Sum Frequency Generation and Coherent Anti-Stokes Raman

Article pubs.acs.org/Langmuir

Sum Frequency Generation and Coherent Anti-Stokes Raman Spectroscopic Studies on Plasma-Treated Plasticized Polyvinyl Chloride Films Jeanne M. Hankett, Chi Zhang, and Zhan Chen* Department of Chemistry, University of Michigan, 930 North University Avenue, Ann Arbor, Michigan 48109, United States ABSTRACT: Polyvinyl chloride (PVC) is a widely used polymer to which various phthalates are extensively applied as plasticizers. PVC materials are often treated with plasma to vary the hydrophobicity or for cleaning purposes, but little is known of the nature of the surface molecular structures after treatment. This research characterizes molecular surface structures of PVC and bis-2-ethylhexyl phthalate (DEHP)plasticized PVC films in air before annealing, after annealing, and after exposure to air-generated glow discharge plasma using sum frequency generation (SFG) vibrational spectroscopy. In addition, we compare the vibrational molecular signatures on the surfaces of PVC with DEHP (at a variety of percent loadings) to those of the bulk detected using coherent anti-Stokes Raman scattering (CARS). X-ray photoelectron spectroscopy (XPS) and contact angle measurements have been used to analyze PVC surfaces to supplement SFG data. Our results indicate that DEHP was found on the surfaces of PVC films even at low weight percentages (5 wt %) and that DEHP segregates on surfaces after annealing. The treatment of these films with glow discharge plasma resulted in surface-sensitive reactions involving the removal of chlorine atoms, the addition of oxygen atoms, and C−H bond rearrangement. CARS data demonstrate that the bulk of our films remained undisturbed during the plasma treatment. For the first time, we probed the molecular structure of the surface and the bulk of a PVC material using combined SFG and CARS studies on the same sample in exactly the same environment. In addition, the methodology used in this research can be applied to characterize various plasticizers in a wide variety of polymer systems to understand their surface and bulk structures before and after systematic applications of heat, plasma, or other treatments.

■

INTRODUCTION Plastics are important polymeric materials in modern life. Human society relies heavily on plastics because of their resistance to chemical, physical, and biological degradation. Plastics pose many potential human health and environmental risks because of their wide usage and the physical properties of the many additives contained therein. Plastics can pollute and disrupt important natural processes, and plastic wastes can be found everywhere, including landfills, drinking water, oceans, air currents, and urban soils.1−4 To impart flexibility, pliability, and elasticity to otherwise rigid polymers, such as polyvinyl chloride (PVC), phthalate esters or phthalates are added to plastics as plasticizers. Phthalates are used as plasticizers because they do not covalently bond to polymers, allowing for their beneficial elasticizing properties.5,6 Produced in large quantities since the 1930s, phthalates are ubiquitous in our society and can be found in industrial plastics, household items, wiring, tubing, and cables, paints, medical devices, children’s toys, and personal care products, including cosmetics, lotion, sunscreen, and perfume.7−13 Bis(2-ethylhexyl)phthalate (DEHP) is one of the most commonly used plasticizers, used in a wide variety of different polymer goods mainly because of its low cost of synthesis and effectiveness of plasticization.5,6 © 2012 American Chemical Society

Extensive research has been carried out to investigate DEHP and other phthalates and their influences on human health and environments. Such research has focused on the biological and medical effects associated with exposure to phthalates, biodegradation mechanisms of phthalates, degradation mechanisms of phthalates in bulk media, especially in solutions, and leaching of phthalates from the bulk media into external environments (e.g., measured by chromatography and mass spectrometry).6,12−25 In fact, DEHP has been observed leaching out of plastics, and the toxicity of this molecule continues to be under debate. DEHP is known to be toxic to rodents, a suspected human carcinogen, and suspected to be toxic to many marine organisms. However, in these previous studies, in situ molecular-level characterizations on phthalates on surfaces and at interfaces are missing to a large part because of a lack of appropriate techniques to probe these surfaces.5 As such, the surface molecular behaviors of DEHP in plastics are widely unknown.5,16−20,26,27 It is vital to understand the surface chemistry of plastic materials and phthalate plasticizers at a molecular level because Received: November 18, 2011 Revised: January 25, 2012 Published: February 7, 2012 4654

dx.doi.org/10.1021/la2045527 | Langmuir 2012, 28, 4654−4662

Langmuir

■

a multitude of organisms come in contact with these surfaces and phthalate plasticizers leach into the environment through surfaces. To probe these interfaces, a truly surface-sensitive technique is required to gather knowledge of the molecular migrational behaviors and the molecular degradation that these materials commonly undergo in use and after disposal. Sum frequency generation (SFG) vibrational spectroscopy is a surface-sensitive nonlinear optical spectroscopy capable of providing molecular vibrational information for molecules where centrosymmetry is broken.28−30 Previous literature has shown that SFG is a powerful technique for probing both environmental and polymeric interfaces.30−45 SFG has also been applied to study polymer surfaces exposed to plasma, radicals, and ultraviolet (UV) light.46−50 In this work, SFG was used to probe the surface of thin films of PVC with various weight percentages of the phthalate plasticizer DEHP (Figure 1), before annealing, after annealing, and after exposure to an

Article

EXPERIMENTAL SECTION

Materials. DEHP (analytical standard) was obtained from Fluka (St. Louis, MO). Tetrahydrofuran (THF; ≥99.9% purity), concentrated sulfuric acid (reagent grade), potassium dichromate, and PVC (Mw, 62 000; in pellet form) were obtained from Sigma Aldrich (St. Louis, MO). Sample Preparation. PVC pellets were dissolved in THF to prepare thin films for study. A 30:1 weight ratio of THF/PVC was used for all films. DEHP was added by weight percent to PVC. Solutions were mixed using a vortex mixer (Vortex-Genie 2T, Scientific Industries, Inc.) until clear. Fused silica windows (ESCO Products, Inc.) were used for SFG measurements and cleaned using a concentrated sulfuric acid bath saturated with potassium dichromate overnight. Windows were rinsed with deionized water and dried with nitrogen gas before sample preparation. Both silica windows and microscope glass slides were used as substrates for polymer films for contact angle measurements; no differences were found when different substrates were used. A P-6000 spin coater (Speedline Technologies) was used to prepare all polymer/plasticizer films. Samples were spin-coated at 3000 rpm for 30 s on windows and glass slides. PVC films were spin-coated with varying weight percentages of DEHP at 0, 5, 10, 25, and 65 wt % of the total PVC/DEHP mass. Selected films were annealed in an Isotemp lab oven (Fisher Scientific) at 76 °C overnight. Instrumentation. Static water contact angle measurements were performed using a CAM 100 optical contact meter (KSV Instruments). At least three samples of each type of polymer blend were used for contact angle goniometry measurements. Four spots were taken per sample on average. Film thickness measurements were taken using a Dektek3 Profilometer (Veeco). Film thicknesses averaged around 150 nm. Air plasma treatment was completed using a homebuilt glow discharge plasma cleaner under low vacuum. The discharge was maintained between two parallel Ti plates using alternating current (AC) high voltage at 700 V. SFG has been widely used to obtain molecular information of a variety of surfaces and interfaces, including polymer surfaces in air.30,31 The details of the SFG theory and setup have been extensively examined in previous papers.28−31,59,60 The SFG system is composed of a pico-second Nd:YAG laser, a harmonic unit with two KD*P crystals, an optical parametric generation (OPG)/optical parametric amplification (OPA) and difference frequency generation (DFG) system based on LBO and AgGaS2 (or GaSe) crystals, and a detection system. The output of the Nd:YAG laser is a 20 Hz, 20 ps, 1064 nm near-infrared (NIR) beam. The visible input 532 nm beam for SFG experiments is generated by frequency-doubling a portion of this 1064 nm beam. The IR beam is generated from the OPG/OPA and DFG system and can be tuned from 1000 to 4300 cm−1. For SFG experiments, the incident angles of the visible and IR input are 60° and 55° with respect to the surface normal, respectively. The diameters of both input beams at the surface are about 500 μm. The SFG signal from the surface is collected by a photomultiplier and processed with a gated integrator. To compare molecular vibrational signals on surfaces to the bulk of plasticized polymers, CARS spectroscopy was also used in this study. CARS capability was added to our SFG system, as we reported previously.61 Instead of using the frequency tunable mid-IR input beam, in the CARS experiments, we used the frequency tunable visible beam generated from the OPG/OPA system (as the Stokes beam) and the 532 nm input beam (as the pump/probe beam). CARS spectra can be collected from a sample using our SFG system in the same environment. The collection of SFG or CARS spectra can be controlled by flipping mirrors. To ensure the temporal overlap on the sample, an additional delay line was used in the CARS experiments. In this study, both SFG and CARS spectroscopies examined molecules in the C−H stretching frequency range (2700−3100 cm−1). SFG spectra presented in this paper were collected using the ssp polarization combination: s-polarized output signal, s-polarized visible input, and p-polarized IR input. CARS spectra were collected using the

Figure 1. Molecular structures of PVC (left) and DEHP (right).

air plasma. To compare the surface structures to the bulk structures of plasticized PVC materials, coherent anti-Stokes Raman scattering (CARS) was used to probe the bulk. CARS is also a nonlinear spectroscopy, but unlike SFG, it is not a surface-sensitive analytical technique.51−53 Previous publications have studied plasma treatment of PVC for a number of applications, showing promise for both crosspolymerization to prevent phthalate migration in plastics and as a means to accelerate PVC degradation for environmentally friendly material disposal.54−58 However, there is little data comparing signatures of molecules on the surface to bulk using a nondestructive truly surface-sensitive technique. Fourier transform infrared (FTIR), attenuated total reflection (ATR)−FTIR, and Raman spectroscopies have been used to probe PVC materials treated with various forms of plasma.54−58 However, they give molecular information beyond surface layers (≥200 nm). X-ray photoelectron spectroscopy (XPS) has also been used to probe surface elements using photoionization, but it must be completed under ultrahigh vacuum and cannot provide in-depth surface structural information (e.g., orientation of surface functional groups). SFG has the capability of probing molecular surface structures with submonolayer surface sensitivity under atmospheric conditions. This study aims to gain a fundamental understanding of surface segregation of DEHP molecules on plasticized PVC surfaces with different bulk DEHP concentrations. In addition, surface structural changes of these materials after plasma treatment were examined at the molecular level. Surface structures were probed using SFG, and bulk structures were studied using CARS for comparison purposes. This study also aims to demonstrate that SFG and CARS are capable of observing the molecular changes in PVC because of plasma application, giving information on the bulk and surface, respectively. Finally, this study will lay a foundation for future research on surface and bulk structures of plasticized PVC materials in different chemical environments, such as water and various solutions. 4655

dx.doi.org/10.1021/la2045527 | Langmuir 2012, 28, 4654−4662

Langmuir

Article

ssss polarization combination. The SFG ssp spectra are sensitive to the functional groups with dipole transitions of the vibrational modes along the surface normal. Usually, ssp spectra from polymer surfaces are much stronger than spectra collected using other polarization combinations and are dominated by signals from symmetric vibrational modes.45 The symmetric vibrational modes dominate the CARS ssss spectra as well, and the intensity of the ssss spectra is typically the strongest of the different polarization combinations.61 Both SFG and CARS spectra were collected using the reflection geometry. For each sample, two SFG spectra were collected per spot and at least five spots on the film−air interface were examined per sample. For CARS, three spectra were collected per spot and at least two spots were examined per sample. In this work, film-thickness-dependent studies were not performed. As we discussed in detail in a recent publication,62 usually SFG signals in the C−H stretching frequency region from the polymer/hydrophilic substrate and polymer/water interfaces are quite weak. Therefore, the signals from the two interfaces may interfere with each other. The polymer-thickness-dependent studies can deconvolute signals of the polymer/water interfaces from the overall signals.62 However, here we focused solely on polymer films in air, which have much stronger SFG signals, and it was not necessary to perform such a thicknessdependent study.62 XPS was performed at the Electron Microbeam Analysis Laboratory of the University of Michigan using a Kratos AXIS Ultra DLD XPS with a monochromatic Al source that gives an energy resolution better than 0.5 eV measured from the pure Ag 3d peak. Charge neutralization was applied during XPS measurements.

as those in polyvinyl alcohol (PVA), which generate similar symmetric stretching signals at ∼2910 cm−1.66 The weak and broad signal at 2880 cm−1 may be due to the methyl end groups (symmetric stretching).65 The SFG spectrum collected from the pure DEHP surface is markedly different from that collected from the pure PVC surface. The DEHP SFG spectrum is dominated by the contributions from methyl groups, including the CH 3 symmetric stretch at 2880 cm−1 and the Fermi resonance (generated from the methyl symmetric stretching and the overtone of the methyl bending mode) at 2945 cm−1. The contribution from the methylene symmetric stretching at 2860 cm−1 can also be observed. Figure 2 displays SFG spectra collected from the surfaces of DEHP-plasticized PVC with the different percent loadings of DEHP in PVC. Even at lowDEHP bulk percentages (5 wt %), it is clear that CH3 signals from DEHP are present on the surface. All of the figures of SFG and CARS spectra in this work have spectra offset. The SFG spectra shown in Figure 2 suggest that the DEHP plasticizer is present on the surfaces of these films even at low weight percent bulk loadings (5 and 10 wt %). At the two higher weight percent loadings (25 and 65 wt %), the DEHP signals dominate the spectra, indicating that the surfaces may be mainly covered by DEHP. In addition, a clear trend is observed, noting the relative peak heights in the spectra from small to large percent bulk loading of DEHP. The 2880 and 2940 cm−1 peaks gradually increase, and the 2920 cm−1 signal gradually decreases. This demonstrates that, with the increase of the bulk loading of DEHP, the surface coverage of DEHP increases, while the surface coverage of PVC decreases. We believe that the 2880 cm−1 signal observed in the PVC−DEHP mixtures is from DEHP because this signal is much narrower than the peak at the same position observed from PVC alone. At lower weight percentages (5 and 10 wt %), the intensity of the 2880 cm−1 CH3 peak is similar to the 2920 cm−1 CH2 peak. Under these two cases, the surfaces are covered by both PVC (evidenced by the 2920 cm−1 signal) and DEHP (evidenced by the 2880 and 2945 cm−1 peaks). At 25 wt % and above, the 2880 and 2945 cm−1 peaks dominate, showing that the surface is mainly covered by DEHP. At 25 wt % DEHP, the PVC peak at 2920 cm−1 can still be observed from the surface as a shoulder in the spectrum, but on the surface of the sample with 65 wt % DEHP, this peak cannot be observed. To compare the presence of DEHP on these film surfaces to the bulk mixture, CARS spectra of the non-annealed films were collected (Figure 4). Unsurprisingly, because the PVC is the dominating component, the CH2 signal of PVC (2915−2920 cm−1, CH2 symmetric stretch) dominates all percent bulk loadings, except 65 wt % DEHP. Because the signals from symmetric modes in the CARS ssss spectra are usually much stronger than the asymmetric modes, this further confirms that the 2915−2920 cm−1 signal is due to a CH2 symmetric stretch. For the 65 wt % DEHP sample, in addition to the PVC CARS peak, DEHP signatures [2860 cm−1, symmetric CH2 stretch; 2880 cm−1, symmetric stretch (as a shoulder); 2945 cm−1, Fermi resonance; and 2970 cm−1, CH3 asymmetric stretch] can be seen. For pure DEHP, these signals can also be detected (Figure 3). Figure 4 demonstrates that the bulk composition of the films may be different from the surface and that the bulk CARS signal is closely related to the percent loadings of DEHP prepared. Plasma Treatment of Non-annealed Films. The effects of plasma exposure to PVC products have been examined a

■

RESULTS AND DISCUSSION Non-annealed Films. SFG spectra were collected from surfaces of pure PVC, PVC with 5, 10, 25, and 65 wt % DEHP, and pure DEHP in air (Figures 2 and 3). The PVC SFG

Figure 2. SFG ssp spectra of PVC films with 0, 5, 10, 25, and 65 wt % DEHP.

spectrum is dominated by a CH2 symmetric stretching signal at 2915−2920 cm−1, which was assigned as such in the literature using isotope-labeled studies.63,64 The symmetric stretching signal of the regular CH2 groups in aliphatic chains usually appears at ∼2850 cm−1, according to the previous studies.65 The differing wavenumbers of the CH2 symmetric stretching signals in different compounds are due to changes in the local environment surrounding these CH2 groups. In usual aliphatic chains (e.g., in DEHP), CH2 groups are next to each other. In PVC, CH2 groups appear in every other position along the backbone. The CH2 groups in PVC have similar environments 4656

dx.doi.org/10.1021/la2045527 | Langmuir 2012, 28, 4654−4662

Langmuir

Article

Figure 3. SFG ssp (left) and CARS ssss (right) spectra of pure DEHP.

Figure 4. CARS ssss spectra of PVC films with 0, 5, 10, 25, and 65 wt % DEHP.

Figure 5. SFG ssp spectra of PVC films after plasma treatment with 0, 5, 10, 25, and 65 wt % DEHP.

number of times in the literature, mainly using bulk analytical techniques to understand the physiochemical changes that the polymer undergoes because of radical reactions.54−58 Using SFG and CARS, we are able to compare molecular structural changes on the surface and the bulk and also compare differences between molecular structural changes in PVC to DEHP. Figure 5 displays the SFG spectra obtained after exposure of individual samples to glow discharge atmospheric plasma for 5 s. After the plasma treatment on the pure PVC film, the SFG spectrum is dominated by a peak at 2930 cm−1, which has a very large peak width. A weak signal at 2875 cm−1 can also be observed. In comparison to the SFG spectrum collected from the PVC surface before the plasma treatment, the main peak for PVC shifts to a higher wavenumber from 2920 to 2930 cm−1 and the 2875 cm−1 signal substantially decreases. The decrease of the 2875 cm−1 signal is due to the surface coverage of methyl end groups decreasing or becoming more disordered. The reasons for the 2920 cm−1 peak shift observed after plasma treatment are not straightforward. The increased peak width at 2930 cm−1 suggests that the methylene groups in PVC undergo some changes because of the plasma treatment and/or that spectral interference is occurring. The spectral interference may be due to an increase in intensity near 2920 cm−1 (i.e.,

2940 cm−1), which, because of peak overlap, is indistinguishable from a peak at 2920 cm−1. Under normal circumstances, the 2920 and 2940 cm−1 peaks can be separated. However, for peaks with larger widths, these two signatures may overlap. In addition, previous research has suggested that plasma treatment of PVC may remove chlorine, which can shift the SFG peak to a higher wavenumber and/or broaden the CH2 signal because of the environmental changes of the methylene groups. It was also suggested that PVC chain scission occurs in radical reactions with plastics, leaving some CH3 groups on the surface, which would result in signals at 2940 cm−1 (Fermi resonance of the methyl group).54−56 Clearly, these newly created methyl groups are different from the original methyl end groups, which may also lead to the decrease of the 2880 cm−1 signal. A second surface-sensitive analytical technique, i.e., XPS, is required to confirm the spectral changes. It can be said, however, that some molecular changes have occurred likely involving bond breaking and/or reformation. The SFG spectra collected from the PVC films with 5 and 10 wt % DEHP after the plasma treatment are similar but different from those detected before the plasma treatment. After exposure to plasma, SFG spectra show the following signals: 2930, 2855, 2880, and 2960 cm−1 (small shoulder). The peak 4657

dx.doi.org/10.1021/la2045527 | Langmuir 2012, 28, 4654−4662

Langmuir

Article

shift of the dominating 2920 cm−1 signal to 2930 cm−1 and the peak width increase are due to the changes of surface PVC, as discussed above. A small dip at around 2900 cm−1 may be due to the interference between the nonresonant signal and the 2930 cm−1 signal. The signal decrease after plasma treatment of the 2880 cm−1 peak is also due to the signal change of the surface PVC. Therefore, we believe that plasma treatment caused similar effects on surface PVC for the pure PVC sample and the mixed PVC samples with 5 and 10 wt % DEHP. For the PVC samples with 25 and 65 wt % DEHP, no noticeable changes were observed in SFG spectra collected after the plasma treatment compared to those collected before. As we discussed above, DEHP dominates the surfaces of these two samples. Therefore, this may indicate that plasma treatment does not cause changes to DEHP methyl groups on PVC surfaces. CARS spectra were also collected from these PVC samples with and without DEHP (Figure 6). No difference was Figure 7. SFG ssp spectra of annealed PVC films with 0, 5, 10, 25, and 65 wt % DEHP.

Figure 6. CARS ssss spectra of PVC films after plasma treatment with 0, 5, 10, 25, and 65 wt % DEHP.

observed in the CARS spectra collected before and after plasma treatment. This indicates that the polymer film bulk does not exhibit structural changes after sample exposure to plasma. By directly comparing SFG spectra to CARS spectra, we were able to confirm that a surface layer reaction occurred, demonstrating that plasma treatment times were short enough to only affect the surface layers of the films, while the bulk remained undisturbed. Annealed Films. The above investigated samples may not have reached equilibrium. To ensure that the sample surface reached equilibrium, sample films were annealed in an oven overnight. SFG (Figure 7) and CARS (Figure 8) spectra were taken to compare the surface structures of these films to nonannealed films (Figures 2 and 4). The SFG spectra in Figures 7 and 2 are markedly different for some samples before and after annealing, while the CARS spectra in Figures 8 and 4 for all of the samples are similar before and after annealing. For the pure PVC, the SFG spectrum collected after annealing shows the disappearance of the 2880 cm−1 signal, indicating that the methyl end groups retreat to the bulk or become random on the surface. The 2920 cm−1 peak intensity also decreases, showing that the annealing causes the surface methylene groups to tilt more toward the surface (i.e., the PVC backbones stand up more on the surface). For the PVC samples with 5 and 10 wt % DEHP, SFG spectra are completely different from those

Figure 8. CARS ssss spectra of annealed PVC films with 0, 5, 10, 25, and 65 wt % DEHP.

collected before annealing. After annealing, only spectral features from DEHP can be observed, demonstrating entire surface coverage of the plasticizer. Even for only 5 wt % DEHP in the bulk, the only surface signals present are those from DEHP at 2880 and 2940 cm−1. This shows that, at equilibrium, DEHP segregates to the PVC surface. SFG spectra collected from PVC films with 25 and 65 wt % DEHP after annealing exhibit similar spectral features compared to those collected before annealing. The surfaces are covered by DEHP. However, similar to the pure PVC surface, the SFG signal intensity decreased after annealing, possibly showing that DEHP methyl groups tilt more toward the surface after annealing or more likely have a broad orientation distribution, as discussed below. It is interesting to observe from Figure 7 that SFG signals of 5 and 10 wt % DEHP samples are stronger than those from the 25 and 65 wt % DEHP samples. As discussed above, before annealing, the surfaces of 5 and 10 wt % DEHP in bulk PVC are covered by both DEHP and PVC. After annealing, these two surfaces are dominated by DEHP. Because the total amount of DEHP in PVC is small (only 5 and 10 wt %), very 4658

dx.doi.org/10.1021/la2045527 | Langmuir 2012, 28, 4654−4662

Langmuir

Article

likely DEHP barely covered the top layer, and therefore, methyl groups can stand up more, generating stronger signals. However for 25 and 65 wt % DEHP samples, both before and after annealing, the top layers are crowded with DEHP. The crowded methyl groups can have different orientations, generating weaker SFG signals. This explains the relative intensity in Figures 2 and 7. CARS spectra of the annealed films are comparable to those of non-annealed films, except for the one with 65 wt % DEHP. The structure of DEHP in the bulk may have slightly varied after annealing, which can be detected by polarized CARS (Figure 8). Plasma Treatment of Annealed Films. Figure 9 displays the SFG spectra collected after plasma treatment for the

these two signals can still be observed, showing some surface coverage of DEHP on these two surfaces. SFG spectra collected from the 65 wt % DEHP sample surface after the plasma treatment are similar to those from the surface before the plasma treatment. This shows that plasma treatment does not substantially alter the surface DEHP structure, as observed from the non-annealed samples. In comparison to 5, 10, and 25 wt % samples, the 65 wt % samples contain more DEHP; therefore, the surface DEHP layer must be much thicker, and the plasma treatment could not expose bulk PVC to the surface. CARS spectra demonstrate no change in bulk molecular vibrational signals before and after plasma treatment for annealed samples, as shown in Figure 10.

Figure 9. SFG ssp spectra of annealed PVC films with 0, 5, 10, 25, and 65 wt % DEHP after exposure to glow discharge plasma.

Figure 10. CARS ssss spectra of annealed PVC films with 0, 5, 10, 25, and 65 wt % DEHP after exposure to glow discharge plasma.

annealed films. After plasma treatment, the SFG spectrum collected from pure PVC has similar spectral features; however, the intensity is enhanced, and the peak width is increased. The peak width increase must be caused by the same means that we discussed for the non-annealed samples. The signal intensity increase indicates that the dominating methylene groups on the surface are more oriented along the surface normal. This means that the PVC backbones lie down flatter relative to the surface such that the CH2 groups are perpendicular to the air interface. SFG spectra collected from DEHP plasticized PVC with 5 and 10 wt % DEHP after plasma treatment are markedly different compared to those collected before the plasma treatment. Before plasma exposure, the surfaces were dominated by DEHP. Because the overall DEHP contents in these two samples are small, the surface-segregated DEHP may have just barely covered the surfaces, as we discussed above. The plasma treatment may have exposed more PVC underneath the surface. Similar to the pure annealed PVC sample, plasma exposure may have oriented CH2 groups along the surface normal in PVC in these samples (the PVC backbones lie flatter on the surface). Evidence of such lies in the fact that, although the PVC signal at 2920 cm−1 was not observed before plasma treatment, after plasma treatment, PVC generates strong signals near 2920 cm−1. Because more PVC is exposed to the surface, the surface coverage of DEHP decreases, resulting in a signal decrease at 2875 and 2855 cm−1. However,

Complementary Techniques. Contact Angle of Films. To substantiate our SFG results, we applied contact angle measurement and XPS to study the pure PVC sample and the PVC sample with 25 wt % DEHP. Table 1 displays the contact Table 1. Contact Angles for Annealed Films with Varying Percent Loadings of DEHP before and after Exposure to Glow Discharge Plasma water contact angle results for annealed films non-annealed 25 wt % DEHP

PVC before plasma after plasma

annealed PVC

25 wt % DEHP

∼80−82°

∼88°

∼86°

∼86−87°

∼52−54°

∼63−66°

∼52−53°

∼48−50°

angles of deionized water on annealed, non-annealed, and plasma-treated films. Before plasma treatment with nonannealed films, an increase in the plasticizer content leads to an increase in the contact angle. This may suggest that the hydrophobic methyl groups of the plasticizer are present on the surface. This is well-correlated to SFG data, where on the surface of non-annealed PVC with 25 wt % DEHP, the SFG signal is dominated by DEHP. There, the characteristic peaks for DEHP are contributed by CH3 end groups with no phenyl ring peak observed (Figure 2). The phenyl molecular group is 4659

dx.doi.org/10.1021/la2045527 | Langmuir 2012, 28, 4654−4662

Langmuir

Article

Figure 11. Fit XPS spectra for O 1s, Cl 2p, and C 1s in PVC films and PVC films plasticized with 25 wt % DEHP before and after exposure to glow discharge plasma.

down flatter on the surface, as mentioned previously. There are major contact angle decreases for both annealed and nonannealed PVC after plasma treatment, but this is more likely induced by more oxygen content on the surface rather than CH2 orientation. The annealed PVC sample with 25 wt % DEHP also generates a very different SFG spectrum after the plasma treatment. However, the non-annealed PVC with 25 wt % DEHP loading exhibits a similar spectrum before and after treatment, and their contact angles have the least difference. Again, the contact angle decrease measured here may be due to oxygen bonding to the polymer chains, which has been observed after plasma treatment in the literature67 (as well as the XPS data shown in this paper) but could not be confirmed

only observable in bulk CARS measurements. SFG results on annealed PVC surfaces indicate structural changes, while the contact angles on annealed and non-annealed PVC surfaces also exhibit differences. The annealed and non-annealed PVC with 25 wt % samples generate similar SFG spectra, showing similar surface structures. The water contact angles on these two surfaces are also similar, compatible to the SFG data. Both annealed and non-annealed films (Table 1) demonstrate changes in the contact angle for both PVC as well as PVC with 25 wt % DEHP after plasma treatment. The SFG spectra change after plasma treatment as well. For annealed and nonannealed samples of pure PVC, there is an increase in the 2920 cm−1 peak because of the orientation change of CH2 groups toward the surface normal; the PVC backbone may be lying 4660

dx.doi.org/10.1021/la2045527 | Langmuir 2012, 28, 4654−4662

Langmuir

Article

surfaces compared to the bulk of the film. It was found that annealing facilitates the surface segregation of DEHP to the PVC surfaces. Only 5 wt % of DEHP loaded to PVC can dominate the PVC surface after annealing. Annealing also seems to affect the bulk structure of PVC with a substantial amount of plasticizers loaded (e.g., with 65 wt % DEHP), revealed by CARS. When using SFG and CARS to probe films after plasma exposure, we have shown that SFG and CARS can successfully be used to demonstrate molecular changes from a surfacesensitive chemical reaction. Probing the films using SFG in the C−H spectral range illustrates that some bond breaking and rearrangement occurs with PVC exposure to plasma, but the vibrational signatures of plasticizer methyl bonds remain similar. CARS signals remain the same after the plasma treatment, indicating that, here, the plasma treatment only affects surface structures. Because the entirety of surface chemical reactions could not be explored using solely the signals in the C−H stretching frequency range with nonlinear spectroscopies, we applied XPS and contact angle measurements to substantiate SFG data. The measured contact angles suggest the addition of hydrophilic groups (e.g., groups with oxygen) to the film surface after plasma treatment. XPS results indicate that chlorine is extracted from PVC and new bonds or functionalities are formed with oxygen atoms on the surface after the plasma exposure. To further investigate this process, experiments using the CO frequency region of SFG and CARS should be completed in the future. Future work will also include studying the molecular behaviors of other common phthalate plasticizers before and after plasma exposure and studying the molecular changes in plasticized PVC from other forms of radical reactions. Finally, this paper demonstrates a platform for the molecular vibrational signatures of lab-synthesized plastic films that can be used for further surface-sensitive studies of plastics exposed to different chemical environments, e.g., water and various solutions, where XPS cannot be used.

solely using the SFG spectra detected in the C−H stretching frequency region. This hypothesis will be studied in a lower frequency region using SFG in the future. XPS. XPS spectra collected from the non-annealed PVC sample and the PVC sample with 25 wt % loading before and after plasma treatment are shown in Figure 11. We studied the O 1s peak, Cl 2p peak, and C 1s peak signals. On the nonannealed PVC surface before plasma treatment, almost no O 1s signal can be detected. Both the C 1s and Cl 2p peaks are visible. The doublet formation of the C 1s peak suggests that there are two forms of carbon present, with one carbon type bonded to a more electronegative element (higher bonding energy). This could be representative of the carbon with carbon−chlorine bonding in the PVC structure, while the second peak is associated with the carbon in the PVC CH2 group. The O 1s signal can be clearly detected from the PVC surface with 25 wt % DEHP loading (Figure 11). Our SFG results indicated that, on this surface, DEHP dominates, which contains CO groups. Therefore, SFG and XPS data are wellcorrelated. After plasma treatment, the removal of chlorine atoms and the addition of oxygen atoms on the surface of PVC and plasticized PVC with 25 wt % DEHP were confirmed by XPS spectra. As shown in Table 2, the intensity ratio of the O 1s Table 2. XPS Peak Area Ratios Calculated for Elements in PVC and Plasticized PVC Films before and after Exposure to Glow Discharge Plasma peak area ratios PVC 25 wt % DEHP

Cl 2p/C 1s Cl 2p/C 1s after treatment Cl 2p/C 1s Cl 2p/C 1s after treatment

0.41 0.30 0.39 0.20

O 1s/C 1s O 1s/C 1s after treatment O 1s/C 1s O 1s/C 1s after treatment

N/A 0.17 0.023 0.22

peak versus the C 1s peak increases substantially for both PVC and PVC plasticized with 25 wt % DEHP after plasma treatment. On the contrary, the intensity ratio of the Cl 2p peak with respect to the C 1s peak decreases after plasma treatment. In addition, after plasma treatment, the lower bonding C 1s energy peak (C in CH2) is more intense than the higher bonding C 1s energy peak (C in CHCl), whereas previously, the higher bonding energy peak is of greater intensity. This C1s peak shape change also suggests that chlorine removal occurred. The appearance of a third C 1s peak suggests that a new form of carbon bonding is present after plasma treatment, perhaps because of the new bonding between C and O on the surface. Therefore, XPS data can interpret the contact angle decrease reported in the previous section and can be correlated to the SFG results as well. In the future, SFG signals will be collected from a lower frequency region, which should be able to be correlated to XPS results even further.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This research is supported by the National Science Foundation (NSF) (CHE 1111000). XPS spectra were obtained at the Electron Microbeam Analysis Laboratory of the University of Michigan. The authors thank Sabrina Peczonczyk for her support with XPS data analysis.

■

■

REFERENCES

(1) Koch, H. M.; Calafat, A. M. Philos. Trans. R. Soc. London, Ser. B 2009, 364 (1526), 2063−2078. (2) Li, N.; Wang, D. H.; Zhou, Y. Q.; Ma, M.; Li, J. A.; Wang, Z. J. Environ. Sci. Technol. 2010, 44 (17), 6863−6868. (3) Moore, C. J.; Moore, S. L.; Leecaster, M. K.; Weisberg, S. B. A. Mar. Pollut. Bull. 2001, 42 (12), 1297−1300. (4) Staples, C. A.; Peterson, D. R.; Parkerton, T. F.; Adams, W. J. Chemosphere 1997, 35 (4), 667−749. (5) Wypych, G. Handbook of Plasticizers; ChemTec Publishing: Toronto, Ontario, Canada, 2004.

CONCLUSION We have probed the surface structures of both annealed and non-annealed films of PVC and plasticized PVC before and after glow discharge plasma exposure using SFG and CARS. SFG spectra have revealed that our plasticizer of choice, DEHP, is present on lab-synthesized film surfaces at low concentrations. Along with the SFG spectra, CARS spectra revealed that there is a difference between molecular signatures on 4661

dx.doi.org/10.1021/la2045527 | Langmuir 2012, 28, 4654−4662

Langmuir

Article

(41) Bisson, P.; Xiao, H.; Kuo, M.; Kamelamela, N.; Shultz, M. J. J. Phys. Chem. A 2010, 114 (12), 4051−4057. (42) Lu, X. L.; Spanninga, S. A.; Kristalyn, C. B.; Chen, Z. Langmuir 2010, 26 (17), 14231−14235. (43) Lu, X. L.; Han, J. L.; Shephard, N.; Rhodes, S.; Martin, A. D.; Li, D. W.; Xue, G.; Chen, Z. J. Phys. Chem. B 2009, 113 (39), 12944− 12951. (44) Lu, X. L.; Li, D. W.; Kristalyn, C. B.; Han, J. L.; Shephard, N.; Rhodes, S.; Xue, G.; Chen, Z. Macromolecules 2009, 42 (22), 9052− 9057. (45) Chen, Z. Prog. Polym. Sci. 2010, 25 (11), 1376−1402. (46) Zhang, D.; Dougal, S. M.; Yeganeh, M. S. Langmuir 2000, 16 (10), 4528−4532. (47) Miyamae, T.; Yamada, Y.; Uyama, H.; Nozoye, H. Appl. Surf. Sci. 2001, 180, 126−137. (48) Miyamae, T.; Nozoye, H. J. Photochem. Photobiol., A 2001, 145, 93−99. (49) Miyamae, T.; Yamada, Y.; Uyama, H.; Nozoye, H. Surf. Sci. 2001, 493, 314−318. (50) Ye, H. K.; Gu, Z. Y.; Gracias, D. H. Langmuir 2006, 22 (4), 1863−1868. (51) Druet, S. A. J.; Taran, J.-P. E. Prog. Quantum Electron. 1981, 7, 1−72. (52) Cheng, J. X.; Xie, X. S. J. Phys. Chem. B 2004, 108 (3), 827. (53) Evans, C. L.; Xie, X. S. Annu. Rev. Anal. Chem. 2008, 1, 883− 909. (54) Audic, J. L.; Poncin-Epaillard, F.; Reyx, D.; Brosse, J. C. J. Appl. Polym. Sci. 2001, 79 (8), 1384−1393. (55) Kumagai, H.; Tashiro, T.; Kobayashi, T. J. Appl. Polym. Sci. 2005, 96 (2), 589−594. (56) Xiao-Jing, L.; Guan-Jun, Q.; Jie-Rong, C. Appl. Surf. Sci. 2008, 254 (20), 6568−6574. (57) Li, R.; Chen, J. R. Chin. Sci. Bull. 2006, 51 (5), 615−619. (58) Wen, X. A. Q. O.; Liu, X. H.; Liu, G. S. Vacuum 2010, 85 (3), 406−410. (59) Wang, J.; Chen, C. Y.; Buck, S. M.; Chen, Z. J. Phys. Chem. B 2001, 105 (48), 12118−12125. (60) Wang, J.; Woodcock, S. E.; Buck, S. M.; Chen, C. Y.; Chen, Z. J. Am. Chem. Soc. 2001, 123 (38), 9470−9471. (61) Zhang, C.; Wang, J.; Khmaladze, A.; Liu, Y. W.; Ding, B.; Jasensky, J.; Chen, Z. Opt. Lett. 2011, 36 (12), 2272−2274. (62) Lu, X.; Clarke, M. L.; Li, D.; Wang, X.; Xu, G.; Chen, Z. J. Phys. Chem. C 2011, 115 (28), 13759−13767. (63) Krimm, S. J. Polym. Sci., Part A: Gen. Pap. 1963, 1 (8), 2621− 2650. (64) Enomto, S.; Asahina., R. I. J. Polym. Sci., Part A-1: Polym. Chem. 1966, 4 (6), 1373−1390. (65) Snyder, R. G.; Strauss, H. L.; Elliger, C. A. J. Phys. Chem. 1982, 86 (26), 5145−5150. (66) Wei, X.; Zhuang, X.; Hong, S. C.; Goto, T.; Shen, Y. R. Phys. Rev. Lett. 1999, 82 (21), 4256−4259. (67) Ghoranneviss, M.; Shahidi, S.; Wiener, J. Plasma Sci. Technol. 2010, 12 (2), 204−207.

(6) Lorz, P. M.; Towae, F. K.; Enke, W.; Jäckh, R.; Bhargava, N.; Hillesheim, W. Phthalic acid and derivatives. In Ullmann’s Industrial Encyclopedia; Wiley-VCH Verlag GmbH and Co.: Weinheim, Germany, 2007. (7) Kamrin, M. A. J. Toxicol. Environ. Health, Part B 2009, 12 (2), 157−174. (8) Latini, G. Clin. Chim. Acta 2005, 361, 20−29. (9) Sathyanarayana, S. Curr. Probl. Pediatr. Adolesc. Health Care 2008, 38, 34−49. (10) Sathyanarayana, S.; Karr, C. J.; Lozano, P.; Brown, E.; Calafat, A. M.; Liu, F.; Swan, S. H. Pediatrics 2008, 121, e260−e268. (11) Buchta, C.; Bittner, C.; Heinzl, H.; Hocker, P.; Macher, M; Mayerhofer, M.; Schmid, R.; Seger, C.; Dettke, M. Transfusion 2005, 45 (5), 798−802. (12) Halden, R. U. Annu. Rev. Public Health 2010, 31, 179−194. (13) Ekelund, M.; Azhdar, B.; Gedde, U. W. Polym. Degrad. Stab. 2010, 95 (5), 1789−1793. (14) Xu, B.; Gao, N.; Cheng, H.; Xia, S.; Rui, M.; Zhao, D. J. Hazard. Mater. 2009, 162, 954−959. (15) Jaeger, R. J.; Rubin, R. J. Lancet 1970, 2 (7664), 151. (16) Rael, L. T.; Bar-Or, R.; Ambruso, D. R.; Mains, C. W.; Slone, D. S.; Craun, M. L.; Bar-Or, D. Oxid. Med. Cell. Longevity 2009, 2 (3), 166−171. (17) Oehlmann, J.; Schilte-Oehlmann, U.; Kloas, W.; Jagnytsch, O.; Lutz, I.; Kusk, K. O.; Wollenberger, L.; Santos, E. M.; Paull, G. C.; Van Look, K. J. W.; Tyler, C. R. Philos. Trans. R. Soc., B 2009, 364 (1526), 2047−2062. (18) Martino-Andrade, A. J.; Chahoud, I. Mol. Nutr. Food Res. 2010, 54, 148−157. (19) Api, A. M. Food Chem. Toxicol. 2001, 39 (2), 97−108. (20) Muncke, J. Sci. Total Environ. 2009, 407 (16), 4549−4559. (21) Fang, C. R.; Yao, J.; Zheng, Y. G.; Jiang, C. J.; Hu, L. F.; Wu, Y. Y.; Shen, D. S. Int. Biodeterior. Biodegrad. 2010, 64 (6), 442−446. (22) Zhao, X. K.; Yang, G. P.; Wang, Y. J.; Gao, X. C. J. Photochem. Photobiol., A 2004, 161, 215−220. (23) Ding, X. J.; An, T. C.; Li, G. Y.; Chen, J. X.; Sheng, G. Y.; Fu, J. M.; Zhao, J. C. Res. Chem. Intermed. 2008, 34, 67−83. (24) Bajt, O.; Mailhot, G.; Bolte, M. Appl. Catal., B 2001, 33 (3), 239−248. (25) Mailhot, G.; Sarakha, M.; Lavedrine, B.; Cáceres, J.; Malato, S. Chemosphere 2002, 49 (6), 525−532. (26) Cao, X. L. Compr. Rev. Food Sci. Food Saf. 2010, 9, 21−43. (27) Kovacic, P. Med. Hypothesis 2010, 74 (4), 626−628. (28) Shen, Y. R. The Principles of Nonlinear Optics; Wiley: New York, 1984. (29) Shen, Y. R. Nature 1989, 337 (6207), 519−525. (30) Chen, Z.; Shen, Y. R.; Somorjai, G. A. Annu. Rev. Phys. Chem. 2002, 53, 437−465. (31) Clarke, M. L.; Wang, J.; Chen, Z. Anal. Chem. 2003, 75, 3275− 3280. (32) Shultz, M. J.; Baldelli, S.; Schnitzer, C.; Simonelli, D. J. Phys. Chem. B 2002, 106 (21), 5313−5324. (33) Geiger, F. M. Annu. Rev. Phys. Chem. 2009, 60, 61−83. (34) Tarbuck, T. L.; Ota, S. T.; Richmond, G. L. J. Am. Chem. Soc. 2006, 128 (45), 14519−14527. (35) Watry, M. R.; Richmond, G. L. J. Am. Chem. Soc. 2000, 122 (5), 875−883. (36) Stokes, G. Y.; Buchbinder, A. M.; Gibbs-Davis, J. M.; Scheidt, K. A.; Geiger, F. M. Vib. Spectrosc. 2009, 50, 86−98. (37) Voges, A. B.; Al-Abadleh, H. A.; Musorrariti, M. J.; Bertin, P. A.; Nguyen, S. T.; Geiger, F. M. J. Phys. Chem. B 2004, 108 (48), 18675− 18682. (38) Callahan, K. M.; Casillas-Ituarte, N. N.; Roeselova, M.; Allen, H. C.; Tobias, D. J. J. Phys. Chem. A 2010, 114 (15), 5141−5148. (39) Liu, D. F.; Ma, G.; Allen, H. C. Environ. Sci. Technol. 2005, 39 (7), 2025−2032. (40) Xu, M.; Liu, D. F.; Allen, H. C. Environ. Sci. Technol. 2006, 40 (5), 1566−1572.

■

NOTE ADDED AFTER ASAP PUBLICATION This article was published ASAP on February 28, 2012. Figure 11 has been updated. The correct version was published on March 1, 2012.

4662

dx.doi.org/10.1021/la2045527 | Langmuir 2012, 28, 4654−4662