Multivariate analysis of chemical markers to distinguish “Ceylon” and "Cassia” cinnamon in the Spanish market

Abstract

Cinnamon adulteration represents a significant threat to food authenticity, consumer health and market integrity. This study aimed to analyse the presence of Ceylon and Cassia cinnamon in supermarkets across the province of Alicante (Spain) and identify markers to distinguish these species using spectrophotometry, chromatography, and sensory analysis. Cinnamomum zeylanicum, was characterized by malic, butyric, sugars and monoterpenes, while C. cassia by aldehydes, sesquiterpenes and coumarin. C. zeylanicum was mainly associated with “citrus” aroma and high intensity, while C. cassia with “sweet” aroma. Four out of 52 volatiles (β-phellandrene, d-3-carene, cryptone, and eugenol) were exclusive to C. zeylanicum and 9 (limonene, eucalyptol, endo-fenchol, Δ-elemene, cyclosativene, 7-epi-sesquithujene, bisabolene, α-calacorene, and cadalene) were exclusive to C. cassia ground cinnamon. Multivariate analysis revealed that only one of the 16 ground cinnamon samples from the Spanish market closely resembled C. zeylanicum, one was a mix of both, and the remaining were identified as C. cassia.

Graphical abstract

Highlights

• •Color is not a good indicator for ground cinnamon species identification.
• •Malic, butyric and sugars are higher in C. zeylanicum than C. cassia.
• •Monoterpenes dominate in C. zeylanicum, while aldehydes dominate in C. cassia.
• •Four out of 52 volatiles are exclusive to “Ceylon” and 9 are exclusive to “Cassia”.
• •Only 1 out of 16 Spanish ground cinnamon samples belong to C. zeylanicum.

1. Introduction

Currently, food adulterations commonly associated with food fraud, which can occur throughout the entire food chain, represent a problem of great relevance as have become increasingly sophisticated. Therefore, detecting foods that have been adulterated is a complex task when the industry requires quick and simple analytical methods to detect these adulterations. According to the European Parliament resolution of 14 January 2014 on the food crisis, spices are among the commodities that are most vulnerable to fraud, for this reason, this is one of the main concerns of this industry, especially when they are sold in ground form (Cantarelli, Moldes, Marchevsky, Azcarate, & Camiña, 2020). Cinnamon is one of the most used spices in the kitchen of almost all cultures and has a very high risk of fraudulent adulterations such as substitution and dilution of the product. According to the Spanish regulations set forth in “Real Decreto 2242/1984, of September 26, approving the technical-health regulations for the preparation, circulation, and trade of condiments and spices”, cinnamon is defined as dried bark, mostly stripped of its outer (epidermal) layer (BOE, 1984), from trees of the genus Cinnamomum (Lauraceae family) that has around 250 species, but four of them are used commercially: Cinnamomum zeylanicum or Cinnamomum verum, known as Ceylon cinnamon or true cinnamon; Cinnamomum cassia or Cinnamomum aromaticum, known as Cassia or Chinese cinnamon; Cinnamomum burmannii, known as Indonesian or Javanese cinnamon; and Cinnamomum loureiroi, known as Vietnamese or Saigon cinnamon (Jamali, Jalali, Saffari-Chaleshtori, Samare-Najaf, & Samareh, 2020).

In the legislation, definitions of cinnamon are specified separately, therefore the UNE ISO 6539 standard regulates C. zeylanicum species and divides it into 3 types: Ceylon or Sri Lanka type which is also known as true cinnamon, Seychelles type, and Madagascar type (ISO 6539:2015, 2015). While the UNE ISO 6538 standard regulates C. cassia and divides it into 3 other types depending on the origin: China type, Indonesia type and Vietnam type (ISO 6538:1997, 1997). Furthermore, cinnamon represents one of the four spices for which substitutes are allowed. Its commercialization must indicate on the label (i) that it is a substitute; and (ii) its organic matter. This substitute cannot be mixed with spices (BOE, 1984). The two species of cinnamon above mentioned are similar to each other but have different physicochemical and sensory properties as well as price on the market. Cinnamon barks of C. zeylanicum are made up of many thin layers that break easily, they are light brown and have a light, sweet taste, and mild aroma. On the other hand, barks of C. cassia are darker and thicker, formed by a single and very hard layer that is difficult to break, with a more intense and spicier flavor and aroma (Pages-Rebull, Pérez-Ràfols, Serrano, & Díaz-Cruz, 2024). When they are in sticks form, they can be easily differentiated, but in powder form it is very difficult to distinguish them, although C. zeylanicum has a lighter color than C. cassia. Despite their sensory characteristics, the main difference between them is their commercial price and coumarin content (Cantarelli et al., 2020). Cinnamomum cassia is cheaper, which makes it increasingly widespread in the European market, while C. zeylanicum cinnamon has a higher commercial cost and, therefore, it is more prone to adulteration with the first one (Wu et al., 2021). Cinnamomum zeylanicum is frequently adulterated with other parts of the plant which are cheaper, as well as with the bark of C. cassia, which is rougher, thicker, and less aromatic. The powder is often adulterated with cinnamaldehyde or sugar flavored beech shells, ground walnut shells or galangal rhizome, among others (Thomas & Kuruvilla, 2012). Beside price, adulteration also result in a health issue, as for example, C. cassia presents a higher content of the hepatotoxic compound coumarin, whose concentration is reported to reach thousands of times higher than in C. zeylanicum (Pages-Rebull et al., 2024). Regarding coumarin, it is a natural secondary metabolite that is found in many plants, and especially in some species of cinnamon. It is regulated by the European Food Safety Authority (EFSA) due to its liver toxicity in humans that established a tolerable daily intake (TDI) of 0.1 mg coumarin kg^-1 body weight (bw). While C. cassia contains higher levels of coumarin (1 %), the concentration is very low (0.004 %) for C. zeylanicum, making it the most beneficial and safest for regular consumption (Lončar et al., 2020).

Color is one of the most important attributes in the acceptance of a food product. The CIELab color space is the most used method for its determination in food. In the case of cinnamon, as explained before, a lighter color is associated to C. zeylanicum, while a darker color is more associated to C. cassia. On the other hand, one of the most prominent functions of sugars is to provide a sweet taste to the products, and the cinnamon is widely used a healthier option to substitute sugar and sweeten the products, for example, in yogurt (Jiménez-Redondo et al., 2022). Cinnamon has antioxidant activity due to the presence of polyphenols, the most frequently used methods to measure this activity in food are ABTS⁺, DPPH^● and FRAP (Martinelli et al., 2021). In the case of volatile compounds, the essential oil of cinnamon, which contains cinnamaldehyde, eugenol, camphor or caryophyllene, can be extracted by solid phase micro extraction, hydro distillation, solvent extraction and microwave-assisted extraction and its composition is usually determined by gas chromatography and mass spectrometry (GC-MS) (Cardoso-Ugarte, López-Malo, & Sosa-Morales, 2016). While there is extensive information in the scientific literature regarding the volatile profile in cinnamon essential oils, there is limited information about the physicochemical and aromatic characterization of cinnamon powder.

Overall food labelling does not allow for identifying the cinnamon species used, as it is not compulsory to include such information. Meanwhile, the available literature on cinnamon adulteration and authentication is scarce, whereas the potential for adulteration and its food safety implications are highly relevant. Given this scenario, it is highly relevant to contribute to the knowledge on the topic.

Therefore, the aim of the present study was to characterize by analytical techniques and sensory analysis a wide range (16) of commercial ground cinnamon available in the Spanish market, together with C. zeylanicum and C. cassia sticks as reference, to determine potential biomarkers to identify “Ceylon” or “Cassia” cinnamon.

2. Material and methods

2.1. Plant material

The samples of cinnamon sticks and cinnamon powder were purchased in different supermarkets in the Vega Baja region (Alicante, Spain) (Table 1, Fig. 1). Visited supermarkets belonging to nationwide companies, which makes selected samples representative of brands and formats available in Spanish supermarkets. Sixteen types of commercial ground cinnamon (M1 to M16) were compared to 2 types of cinnamon sticks (M17Cs and M18Cy). The sticks belonging to the C. cassia (Cs) and C. zeylanicum species (Cy), were grinded with a Moulinex AR110830 grinder (Moulinex, Écully, France) for 20 s and passed through a 0.417 mm sieve. These samples (M17Cs and M18Cy) were used as standards to check the position of the other 16 ground cinnamon samples found in the Spanish market.

Table 1.

Samples details (code, type, weight, price, etc.) and physical parameters (moisture, water activity and instrumental color)

Sample details						Physical parameters
Code	Description	Weight (g)	Best by (D/M/Y)	Shop	Price (€)			Instrumental color
						Moisture %	aw	L*	a*	b*	Chroma	Hue
						ANOVA Test†
						***	***	***	***	***	***	***
						Tukey Multiple Range Test‡
M1	Ground cinnamon: may contain mustard, gluten, sesame, and celery	40	3/2024	A	1.69	10.7 bcde	0.52 abc	36.2 h	14.3 c	25.5 h	29.2 f	60.7 l
M2	Ground cinnamon	38	2/2024	A	0.37	9.67 def	0.48 abc	46.2 ab	12.5 ef	30.1 bc	32.6 bcd	67.4 e
M3	Ground cinnamon	35	10/2023	B	0.65	8.67 ef	0.50 abc	44.9 ab	11.4 g	28.9 bcde	31.0 de	68.5 c
M4	Ground cinnamon: Origin Vietnam	37	12/2024	C	0.99	7.00 f	0.47 bc	44.0 bcd	8.47 h	26.1 gh	27.4 g	72.0 a
M5	Ground cinnamon	47	22/12/2020	D	0.48	12.7 abc	0.54 abc	44.1 cdef	16.2 a	26.6 fgh	31.1 cde	58.7 m
M6	Ground cinnamon: Origin Ceylon and Indonesia	52	08/2022	E	0.51	14.3 a	0.52 abc	47.5 a	13.0 de	29.5 bcd	32.3 bcd	66.2 f
M7	Ground cinnamon	39	07/01/2024	F	2.85	13.3 ab	0.54 abc	44.5 abc	14.6 c	28.5 cde	32.1 bcd	62.9 k
M8	Roasted ground cinnamon: Origin Ho Chi Minh City	40	08/2022	F	3.50	8.67 ef	0.42 c	38.6 fgh	14.5 c	25.7 gh	29.5 ef	60.6 l
M9	Ground cinnamon	18	2023	G	0.55	11.0 bcde	0.50 abc	44.2 abcd	11.1 g	29.8 bc	31.8 bcd	69.6 b
M10	Cassia ground cinnamon	42	09/2023	H	1.75	14.3 a	0.63 a	45.9 ab	13.0 de	32.3 a	34.8 a	68.1 d
M11	Ground cinnamon	38	11/01/2024	H	0.89	11.7 abcd	0.56 abc	42.8 bcde	15.2 b	29.9 bc	33.6 ab	63.0 jk
M12	Ground cinnamon	40	16/12/2023	I	0.39	11.7 abcd	0.58 abc	39.8 efg	14.2 c	28.0 def	31.4 cd	63.2 j
M13	Ground cinnamon	43	11/2023	J	1.46	11.3 bcde	0.54 abc	44.3 abc	13.2 d	29.1 bcd	32.0 bcd	65.6 g
M14	Ground cinnamon	35	07/2023	J	0.85	9.67 def	0.54 abc	40.9 def	11.3 g	27.3 efg	29.5 ef	67.6 e
M15	Ground cinnamon from Ceylon	40	02/2025	K	1.17	8.67 ef	0.48 abc	44.6 abc	12.3 f	30.5 b	32.9 bc	68.0 d
M16	Ground cinnamon: Origin Vietnam	90	30/03/2023	L	2.25	10.0 cde	0.58 abc	37.1 gh	12.3 f	26.0 gh	28.8 fg	64.6 i
M17 Cs	Cassia cinnamon sticks ground after	10	11/02/2024	M	0.65	12.7 abc	0.57 abc	38.1 fgh	16.0 a	28.6 cde	32.8 bcd	60.8 l
M18 Cy	Ceylon cinnamon sticks ground after	22	11/02/2024	M	4.29	12.0 abcd	0.55 abc	39.7 efg	11.6 g	25.2 h	27.8 fg	65.2 h

^†*** significant at p < 0.001, respectively.

^‡Values (mean of three replications) followed by the different letter, within the same column, were significantly different (p<0.05), according to Tukey’s least significant difference test.

Fig. 1.

Photography of 18 commercial samples of cinnamon studied in the present research (A), together with their 2D (B) and 3D (C) distribution in the L*, a*, and b* color coordinate system for a better visualization of the color differences.

2.2. Moisture content, water activity and instrumental color

Moisture content was calculated after 3 measurements were done using an infrared balance for humidity (Gram RH-110 model; L’Hospitalet de Llobregat, Barcelona,  Spain). Furthermore, water activity of cinnamon samples was determined with a water activity meter (a_w) (Novasina aw-Sprint TH500; Pfaffikon, Zurich, Switzerland). Three color determinations were made using a Minolta CR-300 Colorimeter (Osaka, Japan). This spectrophotometer uses an illuminant D₆₅ and a 10° observer as references. Results were given as CIEL*a*b* coordinates, which define the color in a three-dimensional space: L*, a*, b*. Parameters such as Chroma or Hue were later computed using the values of the a* and b* color coordinates, and photography of all samples were done (Fig. 1).

2.3. Organic acids and sugars

The identification and quantification of organic acids (OA) and sugars was carried out using a high-performance liquid chromatography (HPLC) equipment (Hewlett Packard 1100 series; Willmington, DE, USA) and the methodology previously described by Lipan et al. (2020) with 1 g of ground cinnamon. Organic acids and sugars were analyzed in triplicate. The results were expressed as g kg^-1 of cinnamon powder.

2.4. Antioxidant activity and total phenol content

To determine the antioxidant activity (AA), the method of extraction consisted of mixing 0.3 g of cinnamon powder with 10 mL of extractant (MeOH / water (80:20, v / v) + 1 % HCl), the mixture was sonicated at 20 °C for 15 minutes and it was stored for 24 hours at 4 °C. Later, was sonicated again under the same conditions and centrifuged at 10.000 rpm for 10 minutes. The antioxidant activity of the extract was measured by 3 methods: ABTS⁺, DPPH^● and FRAP, as described by Miller, Rice-Evans, Davies, Gopinathan, and Milner (1993), Brand-Williams, Cuvelier, and Berset (1995) and Benzie and Strain (1996), respectively. The results were expressed as mmol of Trolox kg^-1 of cinnamon powder. The total phenolic content (TPC) was measured using the Folin-Ciocâlteu colorimetric method (Singleton, Orthofer, & Lamuela-Raventós, 1999). The results were expressed as gallic acid equivalents (GAE), g kg^-1 of cinnamon powder.

2.5. Volatile compounds

Volatile compounds were extracted using headspace extraction (HS). For the extraction, 1 g of ground cinnamon and 100 μL of 1-octanol (1000 mg kg^-1) that was used as internal standard were added in a hermetic vial with polypropylene cap and PTFE (polytetrafluoroethylene)/silicone septa. The vials were placed in an automatic sampler (AOC-6000 Plus, Shimadzu) at 250 rpm and 40 °C, during 60 min, and the gas phase was injected into the GC-MS for analysis. The identification and quantification of volatile compounds was carried out using a Shimadzu GC2030 gas chromatograph and a TQ8040 NX triple quadrupole mass spectrometer as detector, using only one quadrupole (Shimadzu Scientific Instruments, Inc., Columbia, MD, USA). The column used was an X5MS (silphenylene polymer; Teknokroma, Barcelona,  Spain) with dimensions of 30 m (length), 0.25 mm (internal diameter) and 0.25 μm (film thickness). The GC temperature program was as follows: 50 °C 2 min^-1, 3 °C min^-1 up to 180 °C and, 20 °C min^-1 up to 230 °C. Volatile compounds were identified (GCMS (Shimadzu) by comparing: (i) mass spectra (NIST 17 Mass Spectral) and (ii) linear retention indices (C6-C20 n-alkanes, Sigma-Aldrich, Steinheim, Germany) (National Institute of Standards and Technology (NIST), 2023). The results were expressed in mg kg^-1.

2.6. Sensory analysis

A sensory analysis was carried out with 24 trained panelists whose ages ranged from 21 to 64 years, and the gender ratio was 56:44 female: male. Panelists were previously trained: three sessions of olfactory training were carried out. Reference materials for training were ground C. cassia and C. zeylanicum cinnamon, the main attributes were identified and agreed upon in the first session. The following sessions served to make sure that the panelists were able to recognize the characteristic aroma of each species. After the training sessions with the reference patterns, the tasters began to carry out the olfactory phase of the samples under study. The samples were served in covered 3.1 cm x 4.5 cm; odorless, disposable plastic cups coded with three-digit numbers containing 0.5 g of cinnamon. To reduce the fatigue caused by the high number of samples, these were evaluated in 3 different sessions with 6 samples per each and 15 minutes of break among sessions. The panelists were asked about the attributes they identified in each sample during training (“ceylon” aroma, “cassia” aroma, “sweet”, “pepper” “menthol”, “pine”, “citrus”, “curry”, "woody", "toasty", "earthy", "rancid" and "cardboard"), the intensity of the aroma of cinnamon (low, medium or high) and if they associated it with cinnamon C. cassia or C. zeylanicum. The study had been previously approved by the ethical committee of Universidad Miguel Hernández (procedure authorized with code: PRL.DTA.ESN.02.20). Further details on ethical approval are reported in the section Ethical statement of the manuscript.

2.7. Statistical analysis

The statistical analyses were done by using one-way analysis of variance (ANOVA), and data were submitted to Tukey’s multiple range test to compare means. Moreover, to provide a summary of all data a Principal Component Analysis (PCA) was carried out. While for the results of “Check-All-That-Apply” (CATA) method, the data was recorded in a binary format (0=attribute not checked; 1=attribute checked) and a Cochran`s Q test and Correspondence Analysis (CA) was run to process it. In addition, Agglomerative Hierarchical Clustering (AHC) was also carried out to identify the clusters, and the results are plotted directly on the PCA graphic with colored circles. The data for PCA and AHC analysis was previously standardized using z-score. Statistically significant differences were considered when p < 0.05 and were studied using SPSS Program (IBM SPSS Statistics for Windows, Version 29.0.2.0 Armonk, NY, USA).

3. Results and discussion

3.1. Moisture content, water activity and instrumental color

The results of moisture content and water activity are shown in Table 1. The former ranged between 7.0 % (M4) and 14.3 % (M6) (Table 1). The control samples M17 (C. cassia) and M18 (C. zeylanicum) contained a moisture content of 12.7 and 12.0%, respectively. The commercial samples M6 and M10 (14.3 %), M7 (13.3 %) and M5 (12.7 %) registered the highest moisture content. Regarding water activity, it ranged between 0.4 (M8) and 0.6 (M10), without significant differences among samples except for M10 (0.6) and M8 (0.4). Voelker, Sommer, and Mauer (2020) obtained similar values, for a_w (0.50) and Xie et al. (2021) for moisture content and a_w (9.6 % and 0.50, respectively). The moisture content of M5 was similar to that of M17 (C. cassia) while the moisture content of M11 and M12 to that of M18 (C. zeylanicum).

Fig. 1 shows the image of all 18 samples studied in the present research (A), along with their 2D (B) and 3D (C) distribution in the L*, a*, and b* color coordinate system for a better visualization of the color differences, while Table 1 shows their values. Lightness (L*) of the samples, ranged from 36.2 (dark) for M1 to 47.5 (light) for M6; a* color coordinate indicates colors between green-red and in the present study ranged from 8.5 (M4) to 16.2 (M5); while b* values  which represent colors between blue-yellow ranged between 25.2 (M18) and 32.3 (M10). The color coordinates values for the reference samples M17 (C. cassia) were L*=38.1; a*=16.0; and b*=28.6, while for M18 (C. zeylanicum) were L*=39.7; a*=11.6; and b*=25.2. This means that cinnamon C. cassia was characterized by a darker brown color due to the low values  of L* and high values  of a* and b*, while cinnamon C. zeylanicum presented a lighter cream color due to the highest value of luminosity and lower values  of red a* and yellow b* color coordinates. With respect to the ground cinnamon, M6 (47.5), M2 (46.2) and M10 (45.9) were the samples that presented the greatest lightness, without significant differences between them, indicating a light color of these samples. Regarding the coordinate a*, M5 (16.2) was the one that showed a redder color together with M17 (C. cassia), without significant differences between them, followed by M11 (15.2). The b* coordinate indicated that the sample M10 (32.3) was the one with the most yellowish coloration together with M15 (30.5), M2 (30.1), M9 (29.8), M6 (29.5), M13 (29.1) and M3 (28.8), without showing significant differences between them. Regarding chroma (C*), M10 (34.8) and M11 (33.6) were the most intense. Considering the hue angle (h*), the highest values  were found for M4 (72.0 °), M9 (69.6 °) and M3 (68.5 °) and the lowest for M5 (58.7 °), M8 (60.6 °) and M1 (60.7 °), oscillating between orange and yellow. The most similar sample to C. cassia in terms of luminosity was M8 and to C. zeylanicum, M12. Regarding the coordinate a*, M5 was the most like M17 and M3, M9 and M14 were the most similar to M18. As for b*, M7 and M12 were the most like M17 while M1 and M3 to M18.

As can be seen, there is a great variability in the color of the samples, and this may be due to the different raw materials used in the elaboration of the final product. Food fraud in spices, including cinnamon, where they are adulterated with other inferior quality ingredients or other parts of the tree of less economic value, as well as extenders to increase the volume and artificial colors to mask their real quality, are frequent (Modupalli, Naik, Sunil, & Natarajan, 2021). This makes the color not a good indicator of quality and not a good indicator of the species of cinnamon used in the elaboration of commercial ground cinnamon because the presence of these compounds can interfere with the real color of the samples.

3.2. Organic acids and sugars

The profiles of organic acids (OA) and sugars are represented in Table 2. Organic acids contribute to the organoleptic characteristics of foods by providing flavor and aroma, hence their qualitative and quantitative determination in cinnamon is important. The existence of organic acids in samples could explain the ability of their extracts to accentuate the flavor of food (Duan, Huang, Xiao, Zhang, & Tang, 2020). The main OA found were citric, malic, succinic, propionic, and butyric, at concentrations that ranged between 0.9-3.9; 0.5-15.5; 8.2-30.3; 5.0-76.2; and 0.2-53.6 g kg ^-1, respectively. Of all analyzed samples, M18 was the one with the highest total OA content (124 g kg^-1) followed by M10 (105 g kg^-1). The sample that most closely resembled to cinnamon C. cassia in terms of total OA, was M5 as well as to M2, M3, M4, while M10 was similar to C. zeylanicum. The predominant organic acid in the samples was propionic (311 g kg^-1), followed by succinic (292 g kg^-1) and citric was the least abundant (27.7 g kg^-1) OA. The concentration of citric, malic, succinic, propionic, and butyric in the reference sample M17 (C. cassia) was 1.39, 0.54, 16.7, 19.7 and 4.15 g kg^-1, respectively, while, for the other reference sample, M18 (C. zeylanicum), the concentrations was 2.92, 15.5, 12.6, 39.3 and 53.8 g kg^-1, respectively. In commercial samples, citric, succinic, and malic were significantly higher in M15 (3.92, 30.3 and 13.9 g kg^-1, respectively), butyric in M4 (11 g kg^-1) and propionic in M10 (76.2 g kg^-1). As observed, malic and butyric acid were 29- and 13-fold higher in C. zeylanicum than in C. cassia. However, none of the other samples obtained values similar to C. zeylanicum for these compounds. For instance, M6 and M15 which were labelled as Ceylon, registered lower values for these compounds. Although M15 was the sample with the highest value of malic acid (13.9 g kg^-1) after C. zeylanicum.

Table 2.

Organic acids, sugars, antioxidant activity and total phenolic content (TPC)

	Organic acids (OA)						Sugars (S)				Antioxidant activity and total polyphenol content (TPC)
	Citric	Malic	Succinic	Propionic	Butiric	ΣOA	Sucrose	Glucose	Fructose	ΣS	ABTS+	DPPH●	FRAP	TPC
	g kg-1						g kg-1				mmol Trolox kg-1			g GAE kg-1
	ANOVA Test†
Samples	***	***	***	***	***	***	***	***	***	***	***	***	***	***
	Tukey Multiple Range Test‡
M1	0.98 h	1.42 ij	8.15 i	5.02 e	2.50 j	18.1 j	1.79 i	13.5 e	6.60 g	21.9 f	118 cd	908 def	108 f	55.3 d
M2	1.24 defgh	6.34 de	19.1 c	13.9 cde	5.11 ef	45.7 ef	4.03 bc	14.9 cde	12.6 bc	31.5 c	129 c	913 def	143 de	90.7 b
M3	0.99 gh	6.46 de	15.8 def	18.1 cd	7.98 c	49.4 def	3.06 de	14.1 cde	11.5 cde	28.8 de	86.7 efg	947 a	68.0 hij	30.4 fg
M4	1.04 fgh	5.25 f	15.3 cde	13.8 cde	10.9 b	46.4 ef	2.54 efgh	14.0 de	11.0 de	27.5 e	69.2 gh	880 gef	52.7 j	14.2 h
M5	1.52 cd	2.02 i	15.8 def	19.2 cd	3.90 hi	42.4 efg	2.71 ef	14.2 cde	10.1 e	27.1 e	207 a	906 bcde	235 b	131 a
M6	1.21 defgh	5.10 fg	13.1 fgh	5.01 e	3.02 ij	27.5 ij	2.72 ef	14.3 cde	10.1 e	27.1 e	132 c	922 ab	137 e	66.3 c
M7	1.73 c	3.07 h	23.7 b	31.2 b	0.29 k	60.0 c	2.79 ef	14.3 cde	10.5 de	27.6 e	226 a	940 abcd	274 a	138 a
M8	1.33 def	8.18 c	11.2 h	5.98 e	6.71 d	33.4 ghi	2.06 ghi	13.9 de	7.66 fg	23.6 f	90.2 ef	930 abc	87.8 g	44.4 e
M9	1.19 efgh	6.61 de	18.8 cd	18.7 cd	5.94 de	51.4 cde	2.62 efg	14.7 cde	10.7 de	28.0 de	77.7 fgh	936 abc	77.9 ghi	22.4 gh
M10	1.72 c	3.14 h	23.7 b	76.2 a	0.29 k	105 b	3.55 cde	15.1 cde	11.5 cde	30.4 cd	155 b	892 fg	163 c	70.6 c
M11	1.13 efgh	8.68 c	11.4 h	6.65 e	4.16 gh	32.01 hi	4.16 fgh	13.8 de	10.9 de	27.1 e	81.9 fgh	928 abcde	78.5 ghi	33.7 f
M12	1.11 fgh	8.47 c	11.9 h	6.68 e	4.11 h	32.3 hi	2.73 ef	16.7 de	11.0 de	27.4 e	78.0 fgh	907 def	68.1 hij	28.4 fg
M13	1.81 c	4.21 g	18.4 cd	4.56 e	3.07 ij	32.0 hi	4.16 b	15.5 c	12.7 abc	32.4 c	107 de	941 ab	112 f	47.1 de
M14	1.30 defg	6.78 d	13.4 fgh	12.4 cde	6.50 d	40.4 fgh	3.04 de	14.8 cde	13.6 ab	31.4 c	61.8 h	915 cdef	61.8 ij	17.0 h
M15	3.92 a	13.9 b	30.3 a	5.25 e	5.04 fg	58.4 cd	5.24 a	19.9 a	14.1 a	39.2 a	154 b	938 ab	160 cd	73.3 c
M16	1.14 efgh	5.81 ef	12.9 fgh	9.57 de	5.33 ef	34.8 ghi	2.27 fghi	14.1 cde	7.82 fg	24.1 f	82.6 fg	915 cdef	81.0 gh	27.3 fg
M17 Cs	1.39 de	0.54 j	16.7 cde	19.7 c	4.15 gh	42.5 efg	2.01 hi	13.9 de	8.24 f	24.1 f	105 de	929 abcd	133 e	53.5 de
M18 Cy	2.92 b	15.5 a	12.6 gh	39.3 b	53.8 a	124 a	3.82 bc	17.5 b	13.8 ab	35.1 b	114 cd	950 a	131 e	55.6 d

^†*** significant at p < 0.001, respectively.

^‡Values (mean of three replications) followed by different letter, within the same column, were significantly different (p<0.05), according to Tukey’s least significant difference test.

Considering sugars, whose main function is to add sweetness to the food, but also to improve the sensory profile and the intensity of flavors (Hutchings, Low, & Keast, 2019), in this study, sucrose (1.7-5.2 g kg^-1), glucose (13.5-19.8 g kg^-1) and fructose (6.6-14.1 g kg^-1) were the identified sugars (Table 3). The content was higher in C. zeylanicum (3.82, 17.52 and 13.78 g kg^-1) compared to C. cassia (2.01, 13.89 and 8.24 g kg -1). Regarding the rest of the samples, M15 was the one with the highest content in sucrose, glucose, fructose and so the total sugars (5.24, 19.8,14. 1, 39.2 g kg^-1, respectively). The total sugar content of M1, M8 and M16 was as the same as for M17 (C. cassia), while the total sugar content of M15 was the most similar to that of M18 (C. zeylanicum). Once again, great variability can be observed in the content of organic acids and sugars in the analyzed samples, this must be related to the quality of cinnamon, which depends on factors such as the botanical source, climatic conditions, and harvesting and production methods (Avula, Smillie, Wang, Zweigenbaum, & Khan, 2015).

Table 3.

Volatile profile found in the cinnamon powder from Spanish market, retention index and main odor descriptors (The Good Scent Company, 2021; National Center for Biotechnology Information, 2021).

	Retention Index
Code	Compound	Chemical family	RTa (min)	Experimental	Literaturab	Odor descriptor
V1	n-Hexanal	Aldehydes	5.388	789	787	Rancid
V2	Styrene	ACc	8.554	889	888	Balsamic, floral
V3	α-Pinene	Monoterpenes	10.600	919	917	Pine, herbal, earthy
V4	Camphene	Monoterpenes	11.444	935	929	Woody, pine, herbal
V5	Benzaldehyde	Aldehydes	11.950	952	955	Bitter almond
V6	β-pinene	Monoterpenes	12.954	962	965	Herbal, woody, pine, green
V7	Myrcene	Monoterpenes	13.632	976	983	Spicy, balsamic, woody
V8	α-Phellandrene	Monoterpenes	14.602	995	1003	Pepper, woody, citrus
V9	α-Terpinene	Monoterpenes	15.254	1004	1012	Citrus, woody, thyme
V10	Para-Cimene	Monoterpenes	15.701	1012	1017	Cumin, cilantro
V11	Limonene	Sesquiterpenes	16.080	1015	1020	Citrus, lemon peel, sour
V12	β-Phellandrene	Monoterpenes	16.114	1018	1013	Mint
V13	Eucalyptol	Monoterpenes	16.190	1018	1020	Eucalyptus, herbal, medicinal
V14	β-(E)-Ocimene	Monoterpenes	17.156	1034	1034	Floral, green, vegetal, tropical
V15	γ-Terpinene	Monoterpenes	17.864	1044	1035	Citrus, lime, tropical, oily
V16	Acetophenone	Ketones	18.240	1053	1057	Sweet, marzipan, vanilla
V17	Terpinolene	Monoterpenes	19.646	1071	1079	Pine, sweet citrus, lemon peel
V18	Linalool	Sesquiterpenes	20.639	1088	1090	Floral, roses, orange
V19	Endo-Fenchol	Monoterpenes	21.817	1103	1110	Mint, earthy, woody
V20	Δ-3-Carene	Monoterpenes	22.224	1109	1109	Pepper, juniper, wasabi
V21	α-Campholenic aldehyde	Aldehydes	22.325	1111	1111	Vegetal, cilantro
V22	Camphor	Monoterpenes	23.681	1130	1126	Camphoraceous, pungent
V23	Benzyl acetate	Esters	24.652	1147	1151	Floral, jasmine, fruity
V24	Isoborneol	Monoterpenes	25.383	1156	1156	Balsamic, pine, woody
V25	Terpinen-4-ol	Monoterpenes	26.005	1165	1174	Spicy, mentholated, pepper
V26	Cryptone	Monoterpenes	26.412	1171	1180	-
V27	α-Terpineol	Monoterpenes	26.980	1180	1189	Floral, lilac, pine, citrus
V28	Cis-Pinocarveol	Monoterpenes	27.476	1188	1184	Pine, mint, pungent
V29	Cuminaldehyde	Aldehydes	29.997	1226	1226	Spicy, spiced, cumin
V30	(E-)-Cinnamaldehyde	Aldehydes	31.670	1255	1260	Cinnamon, sweet, spicy
V31	Bornyl acetate	Esters	32.403	1268	1270	Pine, cedar, spicy
V32	Δ-Elemene	Sesquiterpenes	35.186	1318	1326	Lavender, woody, woody
V33	Eugenol	Phenols	35.906	1333	1341	Clove, spicy, cinnamon
V34	Cyclosativene	Sesquiterpenes	36.808	1351	1358	Floral
V35	α-Cubebene	Sesquiterpenes	37.244	1359	1354	Herbal, waxy
V36	α-Funebrene	Sesquiterpenes	38.563	1385	1385	-
V37	7-Epi-Sesquithujene	Sesquiterpenes	39.040	1394	1393	-
V38	(E-)-Caryophyllene	Sesquiterpenes	39.305	1400	1409	Spicy, clove
V39	Coumarin	Phenols	39.767	1410	1406	Tonka bean, hay
V40	α-cis-Bergamotene	Sesquiterpenes	39.970	1414	1415	Spicy, citrus, tea
V41	(E-)-Cinnamyl acetate	Esters	40.385	1424	1418	Spicy, floral, cinnamon, spiced
V42	α-Guaiene	Sesquiterpenes	40.610	1430	1435	Guaiac wood, violets, roses
V43	α-Humulene	Sesquiterpenes	40.934	1436	1432	Woody, pine
V44	γ-Muurolene	Sesquiterpenes	41.836	1456	1461	Oily, woody, spicy
V45	α-Curcumene	Sesquiterpene	42.019	1462	1459	Curry, spicy
V46	Naphtalene	ACc	42.413	1469	1471	Pungent, acrid, resinous
V47	α-Muurolene	Sesquiterpene	42.888	1479	1480	Pungent, tar
V48	Bisabolene	Sesquiterpene	43.321	1487	1482	Green, banana, fruity
V49	Trans-Calamenene	Sesquiterpene	43.846	1500	1510	Hops
V50	α-Calacorene	Sesquiterpene	44.667	1520	1523	Tar
V51	Fokienol	Sesquiterpene	46.365	1561	1568	Ethereal, fruity, grape
V52	Cadalene	Sesquiterpene	49.944	1650	1654	-

^aRT, retention time;

^bNIST (National Institute of Standards and Technology) (NIST, 2021);

^cAC: aromatic hydrocarbons

3.3. Antioxidant activity and total phenolic content

Table 2 shows the results of antioxidant activity (AA) and total phenol content (TPC) analyzed in cinnamon. It has been shown that different parts of cinnamon have an important AA thanks to the presence of a large quantity of phenolic compounds (Almatroodi et al., 2020). In the present study elevated AA values  were found (62-226, 880-947 and 53-274 mmol Trolox kg^-1 for ABTS⁺, DPPH^● and FRAP, respectively). For its part, the TPC ranged between 14 and 138 g GAE kg^-1, this variation agrees with the values  reported by Su et al. (2007) that reported a TPC of 14.8 g kg^-1 and Shan, Cai, Sun, and Corke (2005) of 119 g kg^-1.

Pure cinnamon C. cassia and C. zeylanicum (M17 and M18) showed values of 105, 929 and 133 mmol Trolox kg^-1 (ABTS⁺, DPPH^● and FRAP) and 114, 950 and 131 mmol Trolox kg^-1 (ABTS ⁺, DPPH^● and FRAP), respectively, without significant differences between them. These values  were lower than those reported in the scientific literature (525 and 637 mmol Trolox kg^-1 ABTS and FRAP, respectively) and (4130 and 1880 mmol Trolox kg^-1 ABTS⁺ and DPPH^●, respectively) for C. zeylanicum (Lu, Yuan, Zeng, & Chen, 2011; Vallverdú-Queralt et al., 2014). Factors such as extraction solvent, extraction method, species type and plant parts might affect the composition of bioactive compounds in extracts and, therefore, their antioxidant activity (Tang, Chen, Qin, Hou, & Deng, 2020), which would explain the differences observed in this study. The TPC was also similar between the control samples, 54 g GAE kg^-1 (M17) and 56 g GAE kg^-1 (M18). Considering ABTS⁺ and FRAP methods, the samples M5, M7, followed by M10 and M15 registered the highest AA, while M5 was also the sample with the highest content of TPC130 g GAE kg^-1. The most similar sample to C. cassia was M13 and to C. zeylanicum was M1 for ABTS⁺ method and TPC. Since the antioxidant capacity of a food product is determined by interactions between the compounds that compose it, including organic acids and polyphenols with antioxidant properties, that has exhibited great variation between samples in this study, reflected in the AA. This being able to explain the great variability observed in this parameter.

3.4. Volatile compounds

A total of 52 volatile compounds were identified in the cinnamon powders purchased in the Spanish market and are presented in Table 3 in which their retention time, and retention index used for the identification of aromatic compounds and their odor descriptors, are also presented. Moreover, the concentration of each volatile compound and their significant differences among samples for each compound is presented in Table 1S. Regarding the chemical families (Fig. 2), monoterpenes (n = 20), sesquiterpenes (19), aldehydes (5), esters (3), phenols (2), aromatic hydrocarbons (2) and ketones (1) were found in these cinnamon samples. Using the average of all samples, monoterpenes (5472 mg kg^-1), sesquiterpenes (3707 mg kg^-1) and aldehydes (3730 mg kg^-1) were the major chemical families of volatile compounds found in the present samples, followed by aromatic hydrocarbons (262 mg kg^-1), esters (188 mg kg^-1), phenols (5.,6 mg kg^-1) and ketones (7.60 mg kg^-1). However, this tendency depends on each sample. For instance, regarding the reference samples, M18 which corresponds to C. zeylanicum, followed this trend for the monoterpenes (9347 mg kg^-1), sesquiterpenes (7533 mg kg^-1) and aldehydes (5317 mg kg^-1), while quite different tendency was observed for the lowest chemical families which presented the following order: esters (179 mg kg^-1), aromatic hydrocarbons (34.3 mg kg^-1) and phenols (22.9 mg kg^-1). While the major chemical families found for M17 (C. cassia) were first aldehydes (9987 mg kg^-1), sesquiterpenes (8387 mg kg^-1) and monoterpenes (6307 mg kg^-1), followed by esters (825 mg kg^-1), phenols (116 mg kg^-1) and aromatic hydrocarbons (39.1 mg kg^-1). According to these data, the true cinnamon, C. zeylanicum, is higher in monoterpenes, and lower in all the other chemical families than C. cassia. Is important to highlight that both reference samples C. zeylanicum and C. cassia presented a lower amount of aromatic hydrocarbons and no ketones. As observed in Fig. 2, this chemical family was found in all the other ground samples except, M10 (which was labeled as Cassia), M15 (labeled as Ceylon) and the reference samples commented above (M17 and M18). As seen, sample M10 presented a different behavior than the other samples having a higher amount of mono and sesquiterpenes mainly α-pinene, camphene, β-pinene, limonene, and eucalyptol. This sample was labeled as Cassia, however, if we compare with M17 which is the reference for C. cassia, is very different in terms of concentration of compounds, moreover, has a mix of compounds found both in C. cassia and C. zeylanicum, but in higher amounts. Other authors evaluated the volatile compounds in different constituents of cinnamon as fruit, bark (our case), leaf and root, and cinnamon root highlighted higher amounts of these compounds (Paranagama et al., 2001; Senanayake, Lee, & Wills, 1978). Thus, roots or other parts were higher in pinene and limonene volatiles than bark.

Fig. 2.

Graphical visualization of the cinnamon powders purchased on the Spanish market based on the chemical family of aromatic compounds (mg kg^-1)

Inside these chemical families in authentic C. cassia cinnamon (M17) highlighted (Table 1S) the volatiles alpha-pinene (2311 mg kg^-1), eucalyptol (1591 mg kg^-1), cinnamaldehyde (9862 mg kg^-1), α-cubebene (4706 mg kg^-1) and caryophyllene (2032 mg kg^-1). These volatiles are characterized by aromas of pine and earthy; eucalyptus and medicinal; cinnamon and sweet; herbal and waxy; and clove and spicy, respectively. In authentic C. zeylanicum cinnamon (M18) stood out, p-cymene (1976 mg kg^-1), β-phellandrene (4561 mg kg^-1), linalool (1948 mg kg^-1), cinnamaldehyde (5122 mg kg^-1) and caryophyllene (4381 mg kg^-1), compounds with cumin and coriander aroma, floral and roses notes and, cinnamon and clove aroma, respectively. Regarding the rest of the samples, it should be noted that sample M10 had the highest concentration of limonene (2612 mg kg^-1), a volatile characterized by its citrus aroma and lemon peel, followed by M7 (541 mg kg^-1) and M6 (481 mg kg^-1), among which, there were no significant differences neither with M17 (C. cassia). β-phellandrene compound was found only in M15 (2150 mg kg^-1), M6 (448 mg kg^-1) and in true C. zeylanicum cinnamon (M18). Again, M10 was the commercial cinnamon with the highest concentration of eucalyptol (12626 mg kg^-1). Eucalyptol was absent in M18 and M15 and it should be noted that M1 (1497 mg kg^-1) had no significant differences with respect to M17 (1591 mg kg^-1) for this volatile. Linalool was found at low concentration in commercial samples, the highest quantity was found in M18 (1948 mg kg^-1) and M15 (411 mg kg^-1). High concentrations of α-cubebene were reported in M10 (4939 mg kg^-1), followed by M12 (3095 mg kg^-1), M7 (2947 mg kg^-1), M11 (2846 mg kg^-1) and M6 (2646 mg kg^-1) without significant differences among them. Finally, samples M10 and M7, presented the highest caryophyllene content (1020 mg kg^-1 and 1296 mg kg^-1, respectively). Thus, comparing all the samples, M10 stood out for containing the highest content of limonene, eucalyptus and α-cubebene, being, in addition, the one with the highest concentration of caryophyllene among the commercial samples. In the reference M18 and in sample M15 no limonene or eucalyptol was found. Likewise, these two samples and M6 were the only ones where the volatile compound β-phellandrene was identified.

Other important volatile compound was eugenol due to its high antioxidant activity (Ribeiro-Santos et al., 2017). Authors reported that eugenol is absent or only in traces in C. cassia cinnamon but it is a prominent volatile found in true cinnamon (Jose, Leela, Zachariah, & Rema, 2019). As observed in our results, it was only detected in cinnamon samples identified as C. zeylanicum (M18 and M15) at concentrations of 17. 3 and 1.99 mg kg^-1, respectively.

Fig. 3A represents the total content of volatile compounds, which ranged between 3756 mg kg^-1 (M16) and 59994 mg kg^-1 (M10). The C. cassia cinnamon (M17) had a total content of 25661 mg kg^-1 and C. zeylanicum (M18) a total content of 22433 mg kg^-1. The commercial sample M10 was the one with the highest volatile content followed, although, with a much lower content, by M7 (25126 mg kg^-1) and C. cassia reference M17 (25666 mg kg^-1). Among all the volatile compounds, cinnamaldehyde needs more attention (Fig. 3B), as is considered the most important aromatic compound of cinnamon and responsible for its typical odor (Silvis, Luning, Klose, Jansen, & van Ruth, 2019). This was the most abundant compound in all the analyzed samples, being C. cassia (M17) with the highest concentration (9862 mg kg^-1) while in C. zeylanicum (M18), recorded 5122 mg kg^-1. Among the other commercial samples, M8 (6364 mg kg^-1) was the one with the highest concentration, followed, although in much lower concentration, by M10 (3977 mg kg^-1) and M7 (3898 mg kg^-1). As previously reported cinnamaldehyde content was higher in C. cassia compared to C. zeylanicum (Ananthakrishnan, Chandra, Kumar, & Rameshkumar, 2018) which is consistent with what was found in this study. Another important volatile compound which is necessary to highlight is coumarin (Fig. 3C) with notes of Tonka bean and hay but regulated by EFSA due to its liver toxicity in humans. According to the scientific literature, this compound, is characteristic to C. cassia (Lončar et al., 2020), as well as in the present study, where the highest content was found in M17 C. cassia reference sample (116 mg kg^-1) and in M8 (112 mg kg^-1). This last sample according to its labeling is Saigon cinnamon (C. loureiroi), an origin that the UNE ISO 6538 standard classifies it as Cassia cinnamon (Standardization, 1997). On the other hand, the lowest content was observed in M15 (1.4 mg kg^-1) and in the C. zeylanicum control (5.6 mg kg^-1). This indicates that, except M15, all the other samples belong to the species C. cassia or contain a mixture with this type of cinnamon. Similar results were previously reported for Cinnamomum zeylanicum Blume in accessions from germplasm established at the National Cinnamon Research and Training Center, Palolpitiya, Thihagoda, Matara, Sri Lanka at 6.0263, 80.5623 managed by the Department of Export Agriculture (Liyanage et al., 2021). Moreover, the authors revealed that 360 accessions out of a total of 515 (70 %), did not have a detectable level of coumarin in the bark. Thus, considering only those accessions with detectable amount of coumarin the average value reported was 34 mg kg^-1, compared to 10 mg kg^-1 if all 515 accessions were included. So, in the present study, low concentrations of coumarin were only obtained by M18 and M15 samples. Coumarin is hepatotoxic and its habitual consumption in large concentrations could represent a health risk and may cause liver failure. Thus, C. zeylanicum with low concentration of coumarin could be exempt (Farag, Labib, Noleto, Porzel, & Wessjohann, 2018). Despite its toxicity, EFSA determined that an intake 3 times higher than the TDI (0.1 mg kg^-1 bw) for 1 or 2 weeks did not represent a safety concern.

Fig. 3.

Cinnamaldehyde, coumarin and total volatiles content found in cinnamon commercial samples purchased in Spanish mark

3.5. Sensory analysis

In the sensory analysis, the control sample M17 (C. cassia) was mainly associated with the attributes “Cassia” and “sweet” aroma (71 % of the panelists identified both attributes in the sample), and to a lesser extent, pepper, and woody (33 %). The other reference sample, M18 (C. zeylanicum), was mainly associated with the attributes aroma of "Ceylon", "citrus" and "sweet"; these attributes were selected by 58 %, 54 % and 46 % of the panelists, respectively. The "sweet" aroma detected in both samples is due to the volatile cinnamaldehyde, responsible for the sweet taste of cinnamon, producing a synergistic effect that increases the sweet sensation when combining the sweet taste of sugar and the sweet aroma of cinnamon (Ribeiro-Santos et al., 2017). This odor was perceived to a greater degree in M17, because was the sample with the highest content of cinnamaldehyde, while the "citrus" aroma perceived in M18 could be due to the compound para-cymene with citrus and lemon notes (The Good Scents Company Information System, 2021).

Regarding commercial samples, the attributes selected by a greater proportion of panelists were: in M1 woody (50 %), aroma of “Cassia” (42 %) and "Pine" (33 %); M2 aroma of “Cassia” (42 %) and sweet (33 %); M3 aroma of Cassia (50 %), pepper (42 %) and woody (33 %); M4 woody, roasted (38 %) and aroma of Ceylon (33 %); M5 woody (46 %) and aroma of “Ceylon” (29 %); M6 Ceylon odor (42 %), Cassia and woody aroma (33 %); M7aroma of Ceylon (46 %), pine (38 %) and menthol (33 %); M8 aroma of Cassia (63 %), sweet (46 %) and aroma of Ceylon (33 %); M9 stale (67 %), earthy (42 %), cardboard and aroma of Cassia (38 %); M10 mentholated (54 %), pepper (42 %), aroma of Cassia and pine (38 %); M11 sweet (42 %), aroma of Cassia and woody (38 %); M12 Ceylon and woody aroma (42 %), Cassia aroma (38 %); M13 pepper (50 %), aroma of Ceylon (46 %) and sweet (38 %); M14 Cassia scent, woody and earthy (42 %); M15 aroma of Ceylon (71 %), pepper (46 %), sweet and minty (42 %); M16 woody (63 %), cardboard (54 %) and earthy (46 %). These attributes selected by the panelists are associated with the volatile compounds camphene and α-humulene (woody, pine), Δ-elemene (woody), α-phellandrene (citrus, pepper), Δ-3-carene (pepper), eucalyptol (eucalyptus, minty, herbal), cinnamaldehyde (cinnamon, sweet, spicy) and acetophenone (sweet, vanilla) (The Good Scents Company Information, 2022), among others, and which were identified in the analyzed cinnamon samples. It should be noted that (i) samples M7 and M10 were associated with the attribute “menthol”, coinciding with the two samples with the highest eucalyptol content; (ii) sample M8, which was related to the "sweet" odor, was the commercial sample with the higher content of cinnamaldehyde; and (iii) sample M15, associated with the attribute "pepper" for its content of beta-phellandrene, an attribute also associated with M18.

In accord with the intensity of the cinnamon aroma, panelists considered that most of the samples had a medium intensity, except for samples M2, M3, M4, M7 and M12 where a low intensity was perceived (Fig. 4). In M18, 50 % perceived a medium-high intensity and in M5 the same proportion of panelists (46 %) noted both medium and low intensity. In general, M18 was the sample that was perceived with greater aroma of cinnamon followed by M10, M15 and M17. Conversely, M3, M4 and M12 were the samples where this intensity was lower. Finally, samples M5, M6, M10, M11, M13, M15 and M18 were associated with C. zeylanicum cinnamon and samples M1, M2, M3, M4, M7, M8, M9, M12, M14, M16 and M17 were identified with the C. cassia specie.

Fig. 4.

Rank sum scale of participants scores of the commercial cinnamon aroma intensities.

3.6. Principal Component Analysis (PCA) after Varimax Rotation, Correspondence Analysis (CA) for Check-All-That-Apply (CATA) data, and Agglomerative Hierarchical Clustering (AHC) of cinnamon samples purchased in the Spanish market

To provide a summary of the original data a Principal Component Analysis (PCA) and Correspondence Analysis for CATA data was carried out and the results are presented in Fig. 5. For a better characterization and visualization of the samples, 4 biplots were included in this figure, 3 of them (A, B, and C) are related to physicochemical results, chemical families of volatile compounds and sensory parameters, respectively, while the last one (D) includes all the parameters together. Agglomerative Hierarchical Clustering (AHC) was also carried out to identify the clusters, and the results are plotted directly on the PCA graphic with colored circles, each color representing a different cluster. Moreover, the variables explained by the 1^st principal component (PC-1) were colored in green, while those explained by 2^nd principal component (PC-2) were represented in red color; thus, green vectors on left and right sides are negatively correlated among them, as well as the red variables from the top with the red variables from the bottom side. The first two principal components represented in Fig. 5A, explained 65 % of total variability, PC-1 explained 39 % and PC-2 26 % and allowed us to identify the samples which can be interpreted, in terms of proximity and characterized by specific parameters. The variables significant on the PC-1 were lightness, organic acids, and sugars, and those significant on the PC-2 were a* and b* color coordinates, propionic acid, antioxidant activity, and total phenolic content. As seen, in general, samples located on the right and upper side were more intense in mainly all parameters, and lower intensities were found for the samples grouped on left and mainly bottom side. For example, samples M10, M13 and M15 were characterized by a lighter color (L*) with yellowish notes (b*), while M1, M5, M7, M8, M11, M12, and M17 with more reddish notes. In terms of antioxidant activity (ABTS, FRAP) and total phenolic content (TPC), explained by the PC-2, all samples located on the left side of PC-2 (top and bottom) but also M1 and M18 from the right side, presented a higher antioxidant activity and a TPC. The opposite was observed for the other samples, found on the purple circle. Here is important to highlight the positive correlation between a* color coordinate and these parameters (ABTS, r=0.46; FRAP, r=0.49*; TPC, r=0.52*), this means that samples with reddish notes met a higher TPC and antioxidant activity. Finally, samples M7, M10 as well as M15 and M18 were characterized by a higher content of citric, succinic, propionic (except for M15), and so a total organic acid. Same samples together with M2 and M13 were also characterized by a higher content of sucrose, glucose, fructose, and so total sugars.

Fig. 5.

Principal component analysis biplot of color, organic acids, sugars, antioxidant activity and total phenolic content (TPC) of commercial cinnamon samples found in the Spanish market (A); PCA biplot of chemical families of volatile compounds (B); correspondence analysis symmetric plot of attributes frequency data generated by “Check-All-That-Apply” (CATA) method (C), and PCA of all mentioned parameters together (D).

In Fig. 5B the samples are summarized in terms of chemical families of aromatic compounds. Here the first two principal components explained 69 % of the total variability, and AHC clustered the samples in 5 groups. While PC-1 explained 46 % with the significant variables being esters, sesquiterpenes, total volatiles, monoterpenes on the right side and ketones and hexanal on the left side, the PC-2 explained 23 % of the variability with only 2 significant parameters, phenols, and aldehydes. As observed, at this step, hexanal was separated by the aldehydes family, because it was considered a negative volatile, related to rancid notes. The biplot shows that M7, M10 and M18 were mainly characterized by monoterpenes; M10 and M17 by sesquiterpenes; M8 and M17 by aldehydes; M7, M10 and M17 by esters; M3, M8 and M17 by phenols, M1, M6, M10, and M12 by aromatic hydrocarbons; and M1, M2, M3, M4, M9, M14, M16 were characterized by ketones (acetophenone) and hexanal.

The symmetric plot presented in Fig. 5C summarizes the samples position in terms of sensory parameters selected in “Check all that apply” (CATA) test by the semi-trained participants. For this test, the data was recorded in a binary format (0=attribute not checked; 1=attribute checked) and a Cochran`s Q test and Correspondence Analysis was run to process these data. As explained in the materials and methods section the attributes given to the participants were related to cinnamon volatile compounds and product ingredients (“Ceylon aroma”, “Cassia aroma”, “Sweet aroma", "Black Pepper”, “Minty”, “Pine”, “Citric”, “Curry”, “Woody”, “Toasted”, “Earthy”, “Rancid”, “Cardboard”). All attributes except “pine” (p=0.837) and “toasted” (p=0.173) were associated to significant p-values at 0.05 according to Cochran`s Q test, for this reason these two attributes were not included in the symmetric plot. “Black pepper” (p=0.042) and “curry” (p=0.065), although not significant at 5% level, were included in the map due to their close values to the established threshold. Thus, according to Cochran’s Q test, significant differences (p<0,05) in the frequency with which 11 out of 13 attributes of the CATA choice question were used to characterize the 18 cinnamon samples, indicated that also the participants found differences in the sensory attributes of the evaluated samples. Sample 18 was the same as the reference used to train the “Ceylon aroma”, while sample 17 was the same as the reference used to train the “Cassia aroma”, thus, is not surprisingly that most of the participants checked “Ceylon” and “Cassia” in these samples. For a better visualization of the differences and similarities of the cinnamon commercial samples together with the main sensory properties of each sample, a correspondence analysis was applied to the contingency table built from CATA. This sensory map (Fig. 5C) presents the sample loadings and sensory attributes in the first two dimensions of the symmetric plot of the correspondence analysis. Both dimensions together (F1 + F2) accounted for 65 % of the total data variance, while F1 explained 51 %, F2 explained the rest of 14 %. All cinnamon samples and sensory attributes were scattered around the bisectors of the 1^st and 2^nd quadrants in the plot. Samples loads are represented in triangle and blue color, while the sensory attributes with red dots; the distance among the loading points measures the similarity (near) or differences (far from each other). Thus, it can be shown that there is a cluster of samples formed by M2, M4, M14, M9 and M14 located in the intermediate values of F2 and negative value of F1, characterized by “earthy”, “woody” “rancid”, and “cardboard” sensory attributes. Another small cluster is formed by M7, M10, M13, M15, and M18 located faraway of the 1^st cluster, at the positive values of F2 and close to sensory attributes such as “Ceylon” and “citric”. Last cluster is formed by the rest of nine samples, and were surrounded by parameters such as “woody, “cassia”, “black pepper”, “curry”, “minty”, and “sweet”. Of these, the reference samples which corresponds to M17 was found close to “Cassia” and “sweet” attributes. As observed, “Ceylon” and “Cassia” sensory terms were located on the positive and negative sides of F1 and F2, as well as reference sample for “Ceylon” M18 and “Cassia” M17, which shows the negative correlation between them in terms of sensory attributes. While C. zeylanicum was perceived as more “citric”, C. cassia was perceived as “sweet aroma”, “woody”, and “earthy”. Attributes such as “sweet”, “pine”, “curry” were more frequently checked in M17, while “minty” in M18 although the differences were not significantly different.

Finally, a PCA was applied to all the parameters to get a biplot of the 18 cinnamon sample purchased in Spanish market (Fig. 5D), which enables visualization of the differences and similarities of all samples, as well as the main properties of each sample considering all studied parameters. After preliminary results of PCA with each compound and totals, the organic acids and sugars were used as total, while volatile compounds as the chemical families and total volatiles, for a better visualization on the map. The first and second principal components together accounted for 46 % of the variance of the experimental data, with PC-1explaining 30 % and PC-2 16 %, respectively. Reading this biplot, it can be said that in the Spanish market can be found different groups of samples, one of them (M2, M3, M4, M9, M14 and M16), located in the intermediate values of PC-2 and negative value of PC-1, was definitely characterized by a low-quality cinnamon with dark color, low sensory intensity, low antioxidant activity, low TPC, low organic acids, low sugars, low volatile compounds of great interest, and high content of ketones, hexanal, rancid, cardboard, toasted, and woody notes. There can be seen another group of samples located on the positive side of both components which include both reference samples and other such as M5, M7, M8, M10, M15, M17, M18, characterized by a higher intensity of all these positive parameters. For instance, M17 the reference for C. cassia together with M8 and M5 were characterized by a reddish color, “cassia” and “sweet” sensory aromas, higher content of aldehydes ((E-)-Cinnamaldehyde with "cinnamon”, “sweet”, and “spicy” odor compounds), phenols (coumarin), esters (bornyl acetate with odor compounds “pine”, “cedar”, and “spicy”), sesquiterpenes (cyclosativene, cubebene, sesquithujene, caryophyllene, murolene with odor compounds such as floral, herbal, waxy, spicy, clove, and woody) and total volatiles. Sample M18, the reference for “Ceylon” or C. zeylanicum together with M7, M10 and M15 were closer to yellowish colors and to sensory attributes such as “Ceylon”, “citric”, “pine”, “minty”, “curry”, as well as higher in sugars, organic acids, antioxidant activity, total phenolic content, and monoterpenes (with different odor compound depending on the volatile compound and sample, such as mint, citrus, eucalyptus, pepper, woody, pine, lime, etc.).

3.7. Potential biomarkers to identify “Ceylon” or “Cassia” Cinnamon

Almost all ground cinnamon samples purchased in the Spanish market were labeled in ingredients as “ground cinnamon” except for M1 which included an alert that this sample “may contain mustard, gluten, sesame, and celery”, M4 included “Origin Vietnam”, M6 included “Origin Ceylon and Indonesia”, M8 included “Roasted ground cinnamon: Origin Ho Chi Minh City”, M10 included “Cassia ground cinnamon”, M15 included “Ground cinnamon from Ceylon”, and M16 included “Ground cinnamon: Origin Vietnam”. As seen, samples labeled as “Ceylon” were M15 and M6 although this last was labeled as a mix between 2 origins, while sample labeled as “Cassia” was only M10. To determine the cinnamon typology of each commercial sample belongs and to verify whether those labeled as “Ceylon” or “Cassia” indeed belong to that type, some biomarkers had to be established. For this purpose, out of all the compounds analyzed, only the volatile compounds could be used as biomarkers, because some were found exclusively in the reference sample “Ceylon” (C. zeylanicum) and not in C. cassia, and vice versa. All the other compounds although in different amounts were found in both reference samples. Out of 52 aromatic compounds identified in these samples, only 4 were specific to M18 reference for “Ceylon” including β-phellandrene, d-3-carene, cryptone, and eugenol; and 9 were exclusive to M17 reference for “Cassia” including limonene, eucalyptol, endo-fenchol, Δ-elemene, cyclosativene, 7-epi-sesquithujene, bisabolene, α-calacorene, and cadalene. Once the biomarkers were determined, and after preliminary results, the data was standardized using z-score and a PCA and an AHC was performed to detect patterns and groupings among the samples. The results are graphically represented in Fig. 6, and as can be observed, most of the variation in the data set can be explained by the first two principal components because they were enough to explain 71 % of the cumulative variance (PC-1 explained 43 % and PC-2 explained 28 %). Analyzing the present results, it can be determined that sample M15 was found to be the most similar to the “Ceylon” typology because it contained the same volatile markers as M18, but in a lower intensity, the reason for being positioned not very close to M18. The other sample labeled “Origin Ceylon and Indonesia” M6, was more closely related to “Cassia” typology although also contained compounds of “Ceylon” in a lower amount. This sample can be considered a mix of both with a higher concentration in “Cassia”. Regarding the sample labeled “Cassia ground cinnamon” M10, the results showed that this sample indeed belonged to “Cassia” type, as none of the markers found in “Ceylon” were present in M10. According to the Euclidean distances obtained in AHC processing, the distance between M10 and M17 was lower than M10 and M18, this can be easily seen in the present PCA biplot. Although belonging to typology “Cassia” this sample stands out for having higher content of Δ-elemene, sesquithujene, limonene, eucalyptol and fenchol, which is why it is not positioned just beside M17. As are samples M5 and specially M7, mainly due to the similar amount of these compounds. These samples were only labeled as “ground cinnamon” and the price was very different between them, 0,48 €/47 g and 2,85 €/39 g, respectively. Additionally, all the rest of the samples were closer to “Cassia” type than to “Ceylon”, none of them, beside the above-mentioned ones, contained the compounds found in “Ceylon” type. The rest of the samples also formed 2 different clusters, one composed by M6, M8, M11, M12, and M13 and mainly characterized by α-calacorene and cadalene, while the other cluster was formed by M1, M2, M3, M4, M14, and M16, samples (except M1) with a weaker concentration of the cinnamon specific compounds and higher concentrations of bisabolene. C. zeylanicum (M18) contained compounds from the following families: monoterpenes (alpha-pinene, eucalyptol, p-cymene); sesquiterpenes (α-cubebene, caryophyllene, linalool); aldehydes (cinnamaldehyde); esters (benzyl acetate, bornyl acetate); phenols; (eugenol, coumarin) and aromatic hydrocarbons (styrene, naphthalene), but no ketones (acetophenone) were identified. These compounds are associated with citrus and pine notes; greens, vegetables, and herbs; fruity, green and woody; balsamic, floral and pungent; and sweet, respectively. As can be seen in the Fig. 6A and B, C. cassia cinnamon had a higher content of these chemical compounds compared to C. zeylanicum cinnamon, except for monoterpenes.

Fig. 6.

Principal component analysis biplot of specific volatile biomarkers of “Ceylon” and “Cassia” typology and Agglomerative Hierarchical Clustering (AHC) of commercial cinnamon samples found in the Spanish market

4. Conclusion

The present study highlights the significant variability in moisture content, water activity, colour, organic acids, sugars, and volatile compounds in commercial ground cinnamon samples, including sensory attributes. The colour of cinnamon is not proper for species identification in ground cinnamon. On the other hand, malic and butyric acids as well as sugars, although found in all samples, were always higher in C. zeylanicum specie. C. zeylanicum was also characterized by higher concentrations of monoterpenes, while C. cassia by higher levels of aldehydes and sesquiterpenes. Cinnamaldehyde, the major aromatic compound responsible for cinnamon's distinctive odor, was most abundant in C. cassia. Coumarin, a compound with known liver toxicity, was higher in C. cassia, and only 1 out of 16 ground cinnamon samples, recorded similar values in coumarin content as C. zeylanicum. Four out of 52 volatiles compounds (β-phellandrene, d-3-carene, cryptone, and eugenol) were exclusive to C. zeylanicum and 9 (limonene, eucalyptol, endo-fenchol, Δ-elemene, cyclosativene, 7-epi-sesquithujene, bisabolene, α-calacorene, and cadalene) were exclusive to C. cassia. The sensory analysis confirmed distinct aroma profiles and intensities for C. cassia and C. zeylanicum, influenced by their unique chemical compositions. Moreover, the present study revealed that most of the ground cinnamon on the Spanish market belong to C. cassia specie. Overall, the variability observed across the samples emphasizes the influence of factors such as botanical source, processing methods, and potential adulteration on the chemical composition of cinnamon products.

CRediT authorship contribution statement

Leontina Lipan: Writing – review & editing, Writing – original draft, Software, Investigation, Formal analysis, Data curation, Conceptualization. Marina Cano-Lamadrid: Writing – review & editing, Methodology, Investigation. Hanán Issa-Issa: Methodology, Formal analysis, Data curation. Carmen Muñoz: Writing – original draft, Methodology, Formal analysis. Francisca Hernández: Visualization, Supervision, Methodology, Formal analysis. Ángel Carbonell-Barrachina: Supervision, Resources, Project administration, Funding acquisition. Esther Sendra: Visualization, Validation, Supervision, Resources, Project administration, Investigation, Conceptualization.

Ethical statement

The Ethics Review Committee of Universidad Miguel Hernández, Office of Responsible Research (Oficina de Investigación Responsable), granted ethical approval for the human participants in this study under the reference number PRL.DTA.ESN.02.20 on November 9, 2020. All the participants were given information about the cinnamon samples and a written informed consent was obtained from all the participants prior to participating in the study.

Declaration of competing interest

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

Appendix A. Supplementary data

Supplementary material

Data availability

Data will be made available on request.