ARTICLE pubs.acs.org/Langmuir
Solving the Charging Effect in Insulating Materials Probed by a Variable Monoenergy Slow Positron Beam Wei-Song Hung,† Manuel De Guzman,† Quanfu An,§ Kueir-Rarn Lee,*,† Yan-Ching Jean,†,‡ and Juin-Yih Lai† †
R&D Center for Membrane Technology, Department of Chemical Engineering, Chung Yuan University, Chung-Li, 32023, Taiwan Department of Chemistry, University of Missouri Kansas City, Kansas City, Missouri 64110, United States § Department of Polymer Science and Engineering, Key Laboratory of Macromolecule Synthesis and Functionalization (Ministry of Education), Zhejiang University, Hangzhou 310027, China ‡
ABSTRACT: A variable monoenergy slow positron beam (VMSPB) operating at a high vacuum on insulating materials encounters a problem of significant surface charging effect with time. As a result, positronium formation is inhibited, and the positron annihilation radiation counting rate is reduced; these consequently distorted the experimental positron annihilation and results. To solve such problems, a technique of depositing an ultrathin layer of sputtering noble metals on insulators is developed. We report a successful method of sputtering a few atomic layers of platinum (∼1 nm) on a polyamide membrane to completely remove the charging effect for VMSPB applications in insulators.
1. INTRODUCTION A variable monoenergy slow positron beam (VMSPB) is a powerful probe that is capable of measuring the surface nanostructure and depth profile in metals, semiconductors, and polymers1-4 from scientific information based on positron annihilation signals in defects and free volumes. However, continuously implanting positrons onto insulating materials, VMSPB encounters a severe experimental problem of positive charge accumulating on the surface, commonly called the charging effect. Removal of the charging effect is necessary because it will distort the obtained positron annihilation signals and eventually stop the positron experiments. Two existing attempted techniques to partially overcome this problem are filling and flushing air or dry N2 to the beam (typically several hours)5 and implanting electrons into insulators using an electron gun.6 The former method is very tedious and ineffective, and the latter technique is not applicable for the positron annihilation lifetime spectroscopy (PALS) when one uses secondary electrons as starting signals. This research proposes a new approach: depositing an ultrathin Pt layer (∼1 nm) by means of a sputtering technique commonly used in SEM (scanning electron microscopy), which completely solves the positron charging effect in VMSPB applications in insulating materials. r 2011 American Chemical Society
2. EXPERIMENTAL SECTION 2.1. Materials. Polyacrylonitrile (PAN) polymer was supplied by Tong-Hua Synthesis Fiber Co. Ltd. in Taiwan. N-Methyl-2-pyrrolidone (NMP) of reagent grade was used as a solvent. Triethylenetetraamine (TETA) and trimesoyl chloride (TMC) were purchased from TCI Co. TETA was used as aqueous phase monomer and TMC as an organic phase monomer. These two monomers were reacted to carry out an interfacial polymerization process to form an active layer of poly(thiol ester amide). Distilled water was used in preparing aqueous aminothiol solution, and reagent-grade toluene was used as acyl chloride solvent. In the nitrogen filling method, 99.9% N2 gas was introduced into the VMSPB vacuum chamber. 2.2. Preparation of the Modified PAN (mPAN) Porous Support Membrane. Flat porous support membrane was prepared by casting a 15 wt % PAN-NMP solution onto nonwoven polyester fabrics with the use of a 200-μm casting knife. The cast membrane was precipitated by immersion in a bath of water. The resulting PAN porous membrane was washed in water overnight and was then dried at atmospheric conditions. To improve the hydrophilicity of the PAN membrane surface and to facilitate the spread of an aqueous amine solution on the surface, the PAN membrane was hydrolyzed in a 2 M Received: November 23, 2010 Revised: January 5, 2011 Published: February 18, 2011 3020
dx.doi.org/10.1021/la104660e | Langmuir 2011, 27, 3020–3023
Langmuir
ARTICLE
Table 1. Surface Resistance of a Pristine Membrane and Samples Sputtered with Platinum at Different Times, as Measured by a Multimeter sample
surface resistance (MΩ)
pristine
over limit
5 s Pt-sputtered
over limit
10 s Pt-sputtered
1.28 ( 0.13
100 s Pt-sputtered
0.87 ( 0.09
NaOH(aq) solution at 50 °C. The -CN groups of PAN on the surface of the support membrane could be converted into -COOH groups after hydrolysis with the NaOH(aq) solution. The following may also be formed: -COOH groups as a product of the hydrolysis and COONH2 or -COONH groups as byproducts.7 The resulting mPAN porous membrane support was washed in a water bath for several hours and was then dried under atmospheric conditions.
2.3. Preparation of the Polyamide TETA-TMC/mPAN Composite Membrane. The TETA-TMC/mPAN composite membrane was prepared using a technique of interfacial polymerization between TETA and TMC. The mPAN membrane was first immersed in a 2-wt% aqueous aminothiol solution for 1 min. Then, the excess amount of the solution on the membrane surface was removed. This membrane wet with the aminothiol solution was contacted with a 1 wt% organic acyl chloride solution for 1 min to carry out the process of interfacial polymerization. Finally, the resulting TETA-TMC/mPAN composite membrane was washed in methanol and then dried under atmospheric conditions. 2.4. Positron Annihilation Spectroscopy. The new VMSPB2 in the R&D Center for Membrane Technology in Chung Yuan University in Taiwan was used for this study to determine the mean depth in the membrane between 0 and about 13 μm (the mean depth was determined by converting the positron incident energy from 0 to 30 keV using an established equation8,9). This new radioisotope beam uses 50 mCi of 22Na as the positron source. The Doppler broadening energy spectroscopic (DBES) spectra were measured using an HP Ge detector at a counting rate of approximately 2000 cps. The energy resolution of the solid-state detector was 1.5 keV at 0.511 MeV (corresponding to positron 2γ annihilation peak). The total number of counts for a DBES spectrum was 1.0 million. The PALS data were obtained by taking the coincident events between the start signal detected by a multichannel plate (MCP) from the secondary electrons and the stop signal discerned by a BaF2 detector from the annihilation photons at a counting rate of approximately 200300 cps. A PALS spectrum contains 2.0 million counts. 2.5. Surface Resistance Characterization. The membrane surface resistance was measured with a digital multimeter (FLUKE, model 73IIIA Pt sputtering apparatus, typically applied in conjunction with an SEM instrument (Hitachi, model E-1045), maintained at a 6.0-Pa vacuum, used to deposit Pt particles on the insulating membrane. The deposition rate according to the manufacture’s specifications was 0.1 nm/s.
3. RESULTS AND DISCUSSION 3.1. Surface Resistance of Membrane Samples. The membrane surface conductivity directly influences the charging effect. Thus, we initially determined with a multimeter the surface resistance of a pristine membrane and different Pt-sputtered samples. The results tabulated in Table 1 indicate that the pristine membrane has an infinite surface resistance. At 5 s of Pt sputtering, the surface resistance remained infinite. However, at a 10 s sputtering time, the surface resistance dropped to 1.28 MΩ. With a much longer sputtering time of 100 s, the surface
Figure 1. Counting rate versus time elapsed at 2 keV positron incident energy for TETA-TMC/mPAN membranes: pristine and different samples treated by means of Pt-sputtering and N2-filling methods.
resistance further decreased. The result obtained at 5 s sputtering time could be ascribed to the uneven distribution of the deposited Pt particles, the amount of which was not enough to yield a link between them. At 10 s sputtering time, Pt particles deposited were sufficient to form a network structure, which completely removed the charging effect encountered in VMSPB application (as discussed below). 3.2. Charging Effect and Methods of Its Removal. To examine the charging effect and its solution, we reported the PALS counting rate versus time of data acquisition for pristine and 10 s Pt-sputtered membranes and compared these with the results from filling N2 at 2 h treatments. The results of comparing these two methods are displayed in Figure 1. For the pristine membrane (square symbols), the counting rate from ∼254 s-1 started to decrease rapidly after 300 min and approached 0 s-1 after ∼550 min. This is attributed to the accumulation of positive charge, which hinders positron implantation and thereby inhibits Ps formation. However, after treating the membrane by 10 s Pt sputtering to remove the charging effect, the PALS count (circle symbols) was maintained at a constant rate (∼254 s-1) for days to weeks. A typical PALS spectrum took about 2-3 h of data acquisition for a total count of 1 million. However, for the case of treating the membrane by another method called N2 filling for 2 h and then conducting a PALS test again (upright triangular symbols), the counting rate was restored to only 80% of the initial rate (∼200 s-1). Then, the counting rate decreased rapidly again and approached 0 s-1 after 600 min. For the successive second and third 2 h nitrogen fillings and PALS test again after each filling (inverted triangular and diamond symbols), the respective counting rates from 175 s-1 and 150 s-1 decreased rapidly again, and the PALS test times were prolonged to 650 and 750 min at which a 0 s-1 counting rate was attained. These results suggest that each nitrogen filling cannot completely remove the charging effect, leading to a decreased counting rate. The comparatively low counting rate slowed down the accumulation rate of positive charge, resulting in further extension of the PALS test time. 3.3. PALS Experiments with and without Pt Sputtering Technique. Figure 2 shows a comparison of a pristine membrane and a 10 s Pt-sputtered membrane. These membranes evidently showed different raw PALS data (a PALS spectrum contains 2.0 million counts) at same positron implantation depth (at 2 keV). For the 10-s Pt-sputtered membrane, the obtained PALS spectrum showed good data as did previous results.1,2 However, the pristine 3021
dx.doi.org/10.1021/la104660e |Langmuir 2011, 27, 3020–3023
Langmuir
ARTICLE
Figure 2. Normalized PALS spectra for pristine and Pt-sputtered TETA-TMC/mPAN membranes at 2 keV positron incident energy.
Figure 4. S and R parameters versus positron incident energy data for TETA-TMC/mPAN membranes (pristine and Pt-sputtered) and for pure Pt.
Figure 3. o-Ps annihilation lifetime (τ3) and intensity (I3) versus positron incident energy (depth) for pristine and Pt-sputtered TETATMC/mPAN membranes.
membrane PALS spectrum was distorted: a higher background and a broader full width at half-maximum (fwhm). This comparatively higher background could be attributed to the decreased counting rate associated with the charging effect. Consequently, longer measurement times would lead to noise detected from scattered photons and electrons. The relatively broader fwhm could be attributed to the charging effect intensifying with time at random effective positron incident energy. Figure 3 shows the analyzed results of orthopositronium (o-Ps) lifetime and intensity versus positron incident energy for a pristine membrane and a sample sputtered with Pt for 10 s. All PALS spectra were analyzed in three lifetimes after the longest lifetime component (>12 ns, due to Ps backscattered and annihilated in the beam) was removed using the PATFIT program.10 Only three o-Ps τ3 and I3 data points at different positron implantation energies could be obtained for the pristine membrane within a 0-2 keV incident energy range. From low to high positron implantation energies (0-2 keV), τ3 and I3 for the pristine membrane had relatively larger differences and deviations compared to the corresponding results obtained for the Pt-sputtered sample. Each o-Ps data point for the pristine membrane took about 120 min. Thus, the first two points were obtained after
300 min, at which the charging effect was not yet significant (see Figure 1); however, the total time needed for the third point was 360 min. Hence, this third data point was significantly different from the first two points, and it was not possible to continue with the PALS experiment after the third data point due to the charging effect. However, for the 10-s Pt-sputtered sample, we were able to completely remove the charging effect, and this resulted in uninterrupted PALS data. The sputtering technique for PALS measurements, therefore, made it possible to conduct depth profiling and obtain free volume information. These data are shown in Figure 3, from which it can be deduced that the dense thin polyamide film (TETA-TMC) is in the region between 0 and 2 keV. The transition layer of the composite membrane (dense TETA-TMC film þ dense mPAN skin layer) is located between 2 and 3 keV, whereas the porous support mPAN layer constitutes in the region >3 keV. 3.4. Validating PAS Data with DBES Experiments. The above technique needs further proof that Pt sputtering for 10 s (∼1 nm coating on the surface) does not affect layer and free volume results. We conducted experiments using DBES, which is capable of determining qualitative information about free volumes and layer structure of multilayer membranes.1,2 The results of the S parameter (which is the relative value of the central part of annihilation radiation at 511 keV) and the R parameter (which is the relative valley part of annihilation radiations) versus positron incident energy are shown in Figure 4. In general, S measures free volumes between 0.1 and 1 nm based on 2γ annihilation, and R measures pores in the nanometermicrometer range11 based on 3γ annihilation. In Figure 4, DBES measurement (counting rate ∼2500 cps) times for both samples 3022
dx.doi.org/10.1021/la104660e |Langmuir 2011, 27, 3020–3023
Langmuir are all
