Assessing the Molecular Weight of a Conducting Polymer by Grazing

Grazing emission X-ray fluorescence spectrometry (GEXRF) is a relatively new variant of XRF, capable of detecting elements with a lower atomic number ...
0 downloads 0 Views 40KB Size
Anal. Chem. 2000, 72, 3366-3368

Assessing the Molecular Weight of a Conducting Polymer by Grazing Emission XRF F. Blockhuys, M. Claes, R. Van Grieken, and H. J. Geise*

Department of Chemistry, University of Antwerpen (UIA), Universiteitsplein 1, B-2610 Antwerpen (Wilrijk), Belgium

A spray coated thin film of all-E-poly[(3-methoxy-2,5thiophenediyl-1′,2′-ethenediyl)(4′′-methoxy-2′′,5′′thiophenediyl-1′′,2′′-ethenediyl)], a conducting polymer, was analyzed by grazing emission X-ray fluorescence spectrometry. Measuring the S/Cl ratio in this sulfurcontaining polymer with one terminal chlorine allowed us to determine the chain length and, hence, the molecular weight. The result was in acceptable agreement with that of a well-established method for the determination of the molecular weight of polymeric materials. Grazing emission X-ray fluorescence spectrometry (GEXRF) is a relatively new variant of XRF, capable of detecting elements with a lower atomic number than the more conventional totalreflection XRF (TXRF) setup.1 The main difference between the two techniques is that in the GEXRF setup the emerging characteristic fluorescent radiation is measured under a very small angle (typically a few mrad) over a flat carrier, while in TXRF the exciting radiation is at grazing incidence and the characteristic radiation is measured under 90°. Conventional TXRF with energydispersive X-ray detection is capable of measuring elements from Z ) 14 (Si) upward with a reasonable sensitivity, while GEXRF, in combination with wavelength-dispersive X-ray detection, can measure elements from Z ) 4 (Be). Recently, a study has been performed on the optimization of the GEXRF technique for the analysis of a variety of samples.2-6 We report here on the use of this new technique on a sample with an unknown morphology and a composition rather different from the other samples tested so far.2-6 We were interested in catalyst and reagent residues left after purification in all-E-poly[(3-methoxy-2,5-thiophenediyl-1′,2′-ethenediyl)(4′′-methoxy-2′′,5′′-thiophenediyl-1′′,2′′-ethenediyl)] (abbreviated MOPTV), which was synthesized according to the reaction sequence shown in Figure 1.7,8 The determination of nickel, * To whom correspondence should be addressed. (1) Klockenka¨mper R. Total Reflection X-ray Fluorescence Analysis; John Wiley and Sons: New York, 1997. (2) Claes, M. Ph.D. Thesis, University of Antwerpen (UIA), Belgium, 1999. (3) Claes, M.; de Bokx, P.; Van Grieken, R. X-ray Spectrosc. 1999, 28, 224. (4) Claes, M.; Van Grieken, R.; de Bokx, P. X-ray Spectrosc. 1997, 26, 153. (5) Claes, M.; de Bokx, P.; Willard, N.; Veny, P.; Van Grieken, R. Spectrochim. Acta B 1997, 52, 1063. (6) Claes, M.; Van Dyck, K.; Deelstra, H.; Van Grieken, R. Spectrochim. Acta B 1999, 54, 1517. (7) Jen, K. Y.; Eckhardt, H.; Jow, T. R.; Shacklette, L. W.; Elsenbaumer, R. L. J. Chem. Soc., Chem. Commun. 1988, 215. (8) Jen, K. A.; Elsenbaumer, R. L.; Shacklette, L. W. Int. Pat. WO 88 00945 A1, 1988; 39.

3366 Analytical Chemistry, Vol. 72, No. 14, July 15, 2000

Figure 1. Reaction sequence in the polymerization of dimer 1 to MOPTV (Ø ) phenyl).

chlorine, sulfur, phosphorus, and magnesium would provide information not only directly on the efficiency of the purification method but also indirectly on the average molecular weight, Mw, of the polymer, from the atomic ratio S:Cl (2n:1; see Figure 1). Since we were interested in accurately determining the amount of magnesium (Z ) 12) in the sample and since the determination of this element is beyond the TXRF method, we applied GEXRF to the analysis of MOPTV. EXPERIMENTAL SECTION Measurements were performed with a research prototype GEXRF spectrometer developed by modifying a commercial PW2400 wavelength-dispersive XRF spectrometer (Philips Analytical, Almelo, The Netherlands).9 A CH2Cl2 solution of MOPTV was spray coated as a circular spot, with a diameter of 28 mm and an average thickness of 780 nm, onto a flat silicon substrate and directly analyzed with GEXRF. Measurements were performed at a current of 70 mA and a voltage of 40 kV in technical vacuum, to avoid absorption of the longer wavelengths in air. Figure 2 shows the obtained elemental spectra, applying a 500 µm slit width and an emission angle of 20 mrad. Scans were recorded at a speed of 0.005 deg/s and an integration time of 5 s. Afterward, the elemental intensities were measured at their respective 2 Å channels. Relevant qualitative analysis parameters for each element under study are summarized in Table 1. Typical counting times for each element were 1000 s, including the background off-peak measurements. From the measured net intensities IX (9) de Bokx, P. K.; Urbach, H. P. Rev. Sci. Instrum. 1995, 66, 15. 10.1021/ac990877k CCC: $19.00

© 2000 American Chemical Society Published on Web 05/31/2000

Figure 2. Combined spectra of the five measured elements in the different ranges of 2 Å. Table 1. Relevant Analysis Parameters

X

energy (keV)

cryst type

Mg P S Cl Ni

1.25 2.01 2.31 2.62 7.47

PX1 PE PE PE LiF200

element bkgd 2θ 2θ angle angle (deg) (deg) 22.79 89.50 75.69 65.35 48.67

20.93 91 74 67.68 49.66

detector type flow counter flow counter flow counter flow counter scintillator

measurement emission slit angle width (mrad) (µm) 13.24 8.12 7.14 6.26 2.19

500 500 500 500 100

Table 2. Net Intensity IX (counts/300 s), Sensitivity SX (counts/s‚µg), and Amount mX (µg; See Text) for the Five Elements Considered X

IX

SX

mX

Mg Cl S P Ni

1 554 009 223 347 5 136 550 29 771 66 676

0.347 0.847 0.911 0.826 5.16

14 800 879 20 200 120 43

(corrected for background contribution) and the known sensitivities SX,2-6 the amount mX of element X can be calculated. Table 2 summarizes the obtained data. The accuracy and reproducibility of the technique have been reported.2-6 RESULTS AND DISCUSSION MOPTV was synthesized by a well-established method in the field of conducting polymer research, namely the “nickel phosphine complex catalyzed Grignard cross-coupling” which applies a nickel complex to couple an organomagnesium (Grignard) compound to an organic halide. The most widely used catalytic complex, and the one we used here as well, is trans-dichloro[1,3bis(diphenylphosphino)propane]nickel(II), abbreviated NiCl2(dppp) or [Ni] in Figure 1. The Grignard compound is synthesized by converting an organolithium compound, produced in a standard lithiation reaction using butyllithium, into a Grignard reagent using an etherate complex of magnesium bromide. Of all the reagents

used in this polymerization reaction, both the magnesium salt and the nickel complex are possible contaminations in the final polymer. However, it has never been investigated, by any method in any of the similar syntheses that have been reported on, whether these are removed either by precipitation during the reaction or by the subsequent purification process (usually, a Soxhlet extraction with ethanol). With possible applications of these conducting organic systems in mind, we were very interested in determining the residual presence of magnesium and nickel compounds in the polymer. Since the morphology and composition of the sample differs substantially from what is usual in XRF analysis, we were unable to apply standard analytical and statistical procedures to the investigation, in particular methods in which there is compensation for the effects of the matrix, the composition of the substrate and the thickness of the film using external standards. In this case we apply a relative method in which we use relative amounts of the elements under consideration, within the same matrix. This way, we do not need to consider matrix or substrate effects or the influence of the film thickness, nor are we forced to use external standards. Since, furthermore, the tabulated sensitivities SX have been proven to be independent of the matrix,2-6 the conclusions we draw are as valid as when standard analytical techniques are applied. Impurities in the Polymer. While GEXRF is capable of accurately measuring low-Z elements, the instrument used in the present study was not optimized for heavy atom (Z > 24) analysis and generally high background levels for those elements were observed. Therefore, the measurement on nickel, to determine the amount of residual catalytic complex, must be treated with caution (vide infra). Also, the amount of bromine was not determined, because the amount of residual magnesium compounds can more easily be obtained by directly measuring the amount of magnesium. From the obtained data for magnesium and nickel (Table 2), one can conclude the following. First, the amount of catalytic complex after purification is rather low and can be neglected with Analytical Chemistry, Vol. 72, No. 14, July 15, 2000

3367

respect to the polymer. Second, the amount of magnesium salt is much higher. However, in the polymerization reaction 2 equiv of magnesium is added to 1 equiv of dilithiated dimer (1) in order to obtain a quantitative amount of di-Grignard reagent. Without any purification one would expect an amount of magnesium equal to that of sulfur; in contrast, we find that a great deal has been removed in the purification process. Even though the final polymer is somewhat contaminated with organometallic compounds, preliminary experiments on applications based on the electrical conductivity of MOPTV suggest that these impurities do not greatly affect the polymer’s performance. Determination of the Average Molecular Weight of MOPTV. To determine the average molecular weight of MOPTV we have to start from the relative amounts of phosohorus and nickel in order to arrive ultimately at the atomic ratio of S and Cl in the polymer. According to the stoichiometry of the catalytic complex the relative amounts of P:Ni should be 2:1 but seem (see Table 2) to be rather 3:1. We attribute the deviation of the experimental P:Ni ratio from the expected value to the intrinsic background problems of the GEXRF technique with heavier elements. Thus, the determination of the nickel amount in the sample is slightly off. Fortunately, this is of no consequence in the relative method we apply to estimate the chain length. We know that in the catalytic complex the P:Ni ratio must be equal to the Cl:Ni ratio, and hence, we take the experimental P:Ni ) 3:1 ratio to estimate the amount of chlorine in residual catalytic complex as 3 × 43 units. Then, the amount of chlorine in the polymer, i.e., the total amount of chlorine minus that in the catalytic residue, is estimated as 879-3 × 43)750 units. From this value and that of sulfur it can be seen (Table 2) that the relative amounts of sulfur and chlorine in the polymer are approximately 26 to 1. This means that there is one Cl atom per 13 dimeric units in the polymer, because each dimeric unit contains two thiophene rings and thus two S atoms. The choice of one chlorine atom instead of two (at each end of the chain) is based on the relative amounts of dimer (1) and

3368

Analytical Chemistry, Vol. 72, No. 14, July 15, 2000

E-1,2-dichloroethene used in this polymerization reaction, i.e., 1:1. The structure obtained in this way is displayed in Figure 1 with n ) 13. The fact that only one chlorine is to be expected in the polymer under these circumstances has been verified spectroscopically in similar polymers synthesized by the same method of polymerization.7 The molecular mass of this chain is thus 3624, which may be compared to the value of 3298 obtained by gel permeation chromatography (GPC), which is a standard technique in polymer chemistry to accurately determine the various molecular weights of polymeric materials. The agreement is rather satisfactory, particularly when one takes into account that the GEXRF measuring technique is not yet optimized for these types of samples: on the basis of the GEXRF data, the molecular weight of MOPTV is overestimated by only 10%. Even though this difference is large by analytical chemical standards, both molecular weights have the same order of magnitude and both measurements can be used to unequivocally assign a value to the weight average molecular weight (Mw) and an approximate average chain length of this sample of MOPTV. The satisfactory concordance between these very different analytical methods suggests and substantiates the use of GEXRF (or other techniques based on XRF) in the elemental or compositional analysis of organic materials. ACKNOWLEDGMENT Financial support to the Structural Chemistry Laboratory by the Belgian Organization IWT (Project ITA II) and by the university (GOA-BOF-UA; Project 23; 96-99) is gratefully acknowledged.

Received for review August 4, 1999. Accepted April 19, 2000. AC990877K