Introduction

The term ‘amber’ is the most known and frequently used to indicate resinite or fossil resins formed when natural resin was exuded from the tree wounds and exposed to light and oxygen. However, these fossil resins usually differ significantly in terms of geological age, provenience, botanical designation and chemistry [1]. Natural resins constitute a mixture of organic high molecular mass compounds with polymeric-like materials. Depending on the botanical source, geographical area and temperature–pressure conditions of their diagenesis, different resins occurring worldwide can be distinguished. There are five classes in the classification system of natural resins structure that group various natural polymers, non-polymeric material containing sesquiterpenoids and diterpenoids carboxylic acids [2, 3].

Various species of trees produce natural resins; however, only those resistant to decay may further undergo fossilization. During this process, the resins, preserved in various deposits, change its structure and composition over time by formation of higher molecular mass compounds and conversion of volatile and reactive components into stable species. The main process in fossils formation is polymerization, but also the curing process occurs which is equivalent to the synthetic curing process but much slower. This curing/maturation of natural resins depends not only on the duration of burial but also primary chemical composition and alteration processes in geological environment [4]. Moreover, natural resins even after millions of years still preserve a significant amount of unpolymerized material [5,6,7].

There are only few studies on thermal analysis of natural resins [7,8,9,10,11], where authors tried to correlate thermal properties (glass transition, combustion profiles) of resins with their age. However, there is no agreement whether the thermal analysis is appropriate tool for assessment of natural resins age. Some authors stated that TG and DSC can give information that helps in dating, while another postulate that the glass transition cannot be used for this. Therefore, a good understanding of fossilization process, thermal behavior of natural resins and their structural changes seems to be important in natural resins characterization. Besides, while Baltic amber is widely described and characterized [12], Dominican and Colombian resins still need more interest. On the other hand, to the best of our knowledge, there are no reports concerning Khatanga resin; only Kaliningrad materials were investigated [13, 14].

In this paper, we perform preliminary studies on the thermal properties of differently aged natural resins from Russia (Khatanga), Dominican Republic (El Valle), Colombia and Poland (Jantar). This work is a part of studies on properties of fossil resins, in which a larger set of specimens differentiated by geographical locality and age is investigated by various analytical methods. Initially, this group of four resins was tested according to thermal stability (TG) and behavior under elevated temperatures (DSC, FT-IR) to show that they have different structure and composition. Furthermore, we cross-evaluate our results with those from the literature to show that this complex subject still needs scientific attention.

Materials and methods

The characteristics of the analyzed resin samples are given in Table 1. It provides information on the geographical origin of samples, their age, geological setting of deposit and short description of resins. The samples were denoted with a collection number belonging to the authors.

Table 1 Natural resins characterization
Full size table

Before analysis, the samples of natural resins were cut and grinded in the laboratory mill equipped with a liquid nitrogen cooling system (Retsch CryoMill). First, the material was cooled down in about 15 min and then, three grinding steps were applied, each one for 5 min with frequency of 15 Hz. The total time of grinding was about 20 min.

The obtained powdered sample was then weighed and put in the corundum Al2O3 (TG) or aluminum Al (DSC) crucible to test its thermal behavior. The materials were tested from 25 to 700 °C in TG (TGA2 Mettler Toledo) and from − 60 °C to 200 °C in DSC (DSC1 Mettler Toledo) at 10 °C min−1 heating/cooling rate. DSC instrument was calibrated using indium and zinc standards. Nitrogen atmosphere (50 mL min−1) has been applied in DSC, to avoid oxygen effect on thermal properties of the material. TG experiments were performed under dynamic flow of air (30 mL min−1) to investigate degradation of the material under oxidative conditions.

Structural changes after thermal treatment were investigated by ATR FT-IR measurements. The spectra were recorded on the pristine resins samples and samples after DSC measurements with Bruker TENSOR27 device with a RT-DLaTGS detector and diamond crystal (32 scans, 4 cm−1 resolution). Deconvolution of the spectra and fitting procedure was performed with the Grafity version 0.4.5 open source software, while the spectra were baseline-corrected and processed with OPUS software (version 6.0). In order to compare, the spectra were also normalized with respect to the most intense band in the spectrum. For each vibrational signal, its position, intensity (amplitude) and its full width at half maximum (FWHM) were recorded.

Results and discussion

Natural resins, which are in fact natural polymers originated from different types of trees, were analyzed by thermal analysis methods (TG, DSC), and selected samples were structurally characterized (FT-IR). These kinds of biomaterials are known from their sensitivity to physicochemical environmental factors, e.g., Baltic amber is extremely prone to progressive degradation and eventually to complete disintegration by atmospheric oxidation, accelerated by heat and light [22,23,24].

Thermal stability—Thermogravimetry TG

TG results have shown that resins continuously loose mass on heating through vaporization (low molecular mass products), possible post-curing reactions and decomposition. The slope of TG curves of tested materials allows to divide them into two groups, according to their thermal behavior: (I) Dominican DO/1 and Polish PL/J/1 and (II) Russian RU/Kh/1 and Colombian CO/1 resins.

In general, the first group (I–Fig. 1a) has shown two-step degradation with the main mass loss events in the range of 200–440 °C and 440–600 °C. The DTG profiles suggest that the degradation pathway is less complicated than in the case of other resins under investigation.

Fig. 1
figure 1

Thermal behavior of the natural resins under air atmosphere: a group I and b group II

Full size image

The Colombian (CO/1) and Russian (RU/Kh/1) resins were assigned to second group (II–Fig. 1b), and they are characterized by three-step decomposition with the main mass losses in the range of 220–340 °C, 340–480 °C and 480–600 °C.

In the first decomposition step, most probably low molecular mass components are evolved, such as water and organic species produced during post-curing reactions. The intensive mass loss observed between 200 and 400 °C for the (I) group of resins is related to the decomposition of less stabile structures of the resin. Above 450–580 °C, the main decomposition takes place and this process was registered as more intensive in the (I) group of resins, as compared to the height of DTG signals between low- and high-temperature region. For the (II) group, several extrema were observed on DTG profile in the range of 340–480 °C. This is likely due to melting and subsequent foaming upon heating, that is common for organic systems. It may also suggest that complex reactions of decomposition take place; however, we have found that these signals are not reproducible and this supports our first observation with foaming process. Thermogravimetric step in this range is the most intensive of all, with 49–69 mass% mass loss.

Analyzing the char residue after TG measurements, one can conclude that Russian material seems to be more resistant and produces more solid residue than other resins under investigation. All materials decompose completely or with slight char residue. A comparison of the samples is given in Table 2, where the characteristic temperatures and moss losses for three thermal decomposition regions are presented.

Table 2 Parameters of thermogravimetric analysis of natural resins under oxidative atmosphere
Full size table

The presented results are close to those obtained by Ragazzi et al. [8], who investigated different kinds of natural resins, including Dominican amber, Colombian copal and Baltic amber. The most complicated thermal degradation profile was registered for Baltic amber, but all of those materials expose the main mass loss at about 400 °C. In contrary to those results, for the materials of (I) group, the ‘main’ DTG peak (i.e., the first major thermal event) was observed about 320 °C, while RU/Kh/1 and CO/1 resins from second group showed the main DTG signal above 400 °C. The discrepancy between our results and those reported in the literature may be caused by variability of physico-chemical properties of resin specimens from the same localization or different experimental setups (different sample masses, measurement parameters and procedure for the preparation). This requires more material to be tested; however, the availability of specimens from the same deposit is very limited. Moreover, the resins may undergo structural reorganization when are exposed to air. Nevertheless, the presented here results contain a basic knowledge and will be useful for further more detailed studies.

Despite their geological age, both samples of (II) group followed similar main degradation route, but CO/1 showed additionally low temperature degradation peaks, which can be more meaningful that the ‘main’ DTG peak itself. These thermal events may come from the chemical composition of Colombian copal and also its relatively ‘young’ structure. Moreover, similar observations we have made for the (I) group of resins exposed thermal signals between about 125 °C and 240 °C, which may be related to the presence of reactive species in the resin [9]. This stays in accordance with the earliest reports on natural resins. For example, Poinar [25] suggests that the qualities of copal, as younger resins which contain unpolymerized fraction, may preserve this reactivity for even 3 or 4 millions of years.

As it was previously reported [3], the resin fossilization runs through polymerization of nonvolatile ingredients, cross-linking and isomerization, after low molecular mass compounds have been evaporated. This means that thermal behavior may reflect maturation histories and resin transformations [26].

Here, as a general conclusion from the comparative analysis of the data, one can said that there is no clear dependence on thermal stability vs age of the resin; hence, more complex and statistical studies should be performed. The analytical data cited above show that there is no evident dependence between thermal stability of resins and their geological age. For example, the oldest Khatanga resin (RU/Kh/1) has the ‘main’ DTG peak about 439 °C and further samples, respectively, to their age: PL/J/1 (about 40 Ma) − 327 °C, DO/1 (15–0 Ma) − 323 °C and the youngest one CO/1 (2.5 Ma − 200 y) − 400 °C.

Certainly, these results indicate that larger number of specimens should be investigated to confirm whether there is correlation between thermal stability and geological age of the resin or not. Also, statistical analysis should be performed that will require numerous specimens from the same location in order to assess variability in the structure of resins of the same botanical and geographical origin and age. Besides, more complex studies are needed that also involve a chemical control of the resin composition and detailed physical analysis of the materials structure.

Thermal properties—differential scanning calorimetry DSC

Phase transitions under programed temperature changes were investigated by DSC analysis, and in Fig. 2, DSC curves for analyzed resins are presented. The profile of the first heating scan has shown variation of heat capacity of the materials due to chemical changes (softening, melting, post-curing reactions, vaporization).

Fig. 2
figure 2

Normalized DSC curves recorded with the natural resins under investigation

Full size image

The glass transition Tg of natural resins in the first heating scan is usually overlapped with other effects, such as enthalpy relaxation and post-curing reactions. The course of DSC profile with increasing temperature has shown both endothermal and exothermal effects, which correspond to the softening and partial melting of the resin, evolution of volatiles, possible relaxation and post-curing reactions. In Table 3, the evaluated results are presented.

Table 3 Evaluated DSC results for analyzed resins and samples appearance after DSC analysis
Full size table

The analyzed resins during heating liberate heat due to post-curing reactions, which can be clearly observed on DSC curves as an exothermal effect (Fig. 2). The material also may lose some volatiles, and this evaporation is endothermic and compete with the exothermic reaction of curing. Other ‘endo’ events observed are mainly related to the softening and partial melting.

The distinctive DSC profile was observed for Colombian copal (CO/1), which shows the presence of a sharp endothermal signal at 91 °C corresponding probably to the melting of the material. The appearance of the sample after DSC measurement confirms this observation.

On cooling profiles (Fig. S1), a glass transition can be distinguished, that is overlapped on first heating curves by the melting, vaporization and curing effects. This indicates that conventional DSC analysis is not an appropriate tool for pristine resin properties determination, because the glass transition temperature can only be determined from the cooling or second heating run, so after thermal treatment of the resin. Thermal history of the natural resin is removed in the first heating run, and only subsequent thermal behavior might be observed, while the most important are thermal effects related to the complex resin composition and its age.

In previous reports [10], Dominican and Baltic ambers showed glass transition at 133–163 °C and 117 °C, respectively. The authors did not observe the age-Tg dependence; however, another report by Jablonski et al. [5] showed the relation between the Tg value and age of the material. We also observed such relation, and we have found that the older is the resin and the higher is the glass transition value. However, it should be noted that proper comparison of the results is difficult by the fact that different instrument and measurement parameters were applied. In our studies, Dominican amber (DO/1) presents lower Tg than those from the literature. For Baltic amber, different Tg values were reported ranging from 117 °C [10] to 143 °C [27], which are close to our results.

Feist and coworkers [11] described different resin and copal materials from Baltic Sea area, Dominican Republic, Colombia and Madagascar. The differential thermal analysis (DTA) exhibited weak overlapping phenomena, both endothermal and exothermal character, which authors assigned to evaporation processes and decomposition reactions. On the other hand, Bogdasarov [28] observed in DTA profile of Baltic resins similar endo- and exothermal effects during heating to 200 °C and described them as the partial molecular reorganization inside of the material.

Our preliminary results and those from the literature suggest that for better elucidation of the natural resins characteristics, systematic studies by the use of DSC and DTA combined with chemical and more sophisticated methods should be performed.

Structural changes—FT-IR

FT-IR is one of the most useful and popular methods for amber identification and characterization [6, 12, 29,30,31, 34,35,36,37,38,39,40,41,42]. The prominent features which can be find in the natural resins spectra are methyl/methylene stretching (2860/2920 cm−1), carbonyl group stretching in the region of 1690–1740 cm−1, broad signal due to stretching in the hydroxyl group around 3420 cm−1, deformative O–H vibrations around 1640 cm−1 and weak signal of C=C vibrations around 1610 cm−1 [35].

FT-IR spectroscopy allowed to observe several differences in the structure of analyzed materials (Fig. 3), and selected samples were tested according to structural changes after heating in DSC experiment. The resulted spectra are presented in Figs. 3–5, and their respective assignments are summarized in Tables 4 and 5. The infrared spectra show characteristic indicative bands for different regions, i.e., 2700–3100 cm−1, 3100–3700 cm−1 and 730–1890 cm−1.

Fig. 3
figure 3

FT-IR spectra of the resins under investigation at room temperature in two spectral regions of 4000–2500 cm−1 (up) and 2000–400 cm−1 (bottom)

Full size image
Fig. 4
figure 4

Normalized and deconvoluted FT-IR spectra of the analyzed resins in the spectral range of 1900–1500 cm−1 at room temperature: a group I, b group II

Full size image
Fig. 5
figure 5

Normalized FT-IR spectra of the DO/1 a and CO/1 b resins in the spectral range of 1900–700 cm−1 (after different thermal treatments)

Full size image
Table 4 Main bands observed on the FT-IR spectra of the resins under investigation and their assignments based on the literature data
Full size table
Table 5 List of characteristic absorption bands in FT-IR spectra recorded using the ATR [29,30,31,32,33,34]
Full size table

FT-IR results have shown that pristine Baltic amber (PL/1) is composed of ester part (mainly succinite ester) and acidic part (e.g., succinic acid), which may suggest that it was exposed on humid and acidic conditions. Moreover, possible thermal aging occurred which is indicated by the presence of band at 1652 cm−1, corresponding to C=C bond vibrations. The carbonyl band characteristic for pristine amber was found to be split into two bands at 1717 cm−1 (C=O in ketones) and 1733 cm−1 (C=O in esters) [29].

Carbonyl group vibrations are usually the most intense signals on the FT-IR spectra; however, in some cases are also difficult in the precise interpretation because the signals for different carbonyl compounds overlapped each other. The complex composition of the natural resins can be clearly visible on the FT-IR spectra, especially in the carbonyl stretching region. In order to qualify the signals in the carbonyl region, we followed a spectral deconvolution and fitting approach. The most adequate fitting was achieved with a sum of Lorentzians for vibrational signals, in which intensity differs depending on the resin. The signals observed at around 1702 cm−1 and 1695 cm−1, which correspond to the carboxylic acid groups, can be in fact one signal about 1700 cm−1; however, their presence did not affect the fitting quality. Within the experimental and calculation inaccuracy, the results did not show significant differences to those from the literature. The obtained resolved spectra are presented in Fig. 4. According to the previous reports [29, 42], we have found several signals overlapping in the region of 1500–1900 cm−1, which are presented in Table 6. The resolved spectra showed bands characteristic for esters, ketones, carboxylic acids and double bonds vibrations.

Table 6 FT-IR band positions with intensities after normalization of the carbonyl region of the spectra, and their relevant assignments [32, 33]
Full size table

We have observed (Fig. 4) that while the key band positions in the spectra range of 1500–900 cm−1 did not change significantly, their relative intensities did. This indicates that the composition of the resin differs in terms of carbonyl compounds content and double bonds presence in the structure of the material. In turn, one can conclude that is probably caused by the aging process of tested resins.

For the oldest resin, i.e., Khatanga (RU/Kh/1) and Baltic amber (PL/J/1), the intensity of a band at about 1737 cm−1 corresponding to the C=O vibration in ester groups is clearly higher than for other materials. The same was observed for the intensity of a band at 1714 cm−1 of carboxylic groups in succinic acid which is the dominant marker for Baltic amber [43]. Another interesting observation is that the intensity of band related to the double bond vibration at 1642 cm−1 lowers as the resin is older, suggesting that it can be useful for the age assessment purposes. Some of the authors [44] determine the degree of resin maturity based on the ratio of 1650/1440 cm−1 bands in Raman spectroscopy. Recent investigations [26, 34] did not confirm such relation. Correlation of spectral properties of resins with their age and maturity still requires more investigations.

For the materials after heating in DSC measurement (Fig. 5), we did not observe significant structural changes—the positions of the characteristic bands stayed the same, but the intensity of the band related to carbonyl group vibrations became lower as material was thermally treated. One can point out that the older resins, i.e., RU/Kh/1 and PL/J/1, did not show any changes under thermal conditions applied (200 °C), which can be explained by heat resistance of the resins or low FT-IR sensitivity for observation of such subtle effects.

Conclusions

Actually, advanced analytical methods and techniques allow to shed a new light on the complexed natural products, such as natural resins. The present study was devoted to preliminary investigation of selected samples of resins by means of thermal analysis and structural (FT-IR) studies.

Thermogravimetric results showed that there is no direct correlation between the age of the resin and its thermal stability and decomposition pathway. However, we have observed similarities between Baltic amber (PL/J/1) and Dominican amber (DO/1) under heating in air atmosphere. This was expected as they stand close in the classification of the fossil resins—both resins belong to class I of resins containing labdatriene carboxylic acids in their structure. The course of TG profile indicates that Russian resin (RU/Kh/1) and Colombian copal (CO/1) decompose in a more complex way; however, only the Russian material presents slightly higher thermal stability than other resins under investigation.

For investigated resins, the relative dependence between the glass transition Tg and age was found—as the resin is older the glass transition value is higher. Considering that Tg values were evaluated on the cooling scan, this observation should be precisely confirmed by more advanced techniques. One of them is modulated-temperature DSC, which allows to separate complex heat flow signal and thus determines the extent of overlapping thermal transitions in the raw material.

FT-IR investigations have shown significant differences in the structure between tested materials. Deconvolution of the FT-IR spectra allowed to extract characteristic bands in the region of carbonyl group vibrations of 1500–1900 cm−1, which in turn brought potential information about the amount of particular carbonyl compounds in the resin. It was also observed that the intensity of the band related to the double bond vibrations gets lower as the resin is older, which can be a significant observation in terms of resin age assessment.

In general conclusion, our results showed that systematic thermal studies followed by structural FT-IR studies are recommended for the natural resins to help in classification of this material according to the maturation grade. At the same time, further research is needed on statistically larger number of samples from the same deposit, as well as specimens with different geographical locations, but similar age or botanical origin. We plan to follow this subject in our future papers, where more detailed studies will be presented.