Aroma Profiling Analysis of Peach Flowers Based on Electronic Nose Detection

Abstract

Aroma profiling peach flowers can improve the landscape and ecological value of the surrounding environment. In order to preliminarily explore and screen the flower aromas of peach germplasms, the floral aroma profiles of 50 peach germplasms were analyzed by electronic nose detection. The clustering results from ten sensors showed that it was possible to identify each peach blossom aroma type. ‘Gansutao2’ was separately classified as cluster I; ‘Yuntaishanshantao’, ‘Dazhuanggansutao’, and ‘Zhouxingshantao’ as cluster II; ‘Xinjiangpantao’ and ‘Xinjianghuangrou’ as cluster III; and the remaining germplasms as cluster IV. The flower aromas of clusters I, II, and III were different from that of cluster IV and were mainly highlighted by the significant differences in the response values of the three main sensors, which were W1S (methane), W1W (hydrogen sulfide), and W5S (nitrogen oxides). The principal component analysis and significance analysis of the sensor response values showed that ‘Hua3’, ‘Hongfenjiaren’, and ‘Yuntaishanshantao’ had special flower aromas. The response values of these three varieties produced by the W1S, W1W, and W5S sensors and two aromatic sensors, W2W (aromatic components and organic sulfides) and W3S (methane-aliph), were significantly different from most of the other germplasms. The ‘Yuntaishanshantao’ response values produced by the five sensors’ were low, which showed that it had a light aroma, while the ‘Hongfenjiaren’ and ‘Hua3’ values were high, which showed that they had strong aromas. The results from this study provided basic data that could be used to screen peach germplasms with obvious floral aromas, cultivate new varieties with strong aromas, and aid the development and utilization of peach floral aroma substances.

1. Introduction

Plant flowers can give an unparalleled landscape effect to nature. Flowers also attract pollinators through the volatilization of aroma substances [1] and participate in growth, reproduction, and plant protection mechanisms under various biotic and abiotic environmental stresses [2]. Unique flower aromas can be combined with color to encode highly specific signals [3]. Furthermore, the main components of flower aromas are widely used in food flavoring, the beauty industry, and for medical treatments [4]. Therefore, plant flower aromas have important biological, ecological, and economic benefits. Previous studies have attempted to understand the composition, release, and regulation of volatile compounds in flowers [4].

Peach (Prunus persica (L.) Batsch) is an important fruit tree because it is an economically important producer of fruit, and its flowers have high medicinal and ornamental values [5]. Peach blossom plays an important role in plant product development and utilization, and the trees are brightly colored, have a graceful posture, and are rich in chemical components. Previous research has indicated that peach blossom contains sugars, proteins, phenolic compounds, flavonoids and glycosides, saponin, alkaloids, and other compounds [6]. However, very few studies focused on the detection and analysis of peach floral aroma compounds [7,8]. Previous reports have shown that the major volatile constituents of peach flowers include linolenic alcohol, hexadecanoic acid, cyclohexane, octadecanoic acid, heneicosane, and phytol [7]. To date, the lack of floral aroma data has limited the ornamental and utilization values of peach flowers. However, it is now possible to identify peach germplasms with obvious and rich flower aromas. The qualitative and quantitative detection of flower aroma composition is mainly based on chromatographic technologies [3,9,10]. However, the detection results do not reflect the aroma under natural conditions, which is different from sensory evaluation conditions; this means that the results have little direct significance for practical applications. The electronic nose (E-nose) is a bionic aroma, non-destructive testing instrument. With the help of a specific sensor array and pattern recognition software, it can simulate the human olfactory system for identification and analysis. It can obtain overall information about volatile components in samples and accurately distinguish the odors released by complex samples at a low cost [11,12,13]. Recently, the electronic nose has been applied to orchid, jasmine, lily, and other flowers to detect and analyze the aroma types of different varieties and the same variety at different flowering stages. It plays an important role in evaluating and mining special flower aroma germplasms and characteristic aroma substances and has helped to clarify the aroma release rules for flower aromas [14,15,16,17,18,19]. Messina et al. (2009) [8] reported details about the detection of peach flower aroma by E-nose in four peach varieties. However, there have been no reports about the analysis and screening of flower aromas released by different peach germplasms.

In this study, the peach flower aromas of 50 peach germplasms were determined by E-nose. The purpose was to preliminarily explore the different flower aromas released by peach germplasms and screen peach germplasms with an obvious flower aroma. The results could provide basic data that can be used to cultivate new varieties with strong aromas and aid the development and utilization of peach flower aroma substances.

2. Materials and Methods

2.1. Peach Materials

A total of 50 peach germplasms were obtained from the National Peach Germplasms Repository (Nanjing, China, 32°2′ N, 118°52′ E, 11 m above sea level), with each material coming from two trees. The trees were 5 years old, the spacing in the rows was 2 × 5 m in a Y shape, and conventional cultivation measures were followed. A diverse selection of germplasm resources was tested. The Evaluation of Peach Trait Descriptions and Classifications in Peach [5,20] was used to select materials that covered different countries of origin (China, America, Japan, and Thailand), different species (Prunus persica (L.) Batsch, Prunus kansuensis Rehd., Prunus davidiana (Carr.) Franch, Prunus davidiana var. Potaninii Rehd., and Prunus ferganensis Kost.et Riab.), different germplasms (Wild, Landrace, Improved cultivar, and Breeding lines), and different use types (ornamental, seedling rootstock, and fresh). Details of the 50 peach germplasms are shown in

Table 1. Basic information about the tested peach germplasms.

Number	Name	CO	ST	GT	UT	Number	Name	CO	ST	GT	UT
1	Baibitao1	C	Pp	I	O	26	Jiangtao	C	Pp	L	O
2	Baibitao	C	Pp	L	O	27	Juhuatao	C	Pp	L	O
3	Baihuabitao	C	Pp	L	O	28	Mantianhong	C	Pp	I	O
4	Baihuashanbitao	C	PdF	L	O	29	Taiguohuatao	T	Pp	L	O
5	Baihuashantao	C	PdF	W	S	30	Tanchun	C	Pp	L	O
6	Baihuazhuxingtao	A	Pp	I	O	31	Taohua2	C	Pp	I	O
7	Yuntaishanshantao	C	PdF	W	S	32	Wubaotao	C	Pp	L	O
8	Beizhizhuxingfen	C	Pp	B	O	33	Yingchun	C	Pp	L	O
9	Dazhuanggansutao	C	Pk	W	S	34	Yushuizaihuatao	C	Pp	L	O
10	Danbanbaihua	C	Pp	L	O	35	Sahongtao	C	Pp	L	O
11	Fenhuazhuxingtao	A	Pp	I	O	36	Zhufenchuizhi	C	Pp	L	O
12	Fenrou sebitao	C	Pp	L	O	37	Zhouxingshantao	C	PdF	W	O/S
13	Gansutao2	C	Pk	W	S	38	Galaxy	A	Pp	I	F
14	Gansutao	C	Pk	W	S	39	Zijinhong1	C	Pp	I	F
15	Hehuan erse	C	Pp	L	O	40	Flordaglo	A	Pp	I	F
16	Hongchuizhi	C	Pp	L	O	41	Sunraycer	A	Pp	I	F
17	Hongfenjiaren	C	Pp	I	O	42	Sunblaze	A	Pp	I	F
18	Honghuashantao	C	PdF	W	S	43	Nanshantiantao	C	Pp	L	F
19	Honghuazhuxingtao	A	Pp	I	O	44	Shanganshantao	C	PdP	W	S
20	Hongyetao	C	Pp	L	O	45	Xiahui8	C	Pp	I	F
21	Hua3	C	Pp	I	O	46	Xiahui6	C	Pp	I	F
22	Hua5	C	Pp	I	O	47	Xiacui	C	Pp	I	F
23	Hua6	C	Pp	I	O	48	Xinjiangpantao	C	Pf	L	F
24	Hua8	C	Pp	I	O	49	Xinjianghuangrou	C	Pf	L	F
25	Riyuetao	C	Pp	L	O	50	Riben86	J	Pp	I	F

Note: Country of origin: CO, C: China, A: America, J: Japan, and T: Thailand; Species types: ST, Pp: Prunus persica (L.) Batsch, PdF: Prunus davidiana (Carr.) Franch, Pk: Prunus kansuensis Rehd., PdP: Prunus davidiana var. potaninii Rehd., and Pf: Prunus ferganensis Kost.et Riab.; Germplasm types: GT, W: Wild, L: Landrace, I: Improved cultivar, and B: Breeding lines; and Use types: UT, O: Ornamental, S: Seedling rootstock, and F: Fresh.

.

Flowers were collected from February to April 2020 for detection. The flower samples were collected from the two trees during the blooming period at 10:00 a.m. on a sunny day and immediately brought to the laboratory. Firstly, the flower branches of each sample were collected and quickly brought to the laboratory. Then, the flowers that were flowering on that day were picked for sample preparation and detection.

2.2. Flower Aroma Measurement

Measurements were performed using a portable, commercially available E-nose (PEN3.5, Airsense Analytics GmbH, Schwerin, Germany). The sensor array of the PEN3.5 is composed of ten metal oxide semiconductor type chemical sensors working at high temperatures to permit the classification and identification of different volatile species. Each sensor preferentially responds to a class of organic compounds: W1C for aromatic benzene detection, W5S for broad range and nitrogen oxide detection, W3C for ammonia, W6S for hydrogen, W5C for arom-aliph, W1S for broad-methane, W1W for sulfur-organics, W2S for broad-alcohol, W2W for aromatic components and sulf-chlor compounds, and W3S for methane-aliph. When the sensors are exposed to volatiles, the conductivity (G) to initial conductivity (G0) ratio (G/G0, relative conductivity or response value) changes depending on the change in G conductivity. The greater is the concentration of the volatile, the more the deviation of G/G0 from 1 (greater than or less than 1). The acquired data from the measurements were then properly stored. In this study, 10 g of flowers from each peach germplasm was placed into a 300 mL beaker, which was then sealed with sealing film. The beaker was placed for 30 min at 25 °C and the E-nose data measured. The peach flower aroma measurement method followed Yan et al. (2021b) [13] with some modifications. We inserted the E-nose sampling needle into the sealing film and extracted the gas in the beaker for detection. The volatile gas was pumped over the sensors of the E-nose at a flow rate of 400 mL/min. The E-nose analyses were recorded over a range of 0–60 s and one to three stable signals (response value) in the middle period of the data stabilization time were taken as the analysis time points [21]. After each analysis, the sampling chamber was washed with an air-dried flow for 60 s. All samples were analyzed in triplicate.

2.3. Statistical Analysis

Various software packages, including the WinMuster program in PEN3.5 (Airsense Analytics GmbH, Schwerin, Germany), IBM SPSS Statistics 22.0 (IBM, Armonk, NY, USA), and Minitab 18 (Minitab, State College, PA, USA), were used to analyze the E-nose data. The load analysis method (LO) used by WinMuster was used to statistically analyze the main sensors for peach blossom aromas. SPSS 22.0 was used for the cluster analysis of flower aromas and to analyze significant differences in the response values among the peach germplasms obtained by each sensor. A principal component analysis (PCA) was carried out by Minitab 18 based on the response values produced by the main and aromatic sensors.

3. Results

3.1. Response Values of the Sensors to Peach Flowers

Figure 1 shows the response values of the 10 sensors to the volatile compounds in peach flowers. The data were the ratio of conductivity between G and G0 (the conductivity of the sensors when the sample gas or zero gas blew over, respectively). Each curve represents a different sensor transient. The results show that conductivity increased sharply and then stabilized after 24 s. Therefore, the sensor signals in the intermediate period of the stabilization time (34–36 s) and the signal constant range of 34–36 s were used for the analysis.

Figure 1 and

Table 2. Variations in the aroma response values for the different peach flowers.

Index	W1C	W5S	W3C	W6S	W5C	W1S	W1W	W2S	W2W	W3S
Mean	0.84	2.03	0.91	1.16	0.95	4.02	2.98	1.68	2.14	1.44
Standard deviation	0.05	0.67	0.02	0.06	0.02	1.28	0.77	0.35	0.34	0.21
Variation amplitude	0.8–0.9	1.3–4.7	0.9–1.0	1.1–1.4	0.9–1.0	2.1–7.5	1.6–5.4	1.2–2.8	1.5–3.1	1.2–2.2
Xmax-xmin	0.17	3.34	0.11	0.27	0.05	5.50	3.85	1.55	1.59	1.08
Variable coefficient (%)	5.9	33.0	2.1	5.2	2.1	31.8	25.8	20.8	15.9	14.6

Note: W1C: sensor for aromatic benzene, W5S: sensor for nitrogen oxides, W3C: sensor for ammonia, W6S: sensor for hydrogen, W5C: sensor for arom-aliph, W1S: sensor for methane compounds, W1W: sensor for hydrogen sulfide, W2S: sensor for broad-alcohol, W2W: sensor for aromatic components and organic sulfide, and W3S: sensor for methane-aliph.

show that the different sensors had different response values to each peach blossom aroma. The largest response value was recorded by W1S (methane compounds, 4.02 ± 1.28). The values were followed by W1W (hydrogen sulfide, 2.89 ± 0.77), W2W (organic sulfide, 2.14 ± 0.34), W5S (2.03 ± 0.67, nitrogen oxides), W2S (broad-alcohol, 1.68 ± 0.35), and W3S (methane-aliph, 1.44 ± 0.21). The response values‘ mean of the other sensors were low at about 1. Among the aromatic sensors, which were W2W, W3S, and W1C (aromatic benzene), only W2W had a high response value of 2.14 ± 0.34; this indicated that the peach blossom aromas were generally weak. The three indexes (variation amplitude, xmax-xmin, and the variation coefficient of the sensor response value) used in this study indicated that there were differences in the aromas produced by various substances; these three indexes for the W5S, W1S, W1W, and W2S sensors’ response values were relatively large because the variation coefficients reached 33.0%, 31.8%, 25.8%, and 20.8%, respectively. The three index response values for the W2W, W3S, and W1C aromatic sensors were relatively low with coefficients of variation of 15.9%, 14.6%, and 5.9%, respectively; this showed that the differences in floral aromas among peach germplasms were mainly caused by non-aromatic substances, such as nitrogen oxides, methane, and hydrogen sulfide, and that the differences among the aromatic substances were relatively small.

3.2. Cluster Analysis of the Tested Peach Germplasms Based on Their Blossom Aromas

Figure 2 shows that the response values of the ten sensors and the clustering results produced by the average linkage method were able to divide the 50 germplasms into four clusters when the cluster rescaled distance cluster d = 5. Specifically, ‘Gansutao2’ was separately classified as cluster I; ‘Yuntaishanshantao’, ‘Dazhuanggansutao’, and ‘Zhouxingshantao’ as cluster II; ‘and Xinjiangpantao’ and ‘Xinjianghuangrou’, and Prunus ferganensis Kost.et Riab. as cluster III. All the other germplasms were classified as cluster IV. The results indicated that the peach blossom aromas were relatively simple. There were no obvious difference among the common peach (Prunus persica (L.) Batsch) blossom aroma. However, the blossom aroma of some peach varieties belonged to Prunus kansuensis Rehd., Prunus davidiana (Carr.) Franch and Prunus ferganensis Kost.et Riab. were different from the other germplasms.

3.3. Analysis of the Main Sensors Used for Peach Blossom Aroma Detection

The LO component of the WinMuster software was used to statistically analyze the most important sensors for peach blossom aroma detection. Figure 3a showed that the W1S sensor (methane compounds) made the largest contribution to the first principal component, the W1W sensor (hydrogen sulfide) made large contributions to both the first and second principal components, and W5S (nitrogen oxides) made a large contribution to the second principal component. Only these three sensors were selected for fitting to confirm the contributions made by the sensors used to distinguish peach blossom aroma. The results (Figure 3b) explained 95.06% of the total contribution rate, indicating that the blossom aromas from the different peaches were mainly distinguished by the three sensors.

3.4. Aroma Components Analysis and Special Peach Germplasm Screening

The germplasms with special floral aromas were screened out by analyzing the aroma components of the 50 peach germplasms and the cluster groups based on the response values of the three main sensors [W5S (nitrogen oxides), W1S (methane compounds), and W1W (hydrogen sulfide)] and the three aromatic sensors [W2W (organic sulfide), W3S (methane-aliph), and W1C (aromatic benzene)]. The stable signals from the 34th to 36th time points for the various floral aromas were selected for the PCA analysis (Figure 4) and the significant differences among the sensor response values were compared (Figure 5).

According to the cluster analysis (Figure 2), clusters I, II, and III were different from the flower aromas of most of the other germplasms. Figure 4 showed that these three clusters were clearly distinguishable from each other. Furthermore, cluster IV mainly reflected the response values of the three main sensors. The W1S sensor response value for ‘Gansutao2’ in cluster I was the lowest among the germplasms and was significantly lower than 86% of the germplasms (p < 0.05, Figure 4c). The W1W sensor response value was second only to ‘Honghuashantao’ and ‘Hongfenjiaren’ and was significantly higher than 80% of the germplasms (p < 0.05, Figure 4b). The response values for ‘Yuntaishanshantao’, ‘Dazhuanggansutao’, and ‘Zhouxingshantao’ in cluster II were all low for the W1S, W1W, and W5S sensors; in which the response values of the W1S and W1W sensors were significantly lower than that of 84% and 62% of the other germplasms, respectively (p < 0.05, Figure 4a–c). ‘Xinjiangpantao’ and ‘Xinjianghuangrou’ in cluster III had low response values for sensors W1W and W5S (Figure 4a,b). The response values of these two sensors for ‘Xinjianghuangrou’ were the lowest among the germplasms, in which the response value for the W1W sensor was significantly lower than that of 90% of the other germplasms (p < 0.05, Figure 4b).

The PCA analysis (Figure 5) and the significant difference analysis (Figure 4) of the sensor response values preliminarily showed that ‘Hua3’, ‘Hongfenjiaren’, and ‘Yuntaishanshantao’ (Figure 6) had special flower aromas and that they were significantly different from the other germplasms according to the first and second principal components. There was no significant difference in the response values of the W1C aromatic sensor among the three varieties and the other germplasms. The response values of the three main sensors, as well as the W2W and W3S aromatic sensors for the three varieties, were significantly different from most of the other germplasms (Figure 4). Specifically, the response values produced by the five sensors for ‘Yuntaishanshantao’ were at a low level, indicating that its aroma was relatively weak. However, ‘Hongfenjiaren’ and ‘Hua3’ had relatively stronger aromas, which were reflected in the high response values recorded by the five sensors (Figure 4a–e). The W1S, W2W, and W3S sensor response values for ‘Hongfenjiaren’ were the highest among the germplasms; ‘Hua3’ ranked second, third, and second, respectively, and its W1S sensor response values were significantly higher than 90% of the other germplasms (p < 0.05, Figure 4c,d,e). The W5S sensor response value for ‘Hua 3’ was the highest among the germplasms (p < 0.05), and ‘Hongfenjiaren’ ranked third with significantly higher values than 84% of the other germplasms (p < 0.05, Figure 4a). The W1W sensor response values for both varieties were significantly higher than 72% of the other germplasms (p < 0.05, Figure 4b).

4. Discussion

Plant floral aromas have important biological, ecological, and economic benefits. The detection and identification of floral aroma components are important when attempting to clarify the ecological and biological functions of peach aromas and improve their effective development and utilization [4]. The development and utilization of peach germplasms with obvious, unique, and rich floral aromas can compensate for the poor landscape function shown by peaches. Messina et al. (2009) used the E-nose system (Moses II modular sensor system) and found that doped semiconductive SnO₂ sensors could only differentiate between the ‘Forastero’ cultivar growth stages (anthesis and post-anthesis). Furthermore, there has been no detailed description of peach blossom aroma substance types [8]. Previous studies predominantly analyzed the fat-soluble components of peach blossom using Gas Chromatography-Mass Spectrometer and identified 27 volatile substances. However, due to it’s different detection principles from E-nose system, the results were difficult to compare with the results from this study used by E-nose system. In this study, 50 peach germplasms were evaluated by E-nose (PEN 3.5) detection. The results showed that the peach flower aroma types were relatively simple and that there were no obvious differences in the flower aromas produced by most of the 50 germplasms. However, there were still differences among some of the genotypes, which meant that those varieties with special aromas could be screened out. Peaches belonging to Prunus kansuensis Rehd., Prunus davidiana (Carr.) Franch, and Prunus ferganensis Kost.et Riab. had different flower aromas, which were mainly reflected by their non-aromatic substances, such as methane, hydrogen sulfide, and nitrogen oxides. The flower aromas produced by ‘Hua3’, ‘Hongfenjiaren’, and ‘Baiyunshanshantao’ were also considered to be special. The three germplasms were not only significantly different from most of the other germplasms in terms of non-aromatic substances, such as methane, hydrogen sulfide, and nitrogen oxides, but also significantly different in terms of aromatic components, organic sulfide, and aromatic alkane. ‘Yuntaishanshantao’ produced the weakest aroma, while ‘Hongfenjiaren’ and ‘Hua3’ had the strongest aromas. Therefore, ‘Hongfenjiaren’ and ‘Hua3’ could be used as basic materials for cultivating varieties that have a rich flower aroma and for developing peach blossom aroma substances.

This study substantially contributes to understanding about the important biological and ecological factors associated with the unique compositions of floral aroma substances. Previous studies reported that benzaldehyde could be considered as a bee repellent, while terpenoids were involved in the composition of bee attractants [22,23]. Radice et al. (2010) [24] reported results on the relationship between the flower aroma from plum trees and bee attraction [24]; they found that Prunus armeniaca (L.) Giada was the best attractant for bees. Its flower aroma mainly contained terpenoids, whereas the flower aroma of Prunus salicina Lindl, which contained benzaldehyde compounds, was less attractive to bees. It has been suggested that differences in flower aroma composition might serve as a pollination mechanism for insects. Messina et al. (2009) [8] used the E-nose system (Moses II modular sensor system) to analyze the flower aromas produced by four peach varieties during and after flowering. There was no difference in the flower aromas produced by ‘Barcelo’, ‘Dixiland’, and ‘Summerprice’, whereas the ‘Forastero’ flower aroma was significantly different from the other varieties during flowering. The results might show that ‘Forastero’ is male sterile, whereas the other three are male fertile varieties, so it might release a unique flower aroma during the flowering period to attract pollinators that may improve its reproductive efficiency. In this study, the male sterile germplasms were ‘Yuntaishanshantao’, ‘Dazhuanggansutao’, ‘Hongfenjiaren’, and ‘Hua8’. Among the four varieties, ‘Yuntaishanshantao’, ‘Dazhuanggansutao’, and ‘Hongfenjiaren’ have special flower aromas, which might be partly due to the same reason as that speculated by Messina et al. (2009) [8].

Environmental factors, such as pollution, temperature, and drought, also affect the release of flower aromas [4]. The environmental temperature during the flowering period has an important impact on the natural release of flower aromas [25]. Higher volatility phenylpropanoids were emitted by Trifolium repens L. flowers at 15 °C, whereas their emission was reduced at 10 °C [26]. The maximum emission rate for Petunia axillaris flowers occurred at 30 °C compared with the ambient temperatures of 20, 25, and 35 °C [27]. Phenylpropanoid-based floral-scent production decreased in two Petunia hybrida varieties when the ambient temperature increased [28]. The flowering times of the germplasms in this study varied from early February to early April. ‘Gansutao2’, which was in cluster I, and ‘Yuntaishanshantao’,‘Dazhuanggansutao’, and ‘Zhouxinshantao’ in cluster II, all flower in February when the temperature was about 1–9 °C (Nanjing, China; temperature data taken from the National Meteorological Science Data Center; website: http://data.cma.cn/ (accessed on 21 July 2022)); whereas ‘Xinjiangpantao’ and ‘Xinjianghuangrou’ in cluster III, bloom at the end of March and the beginning of April when the temperature was about 11–20 °C (the National Meteorological Science Data Center). Most other germplasms bloom in March when the temperature was about 5–14 °C (the National Meteorological Science Data Center). Therefore, the differences between the flower aroma clustering groups might not only be related to the genotype differences among the germplasms, but also to the different temperatures during their flowering periods.

The differences in aroma components among genotypes can be further clarified by accurately analyzing the aroma components using appropriate extraction and identification methods [29]. In this study, the flower aromas produced by the peach germplasms were preliminarily evaluated by a combination of methods; they analyzed the natural volatilization of flower aromas and included non-destructive detection using E-nose equipment. In the future, in order to more comprehensively evaluate, extract, and utilize peach flower aroma substances, the existing flower aroma extraction methods, such as steam distillation, simultaneous distillation, supercritical CO₂ extraction, the headspace method, the thermal desorption method and the solid phase microextraction method, should be optimized. E-nose combined with in-situ micro-Raman spectroscopy, gas chromatography, gas chromatography flame ionization detection, and gas chromatography-mass spectrometry have complementary advantages and would be an important technical guarantee of the results.

5. Conclusions

This study used E-nose detection and analysis to investigate the flower aromas from 50 peach germplasms. The results showed that the main substance types in peach flower aromas were methane, hydrogen sulfide, aromatic components and organic sulfide, and nitrogen oxide substances. The peach flower aroma was relatively simple, and the aromas produced by Prunus kansuensis Rehd., Prunus davidiana (Carr.) Franch and Prunus ferganensis Kost.et Riab. were different from the other germplasms. Three special germplasms were screened out. ‘Hua3’ and ‘Hongfenjiaren’ produced strong aromas, whereas ‘Yuntaishanshantao’ produced a weak flower aroma. The results of this study preliminarily clarified the differences in floral aromas from different peach germplasms and provide basic data for screening peach germplasms with obvious floral aromas; they will also help improve the cultivation of new varieties with strong aromas and the development and utilization of peach floral aroma substances. Future research should undertake more in-depth and comprehensive analyses of peach flower aroma profiles and attempt to optimize and combine multiple analysis methods.