Design and Production of an Instant Coffee Product Based on Greek Coffee Oil: Study of the Effect of Storage Conditions on Product Aroma and Quality

Abstract

The objective of this study was to obtain and evaluate a coffee aroma extract/oil with sensorial attributes close to the original brew of Greek coffee for use in an instant Greek coffee powder. The oil was obtained directly from commercial Greek coffee by solid-liquid extraction using hexane as a solvent and treated with a series of hexane-ethanol mixtures (0:10, 1:4, 1:9) to remove the intense roasted flavor of the crude coffee oil obtained by hexane; the de-oiled coffee was used for the recovery of water-soluble compounds, and the produced water extract was freeze-dried. The aromatic volatiles of the coffee oil samples were analyzed by using a purge-and-trap device coupled to GC-MS, as well as sensory analysis. The instant Greek coffee powder was produced by mixing the freeze-dried base (74.4%) with the extract derived after treatment of the crude oil with hexane-ethanol mixture 1:4 (18.2%) and foaming agent (7.4%). Two different materials were studied as bases: instant coffee (F3_Gr-D) and ground Greek coffee (reference sample, C_Gr). The shelf-life stability of the produced powders was examined at three storage temperatures (25, 45, 60 °C). Instrumental analysis (purge-and-trap GC-MS) of aroma and sensory analysis (aroma, taste, staling, total sensory quality on a 1–9 hedonic scale) was conducted. Aroma loss (furfuryl alcohol, furfural, dimethyl pyrazines, ethyl methyl pyrazines) and scores for sensory attributes during storage were modeled using 1st and 0-order reaction kinetics, respectively. The storage temperature effect was expressed by the Arrhenius model (activation energy E_a). According to the results, the developed instant coffee powder presented satisfactorily the aroma characteristics of regular Greek coffee. The shelf life for the instant Greek coffee powder was estimated as 80 days (air packed) (based on 20% retention of furfuryl alcohol that was the most abundant aromatic volatile of Greek coffee aroma, ground as well as extract oil).

1. Introduction

Coffee is one of the most popular beverages worldwide, with a major economic and social factor [1]. World coffee production for 2020/2021 was estimated to be 7.0 million bags (60 kg) higher than the previous year, reaching 175.5 million bags. There are two classes of coffee commercially: Coffea robusta (accounting for 25% of global production) and Coffea arabica (accounting for 75% of global production) [2]. While it comes in many different formulations and preparation modes, the commercially significant coffee product categories are roasted coffee (whole beans and ground), soluble coffee, single serve (pods and capsules), and liquid coffee (concentrates and ready-to-drink) (ICO, 2020, CECAFE, 2019) [3,4]. Among these, coffee brews prepared from roasted and ground coffee (either bulk or single-serve) are one of the most appreciated due to their aroma characteristics [5,6]. The Greek style of coffee known as “Greek coffee” is a mixture of Coffea Arabica and Coffea Robusta beans that are roasted and ground to powder. It is the most popular coffee beverage in Greece and is prepared by mixing the ground coffee and water in a special pot, heating, and directly pouring the heterogeneous mixture into a small cup. The final beverage includes the foam (on the top), the liquid part, and the spent coffee ground (sediment at the bottom, which is not being consumed). Similar coffee beverage is prepared, served and consumed in Turkey (known as “Turkish coffee”), and it is also found in the Middle East, North Africa, the Caucasus, the Balkans and various locations within Eastern Europe [7,8,9]. Research studies show that the consumption of this type of coffee beverage on a regular basis (more specifically, two to four cups of Turkish coffee per day) is related to many health benefits [8,10,11]. It is also reported that the consumption of boiled Greek coffee beverages is associated with improved endothelial function in the elderly [12].

Arabica and Robusta green beans have different chemical compositions (i.e., caffeine, chlorogenic acids, lipids, oligosaccharides, and polysaccharides) and, consequently, varying flavor profiles when roasted [13]. For example, green and roasted Robusta coffee beans have a higher chlorogenic acid content than corresponding Arabica beans. Chlorogenic acids contribute to the bitter taste, acidity, and astringent flavor of the coffee when it is brewed. Additionally, they act as precursors to the formation of phenols and catechols. The process and degree of coffee roasting (defined as light, medium, medium dark, and dark) not only is one of the most important determinants of the flavor, aroma, and volatile organic compounds (VOCs) profile in coffee but also may have a significant impact on the content of bioactive compounds [14]. Roasting profiles are described by many terms, depending on the region of coffee origin, consumer preferences, and the specific experience of the coffee roaster [15]. Beverages made from roasted Coffea Arabica beans are characterized by high acidity (due to higher sucrose content) and fruity aroma characterized by mild and harmonious flavor, whereas coffee brewed from Coffea Robusta beans is stronger and bitter and contains more caffeine characterized by earthy and raw flavor [16].

Coffee exhibits a complex, volatile profile that defines its characteristic aroma. The composition of volatiles is influenced by the coffee bean variety, climatic and soil conditions, cultivation, harvest and post-harvest storage, roasting conditions, and brewing techniques. It evolves during roasting and depends on the conditions of this process [14,17,18]. Reactions of decomposition of non-volatile compounds contained in raw coffee, i.e., pyrolysis, caramelization, and Maillard reactions, yield the final aroma [19]. Over 1000 volatile compounds have been identified in coffee, but only a fraction of them play a significant role in the aroma quality of the product [20]. Two major categories of aroma compounds could be distinguished, i.e., furans and pyrazines. Among the furan derivatives, furfuryl alcohol (2-furanmethanol) plays a pivotal role. Furfuryl alcohol imparts sweet, caramel and oily-like notes to the coffee’s aroma; however, the same compound has been correlated with undesirable burnt and bitter notes of dark-roasted coffees. Furfural is produced during the roasting of coffee and contributes to sweet, bread-like, and caramel flavors. However, as coffee ages, furfural increases since it is derived from the oxidation of furfuryl alcohol. Increased levels of the compound lead to undesirable staling notes, diminishing the beverage’s freshness [21]. Pyrazines, the second significant class of volatile compounds, contribute to its roasted and nutty characteristics. Compounds such as 2,6-dimethylpyrazine, 2-ethyl-3-methylpyrazine, and 2-ethyl-3,5-dimethylpyrazine are formed during the Maillard reaction and Strecker degradation of amino acids and sugars during roasting. Pyrazines impart roasted, nutty, and earthy aromas [5].

The detection and analysis of coffee aroma compounds relies on a range of advanced analytical methods, often employed in combination for improved sensitivity and resolution. Gas chromatography-mass spectrometry (GC-MS) is particularly well suited to address the complexity of coffee aroma for identifying VOCs in coffee due to its high separation efficiency and mass-based identification [22]. In recent years, solid-phase microextraction (SPME) coupled with GC-MS has become a widely used, solvent-free method for extracting headspace volatiles from coffee samples. Other complementary techniques include gas chromatography-olfactometry (GC-O) [23]. Τhose methods, often integrated with chemometric tools, provide a robust framework for profiling aroma compounds across coffee varieties, roast levels, and processing techniques. In addition to solid-phase microextraction (SPME), purge-and-trap (P&T) systems have been employed for the extraction of volatile compounds from coffee. In P&T application, an inert gas, especially nitrogen (N₂), is purged to the sample at a specific flow rate to extract VOCs, which are collected into an adsorbent trap. The trapped compounds are then desorbed using a thermal desorption unit. Different from other techniques, it results in a wider aroma profile, also allowing the extraction of hydrophilic and low vapor pressure volatile compounds even at ultra-trace levels [24].

From the perspective of convenience (easy to transport, store, and prepare), instant coffee products, such as soluble coffee and instant cappuccino, represent a good choice. However, instant coffee products face the drawback that reconstituted coffee beverages lack significant flavor and freshness characteristics due to aroma losses and degradations occurring during production [5,6]. To improve the aromatic quality of instant coffee products, a range of aroma recovery and reconstitution technologies and strategies have been developed [6,25]. An example is the incorporation of aromatic oil into either the liquid coffee concentrates before drying or directly spraying onto the dried soluble coffee [26]. Alternatively, the oil can be encapsulated and mixed as microparticles into the dried instant powder [27]. Nevertheless, roasted coffee oil contains unsaturated fatty acids, which make it susceptible to oxidative degradation. In addition, many of these aroma compounds present high volatility, being easily lost during storage [28,29].

Coffee extracts (coffee oil) are considered high-value products for beverage and alcoholic beverages, bakery, and even instant coffee industries [28,30,31]. Acylglycerols are the major components of coffee oil (80%), followed by unsaponifiable compounds such as diterpenes, sterols, and tocopherols (from 15% to 18%); free fatty acids (0.5 to 4.2%); waxes (1.5 to 2.5%) and hundreds of volatile compounds [27,31] responsible for the pleasant aroma of roasted coffee, such as pyrazines, furans, etc. formed during the roasting process [30,32,33,34]. Its composition depends, among other variables, on the species and cultivar, climate conditions and soil, post-harvest processing, the occurrence of coffee defects and the coffee extraction method [35].

Researchers attempted to isolate roasted coffee aromatic oil with an aroma as similar as possible to that of the original brewed coffee [34,36,37]. Common methods for extracting the oil from coffee beans include Soxhlet extraction with hexane over several hours. However, this procedure has an important drawback: long extraction time, consuming large amounts of solvent and requiring additional concentration steps [38]. Some research has been published on the supercritical extraction of oil from coffee beans as a source of aroma, leading to the most satisfying results [39]. Hydro distillation and steam distillation have been traditionally used for the extraction of essential oils (e.g., aromatic herbs). However, since coffee oil volatiles are the products of thermal processing (roasting), the respective compounds do not possess terpenoid structures and cannot be characterized as essential oils. Processes such as hydrodistillation and steam distillation can induce thermal degradation, hydrolysis and water solubilization of some volatile constituents [40]. No data has been published for oil extraction from ground coffee, and especially boiled coffee, such as Turkish-style, Greek-stylish, and Scandinavian-style coffee.

Instant coffee is the dried, water-soluble portion of roasted coffee. The manufacturing processes of instant coffee powder start with the roasting of green coffee beans, followed by a grinding process. The ground coffee beans produced are then extracted to get their aroma and flavor compounds, and then the extract is dried [3,14,16]. Freeze-drying and spray-drying are the most frequently used methods to produce instant coffee powder. The coffee drying method has also proved to have a significant effect on the aroma and flavor characteristics of instant coffee [41]. Micro and nanoencapsulation for the retention and controlled release of flavor and aroma compounds has been considered [42]. Considering that the roasted coffee oil obtained by extraction has aromatic properties and that the sensory and instrumental analysis is an important tool in the analysis of coffee aroma, the current work aimed to measure the influence of the addition of a mixture of the coffee oil to beverages prepared using samples of freeze-dried instant coffee. Sensory analysis was used to identify if this mixture could enhance the characteristic coffee aroma and flavor of instant Greek coffee beverages and consequently increase consumer preference for beverages prepared with instant coffee. Instrumental analysis was used to identify the volatile compounds of the aromatic profile of Greek coffee and their concentrations.

The present work included (1) the isolation and evaluation of coffee oil through the extraction method, (2) adding this to water-soluble bases in order to create an instant Greek coffee beverage, and (3) the study of the product quality and shelf-life stability through sensory testing and instrumental analysis.

2. Materials and Methods

2.1. Experimental Design

2.1.1. Raw Material

Ground roasted Greek coffee (a blend of Arabica and Robusta coffee varieties, light roasted, fine-grained powder) was purchased from a local supermarket with an expiration date of one year.

2.1.2. Coffee Oil Extraction and Treatments

Aroma extracts were prepared using solid-liquid extraction. A total of 200 g of ground coffee was mixed with 1000 mL of hexane (Sigma Aldrich Chemical Co., St. Louis, MO, USA). The mixture was stirred for 10 min at room temperature, allowed to stand for 45 min, and then filtered. The procedure was repeated 2 more times, and the hexane extracts were combined. The removal of the organic solvent was performed using a rotary vacuum evaporator (Heidolph, Hei-VAP Value Digital, Schwabach, Germany) at 50 °C. The obtained oil extract was dried with N₂ gas. The crude coffee oil was further treated to refine the initially aggressive “roasted” flavor of the extract. Three solvent systems were selected, namely ethanol 96° (BGS Alcohols S.A., Patra, Greece) (100% ethanol), ethanol-hexane mixture 1:4 (20% ethanol) and ethanol-hexane 1:9 (10% ethanol). A total of 10 g of oil extract was mixed with 20 mL of the respective solvent in a separating funnel. The mixture was agitated and allowed to stand for 1 h so that the refined oil would precipitate. The lower phase of oil was collected and subjected once again to the same procedure. Subsequently, the excess solvent was removed from the oil phase using a rotary vacuum evaporator, and the refined coffee oil was purged with N₂ gas to remove the traces of organic solvent. The oil was weighed and refrigerated prior to analysis.

2.1.3. Extraction of Coffee Water-Soluble Compounds

To obtain the water-soluble bases, two materials were used for water extraction: untreated Greek coffee powder (Gr) and coffee powder, which had already been de-oiled (Gr-D). In each case, 10 g of coffee was mixed with 70 mL of drinking water. The mixture was heated to the boiling point of water, cooled to room temperature, and filtered. The coffee residue was extracted as above for a second time. The two water extracts were combined, and the brew was then freeze-dried. The brew was initially frozen at −40 °C and then transferred to a Christ Alpha 1–4 LD Plus freeze-drier (Martin Christ GmbH, Osterode, Germany), where the vacuum pump reduced the pressure to 0.03 mbar. The temperature was lowered to −50 °C, the sublimation of solid water to gas commenced, and the cycle lasted 48 h. After the freeze-drying cycle, the moisture content was determined via the gravimetric method, and the two dried materials were ground to powders.

2.1.4. Formulation of Instant Greek Coffee Samples

The two powders produced by freeze-drying were used as base materials for the formulation of instant Greek coffee powders. The latter included one control sample and four samples with enhanced flavor. The Control sample of instant Greek coffee (C_Gr) contained the water-soluble powder obtained by water extraction of untreated Greek coffee (93%) and foaming agent (7%), while the instant Greek coffee samples with enhanced flavor (Fi_Gr-D, i = 1–4) were composed by the water-soluble powder obtained by water extraction of de-oiled Greek coffee (75%), coffee oil (crude or refined) (18%), and foaming agent (7%). According to the kind of coffee oil used, the samples of enhanced flavor were coded as F1_Gr-D, F2_Gr-D, F3_Gr-D, and F4_Gr-D [where 1: crude coffee oil; 2–4: treated coffee oil with 100% ethanol (2), ethanol-hexane 1:4 (3), ethanol-hexane 1:9 (4)]. Foaming powder was added to the bases, an additive generally found in hot instant drinks as it forms a layer of creamy foam and improves the texture of the drink (Aerion™ FOAM 110, DMW International by, Veghel, The Netherlands, ingredients: glucose syrup, milk solids, vegetable fat, stabilizers).

2.1.5. Storage of the Optimum Instant Greek Coffee

The Control instant Greek coffee (C_Gr) as well as the optimum instant Greek coffee with enhanced flavor (F3_Gr-D) were packaged in polyethylene-polypropylene bags (PE-PP) and stored at three different temperatures T, 25, 45, 60 °C. Sampling was carried out to monitor the product’s shelf life (

Table 1. Sampling plan to monitor product’s shelf life.

Samples	CGr			F3Gr-D
Storage Temperature (T)	25 °C	45 °C	60 °C	25 °C	45 °C	60 °C
Sampling time (t, days)	17	14	5	35	27	10
	35	27	10	55	48	16
	55	40	14			20
	73	48	16
			20

) [29]. Instrumental analysis (Purge-and-trap GC-MS) of aroma and sensory analysis (aroma, taste, staling, total sensory quality) was conducted (The experimental set-up was provided in Supplementary Files; Supplementary Figure S1).

2.2. Analytical Tools

2.2.1. Purge-And-Trap-Gas Chromatography-Mass Spectrometry Analysis

The profile of volatile compounds was obtained directly from ground coffee samples using purge-and-trap (P&T) (OI Analytical, Eclipse 4660, College Station, TX, USA)-gas chromatography (Agilent Technologies 7890A GC, Santa Clara, CA, USA) with HP-5 capillary column (30 × 0.25 mm i.d., coating thickness 0.25 μm)-mass spectrometry (Agilent Technologies 5975 C MSD, Santa Clara, CA, USA) [30]. The ground coffee samples were analyzed in the purge-and-trap concentrator under working conditions (based on the methodology described by Makri et al., 2011) [21]: gas flow 40 mL/min, purge time 20 min, purge temperature 80 °C, system pressure 22.8 psi, trap temperature (purge) 30 °C, trap temperature (desorption) 190 °C, desorption time 1,5 min, 6-port valve temperature 155 °C, transfer line temperature 150 °C, bake temperature 210 °C, water management temperature 130 °C, and septum purge flow 3 mL/min. Analyses in the GC-MS were performed afterward: injector temperature 180 °C, carrier gas nitrogen, flow rate 1 mL/min, split ratio 10:1, oven temperature programmed from 50 °C (0.5 min at constant temperature) to 70 °C at 5 °C/min and then to 160 °C at 10 °C/min, and post-run temperature 240 °C (5 min). Each sample was analyzed in triplicate. The responses of the compounds were quantified according to the total ion count (TIC) recorded for each peak of the chromatograms.

2.2.2. Mathematical Modelling

The most abundant volatile compound in studied Greek coffee samples (oil/freeze-dried base) was furfuryl alcohol [30]. The responses for the selected volatile compounds, furfuryl alcohol, furfural, and groups of compounds, dimethyl pyrazines (2,5-dimethyl pyrazine, 2,3-dimethyl pyrazine), and ethyl methyl pyrazines (2-ethyl, 6-methyl pyrazine, 2-ethyl, 5 methyl pyrazine, 2-ethyl, 3-methyl pyrazine) were calculated in relation to the furfuryl alcohol response according to Equation (1).

$$TIC={\displaystyle \frac{TIC}{{TIC}_{furfuryl alcohol}}}·100\mathrm{\%}$$

(1)

where TIC is the response of the selected volatile compound, and TIC_{furfuryl alcohol} is the response of furfuryl alcohol.

The dependence of the aroma loss on storage temperature is described by the following first-order equation (Equation (2)):

$${TIC}_{furfuryl alcohol}(\mathrm{\%})={exp(k}_{vol}·t)$$

(2)

where TIC_{furfuryl alcohol} (%) is the response of the furfuryl alcohol expressed as a percentage to the initial TIC_{furfuryl alcohol} response at t = 0 d, k_vol the rate of aroma loss and the storage time expressed in days.

The dependence of the model parameter (k_vol) on storage temperature was mathematically modeled using the Arrhenius equation (Equation (3)):

$$k_{vol/sens}=k_{{\displaystyle \frac{vol}{sens}},Tref}\exp \left(-{\displaystyle \frac{E_a}{R}}\left({\displaystyle \frac{1}{T}}-{\displaystyle \frac{1}{T_{ref}}}\right)\right)$$

(3)

where E_a is the activation energy of the parameter k, k_Tref is the deterioration rate at the reference temperature T_ref, and R is the universal gas constant.

The shelf life of each studied product, determined by aroma loss as a function of storage temperature, can be calculated as follows (Equation (4)):

$$S{L}_{vol}={\displaystyle \frac{\ln \left(\frac{{TIC}_{furfuryl alcohol, 0}}{{TIC}_{furfuryl alcohol, L}}\right)}{k_{vol}\exp \left(-\frac{E_{a_s}}{R}\left(\frac{1}{T}-\frac{1}{T_{ref}}\right)\right)}}$$

(4)

where $S{L}_{vol}$ the shelf life based on product aroma (in this case, furfuryl alcohol was selected as the aroma loss index), ${TIC}_{furfuryl alcohol,0}$ and ${TIC}_{furfuryl alcohol,L}$ the response was expressed as a % percentage during sensory evaluation at storage time t₀ (day 0) and at the end of storage (in this study, the limit was 20% retention of the initial).

2.2.3. Sensory Evaluation

Trained panelists of the NTUA sensory laboratory were chosen for the assessment of the sensory attributes of coffee beverage samples (ISO 8586-1,1993) [43]. Assessors (10 members of the sensory panel; female 6 and male 4; ages 22–44) were requested to verify the impact of the added oil on the instant coffee beverage (each sample triple, giving a total of 30 assessments per product). They evaluated the coffee aroma quality, the overall taste, and the overall/total sensory quality) using a hedonic/liking scale from 1 to 9 as well as the staling intensity using a scale from 1 to 9 [44,45].

For the sensory evaluation testing, hot distilled water was added to the soluble coffee together at the time of serving, and each panelist tasted the beverage. The coffee beverages were prepared fresh in a thermostatically controlled flask, and the temperature was kept constant until the sensory evaluation, but the time between the sample preparation and sensory testing was very short, less than 15 min, to minimize aroma loss. The panelists rinsed their mouths with water and waited for 2 min between samples. They were asked to taste the brew and to decide on what they liked compared to the traditional one. The instant drinks were prepared as follows: 70 mL of hot water was added to 2.2 g of instant product powders (F3_Gr-D) and stirred with a hand mixer. The reference sample (C_Gr) was prepared by adding 70 mL of water to 5.25 g of the sample into the kettle and boiling [46].

Shelf-Life Calculation Based on Sensory Quality

The dependence of the total sensory quality on storage temperature is described by the following zero-order equation (Equation (5)):

$$S\left(t\right)=S_0\pm k_{s}t$$

(5)

where S is the sensory score for staling, aroma, taste and/or total sensory quality at time t, S₀ is the respective sensory score at time zero, and k_s is the rate of score deterioration. The sign of k_s denotes the direction in which the deterioration occurs: if the score of an attribute increases with storage time, the sign is positive, whereas when the score decreases with time, the sign is negative.

The dependence of the model parameter (k_s) on storage temperature was mathematically modeled using the Arrhenius equation (Equation (3)), and the shelf life of each studied product, determined by the aroma loss, as a function of storage temperature, can be calculated as follows (Equation (6)):

$$S{L}_s={\displaystyle \frac{S_0-S_L}{k_{s_{Tref}}\exp \left(-\frac{E_{a_s}}{R}\left(\frac{1}{T}-\frac{1}{T_{ref}}\right)\right)}}$$

(6)

where $S{L}_s$ the shelf life based on product sensory evaluation, S and S₀ the scores for total sensory quality during sensory evaluation at storage time t (days) and t₀ (day 0), S_L the limit for total sensory quality (set as 5/9), and k_sTref is the deterioration rate based on sensory evaluation (days⁻¹).

2.2.4. Other Measurements

The color of the developed instant Greek coffee powders, at time zero and during storage, was measured in at least five replicates using an Xrite-i1 portable digital colorimeter (Gretag-Macbeth, Grand Rapids, MI, USA) and expressed in the CIE-Lab scale, where L-value: lightness; a-value: redness and greenness; and b-value: yellowness and blueness (CIE 1978) [35]. Color change during storage was expressed using the total color difference ΔΕ [36], given by the following equation (Equation (7)):

$$\mathsf{\Delta }E=\sqrt{{\left(a-a_0\right)}^2+{\left(b-b_0\right)}^2+{\left(L-L_0\right)}^2}$$

(7)

where L, a, and b are the measured CIELab color parameters and L₀, a₀, and b₀ are the measured color parameters at storage time zero.

Water activity (a_w) of the instant Greek coffee powders was measured using a water activity meter (Rotronic AM3 Hygrometer, Bassersdorf, Switzerland) as 0.3530 (Cgr) and 0.3240 (F3_Gr-D) (at storage time zero). The water activity of Greek coffee was also measured as 0.2614.

2.3. Data and Statistical Analysis

Results were expressed as means ± standard deviation of three experimental replicates. For the estimation of the main interaction effects of the investigated factors, factorial analysis of variance (Factorial ANOVA) was used. As a post-hoc analysis for the separation of means with significant differences (p < 0.05), Duncan’s multiple range test was used. For all statistical analyses, the Statistica 7 software (StatSoft, Hamburg, Germany) package was used. For all the mathematical regressions, the IBM SPSS Statistics Version 19 software package (IBM Corporation, Armonk, NY, USA) was used, and R² and standard errors of model parameters were calculated.

3. Results and Discussion

3.1. Analysis of Greek Coffee Oil Aroma

The volatile compounds determine the characteristic aroma of ground coffee. In ground Greek coffee, according to P&T GC/MS analysis results, over thirty volatile compounds have been detected, composing a very complex chromatographic profile. However, 15 compounds appear as the major volatiles and could be defined as the most important volatile compounds of Greek coffee aroma, while the rest present very low abundance or trace quantities. The major volatiles, as well as retention times and indices (determined according to the results of previous publications), aroma descriptors and odor detection thresholds are presented in

Table 2. The major volatiles of Greek coffee: retention times (r.t.) and indices (r.i.), aroma descriptors and odor detection thresholds (ng/g_water).

No.	Compound	r.t.	r.i.	Aroma Descriptor [c]	Odor Detection Threshold (ng/gwater)
1	2-methyl pyrazine	3.80	902	nutty, cocoa-like	60 [48]
2	furfural	3.94	913	pungent, sweet, bread-like, caramelly, cinnamon-almond-like flavor odor resembles benzaldehyde	3000 [48]
3	furfuryl alcohol	4.35	943	mild, warm, oily, burnt	5000 [48]
4	furfuryl formate	5.25	1010	ethereal
5	2,5-dimethyl pyrazine	5.37	1019	earthy, potato-like	1700 [48]
6	2,3-dimethyl pyrazine	5.54	1032	nutty, cocoa-like	2500 [48]
7	5-methyl furfural	6.38	1095	spicy-sweet, warm and slightly caramelly	6000 [48]
8	furfuryl acetate	6.94	1142	mild, ethereal-floral fruity	100 [48]
9	2-ethyl, 6-methyl pyrazine	7.07	1153	roasted baked potato	40 [48]
10	2-ethyl, 5-methyl pyrazine	7.12	1157	nutty, roasted, grassy	100 [48]
11	2-ethyl, 3-methyl pyrazine	7.12	1157	strong raw potato	130 [48]
12	3-ethyl-2,5-dimethyl pyrazine	8.47	1280		8.6 [48]
13	nonanal	8.88	1321	fruity	1 [48]
14	2,3-diethyl-5-methyl pyrazine	9.79	1418	nutty, roasted vegetable
15	decanal	10.61	1512	floral-orange on dilution	20 [49]

[47]. In all cases, the NIST matching factors ranged between 80 and 92%. The compounds can be classified into two categories, pyrazines and furans, while two aliphatic aldehydes were also detected, namely nonanal and decanal.

The pyrazines include three groups of derivatives, according to their alkylation pattern: the dimethyl pyrazines (2,3-dimethyl pyrazine, 2,5-dimethyl pyrazine), the ethyl-methyl pyrazines (2-ethyl-3-methylpyrazine, 2-ethyl-5-methylpyrazine, 2-ethyl-6-methylpyrazine), and the ethyl-dimethyl pyrazines (3-ethyl-2,5-dimethylpyrazine, 2-ethyl-3,5-dimethylpyrazine). Diethyl-methyl-pyrazines were also detected, but their responses were significantly lower than the rest of the groups of pyrazines, so they were considered as traces. Among the furans, furfuryl alcohol is the major, followed by 5-methyl-furfural, furfural, and furfuryl acetate. Furfuryl alcohol forms in substantial amounts during coffee roasting and corresponds to the most abundant volatile compound in Greek coffee [21,50]. Furfural, which is the oxidative product of furfuryl alcohol, has been used as a Greek coffee staling index [21].

The volatile profile of Greek ground coffee, as monitored by P&T-GC-MS, is presented in Figure 1. It can be observed that the dimethyl pyrazines, as well as the ethyl-methyl pyrazines, are co-eluting, making their quantification difficult according to their individual TIC values. For this reason, the total areas of the peaks were summed, and the respective TIC values correspond to the responses of total dimethyl pyrazines and total ethyl-methyl pyrazines. The use of the P&T equipment is intended to simulate the ideal aroma profile of Greek coffee as perceived by smelling the headspace of a recently opened sachet of the product. The aroma of the beverage differs from the sachet headspace aroma, which is reasonable since the different physicochemical properties of the compounds result in different extractability of the individual components. This generally applies to all kinds of ground coffees and the preparation of the respective beverages. Nevertheless, the aroma of the headspace could be theorized as the ideal and most desirable flavor for the respective beverage. For this reason, P&T was chosen as the equipment that could recover a profile of volatiles from ground Greek coffee that approaches the aroma of the headspace that the consumer smells.

As mentioned, furfuryl alcohol was the most abundant volatile, and this compound was chosen as a marker for determining the relative composition of the aroma components. The ratio between the MS responses of the components to the respective furfuryl alcohol (Response Rate, Equation (1)) provides an estimation of their relative concentration to furfuryl alcohol. The Response Rates of individual compounds and groups of compounds (furfural, total dimethyl pyrazines, 5-methyl- furfural, furfuryl acetate, total ethyl-methyl pyrazines and total ethyl-dimethyl pyrazines) of ground coffee sample, crude coffee oil (F1) and coffee oil samples after treatments with ethanol (Coffee oil F2), hexane-ethanol 1:4 (Coffee oil F3), hexane-ethanol 1:9 (Coffee oil F4) are presented in

Table 3. Response Rate values (mean value ± standard deviation) for ground coffee sample, crude coffee oil (F1), coffee oil samples after treatments with ethanol (Coffee oil F2), hexane-ethanol 1:4 (Coffee oil F3), hexane-ethanol 1:9 (Coffee oil F4) and the final instant coffee product with incorporated the optimal coffee oil (F3_Gr-D).

Compound	Ground Coffee	Crude Coffee Oil F1	Coffee Oil F2	Coffee Oil F3	Coffee Oil F4	Final Product
furfuryl alcohol	1	1	1	1	1	1
furfural	0.423 ± 0.045 * a	0.265 ± 0.023 b	0.249 ± 0.032 b	0.307 ± 0.032 b	0.313 ± 0.036 b	0.25 ± 0.04 b
5-methyl-furfural	0.118 ± 0.024 a	0.239 ± 0.023 b	0.305 ± 0.078 b	0.310 ± 0.017 b	0.305 ± 0.029 b	0.32 ± 0.06 b
furfuryl acetate	0.087 ± 0.008 a	0.113 ± 0.029 ab	0.092 ± 0.015 a	0.162 ± 0.048 ab	0.193 ± 0.032 ab	0.13 ± 0.02 ab
total dimethyl pyrazines	0.837 ± 0.094 a	0.332 ± 0.060 b	0.455 ± 0.098 c	0.535 ± 0.065 c	0.479 ± 0.033 c	0.83 ± 0.19 a
total ethyl-methyl pyrazines	0.496 ± 0.130 a	0.503 ± 0.053 a	0.514 ± 0.066 a	0.803 ± 0.177 b	1.046 ± 0.101 c	0.61 ± 0.18 c
total ethyl-dimethyl pyrazines	0.181 ± 0.018 a	0.329 ± 0.045 b	0.351 ± 0.047 b	0.617 ± 0.092 c	0.512 ± 0.052 c	0.450 ± 0.10 d
Sensory score for the total aroma quality	7.5 ± 0.1	3.2 ± 0.2	6.2 ± 0.4	6.5 ± 0.4	4.7 ± 0.3	7.2 ± 0.5

* Mean value ± standard deviation. Different superscript letters in the same column indicate significant differences between means as calculated by Duncan’s multiple range test for a significance level of p = 0.05. The 3-ethyl-2,5-dimethyl pyrazine was barely detectable.

.

The aromatic profile of the coffee oil sample presents some differences compared to the aromatic profile of the ground coffee sample. The relative total dimethyl pyrazine content of crude coffee oil F1 and coffee oil F2, F3, and F4 samples were lower than the respective of the ground coffee sample (p < 0.05). The relative 5-methyl-furfural content of crude coffee oil and coffee oil samples F2, F3, and F4 were higher than the respective of the ground coffee sample (p < 0.05). The relative furfuryl acetate content of all samples was not statistically significantly different (p > 0.05). The relative total ethyl-dimethyl pyrazines, as well as the total-dimethyl pyrazine content of coffee oil F3 and coffee oil F4 samples, were high compared to ground coffee and crude coffee oil samples. The highest values were calculated for the total ethyl-dimethyl pyrazines (from F1 to F3–F4). The furfural content of all (F2, F3, F4) coffee oil samples showed the least response rate differences compared to the ground coffee sample.

Pyrazines and furans have been used as quality indices for roasted coffee and black instant coffee beverages [34,50,51,52]. Pyrazines are associated with nutty, earthy, roasted, and green aromas [5,51,53,54]. Several pyrazines have been identified as crucial compounds contributing to coffee aroma due to their low sensory threshold values [55,56]. As far as furans are concerned, they have been proposed as responsible for the burnt sugar, burnt, and caramel aromas in roasted coffee [54] and for the caramel flavor in coffee brews [55]. Pyrroles have been identified as furan degradation products and amino acid derivatives, contributing to a distinctively sweet and slightly flared aroma in coffee [56]. Pyridines are formed through the Maillard reaction, resulting in a relatively subtle burnt and smoky aroma [35,38].

Apart from the instrumental analyses, sensory evaluation tests were carried out on an instant coffee base (after freeze-drying) with incorporated the extracted oils crude and treated ones to select the optimum that improves the sensory characteristics of the base. In Table 2, the sensory scores for total aroma quality are provided. Sensory panelists mentioned that the crude coffee samples presented an intense roasted flavor characterized as not desirable (3.2/9.0; F1). The volatile compounds of coffee oil that are water-soluble are isolated as dissolved in the water. For this reason, the use of ethanol, which is a less polar solvent compared to water, was selected for extraction. After extraction using ethanol, the obtained coffee oil had a similar flavor to the ground coffee flavor. According to sensory analysis, the extracted ethanol coffee oil got higher scores for aroma rather than the ground Greek coffee samples (6.2/9.0; F2). The resulting aroma of this oil was sweet and soft despite the intense roasty aroma mentioned before. The treatment with a mixture of hexane-ethanol in a ratio of 1:4 led to a higher score for the total sensory quality compared to the treatment with the respective mixture of hexane-ethanol in a ratio of 1:9 (6.5 vs. 4.7; F3 vs. F4). Both treatments (hexane-ethanol, 1:4 and 1:9) caused an increase in the total amount of furans and pyrazines (approximately 30% of the ground coffee), increasing the toasty aroma of the extracted oil. The optimal treatment was found to be the treatment with a mixture of hexane-ethanol in a ratio of 1:4 (F3), which achieves a balanced aroma. The developed instant product (as a result of F3) scored close to regular Greek coffee by the trained panel (7.2/9.0 compared to 7.5/9.0). Almost all volatile compounds in the instant coffee product had been extracted. The two main groups of furans and pyrazines are higher in the instant coffee product compared to the ground coffee (11–25%).

3.2. Shelf-Life Determination of the Instant Coffee Product

3.2.1. Sensory Quality Loss of the Instant Coffee Product

The evolution of sensory scores for staling, taste, aroma and total sensory quality, with storage time for all the storage temperatures studied for the C_Gr and F3_Gr-D samples, is presented in Figure 2. All tested sensory characteristics exhibited deterioration over storage time. Temperature increase caused significant sensory quality deterioration (p < 0.05). By fitting the zero-order kinetic model to the experimental data, it was possible to determine the rates of sensory deterioration at each storage temperature (k_s). The fitting results for both C_Gr and F3_Gr-D samples are presented in

Table 4. Quality degradation rates (based on total sensory quality, k_s, d⁻¹), and activation energy (E_a) values as calculated from the Equations (3) and (5) fitting to the experimental data, for the instant coffee brew, C_Gr and F3_Gr-D at storage temperatures T, 25, 45, 60 °C.

Sample	CGr		F3Gr-D
T (°C)	ks (d−1) *	R2	ks (d−1) *	R2
	staling
25	0.0218 aA	0.81	0.0753 aB	0.99
45	0.0386 bA	0.87	0.0930 aB	0.99
60	0.132 cA	0.95	0.2399 bB	0.84
Ea (kJ/mol) (R2)	37.2 (0.89)		22.1 (0.77)
	taste
25	0.002 aA	084	0.004 aA	0.99
45	0.006 bA	0.80	0.006 aA	0.99
60	0.019 cA	0.93	0.021 bA	0.79
Ea (kJ/mol) (R2)	49.8 (0.98)		30.1 (0.78)
	aroma
25	0.001 aA	094	0.007A b	0.99
45	0.007 bA	0.99	0.009 aA	0.99
60	0.020 cA	0.95	0.021 bA	0.93
Ea (kJ/mol) (R2)	62.9 (0.99)		22.7 (0.85)
	total sensory quality
25	0.0139 aA	0.99	0.0203 aB	0.92
45	0.0553 bA	0.94	0.039 bB	0.99
60	0.1434 cA	0.83	0.1045 cB	0.87
Ea (kJ/mol) (R2)	54.7 (0.98)		35.2 (0.94)

* Mean value. Different superscript lowercase letters in the same column indicate significant differences between means indicating the effect of storage temperature on k_s as calculated by Duncan’s multiple range test for a significance level of p = 0.05. Different superscript capital letters in the same raw indicate significant differences between means indicating the effect of instant coffee products on k_s as calculated by Duncan’s multiple range test for a significance level of p = 0.05.

.

The shelf life of both C_Gr and F3_Gr-D instant coffee brew products, determined by the total sensory quality (where a score of 5/9 was set as the total sensory quality limit), was calculated by Equation (5) (

Table 5. Shelf life (based on total sensory quality and aroma loss, SL, days) of the instant coffee brew C_Gr and F3_Gr-D at storage temperatures T, 25, 45, and 60 °C.

	Shelf Life (Days)
	Based on Total Sensory Quality SL = 5/9		Based on Aroma Loss 20%
Τ (°C)	CGr	F3Gr-D	CGr	F3Gr-D
25	215 aA	123 aB	268 aB	80 aB
45	54 bA	64 bB	107 bB	50 bB
60	20 cA	23 cA	50 cB	29 cB

Different small superscript letters (a–c) in the same column indicate significant differences, as calculated using Duncan’s multiple range test for a significance level of p = 0.05, expressing the effect of storage temperature on product shelf life. Different capital superscript letters (A,B) in the same line indicate significant differences between products for the same storage temperature.

). The instant coffee product F3_Gr-D presented half the shelf life compared to the Greek coffee (C_Gr, reference) product (T = 25 °C). This is mainly owed to the loss of aroma intensity. This was expected as the developed instant coffee products have resulted from the modification of the dry water-soluble bases to which the extracted coffee oil has been added. The oil has been absorbed into the coffee powder. However, as the two materials (water-soluble base and oil) have different polarities, each microparticle of coffee may be covered externally by a thin layer of oil. Therefore, in contrast to the classic Greek coffee, in which the oil and aroma components are evenly distributed throughout the mass of the bean due to the production method, in the modified instant coffee products, only a superficial dispersion of the aromas is expected. So, the increased shelf life of the reference Greek coffee product (based on aroma intensity) is considered reasonable as the losses of the aromatic substances locked in the microstructure of the bean become more difficult compared to the losses of surface-dispersed volatile substances in the water-soluble grains of instant coffee products. In addition to the reduction in aroma intensity, the overall taste and, consequently, the shelf life of the coffee products are also strongly affected by possible oxidation reactions. In this case, the autoxidation reactions of triglycerides become easier in surface-distributed oils in the grains due to stronger contact with atmospheric oxygen.

3.2.2. Instrumental Analysis of the Instant Coffee Product Aroma

The % of remaining furfuryl alcohol, furfural, dimethyl pyrazines (2,5-dimethyl pyrazine, 2,3-dimethyl pyrazine), and ethyl methyl pyrazines (2-ethyl, 6-methyl pyrazine, 2-ethyl, 5 methyl pyrazine, 2-ethyl, 3-methyl pyrazine) were quantified over time and mathematically modeled as expressed in Equation (2), for C_Gr, and F3_Gr-D instant coffee products (Figure 3;

Table 6. Aromatic compounds’ [furfuryl alcohol, furfural, dimethyl pyrazine (2,5-dimethyl pyrazine, 2,3-dimethyl pyrazine), and ethyl methyl pyrazines (2-ethyl, 6-methyl pyrazine, 2-ethyl, 5 methyl pyrazine, 2-ethyl, 3-methyl pyrazine)] loss rates (k_vol, d⁻¹) and activation energy (E_a) values as calculated from the Equations (2) and (3) fitting to the experimental data, for the instant coffee brew products C_igar and F3_Gr-D.

Τ (°C)	kvol (d−1) (R2)
	CGr	F3Gr-D
furfuryl alcohol
25	0.006 (0.944)	0.020 (0.939)
45	0.015 (0.962)	0.033 (0.998)
60	0.032 (0.949)	0.055 (0.972)
Ea (kJ/mol) kref (d−1)	39.3 (0.996) 0.005	23.5 (0.985) 0.019
furfural
25	0.006 (0.944)	0.030 (0.988)
45	0.015 (0.962)	0.043 (0.960)
0.1160	0.080 (9.89)	0.110 (0.967)
Ea (kJ/mol) kref (d−1)	59.1 (0.923) 0.005	29.3 (0.872) 0.026
dimethyl pyrazines
25	0.002 (0.998)	0.021 (0.968)
45	0.019 (0.928)	0.035 (0.956)
60	0.044 (0.993)	0.079 (0.938)
Ea (kJ/mol) kref (d−1)	74.1 (0.979) 0.002	30.4 (0.941) 0.019
Ethyl, methyl pyrazines
25	0.0004 (0.974)	0.014 (0.975)
45	0.009 (0.858)	0.028(0.961)
60	0.022 (0.960)	0.049 (0.979)
Ea (kJ/mol) kref (d−1)	96.7 (0.961) 0.0004	29.3 (0.997) 0.013

).

In Table 6, the k_vol rates for each temperature are also presented. In the case of 5-methyl furfural, furfuryl acetate and ethyl dimethyl pyrazines, the compounds presented poor correlation with the mathematical model and were excluded from the elaboration. The GC-MS analysis revealed differences in the degradation behavior of key volatile compounds between the two ground coffee samples, C_Gr and F_3Gr-D, during their storage. In all cases, the concentration of furfuryl alcohol, furfural, dimethyl pyrazines, and ethyl-methyl pyrazines decreased over time, with greater loss rates observed at higher temperatures, reflecting the thermally sensitive nature of these compounds. In general, the C_Gr product exhibited significantly lower degradation rates (k_vol) for all volatiles across the tested temperatures compared to the F_3Gr-D. For example, at 60 °C, the loss rate of furfuryl alcohol was 0.032 d⁻¹ in C_Gr, whereas it reached 0.055 d⁻¹ in F_3Gr-D. Similarly, dimethyl pyrazines in C_Gr degraded at a rate of 0.044 d⁻¹, in contrast to 0.079 d⁻¹ in F_3Gr-D. The calculated activation energy (E_a) values also supported this trend. C_Gr showed consistently higher _Ea values, indicating greater thermal sensitivity of volatiles: for instance, the E_a for ethyl-methyl pyrazines in C_Gr was 96.74 kJ/mol versus 29.38 kJ/mol in F_3Gr-D. These differences in chemical stability were reflected in the sensory evaluation results. C_Gr exhibited higher E_a values for both aroma loss (62.9 kJ/mol) and total sensory quality (54.7 kJ/mol), compared to the respective E_a values (22.7 and 35.29 kJ/mol) of F_3Gr-D suggesting a more stable sensory profile. In contrast, F_3Gr-D showed a faster degradation of aromatic and sensory attributes, correlating with the steeper decline in volatile compound levels at all storage temperatures.

The kinetic modeling of scoring for sensorial properties (total sensory quality) of the developed instant Greek coffee product was in good agreement with the modeling of coffee aroma loss (especially for the furfuryl alcohol; furfural and the other groups of compounds showed significantly high E_a, sensory and aroma loss, differences between C_Gr and F3_Gr-D). The activation energy (E_a) ranged from 35.2–54.7 to 23.5–39.3 to kJ/mol for aroma loss of C_Gr and F3_Gr-D, respectively. E_a values of F3_Gr-D calculated by the kinetic modeling of sensory scores for aroma, staling and overall acceptability were 22.7, 22.1 and 35.2 kJ/mol, respectively. The shelf life of both C_Gr and F3_Gr-D instant coffee brew products, determined by the aroma loss (20% retention of furfuryl alcohol), was calculated by Equation (5) (Table 5). The instant coffee product F3Gr-D presented a significantly lower shelf life compared to the Greek coffee (C_Gr, reference) product (p < 0.05) (80 days at the reference temperature T = 25 °C compared to 268 days). As already mentioned, the total sensory quality loss is mainly owed to the aroma intensity loss. In this study, the furfuryl alcohol loss rates of the developed instant coffee product, defined by the k_vol values, were significantly lower for all storage temperatures (25–60 °C).

Both sensory properties (total sensory quality, taste and aroma) and aroma volatiles presented a strong correlation with storage temperature. All studied sensory properties have the same effect as furfuryl alcohol. Staling is positively correlated with storage time, while in relation to the other sensory properties, it is negatively correlated, which indicates that the higher the rating of the classic product in terms of overall liking, taste and aroma, the lower the score it will receive in terms of staling, i.e., the fresher it is. Accordingly, the storage temperature can be negatively correlated with volatile substances, which indicates an opposite behavior of the two variables: as the temperature increases, the concentration of the selected aroma substance decreases.

3.2.3. Other Quality Parameters

The color of the developed instant coffee product F3_Gr-D did not show a statistically significant difference compared to the color of C_Gr at storage time zero (p < 0.05). The same was observed for the color of both products during their storage at temperatures from 25 to 60 °C. In

Table 7. Color parameters L, a, b and color change values ΔΕ for the developed instant coffee products C_Gr and F3_Gr-D at initial storage time (zero) and at the end of storage ) at 25, 45 and 60 °C.

Color Parameters *
Storage Time	L	a	b	ΔΕ
CGr
Day 0	52.09 a	0.60 a	−5.24 a	_
Day 55, 25 °C	51.85 a	0.70 a	−5.18 a	1.94
Day 48, 45 °C	52.33 a	0.63 a	−5.17 a	2.09
Day 20, 60 °C	52.31 a	0.62 a	−5.13 a	2.62
F3Gr-D
Day 0	51.25 a	0.20 b	−6.13 b	_
Day 73, 25 °C	51.66 a	0.42 c	−5.68 ab	5.30
Day 48, 45 °C	51.57 a	0.37 c	−5.96 b	3.26
Day 20, 60 °C	51.41 a	0.25 c	−6.15 b	0.16

* Mean values. Different superscript letters in the same column indicate significant differences between means as calculated by Duncan’s multiple range test for a significance level of p = 0.05.

, the measured color parameters L, a, b, and the calculated color change values ΔΕ were representativity given at initial storage time and at the end of storage. Total color difference (ΔΕ) value, which is calculated based on color differences compared to the initial product, increased during storage of both samples ΔΕ values ranged from 0 to higher than 2, indicating noticeable visual changes in color were perceived at later stages of storage [57].

4. Conclusions

The produced powder enriched with the developed Greek coffee oil extract could lead to a new instant coffee product having the unique aroma and flavor characteristics of Greek coffee. The oil was obtained directly from commercial Greek coffee by solid-liquid extraction using hexane as a solvent, further treated with a series of hexane and ethanol mixtures; the coffee oil-free residue was extracted with water and further freeze-dried (as described in the manuscript). The shelf life of the produced instant Greek coffee powder was estimated as 80 days (air packed) (based on 20% retention of furfuryl alcohol) and more than 123 days (based on the sensory rejection). The sensory analysis results confirmed that the main sensory attribute defining the sensory quality was the aroma loss intensity. Future research will focus on the shelf-life extension of the developed product (i.e., formulation, packaging).