3.1 Experimental design and selection of zones for evaluation

Different saturated salt and aqueous sulfuric acid solutions were utilized to evaluate the effect of RH on the separation of E 472 subgroups. Application of 70% sulfuric acid solution within the humidity control unit of the ADC 2 was not possible due to its high corrosivity, so automatic and manual techniques for setting the RH were evaluated (Sect. 2.7). The first technique (A), the automatic setting of the RH in the humidity control unit of the ADC 2 (which is used for setting the plate activity), was used with saturated salt solutions. Two techniques were investigated for aqueous sulfuric acid solutions: chamber saturation (B), applied for setting the chamber climate, and plate preconditioning (C) for assessing the plate activity. Thus, in this study, the termini chamber saturation and plate preconditioning refer to the application of saturated salt and aqueous sulfuric acid solutions, not for mobile phase usage. Techniques (B) and (C) were also evaluated with saturated salt solutions for a meaningful comparison of the results.

For better clarity and more straightforward evaluation, selected zones for MAG, DAG, and triacylglycerides (TAG) were evaluated. In the first step, zones were assigned to substance groups, according to Oellig et al. 2020 [6], which was possible despite minor changes in the solvent ratios of the mobile phase. For the selection of the zones to be evaluated, differences in the hRF values were considered, leading to three zones for each emulsifier covering nearly the entire migration distance. For E 472a, selected zones were MAG (a1), monoacetic acid esters (a2), and diacetic acid esters (a3) of MAG. For E 472b, selected zones were MAG (b1), DAG and lactic acid esters of DAG (b2), and TAG (b3). For E 472c, selected zones were MAG and citric acid esters of DAG esterified with an additional glycerol (c1), DAG (c2), and TAG (c3), and for E 472e, MAG (e1), diacetic acid esters of MAG (e2), and TAG (e3) (Fig. 1) [6].

Fig. 1

Separation of E 472 emulsifier subgroups on HPTLC silica gel under UV 366 nm illumination after two-fold development with chloroform–methanol–water–formic acid (67:6:1.2:0.2, V/V) to a migration distance of 50 mm and n-heptane–diethyl ether–formic acid (55:45:1, V/V) to a migration distance of 80 mm at a relative humidity (RH) of 33% (set with the ADC 2), and after derivatization with primuline. The standard mix (St) for comparison consists of 1- and 2-monostearate (MSt), 1,2- and 1,2-distearate (DSt), stearic acid (SA), and tristearate (TSt). Zones selected for the evaluation of the influence of the RH on the separation of E 472 emulsifiers are marked with a1 to e3 (a for E 472a, b for E 472b, c for E 472c, and e for E 472e)

The available laboratory for the experiments was not equipped with a humidity control unit or a dehumidifier guaranteeing a constant RH. Hence, during the experiments, RH in the laboratory normally deviated between 30% and 60% and, on some days, even reaching up to 80%. If intralaboratory RH was above 60%, low RH (< 30%) could not be achieved with 60% and 70% sulfuric acid solution or molecular sieve within 15 min. This was demonstrated with an experiment with 60% sulfuric acid solution at intralaboratory RH of 68%. RH was measured inside an enclosed twin-trough chamber. One side was filled with 60% sulfuric acid solution, and after 1 h of preconditioning, 34% RH was reached. Compared with the theoretical RH defined as 15% [16], this is a deviation of > 100%. Consequently, analyses involving low RH were performed on days when RH was < 60% in the laboratory, when theoretical RH values were nearly reachable within 15 min (chamber saturation) or 5 min (automatic humidity control in the ADC 2), respectively (Table 1). To ensure comparable conditions, all salt and sulfuric acid solutions were conditioned for the same duration (15 min) because, for technical reasons, the mobile phase must be added simultaneously with the saturated aqueous salt and sulfuric acid solutions (in the ADC 2).

Table 1 Saturated aqueous salt and sulfuric acid solutions of different concentrations with their theoretical [16, 17] and actual RH measured manually by a hygrometer in a twin-trough chamber after 15 min of saturation (manual procedure of setting the RH) and in the ADC 2 after 5 min of automatic humidity control at a temperature of 25–30 °C and a RH of 30–60% in the laboratory3.2 Effect of chamber saturation and plate preconditioning with mobile phase on the separation of E 472 emulsifiers

Due to technical reasons (by application of the ADC 2), experiments with the saturated twin-trough chamber involved simultaneous saturation with salt or sulfuric acid solution and the mobile phase. In the original method by Schuster et al. [7], saturation or plate conditioning with the mobile phase was omitted. So, the effect of chamber saturation and plate conditioning with the mobile phase was also evaluated. For all E 472 subgroups, no effect on separation was detected, neither for plate conditioning nor for chamber saturation with the mobile phase (data not shown).

3.3 Effect of RH on the chromatographic behavior (hR F)

For all four emulsifier types, equal trends regarding variation in hRF values for the selected signals were observed. Exemplarily, data for the E 472b emulsifier is presented in the discussion, while data for E 472a, E 472c, and E 472e are provided in the Supplementary Information. Using aqueous sulfuric acid solutions in different concentrations for the RH setting, a decrease of hRF values at decreasing RH could be observed for the RH setting technique (B) chamber saturation for 15 min and (C) plate conditioning for 15 min (Fig. 2B and C, blue trend line). Interestingly, the phenomenon was observed only for zones with higher hRF values (zones a2, a3, b2, b3, c2, c3, and e2 and e3). This result aligns with the literature [13, 15], associating higher hRF values at increasing RH with stationary phase deactivation due to increasing adsorbed water content. Strikingly, no trends in hRF value changes were observed across the tested RH range when using saturated salt solutions for all setting techniques (Fig. 2A–C). To estimate and assess the variation of hRF values, reproducibility data determined in the scope of method validation by Schuster et al. (2022) were involved and evaluated [7]. Exemplarily for signal b1, based on the reproducibility, hRF values ranged at 26 ± 1. Variations of the hRF values in the present study were with a few exceptions in this range (hRF 25–28). When the same technique was used for the adjustment as with sulfuric acid solution, and the same RH were achieved within the same time using saturated salt solutions, the results diverged, which could not be explained. One suggestion is that sulfuric acid vapors are partially responsible for these findings. To investigate this matter, 10 mL of 70% aqueous sulfuric acid were placed in one trough of a twin-trough chamber, and the change of pH was monitored through the color of a pH paper placed on the other trough every 5 min. After 15 min, pH was approximately at 3, and after a further 10 min at 1, confirming acidic vapors in the chamber that could lead to an acidification of the stationary phase. Another suggestion is that the highly hygroscopic sulfuric acid can lead to desorption of water from the stationary phase and, thus, influences plate activity. Both possible explanations are not confirmed yet and need further investigation. Berezkin et al. (2006) further described changes in the pH of the mobile phase during elution due to its contact with either an acidic (CO2) or basic (NH3) gas phase, influencing the acid–base speciation of the tested benzoic acid and aromatic amine mixtures [18].

Fig. 2

hRF values of selected signals (b1 (monoacyglycerides)–square, b2 (diacylglycerides and lactic acid esters of diacylglycerides)–triangle, b3 (triacylglycerides)–diamond) of an E 472b emulsifiers plotted against the relative humidity. Setting of the relative humidity (RH) (A) automatically by the humidity control function of the ADC 2, (B) manually by chamber saturation for 15 min, and (C) manually by plate preconditioning for 15 min by using saturated salt solution (black) and aqueous sulfuric acid solutions (blue)

For zones with low hRF values (a1, b1, c1, a2, Fig. 2, and SI), differences between the different techniques of RH setting were seen. For example, for signal b1 of the E 472b emulsifier, mean hRF values increased from 12 [(B) chamber saturation] to 17 [(C) plate preconditioning] to 26 [(A) ADC 2], showing a variance of > 100% between techniques (A) and (B). This indicates that the two developments (Sect. 2.6) might be affected to different extents, possibly due to varying polarity of the mobile phase. Considering these results, the developed method by Schuster et al. (2023) proves to be robust against changes in RH.

3.4 Effect of the technique for setting the RH on the chromatographic behavior

Unlike the RH itself, the RH setting technique resulted in a considerable change in the chromatographic fingerprint. This was exemplarily shown and discussed for the RH of 75% (Fig. 3) and additionally visualized for the RH of 33% in the SI. This was most evident in zones with low hRF values, where the hRF values exhibited variations for these zones between the different techniques of setting the RH. A more detailed signal fine structure was detected for technique (A) ADC 2 in comparison with technique (B) chamber saturation and (C) plate conditioning. Examples of alterations in the chromatographic fingerprint are described in detail for the E 472a and E 472b emulsifiers, but changes were also observed for the E 472c and E 472e emulsifiers. Viewing the E 472b fingerprint for technique (C), an initial separation of zone b1, as indicated by the blurred zones above b1, was observed (Fig. 3, E 472b). Prolonging the conditioning time would probably lead to a more detailed signal fine structure. Zeeuw (1968) performed plate conditioning for 24 h [15], and Suzuki and Matsushita (1967) equilibrated until RH reached a steady state without mentioning a specific time [14].

Fig. 3

Separation of E 472 emulsifier subgroups on HPTLC silica gel under UV 366 nm illumination after two-fold development with chloroform–methanol–water–formic acid (67:06:1.2:0.2, V/V) to a migration distance of 50 mm and n-heptane–diethyl ether–formic acid (55:45:1, V/V) to a migration distance of 80 mm at a relative humidity (RH) of 75%, and after derivatization with primuline. Setting of the RH (A) automatically by the humidity control function of the ADC 2, (B) manually by chamber saturation for 15 min, and (C) manually by plate preconditioning for 15 min

For the E 472a emulsifier fingerprint, considerable variations were detected for signal a2 applying the different techniques (Fig. 3, E 472a). In contrast to the observations for the signal b2 of E 472b, the signal fine structure was more detailed for the techniques (B) and (C) than for (A). Consequently, the different techniques of setting the RH need to be evaluated for every analyte regarding the best result for a respective goal of the analysis.

Despite variations in the TLC fingerprint between the three techniques of setting the RH, the chromatographic fingerprint within the same technique did not vary over the range of tested RH. However, using the sulfuric acid with technique (C) plate preconditioning delivered changes, as depicted in Fig. 4 for the E 472b emulsifier. With decreasing RH (applying increasing sulfuric acid concentration), the signal fine structure became more precise and detailed, resembling the fingerprint obtained by setting of the RH with the ADC 2. Surprisingly, this effect was not obtained for technique (B), the chamber saturation. Respective zones were examined by mass spectrometry to investigate a possible cleavage of the ester bonds of the lactylated glycerides that could be related to sulfuric acid vapors in the chamber. No change in mass-to-charge (m/z) ratios was detected in comparison with RH setting with saturated salt solutions. Nevertheless, potential partial hydrolysis at least to a minor degree of the esters in the presence of the acidic vapors could not be ruled out. The results clearly show that the influence of RH varies depending on analytes and chromatographic systems and needs to be evaluated for every method.

Fig. 4

Separation of an E 472b emulsifier on HPTLC silica gel under UV 366 nm illumination after two-fold development with chloroform–methanol–water–formic acid (67:06:1.2:0.2, V/V) to a migration distance of 50 mm and n-heptane–diethyl ether–formic acid (55:45:1, V/V) to a migration distance of 80 mm at relative humidities (RH) between 10 and 82% set with aqueous sulfuric acid solutions, and after derivatization with primuline. The RH (B) was set manually by chamber saturation for 15 min, and (C) manually by plate preconditioning for 15 min