Subscribe to RSS
DOI: 10.1055/a-0605-3967
Is Low-field NMR a Complementary Tool to GC-MS in Quality Control of Essential Oils? A Case Study: Patchouli Essential Oil[*]
Correspondence
Publication History
received 03 February 2018
revised 27 March 2018
accepted 06 April 2018
Publication Date:
24 April 2018 (online)
Abstract
High-field NMR is an expensive and important quality control technique. In recent years, cheaper and simpler low-field NMR has become available as a new quality control technique. In this study, 60 MHz 1H-NMR was compared with GC-MS and refractometry for the detection of adulteration of essential oils, taking patchouli essential oil as a test case. Patchouli essential oil is frequently adulterated, even today. In total, 75 genuine patchouli essential oils, 10 commercial patchouli essential oils, 10 other essential oils, 17 adulterants, and 1 patchouli essential oil, spiked at 20% with those adulterants, were measured. Visual inspection of the NMR spectra allowed for easy detection of 14 adulterants, while gurjun and copaiba balsams proved difficult and one adulterant could not be detected. NMR spectra of 10 random essential oils differed not only strongly from patchouli essential oil but also from one another, suggesting that fingerprinting by low-field NMR is not limited to patchouli essential oil. Automated chemometric evaluation of NMR spectra was possible by similarity analysis (Mahalanobis distance) based on the integration from 0.1 – 8.1 ppm in 0.01 ppm increments. Good quality patchouli essential oils were recognised as well as 15 of 17 deliberate adulterations. Visual qualitative inspection by GC-MS allowed for the detection of all volatile adulterants. Nonvolatile adulterants, and all but one volatile adulterant, could be detected by semiquantitation. Different chemometric approaches showed satisfactory results. Similarity analyses were difficult with nonvolatile adulterants. Refractive index measurements could detect only 8 of 17 adulterants. Due to advantages such as simplicity, rapidity, reproducibility, and ability to detect nonvolatile adulterants, 60 MHz 1H-NMR is complimentary to GC-MS for quality control of essential oils.
#
Key words
essential oil - quality control - 60 MHz 1H-NMR - adulteration - fingerprinting - chemometrics - patchouliAbbreviations
Introduction
Quality control (QC) of goods by buyers and sellers is a cornerstone of our society. QC is not only necessary to produce a constant quality, but also to detect adulteration of which food ingredient fraud is an example [1]. When quality is dependent on the nature and concentration of contained molecules, like in the case of foods, perfumes, pure chemicals, drugs, herbals, extracts, etc., analysis can be carried out by visual, olfactory, gustatory, physical, or chemical means. More specifically, chemical analysis is performed by spot tests, spectroscopy (UV, IR, MS, NMR, AAS), chromatography (TLC, HPLC, GC), or more powerful combinations thereof (LC-MS, GC-MS). More and more often, spectroscopic techniques like NIR, MS, and NMR are directly applied to complex mixtures without any separation to gain time and save costs. Of those techniques, NMR is, according to Verpoorte et al. [2], not only the “favored overall ʼmacroscopicʼ metabolomics and fingerprinting tool” and “unmatched for structure elucidation of metabolites”, but NMR can also provide absolute quantitative information [3]. Additionally, it is universal for organic molecules due to the ubiquitous presence of hydrogen atoms.
High-field NMR has been used to distinguish between different types of beer [4], [5], detect adulteration of orange juices [6], and used for QC of plant products [7], beer [8], and olive oil [9]. Unfortunately, high-field (≥ 300 MHz) NMR spectrometers are expensive to purchase and maintain, and require expertise to initialise and operate (superconducting magnet and liquid helium). In recent years, commercial low-field (30 – 80 MHz) permanent magnets have entered the market and these are 5 – 20× cheaper to purchase, require much less space, are very cheap to run, and are easy to operate [10], [11], [12]. On the downside, they are less sensitive and, more importantly, their spectral resolution is 5 – 20× lower. In spite of this, several interesting QC or fraud detection applications by low-field NMR exist already, e.g., methyl tert-butyl ether in gasoline [13], QC of polyether polyols [14], raw rubber [15], and (bio)diesel [16], [17], [18], adulteration of olive oil [19] and dietary supplements [20], authentication of beef [21] and edible oils [22], process control [23], distinguishing between arabica and robusta coffees [24], and possibly forensics [25].
Essential oils (EOs) are steam distillates of plants and have considerable value. Therefore, they have been the target of wilful adulteration practices since ancient times [26]. In spite of the development of sophisticated control techniques, adulteration with essential or vegetable oils, other extracts, or pure chemicals still occurs frequently today [27], [28], [29]. In addition to olfactometry and physical methods (specific gravity, refractive index, optical rotation), qualitative or quantitative GC-MS is a powerful technique to detect EO fraud. Although 13C-NMR has been used for EO analysis [30], it never gained popularity. 1H-NMR is even less popular, most likely because many EOs are complex mixtures with mainly overlapping aliphatic signals. Thus, it is not surprising low-field NMR has never been used for the QC of essential oils. However, it does have certain advantages over GC-MS. It is cheaper, faster, requires less sample preparation, is universal (including detection of nonvolatile adulterants, which are invisible with GC-MS), and has the ability to immediately see specific adulterants, such as aromatic compounds in an EO, which should not contain aromatics.
Using PEO, which even today is frequently adulterated, as an example [31], we investigated whether low-field NMR has any merit as a complimentary QC technique for EOs. To this end, a large number of genuine and adulterated PEOs as well as PEOs spiked at 20% by known adulterants were analysed by 60 MHz 1H-NMR, quantitative GC-MS, and refractive index measurements, and, subsequently, the data were processed manually and by chemometrics and the results were evaluated.
#
Results and Discussion
PEO, which is prepared by steam distillation of the leaves of Pogostemon cablin (Blanco) Benth. (Lamiaceae), was selected as a test case for this study for four reasons: (1) we possessed ~ 100 PEOs, partly genuine and partly commercial, (2) PEO is often adulterated, and the adulterants are known and available, (3) PEO is commercially important, and (4) all PEO literature was recently reviewed [31]. To arrive at a good QC comparison of NMR and GC-MS, ~ 75 PEOs, 15 commercial PEOs, 10 other EOs, 17 adulterants, and 1 high-quality PEO deliberately spiked at the 20% level with those 17 adulterants were measured. Twenty percent spiking was chosen as a compromise between monetary gain and avoiding detection. Additionally, one common adulterant (gurjun balsam) was spiked at 5, 10, 20, and 30%. Finally, the refractive index of 35 PEOs was measured to gain insight into the usefulness of a classic, fast, and cheap QC method. For an oil to be treated as non-adulterated and high quality in this study, the data of each method had to be satisfactory.
The low-field NMR part of this study was carried out on a 60 MHz spectrometer. To keep things as simple as possible, initially samples were measured pure, i.e., without solvent and without a qualitative or quantitative internal standard. A spectrum could be obtained after 1 scan (Fig. 1aS, Supporting Information) and a satisfactory spectrum was obtained after 32 scans (Fig. 1bS, Supporting Information; S/N > 3000 for the largest peak). Including the sample change, less than 3 min was needed per sample. Unfortunately, the absolute chemical shift varied up to 0.2 ppm. Although spectra can be aligned afterwards, any misalignment would be detrimental for fingerprinting, therefore, TMS was added and all spectra were aligned at the 0.00 ppm TMS signal. Sample preparation consisted of adding ~ 600 µL PEO to ~ 25 µL TMS in an NMR tube. Spectra of the same sample were highly reproducible (Fig. 1b – cS, Supporting Information), with shift differences ≤ 0.01 ppm.
A problem that was encountered with some commercial PEOs was significant band broadening, naturally diminishing the information content of the fingerprint (Fig. 2aS, Supporting Information). The band broadening was independent of field strength (Fig. 2b – cS, Supporting Information). Broadening was only seen with darker coloured and more viscous samples. High-quality PEOs are light yellow; harshly distilled samples vary from dark yellow via orange-brown to even blackish. Band broadening in NMR spectra can result from the presence of metals (PEOs are known to contain iron ions) or a high viscosity (resulting in short spin-spin relaxation times, T2). To investigate this, a dark-coloured sample was extracted with EDTA and remeasured. No effect was seen, suggesting viscosity to be the cause, and indeed a sharpening of signals could be observed after increasing the temperature from 25 to 75 °C (Fig. 2dS, Supporting Information) or by dilution in CDCl3 (Fig. 2e – hS, Supporting Information). This has also been observed for crude oil samples [32]. Thus, effectively, some PEOs are too viscous to be measured neat. Increasing the temperature is not possible with the 60 MHz instrument, but dilution with chloroform is. Unfortunately, this would considerably complicate the sample preparation step as peak height and chemical shift (solvent effect) will be affected by the PEO-solvent ratio. This means that more standardisation (accurate weighing) will be required for obtaining reproducible fingerprints. In this study, the harshly distilled, i.e., lower quality commercial PEOs, constituted only a minority of the samples and were not taken into account. Still, in spite of the band broadening, it remained easy to see that some viscous, commercial PEOs were severely adulterated (Fig. 3S, Supporting Information).
For successful QC of PEOs by fingerprinting, differences between NMR spectra due to natural variation of genuine PEOs need to be smaller than changes due to adulteration. According to a recent review [31], PEO consists for ~ 50% of sesquiterpene hydrocarbons [SQHCs; main ones: α-bulnesene (1), α-guaiene (2), seychellene (3), and α-patchoulene (4)] and ~ 45% oxygenated sesquiterpenes [main one: ~ 40% patchoulol (5)] ([Fig. 1]). The ratio between all SQHCs was reported to be fairly constant, and the two main causes of variation are (a) the ratio between SQHCs and oxygenated compounds like patchoulol (5), and (b) the concentration of pogostone (6), which varies between 0 – 28%. The latter will impact the fingerprint more because the NMR of pogostone is completely different from that of any sesquiterpenes. Subsequent measurement of ~ 75 genuine PEOs showed that indeed the NMR fingerprints were much alike except when the pogostone concentration was high. The overlaid 60 MHz spectra of 10 genuine PEOs are shown in [Fig. 2], with the spectrum of the high-quality PEO used for deliberate adulteration (vide infra) being at the bottom (in purple).
Apart from methyl signals of the main constituent patchoulol (600 MHz NMR) (Fig. 4aS, Supporting Information) between 0.7 – 1.05 ppm, methyl signals on double bonds of several SQHCs at 1.7 ppm, various aliphatic protons between 0.7 – 2.8 ppm, and olefinic protons at 4.6 – 4.8 ppm, the spectrum was essentially empty. This correctly means the PEO is essentially devoid of constituents with acetyls, primary hydroxyls, and conjugated olefinic, aromatic, or aldehydic protons. The patchoulol concentration does influence the fingerprint but not to such an extent that the fingerprint becomes unrecognisable as that of PEO. The standard oil (bottom spectrum, in purple) has a patchoulol content of 40%. The extremes in [Fig. 2] are the top spectrum (52% patchoulol, in red) and spectrum 5 from the bottom (15% patchoulol, in green). In the latter spectrum, peaks at 4.6, 1.7, and 0.95 ppm are relatively large, and peaks at 1.05 and 0.8 ppm are low due to the high SQHC content and low patchoulol (5) content, respectively ([Fig. 2] and Fig. 7S, Supporting Information). In the top spectrum, it is just the other way around, so the precise fingerprint does give additional information about the approximate patchoulol content. PEOs with a high pogostone (6) content, like stem oils ([Fig. 2], spectrum 6 (green) from bottom), deviate significantly due to extra signals at ~ 5.9, ~ 3.1, and ~ 2.3 ppm. The 600 MHz spectrum of 6 is depicted in Fig. 5S, Supporting Information. The stem PEO also has a high norpatchoulenol (8) content leading to higher signals between 5.25 – 5.6 ppm. It should be noted that 6 and 8 occur in low concentrations in most PEOs and are difficult to observe at 60 MHz (Fig. 1c – dS, Supporting Information). At 600 MHz, they are observable (Fig. 1e – fS, Supporting Information). Overall, the results appeared promising as the presence of any adulterant or impurity showing resonances above 5 ppm, between 4.3 – 2.8 ppm, or below 0.7 ppm, even at 5%, will be immediately spotted. In [Fig. 3], the NMR spectra of nine “easy” adulterants spiked at the 20% level in PEO are shown overlaid on the original PEO (bottom spectrum, in purple). Indeed, in one glance it can be seen that all should be treated as suspicious due to the presence of extra signals. Interestingly, benzyl alcohol was seen to act as an NMR shift reagent, affecting the chemical shifts of various protons of patchoulol (5) through interaction of the hydroxyls (Fig. 4bS, Supporting Information).
Next, it was established if added compounds or extracts lacking resonances in the “empty” regions of a PEO NMR could also be detected at the 20% level. Such adulterants include ricinus oil (castor oil), pepper oil, vetiver oil, amyris oil, paraffin (mineral oil), gurjun balsam, copaiba balsam, and Clearwood ([Fig. 4]). Clearwood is an alternative for PEO produced by genetic engineering and its main constituents are identical to those of PEO. The NMR spectra of the first five adulterants deviated significantly, e.g., vetiver showed larger peaks at ~ 2.85 and ~ 5.3 ppm and paraffin exhibited an intense peak at ~ 1.25 ppm due to its many methylene groups. More difficult to detect were gurjun and copaiba balsams. The PEO with 20% gurjun balsam showed additional signals between 5.2 – 5.4 ppm and a peak at 1.25 ppm, while copaiba balsam adulteration was apparent by differences at 5.2, 2.75, and 0.6 ppm. The latter signal is caused by protons attached to the tricyclic ring of aromadendrane SQHCs, which do not occur in PEO [31]. Adulteration with 20% Clearwood was impossible to detect at 60 MHz. However, at 600 MHz, detection was possible by the presence of patchoulol ethyl ether in Clearwood (Fig. 6a – bS, Supporting Information).
As the results showed that 60 MHz NMR appears to have merit for the rapid detection of adulteration of PEOs, 10 random EOs were measured to get an idea of the variation of EO fingerprints ([Fig. 5]). Additionally, if each profile would be unique, it would prove that low-field NMR can detect any inadvertent mix-up of EOs, e.g., due to wrong labelling. All oils differed greatly from one another and, consequently, also from PEO (bottom spectrum). Although naturally all EOs share peaks at certain ppm values with other EOs, each fingerprint has specific peaks and is indeed unique. However, to be truly useful for QC purposes, the fingerprint of each EO needs to remain fairly constant in spite of the natural variation within the plant material yielding the EO. This was the case for PEO ([Fig. 2]), but it is not known for the 10 investigated EOs. As we possessed a significant collection of ginger oils [33], we preliminary investigated this aspect also for ginger oil ([Fig. 6]). Five ginger EOs from South Africa, Japan, Malaysia, and two different Indonesian islands showed more or less the same reproducibility and variation (especially pronounced between 3 – 4 ppm) as PEO. Similarly, it was possible to distinguish between the various types of buchu oil [34], [35] by 60 MHz NMR (Fig. 8S, Supporting Information). This strongly suggests that QC of EOs by low-field NMR is more widely applicable than just PEO.
Visual analysis of many NMR spectra can be tedious, and thus different chemometric approaches were evaluated for a semiautomated analysis of large numbers of samples. Best results were obtained with a similarity analysis approach based on Mahalanobis distance [36] (SI). Five groups were created: (1) random 50% of genuine PEOs for creating a common pattern as a reference spectrum, (2) other 50% of genuine PEOs used to validate the model created with group 1, (3) commercial PEOs, (4) standard PEO deliberately adulterated at the 20% level, and (5) other EOs. Initially, integration of the whole NMR spectrum was used. Variables investigated were the ppm range and the integration band width. Best results were obtained with a range from 0.1 – 8.1 ppm and integration bands of 0.01 ppm width ([Fig. 7]). The results show that this model is, on the one hand, well capable of recognising a genuine PEO as a genuine PEO (distance > − 1000 for both groups 1 and 2 and good quality commercial oils), on the other hand, it is not well capable of differentiating between genuine PEOs, unlike GC-MS. A different ratio between PEO constituents automatically changes the GC fingerprint due to the chromatographic separation in time. With NMR, there is no such separation, and as NMR spectra of individual SQHCs, or even patchoulol for that matter, do not differ greatly at 60 MHz, the NMR fingerprint does not change too much upon small ratio differences between sesquiterpenes. The exception is the pogostone-rich stem PEO. Low-field NMR could detect all adulterated PEOs, with the exception of vetiver and Clearwood adulteration. Gurjun balsam could already be detected at the 10% level. The fact that the distance values (− 2.6E3, − 7.6E3, − 13E3) corresponding to 10, 20, and 30% gurjun balsam are linear, provides confidence that the model detects real differences. The model clearly uses other criterions than a human observer in view of the differences for gurjun (visually hard, no problem for model, i.e., minor but many linear correlations) and vetiver (other way around). The four adulterated commercial oils were properly recognised. As expected, the 10 other EOs deviated enormously ([Fig. 7]). By changing the integration parameters (more emphasis on 2.95 – 2.75 ppm; SI), the model could be made to distinguish vetiver-adulterated PEO from genuine PEO, but results became worse for other EOs, like amyris and cedarwood oil. Finally, a model focusing mostly on patchoulol-related peaks (SI) was explored, but again overall results were not superior to the generic model (Fig. 9S, Supporting Information).
GC-MS is the golden standard of EO analysis, which is caused by its unprecedented power to give highly detailed information on individual constituents. Qualitative visual inspection of GC profiles of all genuine PEOs gave the fingerprint (Fig. 10S, Supporting Information) of this moderately complex EO [31], but on the downside, this was time-consuming. Furthermore, 100% internal normalisation provided ratios among PEO constituents, including important odour carriers like patchoulol (5), norpatchoulenol (8), and tetracyclopatchoulol, something NMR is not capable of. Measurement of the 17 deliberately adulterated PEOs allowed facile detection of 14 adulterations due to the presence of extra peaks. Adulteration with Clearwood did not change the fingerprint apart from a small additional peak of patchoulol ethyl ether (7). If one knows where to look for, Clearwood adulteration is thus detectable at 20%. Adulteration with nonvolatile 20% paraffin and ricinus oil did not change the fingerprint and was qualitatively impossible to detect. When quantitative GC (internal standard method) was performed, these two adulterations were well detectable ([Table 1]). However, both in terms of sample preparation and calculations, the latter approach requires significantly more time, although this can be partially automated. Even for genuine oils, the percentage of PEO in [Table 1] can reach values higher or lower than 100% depending on the ratio between individual constituents and their RRF (SI) in the mass spectrometer. The factor was based on the standard PEO with a patchoulol content of ~ 40%.
|
Sample info |
% PEO |
Sample info |
% PEO |
|---|---|---|---|
|
Standard PEO |
100 |
commercial PEO |
99 |
|
Genuine PEO 16 |
108 |
commercial PEO |
97 |
|
Genuine PEO 18 |
97 |
adulterated commercial PEO |
74 |
|
Genuine PEO 37 |
93 |
adulterated commercial PEO |
30 |
|
Genuine PEO 45 |
108 |
adulterated commercial PEO |
58 |
|
Genuine PEO 59 |
102 |
adulterated commercial PEO |
32 |
|
Genuine PEO 63 |
100 |
adulterated commercial PEO |
29 |
|
Genuine PEO 70 |
101 |
PEO + 20% Clearwood™ |
103 |
|
Commercial PEO |
94 |
PEO + 20% copaiba balsam |
83 |
|
Commercial PEO |
93 |
PEO + 20% gurjun balsam |
79 |
|
Commercial PEO |
96 |
PEO + 20% paraffin |
79 |
|
Commercial PEO |
97 |
PEO + 20% vetiver oil |
82 |
|
Commercial PEO |
98 |
PEO + 20% benzyl benzoate |
80 |
|
Commercial PEO |
99 |
For GC-MS, various chemometric tools, as for NMR, were evaluated. Using multivariate analysis (PCA), the focus was on genuine PEO constituents, as adulterants are not always known. Components contributing most to the variation were given a weighing factor: patchoulol = 15, α-bulnesene = 20, α-guaiene = 30, seychellene = 40. In the biplot PCA, all adulterated PEOs, except for Clearwood, adulterated commercial oils, and other EOs were separated from the genuine PEOs (Fig. 11S, Supporting Information). The similarity analysis (cosine; SI) showed problems with detecting the nonvolatile adulterants, as well as Clearwood and vetiver (Fig. 12S, Supporting Information). Apparently, the cosine model has more problems dealing with many small extra peaks than one large peak; the Mahalanobis distance model performs better in these circumstances. All commercial adulterated samples and other EOs were well separated from PEO.
Refractometry
In addition to NMR and GC-MS, refractometry was used as a traditional low-tech and cheap QC technique for a number of samples. Genuine PEO is characterised by a refractive index of 1.503 – 1.513 [31]. All PEOs considered genuine according to NMR and GC-MS fell within that range ([Table 2]). However, one of the adulterated commercial PEOs and half of the deliberately adulterated samples escaped detection by this method. Surprisingly, adulteration with Clearwood was detected by refractometry. Apart from this, its complementary value to low-field NMR is limited as it failed to detect the adulterants copaiba balsam, gurjun balsam, and vetiver oil, which were most problematic for NMR. Adulteration with gurjun balsam became detectable only at 30%. Refractometry is highly complementary to qualitative GC, as adulteration with the nonvolatile paraffin and ricinus oil was detected. Thus, this simple and fast QC technique continues to have merit for detecting adulteration of PEO and EOs in general.
|
Sample info |
Refractive index |
Within range |
Sample info |
Refractive index |
Within range |
|---|---|---|---|---|---|
|
PEO used for adulteration |
1.5070 |
yes |
PEO + 20% Clearwood |
1.5009 |
no |
|
Genuine PEO 4 |
1.5070 |
yes |
PEO + 20% copaiba balsam |
1.5050 |
yes |
|
Genuine PEO 6 |
1.5120 |
yes |
PEO + 20% diethyl phthalate |
1.5035 |
yes |
|
Genuine PEO 9 |
1.5061 |
yes |
PEO + 20% dioctyl phthalate |
1.5000 |
no |
|
Genuine PEO 17 |
1.5099 |
yes |
PEO + 5% gurjun balsam |
1.5058 |
yes |
|
Genuine PEO 78 |
1.5076 |
yes |
PEO + 10% gurjun balsam |
1.5035 |
yes |
|
Stem PEO (pogostone) |
1.5070 |
yes |
PEO + 20% gurjun balsam |
1.5025 |
yes |
|
Genuine commercial PEO |
1.5072 |
yes |
PEO + 30% gurjun balsam |
1.5020 |
no |
|
Genuine commercial PEO |
1.5072 |
yes |
PEO + 20% Hercolyn |
1.5061 |
yes |
|
Genuine commercial PEO |
1.5072 |
yes |
PEO + 20% isobornyl acetate |
1.4956 |
no |
|
Adulterated commercial PEO |
1.5015 |
no |
PEO + 20% methyl benzoate |
1.5056 |
yes |
|
Adulterated commercial PEO |
1.5091 |
yes |
PEO + 20% paraffin |
1.4982 |
no |
|
Adulterated commercial PEO |
1.5205 |
no |
PEO + 20% pepper oil |
1.5000 |
no |
|
PEO + 20% amyris oil |
1.5042 |
yes |
PEO + 20% ricinus oil |
1.4980 |
no |
|
PEO + 20% benzyl alcohol |
1.5090 |
yes |
PEO + 20% vetiver oil |
1.5084 |
yes |
|
PEO + 20% benzyl benzoate |
1.5152 |
no |
PEO + 20% myrtenol |
1.5025 |
no |
|
PEO + 20% cedar wood oil |
1.5042 |
yes |
Three QC techniques for EOs were compared, namely, the well-established techniques of GC-MS and refractometry, and, additionally, low-field NMR as a novel outsider technique. As expected, GC-MS performed best. It provides enormously detailed information about each EO in terms of identification of dozens of individual constituents and their ratios in a large concentration range. In addition, it could detect any adulteration of PEO at the 20% level with other oils, extracts, or pure compounds, including Clearwood, which is almost identical to PEO. Most adulteration could be discerned by fingerprinting. An inherent shortcoming of GC is that it cannot detect nonvolatiles, e.g., vegetable and mineral oils. However, by performing quantitative GC such adulterations can be observed. On the downside, certain expertise is needed for (quantitative) GC-MS and the equipment is relatively expensive to run (He, pump, photomultiplier, source cleaning, columns, electricity) and acoustically noisy unless the vacuum pump is encapsulated. Refractometry was found to be a simple, useful technique, especially as a QC tool complementary to GC-MS.
A priori, low-field 1H-NMR was not expected to be of great value for QC of EOs because it can provide almost no information on individual compounds in a complex mixture like an EO, as there is zero chromatographic separation and many signals will severely overlap. Still, NMR fingerprints of EOs were found to be rich in detail even if their shape cannot be explained. Each EO has its own reproducible fingerprint, and within a single EO, it was fairly constant. A significant advantage of NMR over GC-MS was the easier and faster detection of many adulterants, including nonvolatile ones. This process could be successfully carried out by chemometric tools as well. A further advantage is that the deviations of the fingerprint can provide direct information about the type of adulterant. Additionally, low-field NMR was faster, cheaper, and less noisy than GC-MS. Possibly, low-field NMR in combination with automated data processing could serve as an “analytical decision maker” to differentiate samples in “worthwhile” and “not worthwhile” to be further analysed by GC-MS [37]. An advantage over high-field NMR is that it is much cheaper, requires less expertise, and does not need (deuterated) solvents, liquid helium, liquid nitrogen and pressurised air. In remote areas, it might be advantageous that only 100 W of electricity is needed, although temperatures have to be constant.
Like GC-MS, NMR can be made quantitative, but for the QC of EOs, this has little added value as, unlike GC-MS, it is not possible to distinguish which part of the overlapping peaks belongs to the EO and which part belongs to the adulterant. It is better to focus on deviations from fingerprints. An intrinsic shortcoming of NMR is that it cannot be used for too viscous liquids and this will somewhat restrict its use in the EO field unless samples are diluted 1 : 1 with CDCl3. Recently, 80 MHz permanent magnets have become available and application of such magnets will, at a price, certainly increase the richness of the fingerprints. Another way to increase the information content of fingerprints is to apply 2D NMR, e.g., record a COSY spectrum (Fig. 13S, Supporting Information). Gouilleux et al. [22], [38] have shown that spectra similar to a COSY spectrum can be acquired on a low-field NMR in 2 min, i.e., not more than the time used in this study to acquire a 1D spectrum. However, for a quick screening of many such spectra, dedicated software should be available as for the human eye COSY spectra are more difficult to interpret sensibly in a short amount of time. Additionally, the number of cross-peaks will be strongly influenced by how “deep” one looks and it might not be trivial to control this. Characteristics of all three techniques have been summarised ([Table 3]).
|
Attribute |
60 MHz 1H-NMR |
GC-MS |
Refractometry |
|---|---|---|---|
|
*Currently not available but expected soon; when unavailable, score is −. |
|||
|
Purchasing costs |
+ |
± |
++ |
|
Running costs |
++ |
± |
+++ |
|
Space |
± |
– |
++ |
|
Acoustic noise |
++ |
± |
+++ |
|
Training level |
+ |
± |
+++ |
|
Sample preparation |
++ |
+ |
+++ |
|
Adulteration detection |
++ |
+++ |
+ |
|
Detect any compound |
++ |
– |
+ |
|
Overall sensitivity |
+ |
+++ |
+ |
|
Detect trace compounds |
– |
+++ |
– |
|
Viscous samples |
± |
+++ |
++ |
|
Analysis time |
++ |
+ |
+++ |
|
Quantitative |
++ |
+++ |
– |
|
Automation |
* |
+++ |
+++ |
|
Identification |
± |
+++ |
– |
To come back to the question in the title, yes, overall low-field NMR was found to complement GC-MS surprisingly well as a QC technique for EOs. It will neither displace GC-MS nor high-field NMR, but for certain analytical problems, it definitely has merit at a reasonable price. Time will tell whether low-field NMR will find a niche in fragrance industry QC. With more sample preparation, it also has scope beyond EOs, e.g., for QC of herbals and plant-derived drugs.
#
#
Materials and Methods
Oils and adulterants
Most PEOs and two ginger oils from Kalimantan and Flores were a gift from the late Hans Siwon (Surabaya, Indonesia) and were hydro or steam distilled by him in Indonesia. The standard PEO (steam distilled in Madagascar), patchoulol, and pogostone were gifts from Daniel Joulain (SCBZ Conseil, Les Micocouliers-F3, 99 avenue Sidi Brahim, 06130 Grasse, France). Commercial PEOs were bought in shops in The Netherlands, Austria, and England or online via webshops in The Netherlands or via regular suppliers of chemicals: Physalis Patchouli 02/2021 L.1022910; Miaroma patchouli 2557; Jacob Hooy en Co. patchouli; Jacob Hooy Patchouli 8 712053 748557; Absolute aromas patchouli 06177827 06/21; Chi Patchouli 3802 10-2020; Roth Patchouli-Öl 6511 52-7199; Sigma-Aldrich Patchouli oil S70993; Anthemis Patchouli Maleisie E9630 1-4-2017; Anthemis Patchouli Indonesie E1630 1-4-2017; Naproz patchouli; De Tuinen Patchouli 5077 1912; F. E. S. Patchouli; Gezond & Wel Patchouli olie 809-625; Ladrome Laboratoire Patchouli ET190090 v3 NUT/PL 47/319; Primavera patchouli bio 4086900101197. Amyris oil was from Naarden International, benzyl alcohol from VWR Int., benzyl benzoate from Merck, cedarwood oil from FES, Clearwood from Firmenich, copaiba balsam, gurjun balsam, ricinus oil, and pepper oil from Anthemis, myrtenol[**], methyl benzoate, diethyl phthalate, and dioctyl phthalate from Sigma-Aldrich, Hercolyn and isobornyl acetate from Robertet, paraffin from Apotheek de Linge, Buchu oils (Agathosma crenulata, nrs 11 & 14 and A. betulina 4, B, N & O) and the South African ginger oil from Grassroots Natural Products, Ciskei, South Africa, Jatamansi oil, Mentha arvensis oil (Corn mint), Sugandha Kokila oil (Cinnamomum glaucescens) [39] and Zanthoxylum oil from HPPCL, Kathmandu, Nepal, and sandalwood oil (C619 409400) from Carl Roth. The Malaysian ginger was purchased locally and hydrodistilled in The Netherlands. The vetiver oil (Haiti) and one ginger oil (Japan) were of unknown suppliers.
#
NMR measurements and spectrum processing
60 MHz 1H-NMR spectra were recorded on a Magritek Spinsolve 60 carbon. In a fume hood, ~ 25 µL of TMS were transferred to a 5-mm NMR tube after which ~ 600 µL of EO were added. After mixing, the tube was inserted in the NMR and 32 scans were recorded. Acquisition 3.2 s/scan, repetition time 4 s, pulse angle 30°. Every day before actual measurements the NMR was first “power shimmed” (~ 5 min). Next the spectrum.1d spectrum file was imported in MNOVA version 10.0 (MestReNova), the spectrum was reversed horizontally, zoomed in till the desired ppm values, automatically phase and baseline corrected, and TMS was set at 0.00 ppm. Afterwards, the spectrum was saved as a *.jcamp file and imported into ChemPattern.
400 MHz NMR spectra were recorded on a Bruker Avance III.
600 MHz NMR spectra were recorded on a Bruker Avance III. Data were processed and analysed with Topspin 3.5. EOs were prepared as 10% (w/v) solution of essential oil in CDCl3; pure compounds as a 1% solution in CDCl3.
#
GC-MS instrumentation and measurements
Samples were analysed on a gas chromatograph from Agilent technologies 7890A equipped with a mass selective detector 5975C V MSD. Samples were injected with a 7683B Series injector and a 7683 Series autosampler. The column was an HP-5MS 30 m × 0.25 mm × 0.25 µm. Autosampler settings: wash solvent acetone; two washes with A, B, and sample; 1 µL injection; split ratio 1 : 100; carrier gas He 10.5 psi at 0 min (constant flow); injector temperature 250 °C; oven 100 °C (0 min) to 175 °C at 3 °C/min (25 min hold) to 295 °C (0 min hold) at 6 °C/min, run time 45 min; solvent delay 3 min; scan range m/z 40 – 300.
For analysis of the TIC profile of GC-MS runs in ChemPattern, the following parameters were applied to all samples: slope 100; minimal peak width 0.1; minimal height 0; minimal height (%) 0.1; minimal area 0; minimal area (%) 0.1; integration start 3 min; integration stop 40 min.
Sample preparation: ~ 10 mg of essential oil sample, accurately weighed in a 1.8-mL autosampler vial, were dissolved in 1.00 mL of an internal standard solution of 0.250% (w/v) E,E-farnesol (627 mg) in 250.0 mL of methyl tert-butyl ether (MTBE).
#
Refractometry
The refractive index of the EOs was measured on an ATAGO Abbe 1T Refractometer equipped with an ATAGO digital thermometer. The temperature during measurements was 23 °C. The wavelength for analysis was 589 nm.
#
Metal extraction
To investigate the role of metal ions in patchouli oil and their influence on the NMR spectral quality, an extraction with EDTA (ethylenediaminetetraacetic acid) was conducted. Firstly, ~ 1 g of standard PEO was dissolved in 10 mL of methyl tert-butyl ether. This solution was extracted three times (3 × 5 mL) with a solution of 0.1 M EDTA + 0.05 M TRIS buffer, pH = 8, in a separatory funnel. The organic layer was dried over anhydrous Na2SO4. Finally, the solution was filtered through a paper filter and the MTBE was evaporated with a rotary evaporator at 40 °C and 310 mbar. Final solvent residues were removed with N2. The oil was then measured by means of NMR.
#
Data analysis with ChemPattern
For data analysis, the software ChemPattern edition 2.0 was used (Chemmind Technologies).
Integration settings in ChemPattern for Similarity Analysis by NMR, 1st model (generic, values in ppm): δ 8.10 – 0.10 in increments of 0.01 ppm.
Integration settings in ChemPattern for Similarity Analysis by NMR, 2nd model (more emphasis on adulteration with vetiver oil in addition to other adulterants, values in ppm): δ 8.10 – 7.90, 7.60 – 7.10, 5.50 – 4.90, 4.50 – 4.10, 3.825 – 3.725, 3.60 – 3.50, 2.95 – 2.75, 1.77 – 1.52, 1.31 – 1.22, 1.10 – 0.90, 0.84 – 0.80, 0.63 – 0.00 in increments of 0.01 ppm.
Integration settings in ChemPattern for Similarity Analysis by NMR, 3rd model (more emphasis on patchoulol in addition to specific adulteration with vetiver oil, values in ppm): δ 3.00 – 2.80, 1.47 – 1.35, 1.24 – 1.15, 1.12 – 1.02, 0.87 – 0.68 in increments of 0.01 ppm.
#
#
#
Conflict of Interest
The second (Y. W.) and third author (R. T.) are employed by the company (Chemmind Technologies) manufacturing the used chemometric software (ChemPattern).
Acknowledgements
We wish to thank Daniel Joulain for providing us with a generous quantity of freshly distilled high-quality patchouli oil, pure patchoulol and pogostone, and several adulterants, the late Hans Siwon for his kind gift of approximately 100 PEOs as well as two Indonesian ginger oils, Dr. A. Sheak for his gift of several Nepalese essential oils, N. F. Collins and E. H. Graven for their gift of buchu oils and one ginger oil, and Marian van Gessel (Apotheek de Linge) for samples of paraffin. Further, we would like to thank Frank Claassen and Elbert van der Klift for technical assistance, Barend van Lagen for recording the 400 MHz spectra and help with NMR in general, and Pieter de Waard for recording the 600 MHz spectra. The first author (A. K.) thanks the European Union for an Erasmus student exchange scholarship. Finally, the corresponding author (T. v. B.) wishes to thank Rob Verpoorte for being a great M.Sc. and Ph.D. supervisor and teaching him many tricks of the trade, like NMR of natural products, in the period 1978 – 1984.
* Dedicated to Professor Dr. Robert Verpoorte in recognition of his outstanding contribution to natural products research.
** Myrtenol was used instead of (Z)-8-camphene methanol (syn. (Z)-patchouli ethanol) in this study. The latter compound has been used as an adulterant for PEO but was unavailable. Myrtenol has approximately the same NMR spectrum but has never actually been used as an adulterant for PEO.
Supporting Information
- Supporting Information
Fig. 1a – fS: 60 MHz and 600 MHz 1H-NMR spectra of standard PEO used for deliberate adulteration; Fig. 2a – hS: effect of viscosity of EOs on NMR spectra; Fig. 3S: 60 MHz 1H-NMR spectra of a severely adulterated viscous PEO; Fig. 4a – bS: 600 MHz 1H NMR of patchoulol and 60 MHz 1H-NMR spectrum showing shift effects of benzyl alcohol; Fig. 5S 600 MHz 1H-NMR spectrum of pogostone; Fig. 6a – bS: 600 MHz 1H-NMR spectra from 3.2 – 3.95 ppm of standard PEO and Clearwood; Fig 7a – iS: individual 60 MHz 1H-NMR spectra of genuine PEOs depicted in [Fig. 2]; Fig. 8S: 60 MHz 1H-NMR spectra of six buchu oils; Fig. 9S: similarity analysis based on NMR data focusing on patchoulol and vetiver (model 3); Fig. 10S: GC-MS fingerprint of PEO; Fig. 11S: biplot PCA based on GC-MS data; Fig. 12S: similarity analysis based on GC-MS data; Fig. 13S: 60 MHz COSY spectrum of standard PEO.
-
ORCID
- Teris A. van Beek ORCID-🆔 https://orcid.org/0000-0002-9843-7096
-
References
- 1 Moore JC, Spink J, Lipp M. Development and application of a database of food ingredient fraud and economically motivated adulteration from 1980 to 2010. J Food Sci 2012; 77: R118-R126
- 2 van der Kooy F, Maltese F, Choi YH, Kim HK, Verpoorte R. Quality control of herbal material and phytopharmaceuticals with MS and NMR based metabolic fingerprinting. Planta Med 2009; 75: 763-775
- 3 van Beek TA, van Veldhuizen A, Lelyveld GP, Piron I, Lankhorst PP. Quantitation of bilobalide and ginkgolides A, B, C and J by means of nuclear magnetic resonance spectroscopy. Phytochem Anal 1993; 4: 261-268
- 4 Duarte IF, Barros A, Almeida C, Spraul M, Gil AM. Multivariate analysis of NMR and FTIR data as a potential tool for the quality control of beer. J Agric Food Chem 2004; 52: 1031-1038
- 5 Minoja AP, Napoli C. NMR screening in the quality control of food and nutraceuticals. Food Res Int 2014; 63: 126-131
- 6 Le Gall G, Puaud M, Colquhoun IJ. Discrimination between orange juice and pulp wash by 1H nuclear magnetic resonance spectroscopy: identification of marker compounds. J Agric Food Chem 2001; 49: 580-588
- 7 Holmes E, Tang H, Wang Y, Seger C. The assessment of plant metabolite profiles by NMR-based technologies. Planta Med 2006; 72: 771-785
- 8 Rodrigues JE, Gil AM. NMR methods for beer characterization and quality control. Magn Reson Chem 2011; 49: S37-S45
- 9 Dais P, Hatzakis E. Quality assessment and authentication of virgin olive oil by NMR spectroscopy: a critical review. Anal Chim Acta 2013; 765: 1-27
- 10 Singh K, Blümich B. NMR spectroscopy with compact instruments. TrAC Trends Anal Chem 2016; 83: 12-26
- 11 Blümich B. Introduction to compact NMR: a review of methods. TrAC Trends Anal Chem 2016; 83: 2-11
- 12 Blümich B, Singh K. Desktop NMR and its applications from materials science to organic chemistry. Angew Chem Int Ed 2017; 56: 2-17
- 13 Skloss TW, Kim AJ, Haw JF. High-resolution NMR process analyzer for oxygenates in gasoline. Anal Chem 1994; 66: 536-542
- 14 Nordon A, Meunier C, Carr RH, Gemperline PJ, Littlejohn D. Determination of the ethylene oxide content of polyether polyols by low-field 1H nuclear magnetic resonance spectrometry. Anal Chim Acta 2002; 472: 133-140
- 15 Singh K, Blümich B. Desktop NMR spectroscopy for quality control of raw rubber. Macromol Symp 2016; 365: 191-193
- 16 Killner MHM, Danieli E, Casanova F, Rohwedder JJR, Blümich B. Mobile compact 1H NMR spectrometer promises fast quality control of diesel fuel. Fuel 2017; 203: 171-178
- 17 Killner MHM, Garro Linck Y, Danieli E, Rohwedder JJR, Blümich B. Compact NMR spectroscopy for real-time monitoring of a biodiesel production. Fuel 2015; 139: 240-247
- 18 Garro Linck Y, Killner MHM, Danieli E, Blümich B. Mobile low-field 1H NMR spectroscopy desktop analysis of biodiesel production. Appl Magn Reson 2013; 44: 41-53
- 19 Parker T, Limer E, Watson AD, Defernez M, Williamson D, Kate Kemsley E. 60 MHz 1H NMR spectroscopy for the analysis of edible oils. TrAC Trends Anal Chem 2014; 57: 147-158
- 20 Pàges G, Gerdova A, Williamson D, Gilard V, Martino R, Malet-Martino M. Evaluation of a benchtop cryogen-free low-field 1H NMR spectrometer for the analysis of sexual enhancement and weight loss dietary supplements adulterated with pharmaceutical substances. Anal Chem 2014; 86: 11897-11904
- 21 Jakes W, Gerdova A, Defernez M, Watson AD, McCallum C, Limer E, Colquhoun IJ, Williamson DC, Kemsley EK. Authentication of beef versus horse meat using 60 MHz 1H NMR spectroscopy. Food Chem 2015; 175: 1-9
- 22 Gouilleux B, Marchand J, Charriera B, Remaud GS, Giraudeau P. High-throughput authentication of edible oils with benchtop Ultrafast 2D NMR. Food Chem 2018; 244: 153-158
- 23 Meyer K, Kern S, Zientek N, Guthausen G, Maiwald M. Process control with compact NMR. TrAC Trends Anal Chem 2016; 83: 39-52
- 24 Defernez M, Wren E, Watson AD, Gunning Y, Colquhoun IJ, Le Gall G, Williamson D, Kate Kemsley E. Low-field 1H NMR spectroscopy for distinguishing between arabica and robusta ground roast coffees. Food Chem 2017; 216: 106-113
- 25 Singh K, Blümich B. Desktop NMR for structure elucidation and identification of strychnine adulteration. Analyst 2017; 142: 1459-1470
- 26 Browne CA. Adulteration and the condition of analytical chemistry among the ancients. Science 1909; 29: 455-458
- 27 Do TKT, Hadji-Minaglou F, Antoniotti S, Fernandez X. Authenticity of essential oils. TrAC Trends Anal Chem 2015; 66: 146-157
- 28 Boren KE, Young DG, Woolley CL, Smith BL, Carlson RE. Detecting essential oil adulteration. J Env Anal Chem 2015; 2: 132
- 29 Schmidt E, Wanner J. Adulteration of essential Oils. In: Başer KHC, Buchbauer G. eds. Handbook of essential Oils – Science, Technology, and Applications. 2nd ed.. Boca Raton: CRC Press; 2016: 707-745
- 30 Kubeczka KH, Formácek V. Essential Oil Analysis by capillary Gas Chromatography and Carbon-13 NMR Spectroscopy. New York: Wiley; 2002
- 31 van Beek TA, Joulain D. The essential oil of patchouli, Pogostemon cablin: a review. Flavour Frag J 2018; 33: 6-51
- 32 Yunxia F, Xiaoli C, Yupeng X, Songbai T. Corresponding factors influencing crude oils assay using low-field nuclear magnetic resonance. China Petr Proc Petrochem Technol 2014; 16: 34-39
- 33 van Beek TA, Lelyveld GP. Isolation and identification of the five major sesquiterpene hydrocarbons of ginger. Phytochem Anal 1991; 2: 26-34
- 34 Posthumus MA, van Beek TA, Collins NF, Graven EH. Chemical composition of the essential oils of Agathosma betulina, A. crenulata and an A. betulina x crenulata hybrid (Buchu). J Ess Oil Res 1996; 8: 223-228
- 35 Collins NF, Graven EH, van Beek TA, Lelyveld GP. Chemotaxonomy of commercial buchu species (Agathosma betulina and A. crenulata). J Ess Oil Res 1996; 8: 229-235
- 36 De Maesschalck R, Jouan-Rimbaud D, Massart DL. The Mahalanobis distance. Chemom Intell Lab Syst 2000; 50: 1-18
- 37 Sandra P, David F, Tienpont B. From CGC-MS to MS-based analytical decision makers. Chromatographia 2004; 60: S299-S302
- 38 Gouilleux B, Charrier B, Akoka S, Felpin FX, Rodriguez-Zubiri M, Giraudeau P. Ultrafast 2D NMR on a benchtop spectrometer: applications and perspectives. TrAC Trends Anal Chem 2016; 83: 65-75
- 39 Adhikary SR, Tuladhar BS, Sheak A, van Beek TA, Posthumus MA, Lelyveld GP. The oil of Cinnamomum glaucescens (Sugandha kokila). J Ess Oil Res 1992; 4: 151-159
Correspondence
-
ORCID
- Teris A. van Beek ORCID-🆔 https://orcid.org/0000-0002-9843-7096
-
References
- 1 Moore JC, Spink J, Lipp M. Development and application of a database of food ingredient fraud and economically motivated adulteration from 1980 to 2010. J Food Sci 2012; 77: R118-R126
- 2 van der Kooy F, Maltese F, Choi YH, Kim HK, Verpoorte R. Quality control of herbal material and phytopharmaceuticals with MS and NMR based metabolic fingerprinting. Planta Med 2009; 75: 763-775
- 3 van Beek TA, van Veldhuizen A, Lelyveld GP, Piron I, Lankhorst PP. Quantitation of bilobalide and ginkgolides A, B, C and J by means of nuclear magnetic resonance spectroscopy. Phytochem Anal 1993; 4: 261-268
- 4 Duarte IF, Barros A, Almeida C, Spraul M, Gil AM. Multivariate analysis of NMR and FTIR data as a potential tool for the quality control of beer. J Agric Food Chem 2004; 52: 1031-1038
- 5 Minoja AP, Napoli C. NMR screening in the quality control of food and nutraceuticals. Food Res Int 2014; 63: 126-131
- 6 Le Gall G, Puaud M, Colquhoun IJ. Discrimination between orange juice and pulp wash by 1H nuclear magnetic resonance spectroscopy: identification of marker compounds. J Agric Food Chem 2001; 49: 580-588
- 7 Holmes E, Tang H, Wang Y, Seger C. The assessment of plant metabolite profiles by NMR-based technologies. Planta Med 2006; 72: 771-785
- 8 Rodrigues JE, Gil AM. NMR methods for beer characterization and quality control. Magn Reson Chem 2011; 49: S37-S45
- 9 Dais P, Hatzakis E. Quality assessment and authentication of virgin olive oil by NMR spectroscopy: a critical review. Anal Chim Acta 2013; 765: 1-27
- 10 Singh K, Blümich B. NMR spectroscopy with compact instruments. TrAC Trends Anal Chem 2016; 83: 12-26
- 11 Blümich B. Introduction to compact NMR: a review of methods. TrAC Trends Anal Chem 2016; 83: 2-11
- 12 Blümich B, Singh K. Desktop NMR and its applications from materials science to organic chemistry. Angew Chem Int Ed 2017; 56: 2-17
- 13 Skloss TW, Kim AJ, Haw JF. High-resolution NMR process analyzer for oxygenates in gasoline. Anal Chem 1994; 66: 536-542
- 14 Nordon A, Meunier C, Carr RH, Gemperline PJ, Littlejohn D. Determination of the ethylene oxide content of polyether polyols by low-field 1H nuclear magnetic resonance spectrometry. Anal Chim Acta 2002; 472: 133-140
- 15 Singh K, Blümich B. Desktop NMR spectroscopy for quality control of raw rubber. Macromol Symp 2016; 365: 191-193
- 16 Killner MHM, Danieli E, Casanova F, Rohwedder JJR, Blümich B. Mobile compact 1H NMR spectrometer promises fast quality control of diesel fuel. Fuel 2017; 203: 171-178
- 17 Killner MHM, Garro Linck Y, Danieli E, Rohwedder JJR, Blümich B. Compact NMR spectroscopy for real-time monitoring of a biodiesel production. Fuel 2015; 139: 240-247
- 18 Garro Linck Y, Killner MHM, Danieli E, Blümich B. Mobile low-field 1H NMR spectroscopy desktop analysis of biodiesel production. Appl Magn Reson 2013; 44: 41-53
- 19 Parker T, Limer E, Watson AD, Defernez M, Williamson D, Kate Kemsley E. 60 MHz 1H NMR spectroscopy for the analysis of edible oils. TrAC Trends Anal Chem 2014; 57: 147-158
- 20 Pàges G, Gerdova A, Williamson D, Gilard V, Martino R, Malet-Martino M. Evaluation of a benchtop cryogen-free low-field 1H NMR spectrometer for the analysis of sexual enhancement and weight loss dietary supplements adulterated with pharmaceutical substances. Anal Chem 2014; 86: 11897-11904
- 21 Jakes W, Gerdova A, Defernez M, Watson AD, McCallum C, Limer E, Colquhoun IJ, Williamson DC, Kemsley EK. Authentication of beef versus horse meat using 60 MHz 1H NMR spectroscopy. Food Chem 2015; 175: 1-9
- 22 Gouilleux B, Marchand J, Charriera B, Remaud GS, Giraudeau P. High-throughput authentication of edible oils with benchtop Ultrafast 2D NMR. Food Chem 2018; 244: 153-158
- 23 Meyer K, Kern S, Zientek N, Guthausen G, Maiwald M. Process control with compact NMR. TrAC Trends Anal Chem 2016; 83: 39-52
- 24 Defernez M, Wren E, Watson AD, Gunning Y, Colquhoun IJ, Le Gall G, Williamson D, Kate Kemsley E. Low-field 1H NMR spectroscopy for distinguishing between arabica and robusta ground roast coffees. Food Chem 2017; 216: 106-113
- 25 Singh K, Blümich B. Desktop NMR for structure elucidation and identification of strychnine adulteration. Analyst 2017; 142: 1459-1470
- 26 Browne CA. Adulteration and the condition of analytical chemistry among the ancients. Science 1909; 29: 455-458
- 27 Do TKT, Hadji-Minaglou F, Antoniotti S, Fernandez X. Authenticity of essential oils. TrAC Trends Anal Chem 2015; 66: 146-157
- 28 Boren KE, Young DG, Woolley CL, Smith BL, Carlson RE. Detecting essential oil adulteration. J Env Anal Chem 2015; 2: 132
- 29 Schmidt E, Wanner J. Adulteration of essential Oils. In: Başer KHC, Buchbauer G. eds. Handbook of essential Oils – Science, Technology, and Applications. 2nd ed.. Boca Raton: CRC Press; 2016: 707-745
- 30 Kubeczka KH, Formácek V. Essential Oil Analysis by capillary Gas Chromatography and Carbon-13 NMR Spectroscopy. New York: Wiley; 2002
- 31 van Beek TA, Joulain D. The essential oil of patchouli, Pogostemon cablin: a review. Flavour Frag J 2018; 33: 6-51
- 32 Yunxia F, Xiaoli C, Yupeng X, Songbai T. Corresponding factors influencing crude oils assay using low-field nuclear magnetic resonance. China Petr Proc Petrochem Technol 2014; 16: 34-39
- 33 van Beek TA, Lelyveld GP. Isolation and identification of the five major sesquiterpene hydrocarbons of ginger. Phytochem Anal 1991; 2: 26-34
- 34 Posthumus MA, van Beek TA, Collins NF, Graven EH. Chemical composition of the essential oils of Agathosma betulina, A. crenulata and an A. betulina x crenulata hybrid (Buchu). J Ess Oil Res 1996; 8: 223-228
- 35 Collins NF, Graven EH, van Beek TA, Lelyveld GP. Chemotaxonomy of commercial buchu species (Agathosma betulina and A. crenulata). J Ess Oil Res 1996; 8: 229-235
- 36 De Maesschalck R, Jouan-Rimbaud D, Massart DL. The Mahalanobis distance. Chemom Intell Lab Syst 2000; 50: 1-18
- 37 Sandra P, David F, Tienpont B. From CGC-MS to MS-based analytical decision makers. Chromatographia 2004; 60: S299-S302
- 38 Gouilleux B, Charrier B, Akoka S, Felpin FX, Rodriguez-Zubiri M, Giraudeau P. Ultrafast 2D NMR on a benchtop spectrometer: applications and perspectives. TrAC Trends Anal Chem 2016; 83: 65-75
- 39 Adhikary SR, Tuladhar BS, Sheak A, van Beek TA, Posthumus MA, Lelyveld GP. The oil of Cinnamomum glaucescens (Sugandha kokila). J Ess Oil Res 1992; 4: 151-159