Introduction

Ion beams have been used for structural and chemical characterization of materials for decades. Megaelectronvolt ion beams provide access to compositional and structural information from the surface towards the depth of solid targets via the backscattering of the projectile (He+) after collision with the atoms of the sample (Rutherford backscattering spectroscopy, RBS) [1]. The atomic composition of the sample is deduced from the energy lost by the backscattered ion after elastic collision in the solid and the structural information is obtained via the blocking and channelling effects occurring in crystalline materials. In contrast, the penetration depth of kiloelectronvolt ions is limited to the surface region. The backscattering of keV monoatomic ions of noble gas, at the origin of low-energy ion scattering spectroscopy (LEIS, also known as ISS), provides structural and elemental information about the extreme surface of the sample (top atomic layer) because low energy ions have a very high neutralization probability when they penetrate the solid [2]. Another technique using keV ions is secondary ion mass spectrometry (SIMS). As the name suggests, SIMS collects the atomic, fragment, and molecular secondary ions emitted from the surface as a result of the energy deposited by a primary ion in the solid, in order to form a mass spectrum of the interrogated surface. With the use of very well-focused beams, state-of-the-art SIMS instruments provide chemical images with submicronic resolution and a depth resolution of the order of the nanometer [3]. Recently, large gas cluster ion beams (GCIB) have attracted a strong interest because they can desorb surface molecules from organic materials with minimal damage to the sample, provided that their kinetic energy per atom (E/n) is sufficiently low [4]. GCIB have opened the door to less fragmented mass spectra such as in matrix-assisted laser desorption ionization (MALDI-MS) or electrospray ionization (ESI-MS) and to depth profiling with retention of the molecular information. In turn, GCIB using Ar500-5000 + clusters have now become the reference for the chemical and molecular analysis of organic coatings and films (organic electronics multilayers [5], plasma-treated surfaces [6], cells and tissues [3, 7], etc.)

In a recent article, Mochiji et al. have studied the distributions of small Arn + clusters appearing in the secondary ion mass spectra of metallic surfaces as a result of the backscattering of the projectile fragments [8]. They demonstrated that the intensity ratio of Ar2 +/ΣArn + (2 ≤ n ≤ 7) could be correlated to the impulsive stress caused by the impact of the cluster at the metal surface, which depends only on the projectile velocity, the Young moduli of the cluster and of the metal, and their densities (assuming an initially elastic interaction). Therefore, their analysis gives access to the local Young’s modulus of an unknown surface, since the surface nature can be identified with the mass spectrum.

In this work, we show that large gas cluster fragmentation is also sensitive to the physical state of polymer surfaces. In particular, the intensity of backscattered Ar cluster fragment ions displays a pronounced and reproducible variation at the surface transition temperature (TT), in relationship with the bulk Tg of the investigated polymers.

Experimental

Samples

Amorphous polymers were supplied by Sigma-Aldrich Inc (Overijse, Belgium). The investigated polymers were high molecular weight polydisperse polyisobutylene (PIB), polystyrene (homodisperse; Mw = 4000 Da, PS 4K), and polymethyl methacrylate (homodisperse; Mw = 2000 Da, PMMA 2 K and 150,000 Da, PMMA 150 K). Solutions of each polymer in toluene (Sigma-Aldrich, >99.7% purity) were prepared and spin-coated onto clean Si wafers of 1 × 1 cm2 at 5000 rpm with acceleration 20,000 rpm/s for 60 s. The wafers were sonicated in isopropanol prior to the coating and then dried under N2 flux. The solution concentrations (from 20 to 35 g/l) were chosen in order to obtain ≈100 nm thick films. Thicknesses were evaluated via calibration curves previously established for these polymers [9]. Low density polyethylene (LDPE) additive-free film was supplied by Goodfellow (Cambridge, United Kingdom).

DSC Measurements

Differential scanning calorimetry (DSC) measurements were performed with a Mettler Toledo DSC821 for all PS and PMMA samples and with a Mettler Toledo DSC822 for the PIB sample. Scans consisted to heat, cool down, and heat again from 30 °C up to 150 °C at a rate of 10 °C/min, for PS and PMMA; and from –120 °C up to 100 °C at a rate of 10 °C/min, for PIB.

ToF-SIMS Instrumentation

ToF-SIMS spectra were collected using a TOF.SIMS 5 (ION-TOF GmbH, Münster, Germany) time-of-flight secondary ion mass spectrometer. The instrument is equipped with an Ar gas cluster ion beam (Ar-GCIB) mounted at 45° with respect to the sample surface. It was used both as sputtering and analytical source. The time-of-flight mass analyzer is perpendicular to the sample surface. The Ar-GCIB ion source was operated at 10 keV and the cluster distribution was centered on Ar3000 +. For the spectral analyses, an AC target current of 0.037 pA with a bunched pulse width around 70 ns was used. A raster of 128 × 128 data points over an area of 500 × 500 μm2 was used. The total primary ion beam dose per unit area for each analyzed region was kept below 8 × 1011 ions*cm-2, ensuring static conditions. Only positive secondary ion species were analyzed. Mass resolution m/Δm ≈ 280 at 67 m/z (corresponding to C5H7 +) was maintained for spectra acquisition. Charge compensation was ensured using a very low energy electron flood gun (Ek = 5 eV). All the data analyses were carried out using the software supplied by the instrument manufacturer, SurfaceLab (ver. 6.5). Temperature was controlled with a special sample holder called “Holder G” (ION-TOF GmbH, Münster, Germany). This holder allows maintaining any temperature from –150 °C to +600 °C through a combination of simultaneous heating and cooling, with an accuracy of +/– 1 °C. The temperature was stabilized for 20 min before each analysis. Prior to analysis, a pre-sputtering of 2 × 1013 Ar3000 +/cm2 over an area of 1000 × 1000 μm2 was also conducted with the Ar-GCIB in DC mode. This mandatory procedure was done in order to eliminate all the surface contaminations.

Results and Discussion

Figure 1 shows the positive secondary ion mass spectra measured on a PS 4K surface at +30 °C (a) and +110 °C (b). These temperatures have been chosen in order to be well in the glassy state and the rubbery states of PS 4K, respectively (bulk Tg ≈ +83 °C).

Figure 1
figure 1

Positive ToF-SIMS spectra of PS 4K measured at (a) +30 °C, and (b) +110 °C

Full size image

In the case of PS 4K, argon clusters up to Ar7 + (m/z 280), are detected. Backscattered Ar+ is only detected below TT, but with a low intensity. In addition, because of the possible interference with 40Ca+ at the same nominal mass (due to poor mass resolution), the interpretation of this peak is subject to caution.

Figure 2a displays the behavior of the backscattered Arn + peaks (1 ≤ n ≤ 5) for PS 4K as a function of the temperature. The intensities are normalized first by the total counts of the related spectrum, then by the maximum of the curve of the considered cluster. A clear transition is observed for the Ar+ curve. The inflection point of this curve is lower than the bulk Tg of PS. The Ar2 + curve presents a transition too, but less pronounced than the one of Ar+. Unlike Ar+, the intensity of the Ar2 + peak stabilizes after the transition and never levels off. Furthermore, the Ar2 + peak intensity is always higher and there is no interference with other contributions at the same mass, in contrast with Ar+. The other Arn + backscattered clusters also show a transition around TT, but unlike Ar+ and Ar2 +, their intensities increase with T° at the transition.

Figure 2
figure 2

Temperature dependence of (a) the normalized intensities of the backscattered Arn + peaks (1 ≤ n ≤ 5), and (b) the Ar+/(Ar2 + + Ar3 +) ratio, for PS 4K

Full size image

Figure 2b points out a ratio, defined as: Ar2 +/(Ar2 + + Ar3 +), as a function of the temperature. The choice of this ratio is justified by the fact that the Ar2 + and Ar3 + peaks are the most intense in all cases, and no mass interference is encountered. Its variation is very similar to that of Ar2 +/ΣArn + (2 ≤ n ≤7), as defined in [8]. The reason not to use that ratio as such is that the number of detectable Arn + species depend on the analyzed polymer and interferences occurring in some cases for n > 3. Dividing by (Ar2 + + Ar3 +) intensities enhances the variation of Ar2 + and makes it independent of instrumental variations. As was the case with Ar+, the temperature of the inflection point of this curve is also lower than the bulk Tg reported for PS 4K (see Table 1). We propose to consider this ratio rather than the transition also observed on the Ar+ curves because of the low intensity of Ar+ and the possible interference with Ca+.

Table 1 Bulk Tg (Measured by DSC) and T° Measured at the Inflection Point for The Different Polymers Tested in This Study
Full size table

Other polymers, with different bulk Tg, have been investigated with the same methodology. The results are summarized in Table 1. It is remarkable that the temperatures measured at the inflection points are always lower than the reported bulk Tg. These differences between bulk Tg and surface TT were already described by Fu et al. for homodisperse PS [10] and for homodisperse PMMA [11], using a different approach based mainly on the secondary ion spectrum interpretation. No literature was found on this topic for PIB.

Another example of application is shown in Figure S1 in the Supplementary Information, where a droplet of PS 4K has been deposited onto an LDPE film surface. The image (a) displays the chemical contrast between both polymers, while the image (b) shows the ratio Ar2 +/(Ar2 + + Ar3 +).

These images where taken at room temperature. The highest values of the ratio are found in the glassy PS 4K region, whereas the lowest corresponds to the rubbery LDPE (LDPE bulk Tg < –100 °C and PS 4K bulk Tg = + 83 °C).

The mechanistic interpretation of the observed variations, i.e., the explanation of the different distributions of backscattered fragment ions when going from the glassy to the rubbery state of the polymers, remains to be established. It is important to note that the range of variation of the ratio Ar2 +/(Ar2 + + Ar3 +) is the same for all the polymers, showing a certain independence of the effect on the specific polymer chemistry. An interpretation based on the structural properties, resulting in smaller cluster fragments in the glassy state than in the rubbery state, is therefore reasonable. As mentioned earlier, mechanical properties were also the cause of similar variations for metallic surfaces (hardness) [8]. Concerning polymers, molecular dynamics simulations of cluster impacts on crystalline and amorphous polymeric samples [12] with varying free volume will be implemented in order to better understand the dependence of the cluster fragmentation on the polymer surface structure. At this point, it is impossible to completely rule out the influence of ionization or neutralization effects.

Conclusions

A new method to probe the surface mechanical properties of polymers at the molecular level (surface TT) in relationship with the bulk Tg was demonstrated, using the impact of large noble gas cluster ions. Though the detailed mechanistic interpretation remains elusive at this stage, the fact that the method indicates surface TT for a series of polymers with diverse chemistries, in a range between –70 and +130 °C, demonstrates its efficacy and robustness. This new protocol should also be investigated for irreversible structural changes occurring in polymers and induced by heating, such as polymerization and cross-linking via sample curing.