Introduction

Small variations in polymer and filler content can extend the performance of elastomers. Quantitative analysis of blend composition is therefore of a vital matter to the elastomer industry. Varieties of analytical tools have been developed to monitor rubber blend composition. Fourier transformation infrared (FTIR) spectroscopy, thermogravimetrical analysis (TG), pyrolysis gas chromatography–mass spectroscopy (Py-GC/MS), differential scanning calorimetry (DSC) and attenuated total reflectance (ATR) are often used to monitor blend composition [1–21]. These techniques have proven their efficiency; however, many of studies have been analyzed using blend of ratios superior to 10 % [22–30].

Thermogravimetrical analysis (TG) is often used to characterize the composition of blends. In TG, the mass change of a sample is measured while it is heated at constant heating rate or maintained isothermally in either an inert (nitrogen, argon) or an oxidative (air, oxygen) atmosphere.

For a better understanding of the nature of decomposition gases evolved from the TG measurement, Fourier transformation infrared spectroscopy (FTIR) and mass spectrometry (MS) are two common methods often used [31–34]. Simultaneous information about degradation products is obtained during the sample mass change. TG–FTIR and TG–MS are used for identification, qualitative and quantitative analysis of evolved gases as well as for the comprehension and elucidation of the decomposition process. TG–FTIR and TG–MS are very effective techniques only if few, known volatile compounds are evolved or to understand some decomposition routes [35]. These techniques allow the emission profile of compounds among the thermal gravimetric decomposition to be drawn. However, especially with polymers, many decomposition products evolve simultaneously. Consequently, in the best case only the main products can be identified and these can overlap compounds with low concentration which cannot be detected.

A chromatographic separation is therefore necessary to identify the simultaneously evolved products. A well-established method is pyrolysis gas chromatography mass spectrometry (Py-GC/MS) [36]. This technique consists of heating the sample very quickly from room temperature to high temperature in a tube or directly on a filament. The decomposition products can then be trapped by cooling at the beginning of the GC column or directly introduced into the GC capillary column. In most pyrolysis systems, only helium or air can be used as carrier gas. Py-GC/MS allows the detection, identification and even quantification of all thermal decomposition products and is a very sensitive analysis [37–39]. Nevertheless, Py-GC/MS does not allow the monitoring emission of decomposition products as a function of temperature without considerable time and efforts. High and time-consuming maintenance due to blocked capillaries or polluted injection systems by high boiling products is the main disadvantage of this technique [40].

Efforts have been made to couple a TG to a GC or GC/MS system [41–44]. Currently, commercial TG–GC/MS coupling systems are limited regarding the number of GC/MS measurements that can be done during thermogravimetric analysis.

Here a novel approach in which the TG is coupled to the GC/MS by means of a heated storage interface (IST) generally at 250 °C is presented.

The decomposition products from the TG decomposition are transferred to the IST storage oven via a heated transfer line. When the IST is in this “storage” mode, the gases coming from the TG are released through the outlet of the interface. At a predefined time, which corresponds to a defined temperature in the TG, fractions of the volatile decomposition products are stored in a heated loop. Up to 16 loops containing 250 µL of gases can be used. These storage times are selected by the user and are determined during a first TG experiment using the same method.

Once the last loop has been collected, the interface switches to the “inject” mode. The decomposition gases stored in the first loop are injected into the GC via a second heated transfer line by the GC carrier gas. The individual gases are first separated by the GC and then analyzed by the MS detector. Spectral libraries are generally used for their identification. At the end of the first GC/MS measurement, the second loop is then injected into the GC and so on. With this setup, volatile products with masses up to m/z = 250 can be detected.

The emission profile of selected evolved compounds can be followed using the 16 loops as a function of the thermogravimetric decomposition temperature. The storage interface, unlike other evolved analysis method, may be used with any gas allowed in the TG. Decomposition in different matrices may lead to different evolved compounds and therefore different evolved profiles [45].

This study illustrates the feasibility of using TG–IST–GC/MS to estimate the amount of styrene–butadiene (SBR) in natural rubber (NR)/SBR copolymer blends containing very low amounts of SBR.

Experimental

Materials

The rubber samples were supplied by the Malaysian Rubber Board (MRB), 18th Floor Bangunan Getah Asli (Menara), 148 Jalan Ampang, 50450 Kuala Lumpur, Malaysia.

The rubbers were manufactured in the MRB research and development pilot production center. The samples consisted of unfilled natural rubber (NR) and styrene–butadiene (SBR) single elastomer and three different copolymer blends of NR containing low amounts of SBR. The SBR contents were 2.5, 5.5 and 7.5 %, respectively. NR content was 97, 94 and 92 %, respectively. The other constituents consisted of stearic acid (0.5 %), sulfur, accelerator and low volatiles such as oils.

Experimental method

TG

The TG/DSC 3+ from Mettler Toledo was used. In a first experiment, a 20 mg sample was heated in the TG from 25 to 600 °C at 10 K min−1 in a 70-µL alumina crucible. Nitrogen was used as balance purge and sample purge gas at 10 and 30 mL min−1 flow rate, respectively. This first experiment allows the determination of the storage times (corresponding to a certain TG temperature) for the TG–GC/MS analysis to follow.

TG–GC/MS

In a second experiment, the TG was coupled to an Agilent gas chromatography 7890 GC and an Agilent mass selective detector 5975C MSD using the IST16 interface from SRA Instruments (Fig. 1). The amount of sample and the TG method used in this second experiment were the same as for the first.

Fig. 1
figure 1

TG–IST–GC/MS system

Full size image

The IST transfer line and oven temperatures were set at 250 °C. The GC split/splitless injector temperature was set to 280 °C. The GC oven temperature program consisted of an isothermal step at 50 °C for 5 min followed by a heating to 300 °C at 10 °C/min and another isothermal at 300 °C for 5 min. The column was an HP-5 ms of 60 m × 0.32 mm × 0.25 µm. Helium was used as carrier gas in the pressure control mode to ensure a column flow of 1.0 mL min−1. At the GC injector, helium was split 5:1. The MS was operated in scan mode from 33 m/z to 350 m/z. Ionization was effectuated by electron impact (EI) at 70 eV, and an electron multiplier voltage (EMV) of 1 was used. The temperature of the ion source was set at 230 °C and the quadrupole at 150 °C.

The compounds in the GC/MS total ion chromatogram (TIC) were identified using NIST/EPA/NIH Mass Spectral Library 2011 [46].

Results and discussions

TG

The TG and the corresponding DTG curves are presented in Fig. 2. For NR and NR/SBR blend, the first decomposition step up to about 350 °C is caused by the elimination of moisture and volatiles (2 %). Then, the pyrolysis of the elastomer takes place with a typical decomposition profile for NR (97 %). For SBR, the volatization step is about 5 % and the decomposition of the polymer is about 94 %.

Fig. 2
figure 2

TG curves of NR and SBR single elastomers and NR/SBR blend at different ratios

Full size image

NR and SBR as single elastomers clearly show two different decomposition profiles. NR and NR/SBR show similar decomposition profiles. Only a shift in the TG curve with increased concentration of SBR in the blend at the end of the decomposition is observed (indicated by the arrow in Fig. 2).

TG–IST–GC/MS

Based on TG decomposition profiles shown in Fig. 3, storages of the decomposition products for all samples were set according to the TG temperatures as shown in Table 1.

Fig. 3
figure 3

a TIC of loop 10 (370 °C) of 100 % NR; b TIC of loop 12 (400 °C) of 100 % SBR

Full size image
Table 1 Storage of evolved products according to the TG temperatures
Full size table

After collection of the last loop (number 16), each loop was injected individually. The evolved gas compounds were separated by the GC column and then identified by the MS using the NIST/EPA/NIH Mass Spectral Library 2011.

Identification of evolved compounds in NR and SBR

Figure 3 displays the total ion chromatogram (TIC) of loop 10 of the 100 % NR sample in (a) which corresponds to a TG temperature of 370 °C and in (b) the TIC of loop 12 of the 100 % SBR sample which corresponds to a TG temperature of 400 °C. The main five compounds identified using the NIST/EPA/NIH Mass Spectral Library 2011 are summarized in Table 1 together with the retention times (RT) in minutes, the chemical formula, the structure and the main m/z. All of these five compounds had a quality factor above 90 % (Table 2).

Table 2 Selection of some evolved compounds
Full size table

In the 100 % NR, limonene (dimers) is the main evolved decomposition product. Other main evolved compounds are isoprene and m-xylene. Evaluation of the GC/MS analysis results of the 100 % SBR showed that these compounds are not evolved during its decomposition. The main component evolved is toluene. Toluene is not specific to SBR as it is also evolved during NR decomposition. The main evolved decomposition product specific to SBR is styrene.

Identification of evolved compounds in NR/SBR blend

In Fig. 4, the TIC of loop 12 of the NR/SBR containing 2.5 % SBR is shown. The main evolved compounds of NR such as limonene are clearly detected. Styrene which comes from SBR is also clearly identified (Fig. 5). The upper part of Fig. 5 displays the mass spectrum of the corresponding 10.7-min retention time peak of the TIC. The base peak is m/z 104, and the main fragment peaks are m/z 103, m/z 78, m/z 77 and m/z 51. This mass spectrum is identified using the spectral library NIST/EPA/NIH Mass Spectral Library 2011. The best fit shown in the lower part of Fig. 5 is styrene (unique identifier #4830) with a match quality of 94.

Fig. 4
figure 4

TIC of loop 12 (400 °C) of NR/2.5 % SBR blend

Full size image
Fig. 5
figure 5

Identification of styrene using the NIST/EPA/NIH Mass Spectral Library 2011

Full size image

Emission profile

Using the 15 storage loops, decomposition profiles of selected products are determined as a function of temperature.

The response factor Rf i,l of compound i in each individual loop l is calculated by the area of the main m/z peak A normalized by the initial sample mass, m s:

$$Rf_{\text{i,l}} = \frac{A}{{m_{\text{s}} }}$$
(1)

The ratios k i,l are calculated by dividing the individual response factors Rf i,l by the sum of the response factors ∑ Rf i,l:

$$k_{\text{i,l}} = \frac{{Rf_{\text{i,l}} }}{{\sum Rf_{\text{i,l}} }}$$
(2)

The individual ratios k i,l of styrene in the 100 %SBR and NR/SBR blends for all loops are summarized in Table 3.

Table 3 Ratio k l of styrene in the NR/SBR blends for all loops
Full size table

Profiles of styrene (for NR/2.5 % SBR blend and 100 % SBR) and limonene (for NR/2.5 % SBR blend) using k i,l and the TG curves of NR/2.5 % SBR blend and 100 % SBR as a function of temperature are shown in Fig. 6.

Fig. 6
figure 6

TG curves of NR/2.5 % SBR blend and 100 % SBR and emission profile of styrene (for NR/2.5 % SBR blend and 100 % SBR) and limonene (for NR/2.5 % SBR blend) versus temperature

Full size image

From the emission profile, limonene (NR-specific decomposition product) evolved at its maximum at earlier temperature than styrene (SBR-specific decomposition product).

Styrene profiles from the 100 % SBR sample and the NR/2.5 % SBR have the same characteristics, i.e., Gaussian with identical temperatures and the same maximum emission profile. These characteristics are essential for the quantification of SBR in the NR/SBR samples.

Quantification of SBR

Styrene which is a specific decomposition product of SBR in the NR/SBR samples is used to quantify the amount of SBR in the blends.

The response factors of styrene, \(Rf_{\text{i,l,std}}\) for the 100 % SBR are used as a standard.

Using all loops in which styrene is detected, the weighted average amount of SBR, ξ and its weighted standard deviation σ are calculated using k i,l as weighted factor:

$$\xi = \frac{{\sum \left( {\frac{{Rf_{\text{i,l}} }}{{Rf_{\text{i,l,std}} }} \cdot k_{\text{i,l}} } \right)}}{{\sum k_{\text{i,l}} }}$$
(3)
$$\sigma = \sqrt {\frac{{k_{\text{i,l}} \cdot \left( {\frac{{Rf_{\text{i,l}} }}{{Rf_{\text{i,l,std}} }} - \xi } \right)^{2} }}{{\frac{{\left( {N - 1} \right)\sum k_{\text{i,l}} }}{N}}}}$$
(4)

with the N the number of nonzero masses.

Table 4 summarizes the SBR content results of the duplicate experiments in all NR/SBR blend samples. The results show that the SBR content agrees with the formulation within the experimental errors.

Table 4 SBR content results for the respective formulation
Full size table

Conclusions

A novel system to couple a TG to GC/MS by the heated storage interface (IST) offers the possibilities to store fractions of the volatile decomposition products in loops at predefined TG temperature. The stored gases are injected individually to the GC after collection. The gases are first separated by the GC and analyzed by the MS detector. It allows the emission profile of selected evolved compounds as a function of the thermogravimetric decomposition temperature. Here the amounts of very low-content styrene–butadiene (SBR) in natural rubber (NR)/styrene–butadiene (SBR) blends were determined. Investigations by this laboratory by DSC, DMA, TG–MS, TG–FTIR, TG–MS and ATR did not successfully reveal the presence of SBR. Thanks to the chromatographic separation and storage at different temperatures, SBR was detected and its content in individual blend was determined. The results show good agreement with the elastomer formulation. Even the blend containing the lowest amount of SBR content of 2.5 % is clearly quantified.