Influence of Milk on Aroma Release and Aroma Perception during Consumption of Coffee Beverages

Abstract

The aim of this study was to investigate the influence of milk on the aroma release and aroma perception of coffee. The amounts of in-mouth odorants exhaled through the nostrils during the consumption of black coffee and milk coffee were compared using the Retronasal Flavor Impression Screening System (R-FISS) by 3 trained panelists. As a result, it was found that the amounts of most in-mouth odorants including the potent odorants in brewed coffee did not significantly differ between the black coffee and milk coffee. However, the amounts of furfuryl methyl sulfide (FMS), difurylmethane, and furfuryl pyrrole exhaled through the nostrils during the consumption of milk coffee were significantly lower than those of the black coffee regardless of the panelist. It has been previously indicated that FMS could result from the methylation of 2-furfurylthiol (FFT), one of the most important potent odorants in brewed coffee. Based on these results, FFT and FMS were added to milk coffee so that the amount of FMS exhaled through the nostrils was about the same as that of black coffee, resulting in improved intensity of the coffee-like aroma quality for the milk coffee. The present results suggested that the significantly decreased intensity of the coffee-like aroma quality might result from decreased aroma release of a few odorants, including FFT, by the addition of milk to the coffee. Moreover, it was inferred that the difference in aroma release of FFT would have an especially significant impact on the perception of coffee-like aroma quality between the black coffee and milk coffee.

Introduction

Flavor release from food has been acknowledged as an important factor in determining the perceived flavor quality of various foods. It is well-known that aroma perception is the result of odorant/receptor interactions that take place in receptors on the olfactory epithelium in the human nasal cavity. Therefore, in order to understand aroma perception during eating and drinking, it is important to determine the composition and amounts of odorants that reach the olfactory epithelium from foodstuffs or their behavior (Taylor and Roozen, 1996; Taylor and Linforth, 1997). In order to analyze the odorants exhaled through the nostrils, several analytical techniques have been developed, such as nosespace analysis by APCI-MS (Taylor et al., 2000), PTR-MS (Roberts et al., 2003a), EXOM (Buettner and Schieberle, 2000), and R-FISS (Kumazawa et al., 2008). These analytical techniques have been used to reveal the relationship between the intensity of the aroma perceptions and the aroma release during the consumption of foods and drinks (Gwartney et al., 2000; Hollowood et al., 2000; Lethuaut et al., 2004; Linforth et al., 1999; Mestres et al., 2005; Saint-Eve et al., 2009; Weel et al., 2002). In addition, Beauchamp et al. and Fransnelli et al. also developed other analytical techniques capable of analyzing the odorants passing through the nasopharynx and reaching the olfactory epithelium in order to understand the in-mouth release in more detail (Beauchamp et al., 2014; Frasnelli et al., 2005).

Coffee is widely appreciated for its characteristic aroma and taste, and is often enjoyed with milk or creamer. The purposes of adding these products to coffee are to develop a desirable color change, to reduce the bitter and sour tastes, and to reduce the astringency of coffee. In general, adding milk or creamer to coffee significantly affects the aroma perceptions during drinking. Recent studies have demonstrated that the addition of milk products or milk components to coffee reduced the amounts of the aroma compounds released into the headspace of the coffee beverages (Bücking and Steinhart, 1996; Fisk et al., 2012), likely altering the quality of the perceived aroma (Bücking and Steinhart, 1996). It is also well-known that the interaction between fat and odorants can affect the in-mouth release of odorants and their aroma perception (Brauss et al., 1999; Frank et al., 2011; Weel et al., 2004). The influence of fat on aroma release is highly related to the lipophilicity of the odorants. It has been reported that the aroma release of lipophilic compounds, which have high K_o/w and LogP values, generally decreased with increasing fat content (Frank et al., 2011; Frank et al., 2012; Rabe et al., 2004; Roberts et al., 2003b). However, previous studies have not investigated in detail the influence of milk or creamer on the in-mouth release, especially the amounts of potent odorants in brewed coffee exhaled through the nostrils, during coffee consumption.

R-FISS is one of the analytical techniques used to analyze in-mouth odorants exhaled through the nostrils. The major advantages of this technique are the ability to improve the detection limit of odorants by concentrating them on the resin adsorbent and to determine the composition of the mixture consisting of a significant number of odorants in one measurement. Therefore, the aim of this study was to elucidate the influence of milk on the aroma release and aroma perception by investigating in detail the amounts of in-mouth odorants, including the potent odorants in brewed coffee, during the consumption of coffee using R-FISS.

Materials and Methods

Materials    Arabica coffee beans (Columbia Supremo) with a medium roasting degree were supplied by Unicafe, Inc. (Tokyo, Japan), and stored at −20°C until used. The roast degree of the beans was characterized by a color value of 20. The value was obtained by measuring coarsely milled coffee beans with a ZE2000 color-difference meter (Nippon Denshoku Industries Co., Ltd., Tokyo, Japan). Commercial milk (3.5% fat or more, sterilized at 130°C for 2 s; Meiji Co., Ltd., Tokyo, Japan) was purchased from a local market.

Chemicals    Furfuryl methyl sulfide (FMS) was synthesized according to the procedure in the literature (Itobe et al., 2009). The following odorants were purchased from the suppliers shown: compounds (Table 3) 1 – 15, 19 – 22, 24 – 28, and 30 – 38 (Tokyo Chemical Industry Co., Ltd., Tokyo, Japan); compounds 16 – 18, 23, 39, and 2-furfurylthiol (FFT) (Sigma-Aldrich Japan, Tokyo, Japan); compound 29 (Wako Pure Chemical Industries, Osaka, Japan); 2-octanol (Nacalai Tesque, Inc., Kyoto, Japan).

Table 3. Semiquantitative amounts^a and RSD (%)^b of in-mouth odorants exhaled through the nostrils during the consumption of black coffee and milk coffee obtained by the SIM-SCAN mode.

No.	Compound	Odor threshold in waterc (µg/L)	LogPd	Panelist 1						Panelist 2						Panelist 3
				Black Coffeee		Milk Coffee		Ratioe (Effect of Milk)	p valuef	Black Coffeee		Milk Coffee		Ratioe (Effect of Milk)	p valuef	Black Coffeee		Milk Coffee		Ratioe (Effect of Milk)	p valuef
				Amount (pg)	RSD (%)	Amount (pg)	RSD (%)			Amount (pg)	RSD (%)	Amount (pg)	RSD (%)			Amount (pg)	RSD (%)	Amount (pg)	RSD (%)
1	2-methylbutanalg	1–6	1.23	1240	14	1080	19	0.87	ns.	724	54	846	60	1.17	ns.	251	21	268	31	1.06	ns.
2	3-methylbutanalh	0.2–2	1.25	955	30	903	14	0.95	ns.	1100	27	785	35	0.71	ns.	347	41	307	37	0.89	ns.
3	2,3-butanedioneh	2.6–15	−1.34	1730	24	1740	18	1.01	ns.	1020	14	902	10	0.88	ns.	796	38	879	11	1.10	ns.
4	2,3-pentanedioneh	20	−0.85	1670	22	1330	20	0.80	ns.	1050	21	966	30	0.92	ns.	786	50	738	32	0.94	ns.
5	1-methylpyrroleg	–	1.43	713	25	457	24	0.64	<0.01	913	31	719	35	0.79	ns.	458	26	286	47	0.63	<0.05
6	pyridineg	7.9–2000	0.65	22300	20	19400	13	0.87	ns.	11700	21	12700	19	1.09	ns.	19600	14	19200	25	0.98	ns.
7	pyrazineg	50000 – 75000	−0.26	660	18	678	14	1.03	ns.	335	37	383	13	1.14	ns.	678	43	570	29	0.84	ns.
8	2-methylbutanolg	300–4150	1.26	84	30	61	13	0.73	ns.	51	19	43	32	0.86	ns.	36	34	49	43	1.36	ns.
9	isoamyl alcoholg	250–1005	1.26	173	41	126	25	0.73	ns.	86	27	85	16	0.99	ns.	77	51	84	32	1.10	ns.
10	lurfuryl methyl etherg	–	1.14	285	23	211	21	0.74	<0.05	327	25	279	26	0.85	ns.	233	29	166	49	0.71	ns.
11	2-methyltetrahydrofuran-3-oneg	–	−0.20	2470	18	2200	16	0.89	ns.	998	22	963	16	0.97	ns.	1880	23	1500	27	0.80	ns.
12	2-methylpyrazineg	60–105000	0.24	2990	24	2560	12	0.86	ns.	1010	26	975	17	0.97	ns.	2290	31	2120	24	0.93	ns.
13	2,5-dimethylpyrazineg	800–1800	1.03	746	25	632	13	0.85	ns.	261	27	248	15	0.95	ns.	565	29	525	25	0.93	ns.
14	2,6-dimethylpyrazineg	200–9000	1.03	653	30	563	14	0.86	ns.	222	31	211	17	0.95	ns.	451	28	398	37	0.88	ns.
15	2,3-dimethylpyrazineg	400–2500	1.03	258	26	209	12	0.81	ns.	85	22	87	15	1.03	ns.	177	31	171	22	0.97	ns.
16	2-methyl-2-cyclopenten-1-oneg	–	1.26	114	16	100	12	0.88	ns.	55	24	60	13	1.08	ns.	97	22	93	23	0.97	ns.
17	2-ethyl-6-methylpyrazineg	40	1.53	853	24	717	17	0.84	ns.	361	24	339	20	0.94	ns.	740	23	699	26	0.94	ns.
18	2-ethyl-5-methylpyrazineg	16–100	1.53	559	23	478	15	0.85	ns.	238	20	223	18	0.94	ns.	480	21	455	25	0.95	ns.
19	2-ethyl-3-methylpyrazineg	130–500	1.53	194	21	163	14	0.84	ns.	83	19	81	18	0.97	ns.	163	27	152	26	0.94	ns.
20	2,3,5-trimethylpyrazineh	400–9000	1.58	99	28	80	18	0.81	ns.	35	27	32	21	0.94	ns.	76	34	72	15	0.94	ns.
21	furforalg	282∓23000	0.83	4600	34	3360	19	0.73	ns.	1380	27	1120	12	0.81	ns.	2110	40	1710	18	0.81	ns.
22	acetoxy-2-propanoneg	–	−0.19	707	35	574	14	0.81	ns.	206	32	156	17	0.76	ns.	356	52	346	19	0.97	ns.
23	2-ethvl-3,5-dirriethylpyrazineh	0.04–1.0	2.07	50	30	40	18	0.81	ns.	24	38	21	30	0.88	ns.	42	32	40	21	0.97	ns.
24	lurfuryl methyl sulfideb	–	2.00	197	24	103	30	0.52	<0.001	277	36	132	25	0.48	<0.01	145	35	74	34	0.51	<0.01
25	2-acetylfurang	10000	0.80	1760	22	1480	10	0.84	ns.	769	34	675	14	0.88	ns.	1420	20	1320	21	0.93	ns.
26	2.3-diethyl-5-methylpyrazineb	0.09–1.0	2.56	45	36	36	22	0.80	ns.	25	21	20	28	0.79	ns.	42	29	37	30	0.87	ns.
27	pyrroleg	20000–49600	0.75	1070	13	1010	14	0.94	ns.	487	20	400	18	0.82	ns.	430	28	270	55	0.63	ns.
28	2-isobutyl-3-methoxypyrazineh	0.002–0.016	2.86	14	39	5	23	0.38	<0.01	15	92	9	60	0.58	ns.	10	17	6	26	0.63	<0.01
29	1-(2-luranyl)-2-propanoneg	–	0.84	139	22	110	29	0.79	ns.	97	16	73	24	0.76	<0.05	85	25	52	35	0.62	<0.05
30	lurfuryl acetateg	–	1.09	1020	48	676	27	0.66	ns.	528	23	411	28	0.78	ns.	339	30	249	39	0.74	ns.
31	5-methylfixfuralg	6000	1.38	1020	39	747	19	0.73	ns.	374	25	311	12	0.83	ns.	395	42	287	16	0.73	ns.
32	difurylmethaneg	–	2.99	80	23	33	39	0.42	<0.001	128	43	35	33	0.27	<0.01	48	37	18	42	0.37	<0.01
33	2-formyl-1-methylpyrroleg	37	1.18	506	28	402	10	0.79	ns.	166	23	156	15	0.94	ns.	344	33	296	22	0.86	ns.
34	2-acetyl-1-methylpyrroleg	–	1.11	174	26	133	11	0.76	ns.	69	23	59	16	0.86	ns.	129	28	109	22	0.84	ns.
35	lurfuryl alcoholg	1000–2000	0.45	2060	34	1490	22	0.72	ns.	875	27	762	26	0.87	ns.	961	20	955	38	0.99	ns.
36	lurfuryl pyrroleg	100	2.50	165	37	76	43	0.46	<0.01	183	29	69	31	0.37	<0.001	151	25	48	62	0.32	<0.001
37	2-methoxyphenolh	3–21	1.34	109	16	118	34	1.08	ns.	97	19	79	21	0.82	ns.	131	34	132	30	1.00	ns.
38	4-ethyl-2-methoxyphenolh	20–50	2.38	19	18	21	28	1.08	ns.	11	65	7	41	0.58	ns.	10	36	8	50	0.82	ns.
39	2-methoxy-4-vinylphenolh	3–10	2.24	33	29	30	23	0.89	ns.	15	35	10	58	0.64	ns.	19	52	12	51	0.62	ns.

^a  Each amount is the mean value of seven-fold results obtained from each identical panelist.

^b  RSD (%) is calculated using the amounts of the odorants from seven-fold results obtained from each identical panelist.

^c  Odor threshold values are obtained from Flavor-Base 9th Edition – the world's most extensive database for flavoring materials and food additives provided by Leffingwell & Associates.

^d  LogP values are obtained from Chemspider – open access online chemical database hosted by the Royal Society of Chemistry (i).

^e  Each ratio is calculated by dividing the amount of the in-mouth odorant during the consumption of milk coffee by that of black coffee.

^f  Student's t test was performed.; n.s., not significant.

^g  detected in the SCAN mode.

^h  detected in the SIM mode.

Preparation of Coffee Brew    Five hundred grams of hot distilled water (ca. 90°C) were poured on 50 g of the ground Arabica coffee powder in a filter. The filtrate was immediately cooled in an ice water bath.

Preparation of Deodorized Coffee Brew    The coffee brew, which was prepared under the same conditions as described above, was lyophilized and the coffee powder was obtained (the solid content was 1.5%(w/w) of the coffee brew). To remove the remaining odorants in the coffee powder, the powder was distilled using the solvent-assisted flavor evaporation (SAFE) method (Engel et al., 1999) after being dissolved in 100 g of distilled water. The residue was then lyophilized again, and the deodorized coffee powder was obtained. It was then dissolved in distilled water (1.5 g/100 g). The residual odorants were confirmed by Gas Chromatography-Mass Spectrometry (GC-MS) using the adsorptive column method (Shimoda et al., 1995). FMS was completely removed, and almost all the FFT was also removed.

Quantitative Analysis of Furfuryl Methyl Sulfide in Coffee Brew    The aroma concentrate of the coffee brew was prepared using the adsorptive column method. Fifty grams of the coffee brew were passed through a glass column (15 cm × 2 cm i.d.) filled with 5 mL of SP700 resin (Mitsubishi Chemical Corporation, Tokyo, Japan), which was conditioned with distilled water before use, followed by washing with 50 mL of distilled water and eluting with 20 mL of diethyl ether. Fifty microliters of an internal standard solution (10.8 mg/100 mL 2-octanol in methylene chloride) were added to the diethyl ether fraction for the quantitative analysis. The fraction was dried with an excess amount of anhydrous sodium sulfate and then concentrated by rotary evaporation (35°C, 550 mmHg) to about 5 mL, followed by nitrogen stream evaporation to about 100 µL. In order to determine the recovery rate of FMS using this method, the aroma concentrate of the deodorized coffee brew (100 ppb of FMS was added) was also prepared. These aroma concentrates were used as the samples for the GC-MS analysis (a).

Preparation of Coffee Beverages and Deodorized Coffee Beverages.    The experimental design for the coffee beverages is shown in Table 1. FFT and FMS were dissolved in ethanol, and added at 0.1% to each coffee beverage so that the additive amounts of these compounds are the same as the concentrations described in Table 1. In order to prepare beverage 6, the amounts of FFT and FMS to be added to the milk coffee were determined according to the following procedure. The concentration of FMS in the coffee brew was determined to be 44 ppb by GC-MS using the adsorptive column method and the recovery rate of FMS. The concentration of FMS in beverages 1 and 2 (Table 1) was estimated to be 22 ppb because the coffee brew was diluted with the same weight of distilled water or milk. In order to determine the in-mouth release of FMS derived from only FMS in beverages 1 and 2 during consumption of those coffee beverages, 20 ppb of FMS was then added to the deodorized coffee beverages so that the concentration of FMS in beverages 3 and 4 (Table 1) was about the same as in beverages 1 and 2. The in-mouth release of FMS during the consumption of beverage 4 was about two-thirds that of beverage 3. Therefore, based on the different in-mouth release of FMS between beverages 3 and 4, and the concentration of FMS in the milk coffee, the additive amount of FMS to the milk coffee was determined to be 10 ppb so that the in-mouth release of FMS derived from only FMS in beverage 2 was about the same as that of beverage 1. Next, the relationship between the additive amounts of FFT to beverage 5 and the in-mouth release of FMS was investigated. The additive amount of FFT to the milk coffee was determined assuming that the different in-mouth release of FMS between beverages 1 and 5 (Table 1) resulted from the different in-mouth release of FFT in these coffee beverages. All of these beverages were kept at room temperature before use.

Table 1. Experiment design for coffee beverages, additive amount of 2-furfurylthiol (FFT) and furfuryl methyl sulfide (FMS), and concentration of FMS in coffee beverages.

Beverage	Beverage type	Mixed amount (g)				Additive amount (ppb)		Concentration of FMS (ppb)
		Coffee brew	Deodorized coffee brew	Distilled Water	Milk	FFT	FMS	from Coffee brewa	Smnb
1	Black coffee	100	-	100	-	-	-	22	22
2	Milk coffee	100	-	-	100	-	-	22	22
3	Deodorized black coffee	-	100	100	-	-	20c	-	20
4	Deodorized milk coffee	-	100	-	100	-	20c	-	20
5	FMS-added milk coffee	100	-	-	100	-	10d	22	32
6	Aroma-adjusted milk coffee	100	-	-	100	80e	10d	22	32

^a  Calculated from the concentration of FMS in coffee brew and the dilution ratio with distilled water or milk.

^b  Sum of the concentration of FMS in each coffee beverage; Sum = additive amount of FMS + concentration of FMS derived from the coffee brew.

^c  Determined by considering the concentration of FMS in coffee brew so that the concentration of FMS was about the same as in beverages 1 and 2.

^d  Determined by considering the concentration of FMS in coffee brew and the different in-mouth release of FMS between beverages 3 and 4.

^e  Determined so that the in-mouth release of FMS was about the same as in beverages 1 and 2.

Semi-Quantitative Analysis of In-Mouth Odorants Exhaled through the Nostrils    In order to determine the in-mouth odorants, the breath exhaled through the nostrils was trapped according to the following procedure. After 30 mL of the coffee beverage at room temperature was placed in the mouth cavity, trained panelists paused for 1 – 2 seconds and then swallowed all of the liquid in one gulp. Ten breaths after the coffee beverage had been swallowed were passed through a small glass column (6 cm × 5 mm i.d.) filled with Tenax TA (100 mg, 80/100 mesh; GL Science, Tokyo, Japan), which had been heated at 220°C for 2 h prior to the analysis. The end of the glass column was connected to a pump by a silicon tube, and during trapping of the air exhaled through the nostrils, a suction of approximately 1 L/min was applied to the system. The experiments lasted 1 – 3 min. This sampling system allowed the panelists to normally exhale without the need to press air through the Tenax column. After trapping of the in-mouth odorants, 5 µL of an internal standard solution (5 µg/mL 2-octanol in ethanol) was directly added to the Tenax column for the semi-quantitative analysis. The water and ethanol were then removed from the Tenax TA by flowing dry nitrogen (30 min, 100 mL/min). These experiments were carried out at room temperature (25°C ± 2°C). The experiments using beverages 1 and 2 (Table 1) were performed by 3 trained panelists. The Tenax columns were used as the samples for the GC-MS analysis (b). A blank test was also performed under the same conditions using distilled water. In order to determine the additive amounts of FFT and FMS to the milk coffee, the experiments using beverages 1 – 6 (Table 1) were performed by a trained panelist. The Tenax columns were used as the samples for the GC-MS analysis (c).

Gas Chromatography-Mass Spectrometry    (a) The aroma concentrates were analyzed by an Agilent 7890N gas chromatograph coupled to an Agilent 5975C series mass spectrometer (Agilent Technologies, Palo Alto, CA) using the split mode (injection volume of 1 µL; inlet temperature of 250°C; split ratio 1:30). The column was a 60 m × 0.25 mm i.d. DB-WAX fused silica capillary (J&W Scientific, Folsom, CA) with a film thickness of 0.25 µm. The column temperature was programmed from 80 to 230°C at the rate of 3°C/min. The flow rate of the helium carrier gas was 1 mL/min. The mass spectrometer was used with an ionization voltage of 70 eV (EI) and an ion source temperature of 150°C, and operated in the SIM mode. The selected ions of 45 and 81 were monitored for 2-octanol (internal standard) and FMS, respectively. The content of FMS in the volatile fraction of the coffee brew was determined by the internal standard method using a response factor. The response factor of FMS to 2-octanol (internal standard) was 1.5, and this response factor was calculated from the ratio of the selected ion peak area of FMS to the internal standard obtained by mass chromatography of a standard solution containing equal weights of FMS and the internal standard. These selected ion peak areas were the mean values of triplicate results. Moreover, the concentration of FMS in the coffee brew was calculated from the content in the volatile fraction of the coffee brew and the recovery rate on the adsorptive column method. (b) Thermal desorption of the trapped odorants on the Tenax TA was performed using a TDU thermal desorption system (Gerstel GmbH, Muelheim an der Ruhr, Germany) in combination with the ATEX option of an MPS2 autosampler (Gerstel GmbH) and a CIS-4 injector (Gerstel GmbH) according to the literature parameters. Thermal desorption was performed by programming the TDU from 20 to 220°C (held for 3 min) at the rate of 12°C/s in the splitless mode. Cryofocusing was performed with liquid nitrogen at −150°C. Injection was performed with the ramp of 12°C/s from −150 to 220°C (held for 3 min) in the splitless mode. The odorants were analyzed by an Agilent 6890N gas chromatograph coupled to an Agilent 5975 B series mass spectrometer (Agilent Technologies). The column was a 30 m × 0.25 mm i.d. DB-WAX fused silica capillary (J&W Scientific) with a film thickness of 0.25 µm. The column temperature was programmed from 30°C (held for 3 min) to 120°C at the rate of 3°C/min, and then raised at the rate of 5°C/min to 230°C. The flow rate of the helium carrier gas was 1 mL/min. The mass spectrometer was used with an ionization voltage of 70 eV (EI) and an ion source temperature of 150°C, and operated in the selected ion monitoring and scan (SIM-SCAN) mode. The selected ions listed in Table 2 were monitored in the SIM mode for each potent odorant in coffee brew. The semiquantitative amounts of in-mouth odorants were determined by the internal standard method from the total ion or the selected ion peak areas obtained by mass chromatography. The semiquantitative amounts were calculated from the ratio of the total ion peak areas of 2-octanol (internal standard) and those of the in-mouth odorants obtained in the SCAN mode. Based on the ratio of the selected ion to the overall ions of the mass spectra for each reference compound, the total ion peak area of each compound was converted from the selected ion peak area obtained in the SIM mode; total ion peak area = selected ion peak area × conversion factor. The conversion factors are listed in Table 2. The response factor of each odorant to the internal standard was defined as 1. The odorants, whose peak areas were significantly larger than those in the blank test, were quantitated as in-mouth odorants. The results were then analyzed using a Student's t-test. (c) The conditions of the thermal desorption and GC-MS were the same as those in a. The results of beverages 1, 2, and 6 were then analyzed using a one-way analysis of variance (ANOVA) followed by Tukey's test.

Table 2. Selected ions and conversion factors for mass chromatography in the SIM mode.

Odorants	Selected Ion (m/z)a		Conversion Factorb	Previous idendificationc
	cal	ref
3-methylbutanal	44	58	4.5	1
2,3-butanedione	86	43	5.0	1
2,3-pentanedione	100	43	6.1	1
ethyl 2-methylbutyrate	102	85	4.1	2
ethyl iso valerate	88	85	4.4	2
3-methyl-2-buten-1-thiol	102	41	7.3	1
2-methyl-3-furanthiol	114	85	3.5	1
2,3,5-trimethylpyrazine	122	42	2.2	1
2-isopropyl-3-methoxypyrazine	137	152	3.3	1,2
2-methoxy-3,5-dimethylpyrazine	138	109	4.7	2
2-furfurylthiol (FFT)	81	114	2.3	1
methional	104	48	5.2	1,2
2-ethyl-3,5-dimethylpyrazine	135	136	3.0	1
2,3-diethyl-5-methylpyrazine	150	135	6.5	1
(Z)-2-nonenal	41	70	7.7	2
2-isobutyl-3-methoxypyrazine	124	151	2.5	1,2
3-mercapto-3-methylbutyl formate	69	102	3.1	1,3
linalool	71	93	6.7	1,2
(E)-2-nonenal	70	83	8.9	1,2
3-mercapto-3-methylbutyl acetate	69	102	4.7	3
phenylacetaldehyde	91	120	2.0	1
isovaleric acid	60	87	2.0	1,2
3-mercapto-3-methylbutanol	69	75	5.7	1
β-damascenone	121	190	5.3	1
3,4-dimethyl-1,2-cyclopentanedione	126	111	5.1	1
2-methoxyphenol	109	124	3.1	1
anisaldehyde	135	136	2.9	1
4-hydroxy-2,5-dimethyl-3(2H)-furanone	128	85	3.2	1
4-ethyl-2-methoxyphenol	137	152	2.4	1,2
bis(2-methyl-3-furyl)disulfide	113	226	3.5	1
3-hydroxy-4,5-dimethyl-2(5H)-furanone	128	83	11.3	1,2
2-methoxy-4-vinylphenol	150	135	3.5	1,2
5-ethyl-3-hydroxy-4-methyl-2(5H)-furanone	97	142	4.8	1
vanillin	151	152	3.2	1, 2
furfuryl methyl sulfided	81	128	2.8	-
2-octanole	45	97	3.2	-

^a  cal.: quantifier ion, ref.: qualifier ion.

^b  The ratio of the selected ion (quantifier ion) to the overall ions of the mass spectra for each reference compound.

^c  See references 1: Blank et al. (1992), 2: Czerny and Grosch (2000), 3: Kumazawa and Masuda (2003).

^d  Selected as a target compound in the SIM mode to improve the qualitative accuracy.

^e  Internal standard.

Identification of the Odorants    Each odorant was identified by comparing its Kovats GC retention index (RI) and mass spectrum to those of the reference compounds.

Sensory Evaluation    Fifteen milliliters of the coffee beverages (beverages 1, 2, and 6 shown in Table 1) were placed in plastic cups, and these 3 samples were simultaneously presented to panelists in a dark room under red light. Ion exchanged water was provided for rinsing the mouth. The samples were evaluated by 31 panelists (14 males and 17 females) working for Ogawa & Co., Ltd. All panelists had previously received extensive training in the descriptive sensory analysis of coffee beverages and had experience in the sensory profiling of various food samples. Milk coffee (beverage 2) was used as a control, and its intensity of the coffee-like aroma quality was defined as 4 for descriptive purposes. The panelists scored the intensity of the coffee-like aroma quality for the black coffee and aroma-adjusted milk coffee (beverages 1 and 6) using a seven-point scale from 1 (very weak) to 7 (very strong) compared to that for the milk coffee. The experiment was performed one time by each panelist. The sensory results were then analyzed using a one-way ANOVA followed by Tukey's test.

Results and Discussion

Influence of Milk on Aroma Release During Consumption of Black Coffee and Milk Coffee    In order to investigate the influence of milk on the aroma release, the in-mouth release of odorants during the consumption of the black coffee and milk coffee (beverages 1 and 2 shown in Table 1) was compared using R-FISS. The SIM-SCAN mode was applied to a mass spectrometer in order to improve the detection limit of odorants. Thirty-five odorants, except for 2-octanol, listed in Table 2 were selected as the target compounds in the SIM mode. These odorants except for FMS were previously identified as potent odorants in brewed coffee (Blank et al., 1992; Czerny and Grosch, 2000; Kumazawa and Masuda, 2003). There is the possibility that their in-mouth release could have a significant impact on the aroma perception of coffee. As the result of in-mouth release analysis using R-FISS, thirty-nine odorants listed in Table 3 could be detected and their amounts showed individual differences (Table 3). Previous studies have already indicated that the chemical structure of some odorants can be altered by the enzymatic activities of human saliva (Buettner, 2002a; Buettner, 2002b). Based on these results, the different in-mouth release among panelists might have resulted from differences in the oral cavity, volume and composition of the saliva, or swallowing activity. Thus, while there were individual differences in the in-mouth releases, the amounts of most in-mouth odorants did not significantly differ between the black coffee and milk coffee. However, the amounts of 3 odorants (No. 24, 32, and 36 in Table 3) exhaled through the nostrils during consumption of the milk coffee were significantly lower than those of the black coffee regardless of the panelists (Table 3).

Previous studies suggested the possibility that fat would have a greater impact on the aroma release than proteins (Fisk et al., 2012; Roberts and Pollien, 2000). It was also reported that the influence of fat on the aroma release differed depending on the LogP values of the odorants (Frank et al., 2011; Frank et al., 2012; Rabe et al., 2004; Roberts et al., 2003b). In agreement with previous studies, milk significantly decreased the in-mouth releases of relatively lipophilic odorants whose LogP values are equal to or higher than 2 (No. 24, 32, and 36 in Table 3) in the present results. However, not all odorants whose LogP values are higher than 2 (No. 23, 24, 26, 28, 32, 36, 38, and 39 in Table 3) were significantly affected by the milk. These results indicated that the in-mouth releases of some odorants could be easily influenced by the milk, but those of other odorants could not be easily influenced even if those odorants have similar LogP values. Therefore, some factors other than the milk fat and the LogP values of the odorants might have an impact on the aroma release of the milk coffee.

FMS exhaled through the nostrils during the consumption of milk coffee could be derived from not only FMS but FFT, whose LogP value is 3.44 (obtained from Chemspider (i)), in the coffee beverages (Itobe et al., 2009). Especially, it was also reported that nearly 90% of the FFT was methylated to FMS during a short period prior to the odorants being exhaled through the nostrils (Itobe et al., 2009). Therefore, the different in-mouth release of FMS between the black coffee and milk coffee obtained from the present study would include the different in-mouth release of FFT. The interchanging of the sulfhydryl and disulfide groups within the protein or with external thiol groups are well-known reactions in protein chemistry (Whitesides et al., 1983). Thiols would be lost by reacting with the sulfhydryl or disulfide groups on the protein. Therefore, the in-mouth release of FMS might be decreased by the addition of milk to the coffee because some of the FFT in milk coffee might be lost by these reactions with milk proteins, such as casein, and those FFTs could not be exhaled from the nostrils as FMS. The present results suggested the possibility that the addition of milk to the coffee could decrease the in-mouth release of FFT in addition to 3 odorants (No. 24, 32, 36 in Table 3).

Influence of FFT on Differences in Coffee-like Aroma Quality between Black Coffee and Milk Coffee    In order to investigate the influence of the in-mouth release of FFT on the different aroma perceptions of the black coffee and milk coffee, the aroma-adjusted milk coffee (beverage 6 shown in Table 1) was prepared by the addition of FFT and FMS to the milk coffee. The intensity of the coffee-like aroma quality of the aroma-adjusted milk coffee was then compared to those of the black coffee and milk coffee. The additive amounts of FFT and FMS to the milk coffee were determined to be 80 ppb and 10 ppb, respectively, in consideration of the concentration of FMS in the coffee brew and the in-mouth release of FMS during consumption of the coffee beverages (see Materials and Methods). The in-mouth release of FMS during the consumption of the aroma–adjusted milk coffee was then investigated. As a result, the amount of FMS exhaled through the nostrils was roughly the same as that of the black coffee (Fig. 1).

Fig. 1.

Semiquantitative amounts of furfuryl methyl sulfide (FMS) exhaled through the nostrils during the consumption of black coffee, milk coffee, and aroma-adjusted milk coffee. Each amount of FMS is the mean value of triplicate results obtained from one panelist. Error bars show the standard deviations and the letters a and b indicate means that significantly differ at p < 0.01 (Tukey's test).

The intensity of the coffee-like aroma quality during the consumption of the black coffee, milk coffee, and aroma-adjusted milk coffee was then examined by sensory evaluation. As a result, it was demonstrated that the coffee-like aroma quality of the milk coffee was perceived to be significantly weaker than that of the black coffee (Fig. 2). Moreover, the coffee-like aroma quality of the aroma-adjusted milk coffee was perceived to be as strong as that of the black coffee (Fig. 2).

Fig. 2.

Intensity of coffee-like aroma quality during the consumption of black coffee, milk coffee (control, the intensity was defined as 4), and aroma-adjusted milk coffee. Each intensity is the mean value of 31 panelists using a seven-point scale from 1 to 7. Error bars show the standard deviations and the letters a and b indicate means that significantly differ at p < 0.01 (Tukey's test).

These results suggested the possibility that the amount of FFT exhaled through the nostrils as FMS during consumption of the coffee is one of the factors that has a significant impact on the different intensity of the coffee-like aroma quality between the black coffee and milk coffee. The in-mouth release of some odorants (No. 5, 28, 29, 32, and 36 in Table 3) during the consumption of milk coffee was also significantly lower than that of the black coffee in more than one panelist. These odorants could be also presumed to have an impact on the different aroma perceptions between the black coffee and milk coffee. Moreover, the in-mouth releases of trace odorants, which could not be detected by R-FISS in the present study, are likely also affected by the milk. Their different in-mouth releases might have an impact on the different aroma perceptions between these coffee beverages. Furthermore, milk could have an influence on the release kinetics of the in-mouth odorants during the consumption of milk coffee. The different aroma composition of retronasal aroma caused by the different release kinetics might result in the different aroma perceptions between the black coffee and milk coffee.

The present results suggested the possibility that the significantly decreased intensity of the coffee-like aroma quality might result from the decreased aroma releases of a few odorants, including FFT, by the addition of milk to the coffee. Moreover, it was inferred that the different aroma release of FFT would have an especially significant impact on the different aroma perceptions of the coffee-like aroma quality between the black coffee and milk coffee. An important future challenge will be to understand in more detail the relationship between the influence of milk on the aroma release of coffee and the perception of aromas.

References

•  Beauchamp,  J.,  Scheibe,  M.,  Hummel,  T., and  Buettner,  A. (2014). Intranasal odorant concentrations in relation to sniff behavior. Chem. Biodivers., 11, 619-638.
•  Blank,  I.,  Sen,  A., and  Grosch,  W. (1992). Potent odorants of the roasted powder and brew of Arabica coffee. Z. Lebensm. Unters. For., 195, 239-245.
•  Brauss,  M.S.,  Linforth,  R.S.T.,  Cayeux,  I.,  Harvey,  B., and  Taylor,  A.J. (1999). Altering the fat content affects flavor release in a model yogurt system. J. Agric. Food Chem., 47, 2055-2059.
•  Bücking,  M. and  Steinhart,  H. (1996). Headspace GC and sensory analysis characterization of the influence of different milk additives on the flavor release of coffee beverages. J. Agric. Food Chem., 50, 1529-1534.
•  Buettner,  A. and  Schieberle,  P. (2000). Exhaled odorant measurement (EXOM) – a new approach to quantify the degree of in-mouth release of food aroma compounds. Lebensm.-Wiss. Technol., 33, 553-559.
•  Buettner,  A. (2002a). Influence of human salivary enzymes on odorant concentration changes occurring in vivo. 1. Esters and hiols. J. Agric. Food Chem., 50, 3283-3289.
•  Buettner,  A. (2002b). Influence of human saliva enzymes on odorant concentrations. 2. Aldehydes, alcohols, 3-alkyl-2-methoxypyrazines, methoxyphenols, and 3-hydroxy-4,5-dimethyl-2(5H)-furanone. J. Agric. Food Chem., 50, 7105-7110.
•  Czerny,  M. and  Grosch,  W. (2000). Potent odorants of raw Arabica coffee. Their changes during roasting. J. Agric. Food Chem., 48, 868-872.
•  Engel,  W.,  Bahr,  W., and  Schieberle,  P. (1999). Solvent assisted flavour evaporation – a new and versatile technique for careful and direct isolation of aroma compounds from complex food matrices. Eur. Food Res. Technol., 209, 237-241.
•  Fisk,  I.D.,  Boyer,  M., and  Linforth,  R.S.T. (2012). Impact of protein, lipid and carbohydrate on the headspace delivery of volatile compounds from hydrating powders. Eur. Food Res. Technol., 235, 517-525.
•  Frank,  D.,  Appelqvist,  I.,  Piyasiri,  U., and  Wooster,  T.J. (2011). Proton transfer reaction mass spectrometry and time intensity perceptual measurement of flavor release from lipid emulsions using trained human subjects. J. Agric. Food Chem., 59, 4891-4903.
•  Frank,  D.,  Appelqvist,  I.,  Piyasiri,  U., and  Delahunty,  C. (2012). In vitro measurement of volatile release in model lipid emulsions using proton transfer reaction mass spectrometry. J. Agric. Food Chem., 60, 2264-2273.
•  Frasnelli,  J.,  Van Ruth,  S.M.,  Kriukova,  I., and  Hummel,  T. (2005). Intranasal concentrations of orally administered flavors. Chem. Senses, 30, 575-582.
•  Gwartney,  E.A.,  Foegeding,  E.A., and  Larick,  D.K. (2000). The role of texture and fat on flavor release from whey protein isolate gels. In “Flavor Release” ed. by  Roberts,  D.D. and  Taylor,  A.J. American Chemical Society, Washington, DC, Vol. 763, pp. 355-367.
•  Hollowood,  T.A.,  Linforth,  R.S.T., and  Taylor,  A.J. (2000). The Relationship between carvone release and the perception of mintyness in gelatine gels. In “Flavor Release” ed. by  Roberts,  D.D., and  Taylor,  A.J., American Chemical Society, Washington, DC, Vol. 763, pp. 370-380.
•  Itobe,  T.,  Kumazawa,  K., and  Nishimura,  O. (2009). New factor characterizing the in-mouth release of odorants (volatile thiols): Compositional changes in odorants exhaled from the human nose during drinking. J. Agric. Food Chem., 57, 11297-11301.
•  Kumazawa,  K. and  Masuda,  H. (2003). Identification of odor-active 3-mercapto-3-methylbutyl acetate in volatile fraction of roasted coffee brew isolated by steam distillation under reduced pressure. J. Agric. Food Chem., 51, 3079-3082.
•  Kumazawa,  K.,  Itobe,  T.,  Nishimura,  O., and  Hamaguchi,  T. (2008). A new approach to estimate the in-mouth release characteristic of odorants in chewing gum. Food Sci. Technol. Res., 14, 269-276.
•  Lethuaut,  L.,  Weel,  K.G.,  Boelrijk,  A.E., and  Brossard,  C.D. (2004). Flavor perception and aroma release from model dairy desserts. J. Agric. Food Chem., 52, 3478-3485.
•  Linforth,  R.S.T.,  Baek,  I., and  Taylor,  A.J. (1999). Simultaneous instrumental and sensory analysis of volatile release from gelatine and pectin/gelatine gels. Food Chem., 65, 77-83.
•  Mestres,  M.,  Moran,  N.,  Jordan,  A., and  Buettner,  A. (2005). Aroma release and retronasal perception during and after consumption of flavored whey protein gels with different textures. 1. In vivo release analysis. J. Agric. Food Chem., 53, 403-409.
•  Rabe,  S.,  Krings,  U., and  Berger,  R.G. (2004). Dynamic flavour release from miglyol/water emulsions: modelling and validation. Food Chem., 84, 117-125.
•  Roberts,  D. D. and  Pollien,  P. (2000). Relative influence of milk components on flavor compound volatility. In “Flavor Release” ed. by  Roberts,  D.D. and  Taylor,  A.J., American Chemical Society, Washington, DC, Vol. 763, pp. 321-332.
•  Roberts,  D.D.,  Pollien,  P.,  Antille,  N.,  Lindinger,  C., and  Yeretzian,  C. (2003a). Comparison of nosespace, headspace, and sensory intensity ratings for the evaluation of flavor absorption by fat. J. Agric. Food Chem., 51, 3636-3642.
•  Roberts,  D.D.,  Pollien,  P., and  Watzke,  B. (2003b). Experimental and modeling studies showing the effect of lipid type and level on flavor release from milk-based liquid emulsions. J. Agric. Food Chem., 51, 189-195.
•  Saint-Eve  A.,  Deleris,  I.,  Aubin,  I.,  Semon,  E.,  Feron,  G.,  Rabillier,  J.M.,  Ibarra,  D.,  Guichard,  E., and  Souchon,  I. (2009). Influence of composition (CO₂ and sugar) on aroma release and perception of mint-flavored carbonated beverages. J. Agric. Food Chem., 57, 5891-5898.
•  Shimoda,  M.,  Shigematsu,  H.,  Shiratsuchi,  H., and  Osajima,  Y. (1995). Comparison of the odor concentrates by SDE and adsorptive column method from green tea infusion. J. Agric. Food Chem., 43, 1616-1620.
•  Taylor,  A.J. and  Roozen,  J.P. (1996). Volatile flavor release from foods during eating. Crit. Rev. Food Sci. Nutr., 36, 765-784.
•  Taylor,  A.J. and  Linforth,  R.S.T. (1997). Relationship between sensory time-intensity measurements and in-nose concentration of volatiles. In “Flavours and Fragrances” ed. by  Swift,  A.D. Royal Society of Chemistry, Cambridge, pp. 171-184.
•  Taylor,  A.J.,  Linforth,  R.S.T.,  Harvey,  B.A., and  Blake,  A. (2000). Atmospheric pressure chemical ionization mass spectrometry for in vivo analysis of volatile flavour release. Food Chem., 71, 327-338.
•  Weel,  K.G.,  Boelrijk,  A.E.,  Alting,  A.C.,  van Mil,  P.J.,  Burger,  J.J.,  Gruppen,  H.,  Voragen,  A.G.J., and  Smit,  G. (2002). Flavor release and perception of flavored whey protein gels: Perception is determined by texture rather than by release. J. Agric. Food Chem., 50, 5149-5155.
•  Weel,  K.G.,  Boelrijk,  A.E.,  Burger,  J.J.,  Jacobs,  M.A.,  Gruppen,  H.,  Voragen,  A.G., and  Smit,  G. (2004). Effect of emulsion properties on release of esters under static headspace, in vivo, and artificial throat conditions in relation to sensory intensity. J. Agric. Food Chem., 52, 6572-6577.
•  Whitesides,  G.M.,  Houk,  J., and  Patterson,  M.A. (1983). Activation parameters for thiolate-disulfide interchange reactions in aqueous solution. J. Org. Chem., 48, 112-115.
• http://www.chemspider.com (Jan. 30, 2015)