Characterization of vanillin carbon isotope delta reference materials

Stable carbon isotope ratio measurements are used to investigate the provenance of vanillin. In this study, a variety of commercial vanillin samples and vanilla products were analyzed to provide a frame of reference for the variability of carbon isotope delta values in various vanillin samples, with the results ranging from −20.6 to −36.7‰ relative to the Vienna Peedee Belemnite (VPDB). We present information on the development of two synthetic vanillin reference materials, VANA-1 and VANB-1, prepared in 0.75 g units in glass vials, to be used for the calibration of carbon isotope delta measurements of vanillin and other easily combustible organic materials. Characterization of 40 vials each of VANA-1 and VANB-1 was performed by three laboratories over several measurement sequences. The certified carbon isotope delta values are −31.30 ± 0.06‰ (VANA-1) and −25.85 ± 0.05‰ (VANB-1). These uncertainties, for the 95% confidence level, include considerations for measurement uncertainty, coherence of the reference materials used for calibration, batch homogeneity, and stability during storage and transportation. The results are traceable to the VPDB through a set of nine reference materials (IAEA-CH-6, USGS65, IAEA-600, NBS22, USGS61, IAEA-603, IAEA-610, IAEA-611, and IAEA-612). For up to date certified values, users should refer to doi.org/10.4224/crm.2022.vana-1 and doi.org/10.4224/crm.2022.vanb-1. Supplementary Information The online version contains supplementary material available at 10.1007/s00216-022-04322-x.


Introduction
Determining the authenticity and provenance of food items is an integral part of safeguarding food supply chains [1][2][3]. Vanilla, a plant which is cultivated in tropical climates, is one of the most widely used food flavors in the world. The main flavor component of vanilla is vanillin, and this ingredient can be found in a wide variety of processed foods [1][2][3][4]. Vanillin comprises only a few percent of the vanilla bean, and its extraction is a lengthy and expensive process [3]. The demand for vanillin far outweighs the production of vanilla beans, thus resulting in a 100-fold premium paid for vanillin extracted from vanilla pods compared to synthetic vanillin [1][2][3]. While most of the global vanillin supply is a synthetic vanillin made from olefin-based precursor guaiacol [3], it is now common to see "biovanillin" made from natural precursors such as rice [4][5][6].
To minimize sources of error in carbon isotope delta measurements, the reference materials (RMs) used for calibration to the Vienna Peedee Belemnite (VPDB) should have the same matrix as the sample [19][20][21]. Several suites of RMs have been recently developed to provide a wider variety of available matrices [22][23][24]; however, vanillin reference materials are still lacking. This paper describes the measurement of vanillin and vanilla samples to provide a frame of reference for the natural variability of carbon isotope delta values in commercial vanillin. Further, the development of two synthetic vanillin certified reference materials (CRMs), VANA-1 [25] and VANB-1 [26], are described, which are intended for use as a calibrant for carbon isotope delta measurements using IRMS, or as a comprehensive standard for other measurement techniques such as NMR [12,27].

Vanillin samples
Twenty-two samples of vanillin and vanilla products from various sources were sourced by the National Research Council Canada (NRC) and the Department of Chemistry Malaysia (DoCM) as shown in Table 1. These include a variety of pure vanillin samples from either natural or synthetic sources as well as raw natural materials containing vanillin.
Vanillin extracted from vanilla flavor (samples 18 and 19) were prepared at the DoCM. Thirty grams (30 g) of each vanillin sample powder was weighed into a separate 100 mL beaker and dissolved in 25 mL methanol. The solution was stirred for 10 min prior to being filtered. The filtrate was washed with methanol and dried overnight in a water bath at 80 °C until dryness. The dried crystals were re-dissolved using hot DI water, stirred, and filtered to remove residue and oil. The filtrate was immersed in chilled water to re-form vanillin crystals, which were then filtered, washed with chilled water, and placed in the oven to dry overnight.
For CRM production, two batches of high-purity (> 99%) synthetic vanillin were purchased by the NRC from Canadian chemical suppliers: 1 kg from ACROS Chemicals (Fisher Scientific, Waltham, MA, USA) and 2 kg from Sigma Aldrich (St. Louis, MO, USA). Aliquots from materials were stored in 2 mL glass vials capped with black phenolic caps with a polyethylene conical liner (Wheaton, supplied by Fisher Scientific, Waltham, MA, USA), and kept in a drybox at room temperature until analyzed.

Carbon isotope delta measurements
Details regarding the instrumentation and procedures used for carbon isotope delta measurements performed at the NRC; the Ján Veizer Stable Isotope Laboratory at the University of Ottawa, Canada (UO); the stable isotope laboratory at the Research Centre in Earth System Dynamics (Geotop) at the Université du Québec à Montréal, Canada; and DoCM are presented in the Electronic supplementary material. Two different sets of calibrators were used to determine carbon isotope delta values relative to the VPDB for vanillin and vanilla samples listed in Table 1: DoCM used BEET-1, FRUT-1, and GALT-1 [23] whereas UO, Geotop, and NRC used IAEA-CH-6 [28], USGS65 [29], IAEA-600 [20], NBS22 [28], and USGS61 [30]. Reference materials used for calibration of VANA-1 and VANB-1 are discussed later.

Commercial vanillin sample measurements
Carbon isotope delta measurements can be used to differentiate between sources of vanillin [2,[6][7][8][9][10][11][12][13]. In an effort to map the vanillin carbon isotope delta space, the vanilla and vanillin samples described in Table 1 were analyzed ( Fig. 2 and Electronic supplementary material Table S1). The agreement in the average reported δ VPDB ( 13 C) values between the laboratories was within 0.4 ‰ for all vanillin samples, and within 1 ‰ for the vanilla samples.
As vanilla undergoes CAM photosynthesis, the carbon isotope delta values can vary from −34 to −10 ‰, depending on the growing conditions [16,18]. All vanilla samples analyzed for this study exhibited C4-type photosynthesis, indicating a hot and arid climate typical of regions producing vanilla beans. Additionally, while carbon isotope delta measurements can identify natural vanillin, vanillin samples obtained from synthetic or biosynthetic means display largely identical isotopic compositions.
Despite its popularity, the power of carbon isotope delta measurements in provenance studies remains limited. The price of the various precursors of vanillin ranges from a few dollars to thousands of dollars per kilogram. However, the carbon isotope delta values alone are unable to distinguish between the most expensive natural vanillin (1200 USD/kg [3,4,6]) and the least expensive biovanillin obtained from  [4,6]). Thus, additional isotopes are often measured or intra-molecular measurements are performed using 13 C-qNMR [2,9,10,12,13,27].

Development of vanillin CRMs
VANA-1 was prepared from vanillin purchased from ACROS Chemicals, with the appearance of a white powder. The 1 kg bottle was tumbled for 2 h to homogenize the contents, then sieved through a US standard 30 mesh (595 µm) sieve. The sieved VANA-1 was transferred to a 4 L bottle and tumbled for an additional 45 min to further homogenize the sample. VANB-1 was prepared from vanillin purchased from Sigma Aldrich. It appeared to be a mix of white powder and needleshaped larger crystals. This material was ground in batches using a ball mill with a stainless steel grinding jar and balls and then sieved through a US standard 50 mesh (300 µm) sieve. The sieved VANB-1 was transferred to a 4 L bottle and tumbled for 45 min. Aliquots of 0.75 g of the prepared VANA-1 and VANB-1 material were transferred into 2 mL glass vials capped with black phenolic caps with a polyethylene conical liner. The vials were sealed in trilaminate bags and stored at room temperature. In total, 1000 units each of VANA-1 and VANB-1 were prepared.

Stability study
In order to qualify the storage and shipping conditions of a candidate CRM, the stability of the material at various environmental conditions needs to be assessed. Short-term stability studies are intended to model extreme temperature exposures experienced during shipping/transit of the material from the producer to the user laboratories. For each CRM, two unopened vials were stored in the oven (+40 °C) and the freezer (−20 °C). After 16 days, the contents of each vial were measured in triplicate at the NRC using IAEA-CH-6, USGS65, IAEA-600, NBS22, and USGS61 as calibrators (ST1/ST2 vs ST0, Table 2). No effect in the δ VPDB ( 13 C) values was observed.
In addition to the short-term stability experiment, the impact of loss due to the volatilization of vanillin from the open vials was studied. The caps on two vials each of VANA-1 and VANB-1 were removed; the vials were lightly covered and left open in the fume hood at room temperature. After 16 days, each vial was measured in triplicate (ST3 vs ST0, Table 2). No effect in the δ VPDB ( 13 C) values was observed.
To assess the effect of long-term storage of VANA-1 and VANB-1 at room temperature in trilaminate bags, two unopened units of both VANA-1 and VANB-1 were analyzed 23 months after the initial measurement campaign (LT1 vs ST0, Table 2). No effect in the δ VPDB ( 13 C) values was observed. Thus, both VANA-1 and VANB-1 are deemed stable with respect to long-term storage at room temperature, and no special shipping requirements are warranted.
Additional measurements of VANA-1 and VANB-1 were performed at Geotop and NRC using a set of carbonate standards from the International Atomic Energy Agency, IAEA-603, IAEA-610, IAEA-611, and IAEA-612, whose   [31,32]. Since this realization of the VPDB differs from the conventional NBS19-LSVEC definition, we employed the equation given by Hélie et al. [33] to convert the δ( 13 C) values from the VPDB2020 scale to the VPDB2006 scale for these measurements. Table 3 lists the values used for the calibration standards in this work, and Fig. 3 shows the traceability pathways of these materials.

Data reduction and uncertainty evaluation
Data reduction was performed by the NRC. For each measurement sequence (k = 1…6) from either of the three laboratories, a linear multi-point calibration curve was established using a laboratory-specific errors-in-variables statistical model. Thus, the individual isotope delta measurements of the calibration standards made by laboratory against the CO 2 working gas, d k , were modeled as follows: where a k and b k are the laboratory-specific linear regression parameters (intercept and slope), u k is the laboratoryspecific measurement uncertainty, and D i is the true isotope delta value associated with each calibrator: Prior probability distributions for all model parameters were set to be vaguely informative, and the full specification of the statistical model is provided in the Electronic supplementary material.
The vanillin sample measurements were modeled using similar considerations, and the calibrated results were calculated from the appropriate calibration function for each measurement sequence. The obtained carbon isotope delta values for all vanillin samples were then combined with considerations that they arise from a common consensus value affected by random laboratory effect and by the bottle-to-bottle homogeneity. The model fitting was done using Markov Chain Monte Carlo in R using rjags resulting in a single consensus value for carbon isotope delta for each material along with the estimate for the uncertainty due to homogeneity. The results of each individual measurement sequence are shown in the Electronic supplementary material Table S3, and the certified δ VPDB ( 13 C) values for VANA-1 and VANB-1 are shown in Table 4. The standard uncertainty associated with the certified value (u) is obtained by combining the uncertainty of the consensus value (u char ) and homogeneity (u hom ). The expanded combined uncertainty (U 95% CI ) is taken as U 95% = 2u. As noted before, the uncertainties associated with stability during short-and long-term storage were assessed and considered to be negligible.
Recent studies [22][23][24] have provided suites of the same material intended to be used as a set for calibration of samples with similar matrices. In this study, we have provided two CRMs with similar δ VPDB ( 13 C) values, which may be used as a QC material, or as a calibrant for carbon isotope delta Table 3 Re-assigned carbon isotope delta values and the associated standard uncertainties for the reference materials used to calibrate VANA-1 and VANB-1   (10), and propagate its uncertainty to the derived quantities shown in Table 5 using the NIST Uncertainty Machine [37].

Conclusions
To demonstrate the variability of carbon isotope delta measurements in vanillin samples derived from different source materials, we analyzed a variety of commercial vanillin samples and vanilla products, with the results ranging from −20.6 to −36.7 ‰. The results for the vanillin with known provenance were consistent with the previously published values for different vanillin sources. Further, we report the characterization of carbon isotope delta values in two isotope reference materials of synthetic vanillin, VANA-1 and VANB-1, with certified carbon isotope delta values of −31.30 ± 0.06 ‰ and −25.85 ± 0.05 ‰ (k = 2) respectively, and traceable to the VPDB through nine reference materials. The 95 % confidence level uncertainties include components for measurement uncertainty, calibration reference material coherence, and bottle-to-bottle homogeneity. The uncertainty due to short-term stability during transport and storage, as well as long-term storage stability, was deemed negligible. VANA-1 and VANB-1 are the first reference materials on track to be characterized not only for bulk carbon isotope delta values, but also for site-specific carbon isotope delta values [27]. VANA-1 and VANB-1 are intended for use as a calibrant or as a QC material for IRMS measurements of vanillin or samples of other similar matrices. A vanillin reference material with a carbon isotope delta value of approximately −15 ‰ might be of future interest.
Funding Open Access provided by National Research Council Canada.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.  (10) 12.010 682(10) VANB-1 0.010 705 (10) 0.010 821 (10) 12.010 741 (10)