Article pubs.acs.org/ac
Reduction of Matrix Effects in Polystyrene/Poly(methylene methacrylate) Blends by Metal-Assisted Secondary Ion Mass Spectrometry Nora Becker and Tom Wirtz* Department of Science and Analysis of Materials, Centre de Recherche Public−Gabriel Lippmann, 41 rue du Brill, L-4422 Belvaux, Luxembourg ABSTRACT: Secondary ion mass spectrometry (SIMS) is a very surface sensitive analysis technique with low detection limits. The main drawback of SIMS is its inherent incapability of providing quantitative information about sample compositions due to the frequent occurrence of ionization- and sputter-induced matrix effects. Metal-assisted SIMS (MetA-SIMS) is an experimental approach that consists in covering an organic sample with a minute amount of a noble metal prior to a static SIMS analysis, the main objective being an increase of the characteristic secondary ion intensities. We show in this article that MetA-SIMS is also a simple and efficient tool for reducing matrix effects in a set of polymer blend samples containing different relative concentrations polystyrene (PS) and poly(methylene methacrylate) (PMMA). These findings can be explained by diffusion processes leading to a sample surface configuration consisting of individual polymer chains embedded in a common Ag matrix.
S
secondary ions are formed via one or more fragmentation pathway(s) from their precursor molecule, ionization matrix effects can be due to short-range interactions between adjacent functional groups (matrix effect of first type, MEI) and/or longrange interactions between noncovalently bonded groups (matrix effect of second type, MEII).3 Polymer blends and copolymers are thus particularly interesting sample systems for matrix effect studies. Applications of polymer blends and copolymers can be found in a wide range of technological domains, from life sciences4 to photovoltaics.5 Various authors have characterized several copolymer or polymer blend systems by TOF-SIMS (time-of-flight SIMS),6−17 often in combination with XPS. In most cases, the conclusion was that a linear relationship between SIMS intensities and the actual surface composition was observed only for some secondary ions, while the intensities of most characteristic ions were influenced by matrix effects. Inoue and Murase have studied Irganox1010 and silicon oil coatings on Si and polypropylene (PP) substrates by TOFSIMS.18 After deposition of a small amount of Ag onto these samples, the authors found that the matrix effects observed in a traditional SIMS analysis were reduced. MetA-SIMS (metal-assisted SIMS) consists in depositing a small amount of metal (most commonly Au or Ag) onto an organic sample prior to a static SIMS analysis.19−22 Usually the metal is deposited by gas-phase evaporation or sputter-coating, or a submonolayer of metal nanoparticles is deposited onto the sample.23 In comparison with pristine organic samples, positive secondary ion yield enhancements of more than 2 orders of
econdary ion mass spectrometry (SIMS) is commonly used for surface or interface characterization. This analysis technique is characterized by its extreme surface sensitivity (information depth, 1−2 monolayers) and very low detection limits (down to the parts per billion for some applications. For the analysis of organic samples, SIMS is often used in the static mode (S-SIMS): the sample is irradiated by a primary ion fluence inferior to 1013 cm−2, which means that less than 1% of the surface atoms are bombarded (for a surface atomic density of ∼1015 atoms/cm2) and chemical damage in the sample is limited. The analytical potential of SIMS is limited by its inherent incapability of providing quantitative information about the sample composition. This is due to matrix effects: the detection probability of a given secondary ion does not only depend on its concentration (or the concentration of the precursor from which it is formed) in the analyzed sample volume, but it is also strongly influenced by the chemical environment at the sample surface.1 This means that the evolution of a secondary ion intensity as a function of the concentration of its precursor is not necessarily linear, and the use of SIMS intensities alone may easily lead to erroneous conclusions about the sample composition. For this reason, SIMS is often combined with XPS (X-ray photoelectron spectroscopy), which provides quantitative chemical information about the surface composition. However, XPS has a much higher detection limit (0.1−1 atom %) and is less surface-sensitive (information depth around 10 nm) than SIMS. Matrix effects in SIMS can be sputter-induced (i.e., change of the sputter yield with varying sample composition) or ionization-induced (i.e., dependence of the ion formation efficiencies on the chemical environment).2 In the particular case of organic polymer samples, where molecular or fragment © 2012 American Chemical Society
Received: February 14, 2012 Accepted: June 20, 2012 Published: June 20, 2012 5920
dx.doi.org/10.1021/ac300433n | Anal. Chem. 2012, 84, 5920−5924
Analytical Chemistry
Article
On each sample, a survey spectrum was taken, followed by individual spectra of C1s and O1s, and again a C1s spectrum in order to check if there had been any X-ray-induced polymer degradation during the analysis time. This did not seem to be the case since no changes in the C1s peak’s shape or intensity were observed. AFM Experimental Conditions. AFM measurements were acquired with an Agilent 5100 AFM/SPM microscope (Agilent Technologies). Images with a resolution of 512 × 512 pixels were obtained in the tapping mode with the amplitude setpoint corresponding to 80−90% of the free amplitude and with a scanning rate of 0.994 lines/s. TOF-SIMS Experimental Setup. The TOF-SIMS analyses were carried out in a TOF5 instrument (Ion-TOF). A pulsed Bi+ beam with 25 keV impact energy was rastered over a 100 × 100 μm2 large sample surface. The acquisition time per spectrum was 120 s, and the ion current on the sample was 0.88 pA. The polymer samples were also analyzed with polyatomic primary ions in the same instrument: Bi3+ ions were used at an impact energy of 25 keV, 0.64 pA ion current, 100 × 100 μm2 raster size, and 100 s acquisition time. The resulting primary ion fluence (6.6 × 1012 cm−2 with Bi+ and 4 × 1012 cm−2 with Bi3+) was below the static limit.
magnitude have been observed for (quasi-) molecular and fingerprint fragment ions.20 Furthermore, the metallic coating reduces charging effects during the SIMS analysis,20 which makes it possible to obtain molecular information even from thick polymer samples without using an electron flood gun that might significantly damage the sample.24,25 MetA-SIMS can also be used to improve the sensitivity in molecular imaging of organic surfaces.20 Another added value in MetA-SIMS as compared to traditional TOF-SIMS measurements of organic samples is the possibility of detecting metal-cationized fragments (positively charged cluster ions consisting of one or more metal atoms and an organic fragment).26,27 In this paper, we investigate the different types of matrix effects in a series of polystyrene/poly(methylene methacrylate) (PS/PMMA) blends with varying compositions. We show that MetA-SIMS is a simple but efficient approach to eliminate these effects and to make semiquantitative analysis of this sample system possible.
■
EXPERIMENTAL SECTION Sample Preparation. Solutions of polystyrene (PS) (Sigma-Aldrich, Mw = 2000 Da) and poly(methyl methacrylate) (PMMA) (Sigma-Aldrich, Mw = 2000 Da) with different mass ratios of both polymers (0:100, 25:75, 50:50, 75:25, and 100:0) were prepared in toluene. The corresponding molar concentration ratios were very similar since the monomer units of both polymers have approximately the same mass (104 Da for styrene and 101 Da for methyl methacrylate). The total polymer concentration in each solution was 2 wt %. These solutions were used for spin-coating on Si wafers (Siltronix) with an acceleration of 7000 rpm/s and a rotation speed of 3000 rpm during 1 min. The Si wafers were previously cleaned by sonication in demineralized water, acetone, and ethanol (10 min per solvent) and dried under a stream of nitrogen. The thickness of the spin-coated polymer layers is estimated to be 50−70 nm based on ellipsometry measurements. The samples were not annealed. Ag deposition was done by electron beam evaporation under vacuum conditions (base pressure of the chamber, 10−9 mbar; during evaporation, 10 −7 mbar) and monitored by a precalibrated quartz microbalance (Maxtek). Ag pellets of 99.99% purity (Kurt J. Lesker Company Ltd.) were brought to evaporation by an electron beam and deposited onto the polymer samples at a rate of 0.1 nm/s. The nominal thickness of the Ag layer was 2 nm. It is well-known that such small amounts of metal form an island-like structure on polymer samples (initial phase of the Volmer−Weber growth mode).28,29 It should be noted that the topography of the Ag-coated samples is likely to change with time. Therefore, the atomic force microscopy (AFM) and TOF-SIMS measurements were always taken 1 day after the metallization. XPS Experimental Setup. XPS measurements were performed using an Axis Ultra DLD spectrometer (Kratos Analytical). The instrument is equipped with a monochromatized aluminum X-ray source powered at 10 kV and 50 mA that delivers an X-ray beam of 300 × 700 μm2. Charge compensation was obtained with the built-in charge neutralization system. The pass energy was set to 160 eV for the survey spectra and to 20 eV for the high-resolution spectra. The binding energies were calculated with respect to the C−(C, H) component of the C1s peak, and a Shirley background subtraction was used.
■
RESULTS AND DISCUSSION Surface Morphology and Surface Compositions of the Polymer Blends. It is well-known that PS and PMMA are immiscible. In most polymer blends, phase separation is observed upon annealing, leading to the formation of microdomains and often an enrichment of one component near the polymer/air interface. Kailas and co-workers have published a detailed study about the morphology of PS/PMMA blends and copolymers.30−32 These authors reported that the as-cast blend samples (without annealing) were smooth and without any specific surface topography. The samples used for this work were not annealed, and no specific morphology could be identified on NanoSIMS images (not shown). In the XPS survey scans, no Si signal was detected, which confirms that the thickness of the polymer layer exceeds the information depth of XPS (∼10 nm). In the XPS spectrum of the pure PS sample (a polymer consisting only of C and H atoms), the O1s peak was observed, suggesting that a small amount of oxygen-containing contaminants was present on the sample surface. For the quantification of the blends, only the C1s signal was used. The different contributions in the C1s core-level shift associated with PS are C−(C, H) (binding energy, 284.7 eV) and the shakeup (291.4 eV). For PMMA, C−(C, H) (285.0 eV), C− C(O)−O (285.8 eV), C−O (286.9 eV), and OC−O (289.1 eV) can be identified. Table 1 shows the composition in the bulk and at the surface of the PS/PMMA blend samples. The bulk compositions are Table 1. Composition in the Bulk and at the Surface of the PS/PMMA Blend Samples bulk composition (PS/PMMA)
surface composition (PS/PMMA)a
25:75 50:50 75:25
45:55 66:34 83:17
In this context, “surface” refers to the topmost ∼10 nm of the sample, corresponding to the sampling depth of XPS.
a
5921
dx.doi.org/10.1021/ac300433n | Anal. Chem. 2012, 84, 5920−5924
Analytical Chemistry
Article
Furthermore, the total positive secondary ion intensity linearly decreases with increasing PS concentration (not shown), the value for pure PS being 10 times lower than that for pure PMMA. It is possible that the secondary ion formation efficiencies in the studied polymer blend system are affected by a combination of sputtering- and ionization-induced effects. For the ionization contribution, only MEII (long-range interactions) can be considered for this type of sample because the styrene and methyl methacrylate units are not covalently bonded and the end groups are always the same for all the polymer chains (butyl end groups). The origin of this matrix effect is probably hydrogen transfer between spatially close styrene and methyl methacrylate units. It is interesting to note that the evolution of the absolute secondary ion intensities with the sample composition is qualitatively exactly identical with those obtained by Vanden Eynde and co-workers for random copolymers with varying styrene and methyl methacrylate content analyzed with 15 keV Ga+ primary ions.10,11 This similarity was observed for all the positive and negative secondary ions considered by these authors. They concluded that hydrogen transfer from MMA units to covalently bonded styrene groups was the main reason for the observed matrix effects. However, the results presented in this section about polymer blends suggest that long-range ionization-induced effects (MEII) may also have played a significant role in the copolymer system studied by these authors. Polyatomic Primary Ions. The polymer samples were also analyzed with Bi3+ ions. Figure 2 shows the evolution of the
derived from the relative concentrations in the solutions used for spin-coating, and the surface composition was calculated from the XPS results taking into account the C1s contributions of both polymers (except for the shakeup). According to the XPS measurements, the surface of the PS/ PMMA blend samples is enriched with PS, which can be assigned to the difference in surface free energy of PS and PMMA.32 In the following sections of this article, the surface composition values determined by XPS were used. However, it is important to note that the sampling depth of XPS (∼10 nm) is much larger than that of TOF-SIMS and that the actual composition in the topmost monolayer (which constitutes the main source of information for the latter analysis technique) might be different from the values obtained by XPS. Analysis of PS/PMMA Blends by TOF-SIMS. Monoatomic Primary Ions. In order to check for possible matrix effects, the polymer blends were first analyzed with Bi+ ions, and the intensities of characteristic positive secondary ions were plotted as a function of the PS content at the sample surface (as measured by XPS). Figure 1 displays the evolution of the
Figure 1. Normalized intensities of positive characteristic PS (red) and PMMA (black) ions as a function of the PS concentration at the sample surface (Bi+ primary ions).
normalized intensities of some secondary ions: the tropylium ion C7H7+ (m/z = 91) and the protonated monomer C8H9+ (m/z = 105) are characteristic for PS, and CH3O+ (m/z = 31, the methyl ester function), C2H3O2+ (m/z = 59, the methyl methacrylate group), and C5H9O2+ (m/z = 101, the protonated monomer unit) can be associated with the molecular structure of PMMA. The absolute intensities in each figure were normalized to the highest intensity measured for each series for a better comparison of the intensity−concentration relationships obtained for different secondary ions. It should be noted that C7H7+ and C8H9+ ions are also detected from a pure PMMA sample and are thus not entirely specific for PS. It is, however, not possible to estimate the contribution of PMMA fragments to these secondary ion intensities for the blend samples because the emission of these ions from both PS and PMMA may be influenced by (possibly different) matrix effects. The yields of all the considered secondary ions, but especially those associated with PS, are nonlinear with respect to the concentration, and they are thus obviously influenced by matrix effects in the blend samples. The characteristic PS peaks are higher for the blend samples than for the pure PS sample (Figure 1). Since the observed intensity−concentration relationship is not a bijection, semiquantitative analysis based on the intensity of these ions is impossible. The positive peaks of PMMA are all lower in presence of PS than they should be for a linear intensity−concentration relationship (Figure 1).
Figure 2. Normalized intensities of positive characteristic PS (red) and PMMA (black) ions as a function of the PS concentration at the sample surface (Bi3+ primary ions).
normalized intensities (normalized to the maximum intensity recorded for each measurement series) of some positive characteristic PS and PMMA ions as a function of the sample surface composition. The evolution of the normalized intensities of the secondary ions associated with PMMA is similar with Bi+ and Bi3+ bombardment. The situation is very different for the characteristic ions of PS: with Bi3+ bombardment, the C7H7+ curve is much closer to linearity than for Bi+, while the C8H9+ intensities are still higher in the blend samples than they should be. It is likely that there is also a sputter-induced contribution to the matrix effect and that mono- and polyatomic ions affect the sputter rates of the blends in different ways. Another possible explanation for the better correlation of the Bi3+ secondary ion 5922
dx.doi.org/10.1021/ac300433n | Anal. Chem. 2012, 84, 5920−5924
Analytical Chemistry
Article
intensities with the XPS data would be that the sampling depth of the polyatomic projectile is closer to the sampling depth of XPS. MetA-SIMS: A Simple Way to Reduce Matrix Effects in PS/PMMA Blends. The normalized intensities of some characteristic PS and PMMA peaks measured on the Ag-coated blend samples are shown in Figures 3 and 4.
Figure 5. Normalized intensities of the Ag-cationized monomer units as a function of PS content (Bi+ primary ions). The data points represent the average values resulting from three experimental series. The straight lines correspond to linear fits, and R2 is the coefficient for linear regression. Error bars are not shown since they are smaller than the data symbols (relative error
