Technical Note Cite This: Anal. Chem. XXXX, XXX, XXX−XXX
pubs.acs.org/ac
Quality Control of a Functionalized Polymer Catalyst by Energy Dispersive X‑ray Spectrometry (EDX or EDS) Daphne ́ Hector,† Sandra Olivero,† François Orange,‡ Elisabet Duñach,† and Jean-François Gal*,† †
Université Côte d’Azur, CNRS, Institut de Chimie de Nice, UMR7272, 06108 Nice, France Université Côte d’Azur, Centre Commun de Microscopie Appliquée (CCMA), 06108 Nice, France
‡
Anal. Chem. Downloaded from pubs.acs.org by IOWA STATE UNIV on 01/09/19. For personal use only.
S Supporting Information *
ABSTRACT: Energy dispersive X-ray spectrometry (EDX or EDS) is a technique often implemented on scanning electron microscopes and a regularly used method for qualitative characterization of solid catalysts. This Technical Note reports a method for the determination of the metal content in a sulfonated polyether ether ketone in the form of an indium(III) salt. The possibility of quantitative determination of the sulfur/ indium ratio by EDX was assessed by calibration with two indium salts (sulfide and sulfate) readily available in good purity. The accuracy of the uncorrected instrument response was better than 1% under our conditions. A protocol for investigating the metal content of the solid catalyst is proposed, also providing information about the homogeneity of the metal distribution. Because of the simplicity of the sample preparation, the small quantity of material needed, and the rapidity of the EDX measurements, the method appears to be promising for quantitative characterization of solid catalysts.
E
(TEM), also used in EDX, can be more demanding on sample preparation when it requires that ultrathin sections be obtained (50−100 nm). On modern instruments, EDX analytical capabilities apply to most elements, typically from Be to U, with better efficiency for heavy elements. From the viewpoint of elemental analysis, the results are therefore expected to be fast and easily acquired for a wide variety of samples, although low detection limits cannot be expected. One drawback of EDX is the limited energy resolution (∼100−150 eV), leading some elements to have partially overlapping characteristic wavelengths. When interference appears to be possible, the problem can be minimized by a careful choice of analytical lines. Current EDX acquisition software also includes peak deconvolution algorithms, so quantitative peak deconvolution may be achieved, at the price of added uncertainties. EDX probes a volume of a few cubic micrometers, depending on the incident electron beam energy and the material.3 Geometric effects may arise when the sample size is on the same order of magnitude or smaller than this sampled volume, and corrections may be applied.6,7 The SEM/EDX technique is typically applied to inorganic analysis.8 The conditions for obtaining quantitative results from SEM/ EDX, considering only inorganic samples, were recently reviewed.9 The most recent reviews on polymers and supported catalysts focused on uses of electron microscopy
nergy dispersive X-ray spectrometry (EDX or EDS) is often cited as a characterization method for solid catalysts, essentially applied to metal and inorganic materials,1 but seldom applied to quantitative determinations using some validation criteria. As an X-ray microanalysis technique, EDX is readily available on scanning and transmission electron microscopes, under the generic name of analytical electron microscopy or AEM.2−5 Scanning electron microscopy (SEM) is a widespread technique, giving contrasted images of specimen surfaces displaying a three-dimensional aspect with an ultimate resolution of a few nanometers. The high-energy electron beam utilized for imaging the sample induces ionization of inner electrons of the elements. X-ray photons are emitted when electrons from the upper levels release energy by filling holes left by ionization and are analyzed by counting the pulses generated by an energy-sensitive detector. A large part of the literature on AEM deals with metal alloys, semiconductors, minerals such as oxide and salts, often for material sciences or geological or archeological purposes. As already mentioned, in the domain of catalysis,1 AEM techniques are consequently mostly applied to metal surfaces and nanoparticles. The primary application of SEM is the characterization of the surface morphology (imaging and mapping) of solids, and the identification of the constituent elements is a valuable benefit of the technique. Sample preparation for SEM investigation is relatively simple by comparison with those of other imaging techniques and is practically reduced to sample metallization when nonconducting materials are examined. Transmission electron microscopy © XXXX American Chemical Society
Received: September 12, 2018 Accepted: December 22, 2018 Published: December 23, 2018 A
DOI: 10.1021/acs.analchem.8b04170 Anal. Chem. XXXX, XXX, XXX−XXX
Technical Note
Analytical Chemistry
Standards used for the S/In ratio were indium sulfide (99.995%) and indium sulfate (99.99%) from Alfa Aesar (Schiltigheim, France). Instrumentation and Methods. Polymers and indium salts were observed and analyzed with a Tescan Vega 3 XMU scanning electron microscope (TESCAN FRANCE, Fuveau, France) equipped with an X-MaxN 50 EDX detector (Oxford Instruments, Tubney Woods, Abingdon, Oxfordshire, U.K.). The EDX detector is equipped with a superatmospheric thin window (SATW) and a 50 mm2 silicon drift detector (SDD), placed at a takeoff angle of 30°, with a resolution of 129 eV at the Mn K-L3 line (in IUPAC notation; Siegbahn notation Mn Kα1; see the Supporting Information for the equivalence) under the conditions used for this study. The EDX detector is monitored by AZtec (version 3.1, Oxford Instruments) for acquisition and processing of EDX spectra. Monte Carlo simulations were performed with CASINO20 (version 2.5.1) to assess the maximum depth reached by the electron beam (20 kV, 104 electron trajectories, beam radius of 170 nm, 50 nm carbon coating) in the different materials analyzed and to adapt accordingly the choice of the tested particles. Samples were carbon-coated prior to analyses. All analyses were performed under identical conditions: under vacuum, at an acceleration voltage of 20 kV, a 10 mm working distance, a 1500× magnification, and a beam size of 170 nm. A few tests on the effect of acceleration voltage (15 and 30 kV) and Au/ Pd metal coating are reported in the Supporting Information. Under these conditions, probe currents of 1.1−1.5 nA were measured by pointing the beam in a Faraday cup. This allowed spectral acquisition rates of 10000−15000 counts per seconds (cps) with a dead time of ∼10%. The total acquisition time was automatically set by AZtec and usually did not exceed 30 s with total numbers of counts of ∼310000−320000. Prior to each analysis, beam measurements were performed using a Mn standard (Calibration stub #5, Ardennes Analytique, Stavelot, Belgium) using the same beam parameters described above. For EDX spectral processing, AZtec uses a filtered leastsquares (FLS) approach for background removal, fitting, and deconvolution of the peaks, which also includes a pulse pile-up correction. Quantitative data are obtained from the spectra after matrix correction, which follows the XPP exponential model procedure.21,22 All of the details of EDX spectral processing software can be found in the AZtecEnergy In Depth application note.23 Because of the large amounts of light elements in the samples, we did not seek to obtain full quantitative data but instead focused on the S/In ratio. Using AZtec, this was made by indicating that the sample was coated with a 50 nm carbon layer, selecting all the elements other than sulfur and indium as “deconvolution elements” (which means that their presence is taken into account for the calculation but that they are not displayed in the final result), and using the “Quant Standardization (Extended Set)” provided with the software. Ratios were calculated with the sulfur K-series peaks and the indium L-series peaks. S/Na ratios were obtained using the same procedure. EDX spectra showed clear and well-separated sulfur and indium peaks, with the main S K-L3 (S Kα1) peak at 2.31 keV, followed by a smaller S K-M3 (Kβ1) peak at 2.46 keV, and the main In L3-M5 (Lα1) peak at 3.29 keV, followed by decreasing other L-series peaks [L3-M4 (Lβ1), L3-N5 (Lβ2), and L2-N4 (Lγ1) at 3.49, 3.71, and 3.92 keV, respectively].
for morphology and imaging, and elemental quantitation by EDX does not appear to be currently advanced.1,10−13 In the works cited in these reviews, metal loadings of polymersupported catalysts are always reported as uncalibrated values. As EDX is more efficient on heavy elements, quantitation of polymers is indeed made difficult by their high content of light elements (such as C, N, O, etc.). We note only one TEM study of organic and organometallic thin films explicitly devoted to quantitative EDX.14 This note focuses on quantitative aspects of SEM/EDX analysis of a functional polymer loaded with indium(III) as the catalytically active species. Indium salts are known as catalyst active on a wide variety of reactions.15,16 Recently, metal salts of sulfonated poly(ether ether ketone) (SPEEK) were proposed as catalysts.17 One of the important advantages of these catalysts is that they can be easily recovered from the reaction medium for further use. On the other hand, they are more difficult to analyze by methods applied to soluble salts. Because of the promising properties of the SPEEK In(III) salt solid catalyst,17 we chose to investigate the possibilities of EDX analysis on this material. Our aim was to propose a simple and robust quantitative analytical protocol, to assess the composition, especially the metal loading of the catalyst, as a complementary addition to the imaging and morphology characterization of the solid material.
■
EXPERIMENTAL SECTION Materials. The monosulfonic acid derivative of poly(ether ether ketone) (PEEK, KetaSpire KT-820 NT, Solvay Specialty Polymers, average molecular weight of 51900; weight-average molecular weight of 153600)18 was prepared by chlorosulfonation, followed by hydrolysis to give the sulfonic acid SPEEK (Scheme 1).17 Scheme 1. Structure of the Acid Form of Sulfonated PEEK (SPEEK)
The indium SPEEK salts were obtained by reaction with indium metal (100 mesh, 99.99%, Sigma-Aldrich-Merck, SaintQuentin Fallavier, France) under ultrasonic activation, as described previously,17 using a 3/1 SPEEK/In stoichiometric ratio. Alternatively, indium SPEEK salts were prepared by exchange reaction with indium acetate (Sigma-Aldrich-Merck, 99.99% trace metals basis), by mixing a 3/1 SPEEK/In(OAc)3 mixture in water, stirring at room temperature for 3 h, and further evaporating the displaced acetic acid. A different SPEEK polymer salt containing both In3+ and Na+ cations was prepared with an equimolar ratio of indium acetate and sodium hydroxide. A first treatment of SPEEK (4 equiv) with an aqueous solution of sodium hydroxide (1 equiv) was followed by the addition of indium acetate (1 equiv) with stirring at room temperature for 3 h and further evaporation of acetic acid and H2O. The analytical method was also applied to a SPEEK/In used three times as a catalyst in the addition of thiols to olefins.19 This sample is called “recycled” hereafter. B
DOI: 10.1021/acs.analchem.8b04170 Anal. Chem. XXXX, XXX, XXX−XXX
Technical Note
SPEEK-In/Na [In(OAc)3 + NaOH] SPEEK-In [In(0)] SPEEK-In [In(OAc)3] SPEEK-In [In(OAc)3] recycled
The elemental composition is given in mass percent, and the S/In ratio in mole percent. bThe synthesis method is detailed in the Experimental Section. cCalculated for a perfect In(III) polymeric salt [C57H33S3O18In]n neglecting the polymer end group. dElements C, H, and S assessed by combustion; In assessed by ICP-AES. eCalculated for a perfect equimolar mixture of [C57H33S3O18In]n and [C19H11SO6Na]n. The remainder is Na (1.43%). fInclude also Na for this mixed salt. gRepeatability test. a
S/In % O (by difference)
36.58f 35.38 33.05 (32.96) 33.41 (33.38) 63.42 64.62 66.95 (67.04) 66.59 (66.62)
∑% of elements In, C, H, and S %S
7.96 6.02 7.12 7.73 3.53 3.60 3.77 (3.80)g 3.68 (3.65)g 46.59 46.07 45.37 (45.43)g 45.62 (45.68)g
%H % In
5.34 8.93 10.69 9.56 4 3 3 3
S/In %O %S %C
56.79e 56.26 56.26 56.26
% In
7.14e 9.44 9.44 9.44
sample (synthesis method)b
%H
23.89e 23.66 23.66 23.66
%C C
7.98e 7.91 7.91 7.91
measured compositiond
RESULTS AND DISCUSSION This study was prompted by the difficulties of characterizing metal salts derived from SPEEK. These salts appear to be relatively intractable solids, which should be mineralized under harsh conditions to be analyzed for metal content. The composition of organic elements (C, H, O, and S) is also difficult to obtain. As a solid recyclable catalyst, the indium(III) salt of SPEEK is currently being studied in our laboratory.17 In this context, the metal content of the SPEEK salt is especially important. Preliminary analytical data came from elemental analyses reported in Table 1. It appears that the metal content is significantly different from what is expected from a basic application of the stoichiometry of reagents used in the preparation of the SPEEK salts. An In(III) sulfonate should have a theoretical S/ In ratio of 3, and the ratio found for the two SPEEK-In salts was ∼2.4 (see Table 1), with the C and S composition remaining close to the stoichiometric value. The composition of SPEEK-In/Na and recycled SPEEK-In will be considered
stoichiometric compositionc
Table 1. Elemental Analysis of SPEEK Salts Obtained by Combustion and ICP-AESa
■
2.76e 2.73 2.73 2.73
The quality of the signal collected by the EDX detector for each analysis was evaluated using the extent of the background noise and the value of the un-normalized total mass ratios. Only analyses with a background noise at energies lower than 20 keV and total un-normalized mass ratios between 96 and 104% were considered [total of the mass ratios of all of the elements present; 100% being the value of the signal obtained on the Mn standard on the Mn K-L3 (Kα1) and Mn K-M3 (Kβ1) peaks]. Outside this range, the S/In ratios become less consistent (see the Supporting Information). The effect of particle shape and size on quantitative measurements was verified on both standard indium salts and SPEEK salts after being addressed by a reviewer. In particular, in the case of the In2S3 standard, it was presumed that the smaller particle size rough particle shape of the salt (used as received) may induce bias in the spectral intensities. This hypothesis was tested through random analyses of the material, and results are listed in the Supporting Information. It appears that S/In ratio may vary significantly when the signal is too weak or too intense but also that variations are minimal when the signal intensity remains close to that measured on the standard (90−110% range), thus validating the 96−104% criterion. A similar test on polymers shows the same trends for valid analysis selection. Ten spectra were acquired on each sample on different particles. In addition, the reproducibility and repeatability of analyses were verified on one catalyst and one indium salt standard by performing a set of 10 analyses at several-week intervals and recording three successive analyses at the same locations (see the Supporting Information). Elemental analyses were performed by the “Service Central d’Analyses” (ISA, Villeurbanne, France) for C, H, and S and by CREALINS (Genay, France) using inductively coupled plasma atomic emission spectrophotometry (ICP-AES) for indium. Safety Considerations. The manipulation of the chlorinated reactants used in the SPEEK synthesis (chlorosulfonic acid and thionyl chloride), which are highly corrosive and irritating, should be manipulated with caution under a wellventilated hood. Insufficient data are available on the effect of indium derivatives on human health. Indium metal is considered as relatively nontoxic, but indium compounds may be regarded as toxic. Therefore, inhalation of indium-containing powders must be avoided.
5.34 2.41 2.39 2.90
Analytical Chemistry
DOI: 10.1021/acs.analchem.8b04170 Anal. Chem. XXXX, XXX, XXX−XXX
Technical Note
Analytical Chemistry later in light of complementary EDX data. For a more efficient method aimed at the characterization of samples of SPEEK-In salts prepared for catalytic tests, we foresee EDX energy dispersive X-ray spectrometry as a microanalytical technique, readily available on electron microscopes, and applicable to solids with minimal sample preparation. The indium SPEEK salts were analyzed by SEM in the form used for catalytic assays. The powder samples appear in SEM images as fragmented glassy solids with particles (chunk or flakes) roughly 20−200 μm in size. Sulfur, oxygen, and indium appeared to be evenly distributed in EDX maps. In some cases, we observed on the solid a few
