		Country of origin
		Rtime*	Ethiopia (%)	Peru (%)	Costa Rica (%)	Guatemala (%)
1	2-Buten-1-ol	1.12	2.12±0.03	2.44±0.10	2.37±0.13	2.76±0.20
2	2-oxopropanal	1.20	2.90±0.18	15.02±1.82	8.27±0.96	7.97±0.82
3	Methyl-D3 1-diderterio-2-propenyl ether	1.38	11.66±0.57	6.51±0.30	8.66±0.29	8.91±0.44
4	Acetalaldechyde	1.50	8.04±0.73	5.96±0.54	5.78±0.23	2.61±0.35
5	Pirydine	1.66	16.87±0.15	20.01±0.09	15.28±0.40	14.97±0.90
6	Butan-2-one	2.35	9.76±0.88	7.67±0.52	10.80±0.90	8.54±0.51
7	2-methylpirimidine	2.59	9.35±0.8	8.07±0.27	9.54±0.28	9.88±0.80
8	2-furanccarboxaldehyde	2.74	10.22±0.64	5.75±0.51	7.98±0.65	7.39±0.07
9	2-furanmethanol	2.82	7.94±0.50	8.07±0.29	8.05±0.29	7.60±0.87
10	Acetic acid ethenyl ester	3.26	3.23±0.06	4.08±0.25	4.58±0.25	4.84±0.11
11	4,6-dimethylpyrimidine	4.35	5.96±0.35	4.47±0.41	5.47±0.41	7.65±0.76
12	2-pethylpyrazine	4.47	2.72±0.73	2.11±0.37	3.38±0.37	3.61±0.22
13	Cis-ocimene	4.77	0,01±0.001	0.01±0.001	0.01±0.001	0.1±0.01
14	5-methylfuran-2-carbaldehyde	5.69	3.10±0.19	2.62±0.23	2.62±0.23	3.73±0.33
15	2-furylmethyl acetale	6.89	2.22±0.20	3.75±0.12	3.75±0.12	3.43±0.31
16	N,N-dimethylpyridyn-4-amine	7.11	1.55±0.09	1.72±0.22	1.72±0.22	2.82±0.25
17	1-methyl-2-cyano-2-piperidine	7.29	1.85±0.01	1.53±0.26	1.53±0.26	2.83±0.22
18	2,5-dimethyl-3-ethylpyrazine	9.99	0.30±0.03	0.21±0.001	0.21±0.001	0.79±0.03
Sum, Σ (%)			100	100	100	100

Table 3.

Mean values and standard deviation (±) of the maximum response - DR/R_max obtained using the Agrinose

Country of coffee bean origin
Sensor description	Ethiopia (mV/V)	Peru (mV/V)	Costa Rica (mV/V)	Guatemala (mV/V)
TGS2602	3.17±0.06	2.22±0.04	1.88±0.04	1.17±0.02
AS-MLV-P2	2.13±0.04	1.33±0.03	0.86±0.02	0.57±0.01
TGS 2603	1.77±0.04	1.36±0.03	1.05±0.02	0.65±0.01
TGS 2612	0.30±0.01	0.05±0.01	0.07±0.01	0.02±0.001
TGS 2610	0.58±0.01	0.28±0.01	0.22±0.01	0.14±0.01
TGS 2611	0.49±0.01	0.26±0.01	0.20±0.01	0.12±0.01
TGS 2620	1.12±0.02	0.54±0.01	0.42±0.02	0.25±0.01
TGS 2620	1.08±0.02	0.50±0.01	0.47±0.01	0.27±0.01

Analysis of volatile compound emission

Figure 8 shows the projection of the variables on planes PC1 (47.97%) and PC2 (25.90%), which describe 73.87% of the relationships between volatile compounds, caffeine, tocopherols, phenolic compounds, and the VOC emission level. The first important conclusion is the strong positive correlation of the level of volatile compound emission determined by the electronic nose (TGS2603, TGS2610, TGS2602, AS-MLV-P2, TGS2611, TGS2620, TGS2600, TGS2612) with the individual bioactive compounds (4-caffeoylquinic acid 4-CQA, 3-caffeoylquinic acid 3-CQA, 5-caffeoylquinic acid 5-CQA) and the content of total phenolic compounds (CQA)³⁶. There was a strong negative correlation of the level of VOC emission with tocopherols, 4-feruloylquinic acid 4-FQA, and, to a lesser extent, with the caffeine content. Similarly, a negative correlation was observed between some volatile compounds (cis-ocimene, 4,6-dimethylpyrimidine, 2,5-dimethyl-3-ethylpyrazine, 2-methylpirimidine, 5-methylfuran-2-carbaldehyde, 1-methyl-2-cyano-2-piperidine, 2-pethylpyrazine, N,N-dimethylpyridyn-4-amine, 2-buten-1-ol, acetic acid ethenyl ester, 2-furylmethyl acetate, 2-oxopropanal) and their emission intensity³⁷.

Fig. 8.

Projection of variables (volatile compounds, caffeine, bioactive compounds, tocopherols, and VOC emission levels) on planes PC1 and PC2

Figure 9 shows the projection of the variables on planes PC1 and PC2. Taking into account the VOC emission level determined by the electronic nose, it can be concluded that the first main component describes the aroma of coffees in terms of emission intensity in 47.97%. This classifies the coffee types from the lowest to highest emission level in the following order: Guatemalan, Costa Rican, Peruvian, and Ethiopian coffee. This conclusion is important in terms of consumer attractiveness³⁸as the sense of smell is one of the first senses assessing the attractiveness of coffee³⁰. This is followed by the consumer assessment of the aroma bouquet, i.e. the profile of volatile compounds.

Fig. 9.

Projection of cases (coffee beans from Ethiopia, Peru, Costa Rica, and Guatemala) on planes PC1 and PC2

Analysis of the effect of cultivation altitude on bioactive compounds

Figures 10 and 11 present the projection of variables on planes PC1 and PC2 (68.69% and 25.12%, respectively) explaining the relationships between caffeine, bioactive compounds, and cultivation altitude above sea level at 93.81%. The analysis indicates that the caffeine content in the coffee beans generally decreases with the increasing cultivation altitude. This trend has been observed for different coffee varieties grown in different regions^39–41. As shown in Fig. 10, the caffeine content is negatively correlated with the increasing cultivation altitude above sea level. A similar correlation can be found in the case of the tocopherol content. In contrast, a significant positive correlation is evident between the increasing cultivation altitude and the total phenolic content. As suggested in literature reports, this is associated with the slower rate of ripening processes occurring at higher altitudes above sea level⁴². In summary, the caffeine content is usually higher in coffee beans produced at lower altitudes and decreases with the increasing altitude but may vary depending on the geographical origin and coffee variety⁴³.

Fig. 10.

Projection of variables (caffeine, bioactive compounds, tocopherols, and altitude above sea level) on planes PC1 and PC2

Fig. 11.

Projection of cases (coffee beans from Ethiopia, Peru, Costa Rica, and Guatemala) on planes PC1 and PC2

The projection of cases on planes PC1 and PC2 presented in Fig. 11 shows that the first principal component PC 1 explains 68.69% of the relationship of the total phenolic content in the four analyzed coffee types with the cultivation altitude (see Table 1), with the lowest content in the samples from Guatemala, Costa Rica, and Peru and the highest content in the coffee from Ethiopia. As mentioned above, this relationship is positively correlated with the altitude of cultivation, which suggests that the altitude has a positive effect on the synthesis of phenolic compounds in coffee plants⁴⁴as stress conditions stimulate the production of these compounds.

Investigations reported by other researchers indicate that antioxidant activity in coffee tends to increase with altitude, which is closely associated with the presence of phenolic compounds. This further supports the thesis that higher altitudes increase the phenolic content in coffee (Abubakar, Alya et al., 2024). Some studies focus on specific phenolic compounds, e.g. chlorogenic acids, whose content declines with the increasing altitude. Nevertheless, the total phenolic content exhibits a general trend towards an increase at higher altitudes^39,45. This indicates a complex interaction between the altitude and the content of various phenolic compounds. Other factors, e.g. shading and post-harvest processing (Table 1), interact with the altitude as well and have an impact on the phenolic content, but primarily higher altitudes generally contribute to higher total phenolic content. Abubakar, Alya et al. (2024) found that coffee plantations located at higher altitudes tend to produce coffee with fewer defects and better quality attributes. Fig 7,8.

Materials and methods

Research material

Table 3 shows the cultivation and process data for the Typica variety of Arabica coffees from Peru, Costa Rica, Guatemala, and Ethiopia, i.e. parameters analyzed in this study. The data were provided by coffee producers, coffee suppliers, and the roastery where the coffee beans used in the study were roasted. The coffee beans were roasted at the Rovigo Caffee roastery in a Coffed SR 5 roaster equipped with a double-walled drum and coffee bean and exhaust temperature sensors. The computer-aided control and monitoring of the temperature inside the roaster, the temperature of roasted coffee beans, and the increase in coffee bean temperature, referred to as ROR (rate of rise), ensured achievement of the expected roasting degree.

All determinations described below in the methodology were performed for each coffee, taking at least three independent samples after roasting. Each sample taken was determined independently. In the event of a suspicion of a gross error (significantly outlying results), another sample was taken to repeat the determination.

Preparation of samples for determination of bioactive compound content

Coffee bean samples weighing 150 g were ground in a Russell Hobbs grinder to a grinding degree of 250 μm to 380 μm. Particles in the range of 280 μm to 320 μm (measured using the sieve method) accounted for the largest share (75%). The samples were stored in tightly sealed polyethylene bags at a temperature of approximately 4 °C.

Preparation of crude extracts of phenolic compounds

Phenolic compounds were extracted from the samples three times using an 80% aqueous methanol solution with a ratio of 1:3 (v/v). The samples were shaken for 30 min each time. The extracts were pooled and the solvent was evaporated using a rotary vacuum evaporator (Rotavapor-EL, Büchi Labortechnik AG, Flawil, Switzerland). The residue was quantitatively transferred using an 80% aqueous methanol solution into volumetric flasks with a capacity of 50 cm³ each.

Determination of total phenolic content

The content of phenolic compounds in the extracts and methanol infusions was determined with the colorimetric method based on the use of Folin-Ciocalteu reagent (Sigma-Aldrich, St. Louis, MO, USA). 5 cm³ of distilled water and 0.5 cm³ of Folin-Ciocalteu reagent were poured into a 10 cm³ volumetric flask. Next, 0.2 cm³ of the extract was added and mixed thoroughly. The solution was left for three minutes at 22 °C; afterwards, 1 cm³ of a saturated Na₂CO₃ solution was added and the flask was filled with distilled water to the mark. After one hour, absorbance was measured at λ_max = 725 nm. The total phenolic content in the coffee extracts and infusions was determined from the standard curve, and the results were expressed in gallic acid equivalents.

Isolation of phenolic compounds from extracts

The phenolic acid fraction was isolated from the crude extracts using the Chromabond^® system (Macherey – Nagle, Germany) with SPE Bakerbond spe™ columns filled with quaternary amine (500 mg). The process consisted of four stages: the first stage involved conditioning (10 cm³ of methanol, 10 cm³ of distilled water, and 10 cm³ of a 0.15% NaHCO₃ solution). The second stage involved loading the sample (5 cm³ into the columns. This was followed by the third step, i.e. washing the columns with the loaded sample (15 cm³ of a 0.15% NaHCO₃ solution), and elution of phenolic acids with a mixture of 0.2 M H₃PO₄ and 10 cm³ of methanol (2:1 v/v) in the fourth step. The pH value of the eluate was adjusted to approximately 3 by addition of 1 M NaOH Fig 9.

Identification of phenolic compounds

Figure 10 Phenolic compounds were identified and separated using high-performance liquid chromatography (HPLC – Waters Milford, MA, USA)¹⁵. The reversed-phase separation was carried out using an XBridge C-18 column (4.6 mm x 100 mm; 3.5 μm). The mobile phase consisted of aqueous acetonitrile (50% v/v) [A] and H₂O acidified with orthophosphoric acid to pH 2.7 [B]. The process was performed in the gradient mode. The concentration of phase [A] increased to 50% within 50 min; afterwards, its concentration returned to 1% over the next 10 min. The mobile phase flow rate was set at 1 ml/min. The measurements were made using a UV-Vis photodiode detector (Waters 2998) at a wavelength of 320 nm. The identification of CGAs was mainly conducted by comparison of their retention times with those of corresponding standards. These compounds were quantified using the calibration curve for the 5-CQA standard.

Determination of tocochromanol content in coffee beans

Two grams of ground coffee beans and 0.5 g of pyrogallol were placed in a flask. Next, 20 cm³ of ethyl alcohol (96%) and 2.5 cm³ of a 60% aqueous KOH solution were added. The sample was subjected to a 30 min saponification process under reflux at the solvent boiling point (78.4 °C). Afterwards, 50 cm³ of a 1% NaCl solution was added to the saponified sample, the mixture was cooled under running water, and 50 cm³ of n-hexane with 10% ethyl acetate was added. The mixture was shaken in a tightly sealed flask for 30 min with rotation (at 300 rpm). Next, 2 cm³ of a saturated NaCl solution was added to thoroughly separate the organic fraction containing unsaponifiable matter. After 15 min, an appropriately sized sample was taken from the upper organic layer with the extracted unsaponifiable substances for chromatographic analyses. The content of tocochromanols was determined using high-performance liquid chromatography HPLC (Waters 600 Asc. Milord)^15,46,47. A LiChrosorb Si 60 column (250 mm x 4.6 mm, 5 μm) and a fluorometric detector (Waters 474) were used for the analysis. A mixture of n-hexane and 1.4-dioxane (96:4 v/v) at a flow rate of 1.0 cm³/min was the mobile phase. The measurements were conducted at an excitation wavelength of λ = 295 nm and an emission wavelength of λ = 330 nm. The compounds were identified by comparison of the retention times of individual peaks in the chromatograms of the analyzed samples with those of standard solutions.

Determination of caffeine content

The caffeine content was determined in crude extracts and coffee infusions using the same chromatographic system as for the determination of the content of phenolic compounds⁴⁸ (Sect. 2.6). The measurements were made at a wavelength of λ = 270 nm with the use of a UV-Vis photodiode detector (Waters 2998). The caffeine content was calculated from the peak area and the standard curve for caffeine.

Electronic nose

An electronic nose constructed at the Institute of Agrophysics of the Polish Academy of Sciences in Lublin^49,50 was used for the analyses. The device consists of eight MOS sensors for detection of the following compounds: AS–MLV-P2 – designed specifically to detect volatile organic compounds: CO, butane, methane, ethanol, and hydrogen; TGS2602 – ammonia and hydrogen sulfide (high sensitivity to VOCs and odorous gases); TGS2603 – odors generated by spoiled food; TGS2612 – methane, propane, and butane; TGS2610 – liquefied petroleum gas and butane; TGS2611 – natural gas and methane; TGS2620 – solvent vapors, volatile vapors, and alcohol; TGS2600 – general air contaminants, hydrogen, and carbon monoxide). In the study, the maximum responses of the sensors to emissions of VOCs (volatile organic compounds) were analyzed⁵¹. Each coffee sample was placed in a special container for analyses of volatile compounds (Figaro, Japan). The measurement cycle and the sampling protocol included 10 s baseline generation, 60 s VOC sampling, and 140 s matrix purification from the volatile compounds. Analog signals were converted to digital signals using DasyLab software. The sensorgram obtained was converted to the ∗.xls format and analyzed with the use of Statistica software (version 12.0, StatSoft Inc., USA).

GC-MS analysis

Figure 11 The intensity of signals of the volatile organic compounds contained in the coffee was determined with the use of a Trace GC Ultra gas chromatograph coupled with an ITQ 1100 mass spectrometer (ThermoFisher Scientific, Waltham, MA, USA). SPME fibers with an absorbent (50/30 µm Divinylbenzene/Carboxene/Polydimethylsiloxane (DVB/CAR/PDMS), Stableflex (2 cm) 24 Ga (Sigma Aldrich, Poznań, Poland)) were placed (solid phase microextraction) together with VOC-emitting coffee samples in the measuring chamber for 30 min. Next, the fibers were transferred to the GC injector for 5 min for desorption of VOCs. A Zebron ZB-5Msplus Capillary GC column 30 m x 0.25 mm x 0.25 μm was used for the analysis. The injection temperature was maintained at 60 °C for 5 min and then increased from 60 to 250 °C at a rate of 5 °C/min and from 250 °C to 270 °C at 10 °C/min. The final temperature was maintained for 5 min. The helium flow rate was kept constant at 2.2 ml/min. The compounds were identified using the Wiley library⁵².

Statistical analysis

Statistica software (version 12.0, StatSoft Inc., Tulsa, OK, USA) was used for statistical analyses. Principal component analysis (PCA), analysis of variance, and determination of correlations were performed at a significance level of α = 0.05. The PCA data matrix for the statistical analysis of the results of the tests had 29 columns (volatile compounds, tocopherols, caffeine contents, volatile emission intensity, bioactive compounds, a.s.l.) and 12 rows: coffee from Ethiopia, Peru, Costa Rica, and Guatemala. The input matrix was scaled automatically. The optimal number of principal components obtained in the analysis was determined based on the Cattel criterion.

Using Statistica software, an analysis of variance (ANOVA) was performed to determine significant differences in the content of phenolic compounds among the coffee samples from the different regions. In cases where significant differences were detected, Tukey’s post-hoc test was applied. Additionally, Principal Component Analysis (PCA) was conducted to identify variables that contributed most to the sample differentiation. Further statistical calculations and visualizations were performed in Python, utilizing pandas, scipy, statsmodels, seaborn, and matplotlib libraries. For graphical representation, box plots were used to illustrate the distribution of data, medians, and outliers, while a heatmap was employed to visualize the differences in phenolic compound concentrations across the regions. The heatmap was generated using the “coolwarm” color scale, where warm colors indicated higher concentrations and cool colors represented lower values.

Conclusions

The highest levels of tocopherols α-T and β-T and their total content ​​were determined in the Guatemala coffee bean from Huehuetenango SanPedro Necta region, cultivated at 1650 m. a.s.l. Lower values were exhibited by Costa Rica coffee bean from La Pastora Estate Tarrazu Canton, cultivated at 1540 m. a.s.l. and Peru coffee bean from Amazonas region, cultivated at 1600 m. a.s.l., while the lowest tocopherol content was found in Ethiopia premium cherry coffee bean, cultivated at 2065 m. a.s.l., which proves negative correlations with the cultivation altitude.

The present study indicated that the caffeine content in coffee beans generally decreased with the increasing altitude. In turn, a significant positive correlation was found between the increasing altitude and the total phenolic content. Other factors, such as shading and post-harvest processing, also interact with the altitude and have an impact on the phenolic content. Nevertheless, higher altitudes generally tend to contribute to higher total phenolic content. Noteworthy, however, the phenolic content in coffee is also a result of complex interactions between environmental conditions, ripeness at harvest, harvesting methods, and post-harvest processing techniques.

The intensity of VOC emissions varied depending on the country of origin and coffee cultivation conditions. The analyzed samples were ranked from the lowest to highest emission levels as follows: Guatemala Standard, Costa Rica Terrazu, Peru El Palto Organic, and finally the Ethiopian coffee. This indicates high aromatic potential and diversity of the aroma of the Ethiopian Sidamo coffee, compared to the other coffee types.

Data Availability

Some of the data used to develop the research results are confidential due to suggestions from Coffee Roaster Rovigo Caffee, the research partner. Some of the data that support the findings of this study are available from the authors upon reasonable request. Author to contact: Prof. Robert Rusinek (r.rusinek@ipan.lublin.pl).

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Supplementary Information The online version contains supplementary material.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

R. Rusinek: Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. A. Siger: Writing – original draft, Visualization, Methodology, Investigation. B. Dobrzański: Writing – original draft, Conceptualization. M. Gawrysiak-Witulska: Methodology, Investigation, Formal analysis. M. Gancarz: Investigation, Formal analysis. A. Oniszczuk: Data curation, Conceptualization, S. Tabor: Data curation.

Funding

This research received no external funding.

Correspondence and requests for materials should be addressed to R.R.

Data availability

Some of the data used to develop the research results are confidential due to suggestions from Coffee Roaster Rovigo Caffee, the research partner. Some of the data that support the findings of this study are available from the authors upon reasonable request. Author to contact: Prof. Robert Rusinek (r.rusinek@ipan.lublin.pl).

Declarations

Competing interests

The authors declare no competing interests.