Introduction

Synthetic rubber materials are widely used as sealing materials for various fluids. Rubber materials are designed in terms of the properties of target fluids for seals. For example, in the case of seals for gasoline, polar rubber should be used. In the case of gas seals, the transport of gas through sealing materials should be taken into account. From the perspective of the molecular design of rubbers as sealing materials, the clarification of the relationship between the structure of the rubbers and gas permeation phenomena, the effect of gas on the structure of the rubbers and the physical properties of the rubbers have become increasingly important. The molecular mobility of rubber molecules is one of the most important parameters to consider in designing suitable materials for rubber seals. There are several reports on the molecular mobility of rubber composites evaluated from the spin–spin relaxation time (T2) of protons (1H) in rubber molecules determined by free induction decay (FID) signals from pulsed proton nuclear magnetic resonance (pulsed 1H-NMR) measurements.1, 2, 3, 4 Several methods for the evaluation of T2 with different pulse sequences are available. The solid-echo pulse sequence method can evaluate relatively short T2 5. Solid-echo pulse 1H-NMR is widely used to analyze the crystallinity of polymer materials6. Information regarding the inhomogeneous structure of the bulk state of polymers can also be obtained using solid-echo pulse 1H-NMR as the difference in T2 for the components of the inhomogeneous state. Solid-echo pulse 1H-NMR analysis can be used to understand the effect of pulsing on the permeation of fluids through sealing rubber materials. The time course of T2 can be measured because of the short data acquisition time for FID signals. We have already reported on the solid-echo pulse 1H-NMR analysis of the time course of T2 for vulcanized acrylonitrile (AN)–butadiene rubber (NBR) composites, which are typical rubbers used for hydrogen gas sealing, after high-pressure hydrogen exposure to understand the effect of hydrogen on molecular mobility. We identified more than two components with different T2 values in crude NBR using solid-echo pulse 1H-NMR FID signals, and those T2 values and the 1H ratio of each T2 value were determined to depend on the AN content of NBR. To understand the transport phenomena of hydrogen through sealing rubber materials according to the mobility of the molecular chains, each component of vulcanized NBR with a distinct T2 value should be assigned by taking into account the inhomogeneous structure of crude rubber. Fukumori et al.7 reported the solid-echo pulse 1H-NMR analysis of vulcanized NBR. They found two components with different T2 values, a short T2 and a long T2, which could be assigned to the rubber chains around crosslinks and the chains between crosslinks, respectively. However, there are no reports on the concrete assignment of these signals, even for crude rubber materials. NBR is generally polymerized by emulsion polymerization as a random copolymer. Although almost all random copolymers show homogeneous morphology, it has been reported that some random copolymers show an inhomogeneous phase structure.8, 9 In the case of NBRs, the existence of phase-separated structures was suggested by the spin–lattice relaxation time (T1ρ) analysis of CH2 via the 13C-NMR inversion recovery method.10 According to these reports, the inhomogeneity was caused by an intramolecular inhomogeneous structure composed of segments with different monomer sequences and by large differences in the segmental interaction parameters. However, there are no reports on the relationship between the inhomogeneity in the molecular mobility and in the microstructure of copolymer chains. The inhomogeneity of crude NBRs can be confirmed by small-angle X-ray scattering (SAXS), differential scanning calorimetry (DSC) and dynamic mechanical analysis (DMA). In this study, to identify definitively a wide range of T2 components from solid-echo pulse 1H-NMR FID signals of crude NBRs, we tried to understand the inhomogeneity in crude NBRs in terms of their microstructures and molecular mobilities.

Experimental procedure

Crude NBRs with five different AN contents, low (L), middle (M), middle high (MH), high (H) and very high (VH), were obtained from the Zeon Corporation (Tokyo, Japan): Nipol DN401 (L, AN: 18%), Nipol 1043 (M, AN: 29%), Nipol 1042 (MH, AN: 33.5%), Nipol 1041 (H, AN: 40.5%) and Nipol DN003(VH, AN: 50%), respectively. The details of the samples are shown in Table 1. The AN contents described above were specified by the supplier. Me, the molecular weight between entanglements, was estimated from the storage elastic modulus at 30 °C analyzed by DMA. All measurement samples were used without any further purification.

Table 1 Sample information
Full size table

Solution 1H-NMR (ECP-400 400 MHz; JEOL Ltd., Akishima, Japan) measurements of the crude NBRs were performed in chloroform-d. The measurement of the crude NBRs was conducted with a single-pulse sequence (pulse width: 23 ms; pulse power: 34 W; acquisition time: 6.87 ms; observation center: 5.0 p.p.m.; sweep width: 20 p.p.m.; scan repetition: 64 times) using a φ5-mm solution probe.

DSC (DSC 204 HP; Netzsch, Selb, Germany) was performed from −100 to 150 °C in air. The heating rate was 10 °C min−1.

DMA (itk DVA-225; IT Keisokuseigyo, Osaka, Japan) was performed from −100 to 150 °C. The shear stress frequency was 2.5 Hz. Me was calculated by using equation (1):

where ρ is the sample density, R is the universal gas constant, T is temperature and G′ is the storage elastic modulus. G′ values at 30 °C were used; thus, T was 303.15.

SAXS (Nanostar; Bruker AXS, Karlsruhe, Germany) was performed on crude NBR films measuring 5 mm in thickness. The wavelength of the Cu Kα X-ray beam was 0.154 nm, and the camera length (sample detector) was 1053 mm. Scattering patterns were collected with a two-dimensional detector (Vantec-Hi Star; Bruker AXS).

The solid-echo (90°x−τ−90°y) FID signals of the samples were analyzed by using pulsed 1H-NMR (minispec mq20, 20 MHz; Bruker BioSpin, Karlsruhe, Germany) at 30 °C. Samples (φ10 mm × 10 mm) were placed in φ10 mm Pyrex tubes. The T2 values and 1H ratios of the T2 values were calculated from the obtained FID signals by the converged equation (2) using the least-squares method,

where Ai is the intensity at time 0, t is time, T2i is the T2 value of ith component and αi is the FID signal shape constant. The 1H ratios of T2i were evaluated using Mi, as shown in equation (3):

Results and discussion

Monomer sequence distribution of crude NBRs

To determine the monomer sequence distribution of the crude NBRs, the solution 1H-NMR spectra of these samples were measured. Figure 1 shows the spectrum of MH. The protons of methine (CH) at the center of the triad were observed in the regions of 2.5–2.9 and 5.2–5.7 p.p.m. The signals of CH were divided into five regions: 5.5–5.67, 2.85, 2.58, 2.85 and 5.2–5.43 p.p.m., which were assigned to the triads of NBR, ABA, BAA, BAB, BBA and BBB, respectively,11 where A is an AN unit and B is a butadiene (BU) unit. The CH of AAA, which were observed at 3.1 p.p.m.12 in polyacrylonitrile, were not detected in the spectra of the crude NBRs. The average contents of AN, BU and triad distributions were calculated from the area of these signals.

Figure 1
figure 1

Proton nuclear magnetic resonance (1H-NMR) spectrum of crude acrylonitrile–butadiene rubber (NBR) (MH) in chloroform-d.

Full size image

The average chemical composition and triad distribution of the crude NBRs determined by solution 1H-NMR are shown in Table 2.

Table 2 Average chemical composition and triad distributions of crude NBR
Full size table

The contents of ABA and BAB increased with the AN content. The contents of BBA and BBB clearly decreased with increasing AN content. These results suggest that in crude NBRs, AN units prefer alternate or random copolymerization with BU units; however, BU units prefer the formation of BU blocks.

The average sequence lengths of AN (Ln[AN]) and BU (Ln[BU]) in each crude NBR were calculated using equations (4) and (5),13

where FAAA, FBAB, FBAA, FBBB, FBBA and FABA are the molar fractions of the triads. The relationship between the AN content and Ln[AN], Ln[BU] is illustrated in Figure 2. As shown in Figure 2, Ln[BU] decreased proportionally with the increase in the AN content. In contrast, Ln[AN] remained constant. The relationship between the mole fraction of BU units in the polymer (FBU) and the mole fraction of feedstock AN monomer (fAN) was calculated using equation (6),14

Figure 2
figure 2

Relationship between acrylonitrile (AN) content and average sequence length of AN (Ln[AN]) and butadiene (BU) (Ln[BU]). A full color version of this figure is available at Polymer Journal online.

Full size image

where fAN and fBU are the mole fractions of feedstock AN and BU monomers, respectively, and rAN(=0.03) and rBU(=0.2) are the AN and BU monomer reactivity ratios, respectively.15 The calculated results are shown in Figure 3. According to this figure, in the region of fAN=0.2–0.8, fBU was nearly 0.5. During the early stages of copolymerization, an AN–BU alternate copolymer was produced, and during the later stages, BU blocks were produced by the remaining BU monomer.

Figure 3
figure 3

Relationship between molar fraction of butadiene (BU) in polymer chains and molar fraction of feedstock acrylonitrile (AN) monomer.

Full size image

According to these results, crude NBR chains consist of two components, a BU block sequence and an alternately copolymerized AN–BU sequence, at low and high AN contents components, respectively.

Morphological analysis of crude NBRs by SAXS

We tried to estimate the size of aggregated AN–BU chains by using SAXS, and the obtained scattering profile was analyzed by Fankuchen’s method.16 For monodisperse aggregates with a radius of gyration R, the SAXS intensity can be described by Guinier’s method using equation (7),

where I(q) is the SAXS intensity, I(0) is the intensity of q=0, q is the scattering vector and Rg is the radius of gyration. However, for polydisperse aggregates with radii of gyration Rg, the distribution of Rg can be calculated by Fankuchen’s method using equation (8):

Equation (8) is a summation of equation (7). Thus, equation (8) can be described by a double logarithm as follows:

Thus, according to equation (7), Rgi can be calculated from the slope of the plot of ln(I(q)) vs q2, the so-called Guinier plot. The relationship between the slope (Si) and Rgi and the relationship between Rgi and the particle radius (Ri) are described by equations (10) and (11), respectively. The mass ratio of components with Rgi can be calculated by equation (12):17

Guinier plots of the crude NBRs are shown in Figure 4, and sample slopes of the Guinier plots are shown in Figure 5. The estimated values of Ri and W(Ri) are shown in Table 3. Three slopes were obtained from the Guinier plots. As shown in Table 3, the obtained Ri were 0.13–0.37, 3.62–5.21 and 11.7–14.6 nm. The two larger domains exhibited quite small mass ratios. According to the results of SAXS, the crude NBRs were regarded as having a single-domain morphology, that is, nanodomain aggregates with Ri =0.13–0.37 nm. In other words, the crude NBRs exhibited a homogeneous morphology on the nanoscale.

Figure 4
figure 4

Guinier plot and MH slope. A full color version of this figure is available at Polymer Journal online.

Full size image
Figure 5
figure 5

Guinier plot of crude acrylonitrile–butadiene rubbers (NBRs). A full color version of this figure is available at Polymer Journal online.

Full size image
Table 3 Radius of domain (Ri) and weight fraction of domain with Ri of crude NBRs
Full size table

Thermal analysis for glass transition temperature

From the results of morphological analysis, it was concluded that the crude NBRs were single phase. To confirm the uniformity of the materials, DSC and DMA measurements were performed on crude NBR samples to evaluate their glass transition temperatures (Tg).

Figure 6 shows the DSC curves of the crude NBRs. The curves show two endothermic peaks for all samples. The Tg of each sample was determined as the downward peak of the first-order derivative of the corresponding DSC curve. The main endothermic peak temperatures below 0 °C were designated as the Tg values of the crude NBRs.

Figure 6
figure 6

Differential scanning calorimetry (DSC) curves of crude acrylonitrile–butadiene rubbers (NBRs). A full color version of this figure is available at Polymer Journal online.

Full size image

Chandler et al.18 reported that the results of differential thermal analyses for crude NBRs with <36% AN showed two Tg values below 0 °C because of the existence of a phase-separated structure in the samples. 18 In the case of emulsion-polymerized NBR, the monomer sequence distribution was dominated by the polymerization temperature and conversion rate.19 However, our DSC results for crude NBRs show only one Tg.

Figure 7 shows the temperature dependence of tan δ measured by DMA for crude NBR samples. Tg was determined as the peak temperature of tan δ. The tan δ values showed a single peak in all samples. These peak temperatures were designated the Tg values. Tg increased as the AN content increased. Thus, only one Tg was observed in not only the DSC profiles but also the DMA profiles.

Figure 7
figure 7

Tan δ of crude acrylonitrile–butadiene rubbers (NBRs). A full color version of this figure is available at Polymer Journal online.

Full size image

Judging from our SAXS, DSC and DMA results, the crude NBRs used in this study showed a single-phase and homogeneous morphology.

Pulsed 1H-NMR analysis of crude NBRs

The observed FID signals and fitted FIDs for crude NBRs at 30 °C are shown in Figure 8. The FID signals obtained were fitted by equation (1) using the least-squares method. Ai, N, αi and T2i in equation (1) were treated as fitting parameters. As indicated by the fitting results, the number of components N for each crude NBR was 3. The residual sums of the squares of the fitting results were below 10−4. The FID signals of all of the samples could be separated into three T2 components denoted T2s, T2m and T2l. The relationship between AN content and T2 analyzed by pulsed 1H-NMR is illustrated in Figure 9. T2s, T2m and T2l decreased proportionally with increasing AN content. Figure 10 shows the amount of each component as the 1H ratio. As shown in Figure 10, the 1H ratio of the component with T2s increased and that with T2m decreased as the AN content increased. The 1H ratio with T2l was quite small and remained constant. These results indicate that there were two components with different molecular mobilities in the crude NBRs. The mobility of each component depends on the AN content of the component. The number of polar monomer units increases with AN content. The increase in the number of polar monomer units causes a decrease in molecular mobility. Kwak and Kim10 reported the observation of local phase separation into sea–island structures in crude NBRs from the analysis of the spin–lattice relaxation time T1ρ of CH2. Faghihi et al.8 reported that random copolymers of butyl acrylate and methyl methacrylate showed nanophase segregation according to DMA, DSC and 13C-NMR spin diffusion experiments. Kipper et al.9 reported that random copolymers of 1,6-bis(p-carboxyphenoxy)hexane and sebacic acid undergo microphase separation because of large differences in the segment–segment interaction parameters. In this study, the crude NBRs did not show a phase separation structure. Two components with different molecular mobilities were observed in the crude NBRs.

Figure 8
figure 8

Free induction decay (FID) signals of pulsed proton nuclear magnetic resonance (1H-NMR) spectra with fitting results for crude (NBRs): (a) L, (b) M, (c) MH, (d) H and (e) VH. A full color version of this figure is available at Polymer Journal online.

Full size image
Figure 9
figure 9

The relationship between acrylonitrile (AN) content and T2. A full color version of this figure is available at Polymer Journal online.

Full size image
Figure 10
figure 10

The relationship between acrylonitrile (AN) content and 1H ratio with T2. A full color version of this figure is available at Polymer Journal online.

Full size image

Assignment of components with T2s and T2m of crude NBRs

According to the results of solution 1H-NMR, crude NBR chains consist of two components, a BU block sequence and an alternately copolymerized AN–BU sequence, at low and high AN contents, respectively. According to the results of SAXS, the crude NBRs showed a single-domain morphology, that is, the aggregation of nanodomains. The DSC curves and tan δ peaks of the crude NBRs show one Tg. The crude NBRs were determined to exhibit homogeneous morphology on the nanoscale. The spin–spin relaxation time, T2, determined by pulsed 1H-NMR for crude NBRs showed mainly two components with different molecular mobilities, even in homogeneous samples. The 1H ratio of T2s increased and that of T2m decreased with increasing AN content. As shown in Table 1, the molecular weights between entanglements, Me, for all samples were all close to 2.0 kg mol−1, despite the differences in the AN content among the samples. Judging from the dependence of the 1H ratio on the AN content, T2s and T2m were assigned to the alternately copolymerized AN–BU sequence, which is strongly constrained by intermolecular interactions between polar nitrile groups and BU block sequences, which are weakly constrained.

Conclusions

To understand the transport phenomena of hydrogen through sealing rubber materials by examining the mobility of their molecular chains, each component with a distinct T2 in vulcanized NBR should be assigned by taking into account the inhomogeneous structure of crude rubber. In this study, to identify definitively a wide range of T2 components from the solid-echo pulsed 1H-NMR FID signals of crude NBRs, we tried to understand the inhomogeneity of the crude NBRs in terms of their monomer sequence distribution and molecular mobility. The crude NBRs show homogeneous morphology on the nanoscale; however, the spin–spin relaxation time, T2, determined by pulsed 1H-NMR for the crude NBRs showed three components with different molecular mobilities, even in homogeneous samples. The content of the highest-mobility component with T2l was negligibly small. Judging from the dependence of the 1H ratio on the AN content in these components, the low-mobility component with T2s and the high-mobility component with T2m were assigned to the alternately copolymerized AN–BU sequence and BU block sequence, respectively. The inhomogeneity of crude NBR may be attributed to its monomer sequence distribution.