Original Paper
The clean taste of coffee depends on the sodium content of its extraction water
2024 Volume 30 Issue 2 Pages 213-221
Details
2024 Volume 30 Issue 2 Pages 213-221
It is known that using high-purity water such as distilled water or Milli-Q water for extracting coffee changes the taste. Further, since the purity of ice manufactured by an ice-making system is also high, we investigated whether the taste would change when coffee was extracted using water produced from melted ice. Sensory evaluation showed that coffee brewed with the melted-ice water had an improved clean taste. Content analysis of chlorogenic acid, caffeine, cations and anions showed that the coffee brewed with melted-ice water had a low sodium content. When sodium was added as an organic-acid salt to the coffee extraction water, taste sensor analysis revealed that the salty taste increased and the clean taste of coffee decreased. Therefore, this suggests that the clean taste of coffee depends on the sodium content of its extraction water.
Coffee is one of the most popular beverages around the world; therefore, much research has been conducted on the water used for coffee extraction. Coffee extracted with water containing various ions (CO32−, HCO3−, Cl−, Mg2+, K+, Fe3+, Ca2+) shows a change in taste compared to that extracted with distilled water (Lockhart et al., 1955). A sensory test was performed on coffee extracted using distilled water containing 750 ppm of various minerals; the most bitter taste was found when the water contained NaHCO3 and the sourest taste when the water contained CaCl2. The coffee extracted with water containing NaCl became slightly salty; however, no significant difference was observed with the water containing CaCl2 or MgCl2 (Pangborn and Trabue, 1971). It has been reported that espresso coffee prepared with Milli-Q ultra-pure reagent grade water had a lower foam volume compared with Ca2+- and Mg2+-containing water. In addition, the espresso coffee foam volume depended on the bicarbonate ion content of the extraction water (Navarini and Rivetti, 2010). The Specialty Coffee Association of America (SCAA) Standard defines an optimum CaCO3 concentration of 17–85 ppm and a pH range of 6.5–7.5 as the common standard in the coffee industry. It is stated that total chlorine should be zero, since chlorine gas and hypochlorite impart an unpleasant flavor to coffee (Wellinger et al., 2017). Thus, it is known that the hardness and alkalinity of the extraction water affect the taste or foam formation of coffee. Generally, it is preferable that the coffee extraction water does not contain chlorine. In an experiment examining other beverages, Fuding white tea brewed in pure water had a higher catechin content than that brewed in tap water (Zhang et al., 2017). Green tea brewed in deionized water also showed a high epigallocatechin gallate content (Franks et al., 2019). These results indicate that water purity affects the extraction rate of the beverage composition during the brewing process. In past studies, distilled water, deionized water or Milli-Q ultra-pure grade water have been used in experiments; however, another pure and inexpensive water supply is required because of the expense of those waters for industrial application.
In the ice-making industry in Japan, ice is manufactured using tap water or ground water, which contains many minerals and impurities. The traditional method of making ice is known to be effective in removing impurities with air agitation in the water during ice manufacturing (Ponce-Cantōn and Sabirana, 1922; Cunningham and Mills, 1956). It is known that the central part of the ice block becomes cloudy due to the inclusion of air and impurities by the icemaking system. On the other hand, the peripheral part of the ice block is transparent and is composed of higher purity water than the raw water. Then, the cloudy central part of the ice block is removed, and the transparent part is processed into ice for cooling beverages in the Japanese ice-making industry. The content of impurities in the transparent part of the ice block is low; therefore, the purity of the water from this melted ice is considered to be high. This melted-ice water could be used to replace distilled water or ion-exchanged water for coffee brewing.
The evaluation of coffee taste has mainly been performed via sensory tests by assessors (Abdulmajid, 2014; Jaimes et al., 2015; Stokes et al., 2016). However, even for the same beverage, the evaluation of human sensory function depends on individual skill. Moreover, even for the same person, the results of sensory tests vary depending on daily physical and mental conditions. A multichannel test sensor whose transducer consists of lipid membranes can detect the taste of coffee in a manner similar to the human gustatory sensation (Fukunaga et al., 1996; Wu et al., 2020). The bitterness value of instant coffee detected by a taste sensor showed a good correlation with a human sensory bitterness study (Wu et al., 2020). This means that the taste sensor containing polymer membranes serves as the human tongue. The taste sensor is an important tool for objectively evaluating the taste of beverages. In this study, a taste test of coffee brewed using water from melted high-purity ice produced by an ice-making system was evaluated by a sensory test and a taste sensor.
Ice-making system A total of 146 L of well water was placed in a JIS standard ice can (H 1050 × W 560 × D 260 mm). The brine tank was adjusted to a Baume degree of 26 with calcium chloride, and the freezing temperature was −14 °C. An ice can filled with well water was placed in a brine tank, 30 kPa of air was blown, and ice was made over 12 hours with continuous blowing. The weight of the obtained ice was 125–130 kg. Since the center of the obtained ice was cloudy due to contamination with impurities, it was cut with a cutter and only the clear ice was recovered. The ice whose center part had been removed was crushed by a crusher. Crushed ice with particle sizes of less than φ10 mm was collected by passing through a 10 mm mesh. The collected crushed ice was thawed overnight at room temperature to make the melted-ice water. The recovery rate of the melted-ice water from the obtained ice was 86–90 %.
Determination of water physicochemical properties The physicochemical properties of the water types used for coffee brewing were determined by conductivity and pH analysis, including the main cation contents and chloride ion content. Conductivity and pH were determined by a conductivity meter and pH meter, respectively. The Na+ content and hardness (Ca2+ and Mg2+ content) were analyzed by inductively-coupled plasma atomic emission spectroscopy, and the Cl− content was analyzed by ion chromatography according to the Japanese drinking water quality standard methods (Ando, 1994).
Coffee brewing Ground roasted coffee beans (Coffea arabica from Brazil and Ethiopia, mocha blend.) was supplied by Key Coffee Inc. (Tokyo, Japan). A 28 g portion of ground roasted coffee beans was extracted at 78 °C with 515 mL of water using a EC-TC 40 coffee maker (Zojirushi Corp., Osaka, Japan). Four types of extraction water, melted-ice water, ground water (Iida, Japan), tap water (Iida, Japan), melted-ice water with 0.5 ppm NaClO, and ultra-pure water produced by a Milli-Q system (Merck Millipore, Burlington, MA, USA) were prepared. The sensory evaluation was performed immediately after coffee extraction, and the analysis of coffee components was performed after freezing at −20 °C and thawing.
Sensory evaluation Sensory evaluation was performed according to the Japan Coffee Qualification Authority (JCQA), which consists of four positive (round, clean, sweetness, astringency such as tea) evaluations and 3 negative (acrid, miscellaneous flavor, astringency) evaluations (Ishiwaki, 2007). Six volunteers (two male and four female) who underwent sensory test training by a private institution evaluated the taste of coffees brewed with different waters by a blind taste test. During the sensory evaluation, any behavior that may affect the evaluation was prohibited. Each tester recorded their evaluation points on a score sheet. Coffee evaluation terms as defined by JCQA are listed in Table 1. Scores indicate the number of positive evaluations as a plus and negative evaluations as a minus. The total score was calculated by adding all positive and negative scores.
| Evaluation | Taste | Definition |
|---|---|---|
| Positive evaluation | Round | Soft and pleasant on the tongue |
| Clean | The refreshing taste without any miscellaneous taste | |
| Sweetness | Feeling the sweetness on the tongue | |
| Astringency such as tea | Pleasant astringent taste of high-quality black tea | |
| Negative evaluation | Acrid | Stinging sensation on the tongue |
| Miscellaneous flavor | Leaving a rough feeling on the tongue | |
| Astringency | Unpleasant astringent taste |
Taste sensor analysis Taste sensor analysis was performed using a TS-5000Z taste sensing system (Intelligent Sensor Technology Inc., Atsugi, Japan). The sensor measurements were performed automatically according to the instruction manual with 8 different sensor probes (sourness, bitterness, astringency, umami, saltiness, after-taste from bitterness, after-taste from astringency, richness from umami). Sensor measurements were performed in quadruplicate from the standard solution to each coffee sample in sequence. A 30 % aqueous solution of ethanol containing 100 mM hydrochloric acid was used as the washing solution for the sensor chip. Each sensor chip was washed 5 times after the measurement and used in the next step. The intensity of each taste was defined as the potential difference between the coffee sample and the standard solution. Calibration was performed by confirming that there was a potential difference between the coffee sample and the standard solution during the measurement. Sensor outputs were obtained by averaging three measurements, with the exception of the first measurement taken to stabilize the sensor response. The data were analyzed for statistical significance by Student’s t test, where p < 0.05 was considered to be statistically significant.
HPLC analysis of chlorogenic acid Each coffee sample was stored at −20 °C until HPLC analysis. Thawed 6.0 mL coffee samples were mixed with 3.0 ml acetone and filtered through a 0.45 μm pore size filter. The filtered 10 μL samples were directly injected into the HPLC system. The HPLC system consisted of an HPLC 1525 (Waters Corp., Milford, MA, USA) and a YMC-Pack Pro C18 column (YMC Co., Ltd., Kyoto, Japan). The mobile phase consisted of methanol : water : acetic acid = 30 : 70 : 1 volume at a flow rate of 0.5 mL min−1. Chlorogenic acid detection was performed using the UV/VIS detection system at 330 nm at room temperature. The concentration of chlorogenic acid was compared with the standard reagent (Fujifilm Wako Pure Chemical Corp., Osaka, Japan).
HPLC analysis of caffeine The sample preparation method and injection volume were identical to those used for chlorogenic acid analysis. Thawed 5.0 mL coffee samples were mixed with 5.0 mL methanol and filtered through a 0.45 μm filter. The HPLC system consisted of an HPLC 1525 and InertSustain C-18 column (4.6 × 250 mm, GL Sciences Inc., Tokyo, Japan). The elution system involved two solvents: solvent A (0.1 % phosphoric acid, pH 2.7) and solvent B (acetonitrile: 0.1 % phosphoric acid, pH 2.7 = 40 : 60 volume). The gradient was as follows: 0 min, 0 % solvent B; 0–10 min, 20 % solvent B, linear; 10–60 min, 75 % solvent B, linear at a flow rate of 1.0 mL min−1. Caffeine detection was performed using the UV/VIS detection system at 272 nm at 40 °C. The caffeine standard was prepared from caffeine monohydrate (Fujifilm Wako Pure Chemical Corp.), which was dissolved in methanol. The concentration of caffeine in coffee samples was compared with the standard solution.
Cation analysis An atomic absorption photometer (Shimadzu AA-6200, Shimadzu Corp., Kyoto, Japan) and hollow cathode lamps were used (Na lamp and Mg-Ca lamp, Hamamatsu Photonics K. K., Hamamatsu, Japan). Hydrochloric acid (20 %) was added to the coffee samples to a final concentration of 1 % for the preparation of sodium measurement samples. Then, for the preparation of magnesium and calcium measurement samples, 20 % hydrochloric acid and 5 % strontium chloride in 20 % hydrochloric acid were added to coffee samples to a final concentration of 1 % hydrochloric acid and 0.5 % strontium chloride. The measured concentration values by an atomic absorption photometer required correction because 20 % hydrochloric acid and 5 % strontium chloride were added to the coffee samples. For conversion of the sample concentration, the sodium concentration was determined 20/19 times, and the magnesium and calcium concentrations were determined 20/17 times. Standard cation concentrations were 1.0, 3.0 and 5.0 ppm. The coffee samples were diluted 5 or 25 times to fall within the standard cation concentration range.
Anion analysis The ion chromatograph Dionex Integrion RFIC System was supplied by Thermo Fisher Scientific Inc. (Stoughton, MA, USA). Thawed coffee samples (10 μL) that were filtered using a 0.2 μm filter were directly injected into the ion chromatograph system. The eluent consisted of 15 mmol L−1 potassium hydroxide solution, and the flow rate was 0.25 mL min−1. The analysis column was a Dionex IonPac AS20 (Thermo Fisher Scientific Inc.), and anion detection was performed with a conductive detector. Cl−, NO3−, SO4− and PO4− concentrations were calculated by comparison with standards.
Organic acid analysis The organic acid concentration of the coffee sample was determined by a Prominence organic acid analysis system (Shimadzu Corp.). Each coffee sample was diluted with 0.5 % perchloric acid solution and directly injected into an organic acid analysis system (10 μL) after filtration through a 0.2 μm filter. The flow rate was 0.8 mL min−1 for both the mobile phase and the buffer phase. The mobile phase consisted of 5 mM p-toluenesulfonic acid. The buffer phase consisted of 5 mM p-toluenesulfonic acid, 0.1 mM EDTA and 20 mM Bis-Tris. Organic acid detection was performed using a conductivity detector CDD-10AVP and Shim-pack SCR102H column (both from Shimadzu Corp.) at 40 °C.
Water physicochemical properties for coffee brewing The physicochemical properties of melted-ice water, ground water (Iida City, Japan), tap water (Iida City, Japan) and melted-ice water with 0.5 mg L−1 NaClO are shown in Table 2. The water used for ice making is identical to the ground water in Iida City. NaClO (0.5 mg L−1) was added to the melted-ice water, as it is contained in the tap water and ground water for drinking. The purity of the melted-ice water was higher than that of the tap water and ground water, as the conductivity of the melted-ice water was lower than that of the tap water and ground water. Furthermore, the Na+, Mg2+ plus Ca2+ and Cl− concentrations of the melted-ice water were all below the detection limit, which also indicated that the melted-ice water is of high purity.
| conductivity (μS/cm) | pH | Na+ (ppm) | Mg2+ + Ca2+ (ppm) | Cl− (ppm) | |
|---|---|---|---|---|---|
| The melted water of ice | 4 | 5.95 | < 0.1 | < 1 | < 0.2 |
| Ground water | 102 | 7.69 | 11 | 30 | 1.7 |
| Tap water | 43 | 6.78 | 2.7 | 12 | 2.7 |
| the melted water of ice added with 0.5 mg L−1 NaClO | 4 | 7.00 | 0.2 | < 1 | 0.3 |
| Ultra-pure water | < 1 | 6.81 | N.D. | N.D. | N.D. |
N.D.: Not Detected.
Sensory evaluation of coffee extracted with different types of water In the sensory evaluation, the melted-ice water scored 11, which was the highest value of all the water types used for coffee extraction (Table 3). Comparing the positive evaluation of sensory variables, the melted-ice water had a clean score of 6, which was extremely high. In contrast, the addition of 0.5 mg L−1 NaClO to the melted-ice water for coffee extraction increased the negative evaluations of acridity and astringency (Fig. 1). It has been reported that chlorine, including when present in extraction water, negatively impacts the taste of coffee (Wellinger et al., 2017), and the present experiment showed similar results. The total scores for ground water and tap water were 3 and −2, respectively. The total score for the ultra-pure water was 4, which was lower than the total score of 11 for the melted-ice water. In a study using green tea, it was reported that when purified water is used to brew tea, the amount of catechin extracted increases and the flavor decreases (Franks et al., 2019).
| Water types | ||||||
|---|---|---|---|---|---|---|
| Evaluation | Taste | MW | GW | TW | MWH | UPW |
| Positive evaluation | Round | 3 | 3 | 2 | 3 | 3 |
| Clean | 6 | 1 | 3 | 3 | 3 | |
| Sweetness | 2 | 1 | 0 | 1 | 1 | |
| Astringency such as tea | 1 | 1 | 0 | 1 | 2 | |
| Negative evaluation | Acrid | 0 | 0 | −2 | −4 | −2 |
| Miscellaneous flavor | −1 | −2 | −2 | −1 | −1 | |
| Astringency | 0 | −1 | −3 | −4 | −2 | |
| Total score | 11 | 3 | −2 | −1 | 4 | |
*Each alphabetic characters indicates: MW, melted-ice water; GW, ground water; TW, tap water; MWH, melted-ice water with 0.5 mg L-1 NaClO; and UPW, ultra-pure water.
The sensory evaluation score of the coffee brewed with different waters.
For each alphabetic character indication, see the legend for Table 3.
Taste sensor analysis of different coffee extraction water samples Taste sensor analysis was conducted for the four coffees brewed with different extraction waters (Fig. 2). There was a significant difference (p < 0.01) in saltiness when comparing the melted-ice water with the other three extraction waters. On the other hand, in a comparison of the melted-ice water and the other three extraction waters, there were significant differences in after-taste from astringency (p < 0.05) and in richness from umami (p < 0.01) for the tap water and the NaClO-added water. A significant difference in after-taste from bitterness was observed only between the melted-ice water and the extraction water to which NaClO was added. Thus, only saltiness showed a highly significant difference between the melted-ice water and the other extraction waters.
Taste sensor analysis of the coffee brewed with different water types.
For each alphabetic character indication, see the legend for Table 3. ** indicates p < 0.01; * indicates p < 0.05.
HPLC analysis of chlorogenic acid and caffeine concentrations in coffee The concentrations of chlorogenic acid in coffee brewed with each extraction water (MW, GW, TW and MWH) were 0.86, 0.84, 0.91 and 0.97 mg mL−1, and the concentrations of caffeine were 0.91, 0.89, 0.94 and 1.01 mg mL−1, respectively; therefore, no significant differences were observed (Fig. 3). Chlorogenic acid and caffeine are typical bitter components in coffee (Campa et al., 2005; Fujioka and Shibamoto, 2008; Lugwig et al., 2014). Although there was a difference in taste between the coffees, there were no differences in chlorogenic acid and caffeine concentrations, indicating that the bitter-tasting substances did not affect the coffee taste. In addition, the concentrations of chlorogenic acid and caffeine in each coffee were consistent with the contents of typical regular coffee (Fujioka and Shibamoto, 2008).
The concentrations of chlorogenic acid and caffeine in coffees. For each alphabetic character indication, see the legend for Table 3.
Anion and cation contents in coffees brewed with different waters The cation and anion contents of the melted-ice water, ground water, tap water and melted-ice water with 0.5 mg L−1 NaClO are shown in Fig. 4. Cation analysis showed significant differences in sodium content (0.49, 7.02, 2.06 and 0.86 ppm, respectively). In contrast, no significant differences in magnesium (83.44, 77.19, 78.96 and 82.07 ppm, respectively) and calcium (14.55, 13.60, 13.49 and 14.24 ppm, respectively) contents were observed between the brewed coffees. It is suggested that the sodium content depends on that of the extraction water (Table 2). In contrast to the sodium content, the contents of magnesium and calcium were not significantly different between the brewed coffees, even though these cations had different contents in the extraction waters. Therefore, the majority of magnesium and calcium in coffee is thought to be derived from the coffee beans. In fact, calcium and magnesium are known to be extracted from roasted coffee powder to infuse the coffee (Stelmach et al., 2013). Anion analysis showed that the chloride ion content of brewed coffees (10.45, 11.96, 13.05 and 11.45 ppm, respectively) was dependent on that of the extraction water (< 0.2, 1.7, 2.7 and 0.3 ppm, respectively). When 11 ppm was added to the chloride ion content of the extraction water, it became almost equal to that of coffee. It is suggested that the 11 ppm chloride ions is derived from the coffee beans. The high content of nitrate ions in the coffee brewed with the ground water (4.35, 8.65, 4.72 and 4.65 ppm, respectively) simply reflects its content in the intact ground water. The sulfate ion content did not differ between coffees (17.82, 17.40, 19.17 and 17.90 ppm, respectively).
Cation and anion analysis of the coffee brewed with different waters.
The upper row shows the cation analysis, and the lower row shows the anion analysis.
For each alphabetic character indication, see the legend for Table 2.
Effect of different sodium ion concentrations in coffee Since the sensory test score was high for the melted-ice water, which was observed to have a low sodium content, following comparison of the coffees brewed with each different water type, we attempted to change the sodium ion content of the coffee. It is known that brewed coffee contains some organic acids (Jham et al., 2002). Since the brewed coffee originally contained a large amount of citric acid, sodium citrate was added to the extraction water to change the sodium content of the coffee. Melted-ice water with 0.5 mg L−1 NaClO containing 0.2 ppm Na+ tastes different from the melted-ice water (Tables 2 and 3). The sodium citrate concentration containing 0.2 ppm Na+ is 0.75 ppm. Therefore, the coffee was extracted by adding 0, 0.75 and 1.5 ppm of sodium citrate to the melted-ice water. Although the sodium ion content increased to 0.30, 0.52 and 1.06 ppm, the citric acid content was 390, 410 and 402 ppm, respectively, showing no significant change in the brewed coffees as a result (Fig. 5). This indicates that we succeeded in changing only the sodium ion content, without changing the citric acid content of the brewed coffee. Thus, sensory evaluation could be performed with coffee in which only the sodium ion concentration was changed.
Sodium ion and citric acid content in coffee.
The coffee brewed with the melted-ice water containing 0, 0.75 and 1.5 ppm of sodium citrate was analyzed with the taste sensor (Fig. 6). A significant difference (p < 0.01) was observed in saltiness between the water without sodium citrate and the water with sodium citrate added prior to coffee brewing. No differences were observed in other tastes (sourness, bitterness, astringency, umami, after-taste from bitterness, after-taste from astringency and richness from umami). In a sensory evaluation with 6 volunteers, the clean score decreased from 4 to 2 when 0.75 ppm sodium citrate was added to the melted-ice water for brewing coffee (Table 4). It is suggested that the sensory evaluation clean score depends on the sodium ion content, since the clean score was high when the sodium ion content was low. In fact, Tables 2 and 3 show that the waters that showed low clean scores in the sensory evaluation have high sodium ion contents. JCQA defines “clean taste” as “the refreshing taste without any miscellaneous taste”. In the sensory test, it is considered that the coffee was perceived as clean because the saltiness of sodium was perceived as a miscellaneous taste. When sodium citrate was added to the extraction water, the total score also decreased from 7 to 2. There was no difference in negative evaluation scores even following sodium citrate addition (Fig. 7). The difference in the values between MW (water types) of Table 3 and 0 ppm (sodium citrate addition amount) of Table 4 is thought to be due to differences in the physical condition of the sensory testers. In this study, we conducted not only sensory evaluation but also sensor analysis at the same time to examine differences in taste in the context of the physical condition of the tester.
Taste sensor analysis of coffee brewed with sodium citrate-added water.
Square boxes indicate the concentration of sodium citrate added to the extraction water. * indicates p < 0.05.
| Sodium citrate addition amount | |||
|---|---|---|---|
| Evaluation | Taste | 0 ppm | 0.75 ppm |
| Positive evaluation | Round | 4 | 3 |
| Clean | 4 | 2 | |
| Sweetness | 1 | 0 | |
| Astringency such as tea | 1 | 0 | |
| Negative evaluation | Acrid | 0 | 0 |
| Miscellaneous flavor | −1 | −1 | |
| Astringency | −2 | −2 | |
| Total score | 7 | 2 | |
The sensory evaluation score of the coffee brewed with sodium citrate added water.
For meaning of square boxes, see legend for Fig. 6.
In this study, since the concentrations of caffeine and chlorogenic acid in the coffees brewed with four types of extraction water were similar, it was possible to discuss the difference in taste based only on the difference in cation and anion concentrations. It is known that coffee brewed with distilled water containing NaCl is perceived as salty (Pangborn and Trabue, 1971). However, it was not known whether the saltiness came from sodium or chlorine. Since tap water contains chlorine, it was possible to examine the taste of coffee by removing the chlorine from the extraction water. For this reason, it is recommended that the water used to brew coffee be chlorine-free (Wellinger et al., 2017). Since pure sodium reacts violently with water, it is difficult to add this simple substance to the extraction water. In this study, we successfully added sodium to the extraction water as an organic-acid salt, since organic acids are abundant in coffee. As a result, the content of sodium ions could be changed without changing the organic acid content of the coffee. Taste sensor analysis showed that the taste of coffee became saltier when sodium ions were added to the extraction water. It was newly found that the saltiness of coffee increases not only when chlorine is present in the extraction water but also when sodium ions are contained. In addition, the addition of sodium ions to the extraction water was found to reduce the clean taste of coffee by sensory evaluation. Furthermore, the clean taste was found to depend on the sodium ion content, as coffee brewed with the melted-ice water, which has a low sodium ion content, scored higher in clean taste.
The melted-ice water manufactured by an ice-making system was used as pure water in this study. Distilled water and Milli-Q water are generally expensive; however, the melted-ice water can be produced at a low cost. This study demonstrated that the melted-ice water is suitable as a low-sodium-content water for coffee extraction.
Acknowledgements We thank the Nagano Prefecture General Industrial Technology Center for providing the taste sensor, HPLC system, atomic absorption photometer and organic acid analysis system.
Conflict of 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.
Funding sources This research did not receive any specific grants from funding agencies in the public, commercial or not-for-profit sectors.
CRediT authorship contribution statement Koji Imamura performed the experiments and wrote the paper.
