J. MATER. CHEM., 1991, 1(4), 525-529 Characterization of Conducting Polymer-Quartz Composites Steven P. Armes,*a S. Gottesfeld,*J. G. Beery,* F. Garzon,bC. Mombourquette,*M. Hawley," and H. H. Kuhd a School of Chemistry and Molecular Sciences, University of Sussex, Brighton, BN I 9QJ, UK " Materials Science and Technology Division, Electronics Research Group, Los Alamos National Laboratory, Los Alamos, NM 87545. USA Los Alamos National Laboratory, Los Alamos, NM 87545, USA Milliken Research Corporation, P.0. Box 1927, M-405, Spartanburg, SC 29304, USA We have characterized several polypyrrole-quartz and polyaniline-quartz composites over a range of con- ducting-polymer loading levels by thermogravimetric analysis (TG), Rutherford backscattering spectrometry (RBS), scanning electron microscopy (SEM), scanning tunnelling microscopy (STM) and conductivity measure- ments.It is shown that the conducting polymer overlayers are remarkably thin and uniform; film thicknesses were determined independently by TG and RBS and are in close agreement. The film thickness of one of the polypyrrole-quartz samples was used to obtain the first direct measurement of the conductivity of the polypyrrole component (ca.35 Q-' cm-'). Keywords: Polypyrrole; Conducting polymer; Organic-inorganic composite It is well known that the successful commercial exploitation of organic conducting polymers, such as polyacetylene, poly- pyrrole, or polyaniline, has been frustrated by these materials' poor environmental stability and/or intractability.Previous workers have reported many novel organic and inorganic composites and thin films containing a conducting-polymer component.' Recently, the Milliken Research Corporation reported a new technique that involves the in situ deposition of polypyrrole or polyaniline onto various textile substrates such as nylon or polye~ter.~-~ Although not yet fully under- stood from a fundamental viewpoint, this process has been ~atented,~and it now seems likely that the much-vaunted potential of conducting polymers as materials with novel electronic applications will finally be realized. These conducting polymer-textile composites have pre- viously been characterized by various techniques, including SEM, resistance measurements and X-ray photoelectron spec- troscopy (XPS).2-4 One problem that emerged during these earlier studies was the difficulties encountered in estimating the mass and average thickness of the deposited conducting polymer overlayer.The latter parameter is essential if com- posite conductivities (Q-' cm-') rather than resistances (Q/ 0)are to be determined. By using thermally stable quartz or glass fibres rather than conventional textile materials as the substrate, it was shown that the mass loading of the con- ducting polymer could be directly determined by TG.4 In this report we have characterized several polypyrrole- quartz and polyaniline-quartz composites by TG, RBS, and SEM. The combined use of these techniques has enabled us to determine accurately the average overlayer thicknesses of the conducting polymer coatings by two independent methods and to confirm their remarkable uniformity.By using these coating thickness values, and by measuring the resistance of a single filament of these conducting polymer-quartz com-posites, we have succeeded in directly determining the intrinsic conductivity of the conducting polymer overlayer. Finally, STM techniques were used to examine the nano- morphology and structure of the conducting-polymer overlayer. Experimental Preparation of the Conducting Polymer-Quartz Samples The quartz fabric used in these studies is Astroquartz 11, which was purchased as one batch (J. P. Stevens Company). According to the suppliers' specifications, it has a density of 2.20 g cm-, and an average fibre diameter of 9 pm, and it consists of at least 99.9% silicon dioxide.The quartz fibres were coated with polypyrrole or polyaniline according to previously described procedure^.^ For polypyrrole, this pro- cess utilized FeC1, as an oxidant, 5-sulphosalicylic acid as a polymerization retarder (which forms a less reactive 1:1 complex with the Fe"' species), and napthalene- 1,5-disulphonic acid (used as the disodium salt) as an additional doping agent. The reaction was carried out in aqueous media. For polyani- line, the oxidant used was (NH,),S20s in toluene-p-sulphonic acid solution. Characterization of the Conducting Polymer-Quartz Samples Scanning electron microscopy images were obtained using a Philips 505 instrument.Owing to the conductive nature of the samples, no gold sputter-coating techniques were necessary for sample preparation. The conducting polymer-quartz samples were simply affixed to aluminium sample stools using a conductive carbon paste. Scanning tunnelling microscopy studies were carried out using a Nanoscope I1 instrument. Full experimental details will be published elsewhere.6 Thermogravimetric analyses were run in an oxygen atmos- phere to ensure complete combustion using a Perkin-Elmer TGA-7 instrument at a scan rate of 20 "C min-'. Rutherford backscattering spectrometry experiments were run using a 2.2 MeV He' source at the Ion Beam Facility, Los Alamos National Lab~ratory.~ Briefly, the experimental conditions were: ion-beam current =60 nA; total charge focused on sample =40 pC; and single backscattering detector placed at 13" to the incident He' beam.Conductivity measurements were carried out as follows: a single 9 pm filament was carefully removed from a polypyr- role-quartz composite (sample 3) with the aid of an optical microscope and was placed on the surface of a strip of double- sided adhesive tape mounted on a glass slide. Four electrical contacts were made to the filament using silver epoxy glue, which was allowed to cure at room temperature overnight. A low current (clpA) was passed between the two outermost contacts using a Keithley 227 current source and the resulting potential difference between the two inner contacts was meas- ured using a Keithley 177 digital microvoltmeter.Ohmic contacts to the silver epoxy spots were made using four needle pressure contacts in conjunction with the optical microscope. In a control experiment, no current or voltage could be detected between four similarly spaced silver epoxy spots with no polypyrrole-quartz filament between them. Thus, the adhesive tape substrate makes no contribution to the meas- ured electrical conductivity of the filament. Results and Discussion Some of the conducting polymer-quartz samples studied in this work have been characterized previously by TG and resistance measurements by workers at Milliken Research C~rporation.~We have repeated the thermogravimetric analy- ses, and in our control experiments, we found that the untreated quartz fabric exhibits a weight loss of 0.4 0.1YO under the combustion conditions (oxygen atmosphere).This weight loss is attributable to surface moisture and/or certain lubricants and binders present on the fabric' and is well within the manufacturer's specification (0.25-0.90%). Thus, the observed weight losses of the conducting polymer-quartz composites require a small correction to obtain the actual weight of volatilized conducting polymer. Assuming that the conducting polymer overlayer is com- pletely uniform and of known density and that the quartz fibres are all long, thin cylinders with the same cross-sectional area, we may easily derive an equation that allows us to calculate the average thickness of this overlayer from the TG data.From simple geometric considerations, then, we have 6=u{[ 1+($) (3lli.I) where M1, p1and M2, p2 are the masses and densities of the quartz substrate and the conducting polymer overlayer, respectively, a is the mean radius of the quartz filament and 6 is the thickness of the conducting polymer overlayer (see Fig. 1). For the quartz fibre, we have p1 =2.20 g cm-3 and a = 4500 nm & YO.^ For chloride-doped bulk polypyrrole powder, p2 has been estimated to be 1.50 g~m-~ from flotation Fig. 1 Schematic representation of a cylindrical quartz fibre (of length L and radius a coated with a uniform overlayer of conducting polymer of thickness 6 J. MATER. CHEM., 1991, VOL.1 measurements in chlorinated solvent^.^ This value was used to calculate the film thicknesses of the polypyrrole-quartz composites. However, XPS studies of these materials by Gregory et al. suggest that the conducting polymer overlayer contains fewer aj? defects than does the conventional bulk polypyrrole po~der.~ In addition, recent neutron-scattering experiments by Mitchell et aE. on deuterated thin films of electrochemically synthesized polypyrrole indicate that the toluene-p-sulphonate doped material is significantly more ordered than films doped with small inorganic anions such as SO:-or CIO;.'o We have shown that the polypyrrole overlayer on both textile and quartz substrates is almost exclusively doped with similar aromatic sulphonate anions (5-sulphosalicylate and/or napthalene- 1,5-disulphonate) by energy dispersive analytical X-ray (EDAX) techniques.6 Thus, if these observed increases in local molecular order result in more efficient polymer chain packing, then it is likely that the assumed value of 1.50 g cm-3 for p2 may be an underestimate of the true density of the polypyrrole overlayer. This would naturally result in an overestimate of 6 by both TG and RBS techniques.For the polyaniline-quartz composite (sample 6) p2 was taken to be 1.50 g ~m-~,which is the density reported by Stilwell and Park for polyaniline sulphate in sulphuric acid media by flotation measurements." The ratio M2/M1 in eqn. (1) is obtained directly from the TG data. Calculated values of the overlayer thickness using eqn.(1) for all the conducting polymer-quartz composites are presented in Table 1. In the preparation of our polypyrrole-quartz composites, the conducting polymer loading was varied simply by increas- ing the concentrations of monomer and oxidant (and, hence, conducting polymer) relative to the weight of the quartz textile ~ubstrate.~From Table 1, it is evident that the sheet resistance of the polypyrrole-quartz composites decreases non-linearly with increasing thickness of the polypyrrole coating. This relationship is depicted in Fig. 2. RBS has been used by the semiconductor industry for many years to characterize thin film samples. l2 Efficient backscatter- ing mechanisms require substrate atoms with relatively heavy nuclei, so this technique is suitable for probing the silicon atoms in the quartz fibre but not the lighter atoms in the conducting polymer coatings.It is rather unfortunate that neither the RBS nor the TG techniques can be applied to the characterization of more technologically interesting textile substrates such as polyesters or nylons. However, by studying the quartz-fibre based materials, we hoped to gain consider- able insight into the synthesis-structure-property relation-ships of conducting polymer-textile composites in general. Our RBS results for the polypyrrole-quartz samples 1, 2, 4 and 5 are presented in Fig. 3. A similar RBS spectrum was obtained for the polyaniline-quartz sample 6 (not shown). In all five samples, the shift in the silicon edge caused by the conducting polymer overlayer is essentially parallel to the silicon edge of the uncoated quartz fabric.This is direct evidence for the remarkable uniformity of the conducting Table 1 sample ~~ sample number sheet resistance (SZji7) corrected TG weight loss (YO) 6,,'/nm 6,,,d/nm pol ypyrrole-quartz 1 800" 1.34 45f3 47 f4 2 210" 2.15 72+4 68f5 3 120 2.24 80f3 7525 4 75" 2.94 99f4 94+5 5 25" 3.78 128f4 120f5 pol yaniline-quartz 6 - 10.41 28f3368 f8 "From ref. 4; bweight loss of each sample corrected by subtracting the average weight loss of uncoated quartz (0.4 f0.1%); 'assuming a quartz fibre diameter of 9.0pm, density of quartz fibre to be 2.20gcm-3 and the density of both polypyrrole and polyaniline overlayers to be 1.50 g cm-3; derror bars estimated from computer simulation of the energy shift spectrum ?! J. MATER.CHEM., 1991, VOL. 1 900 800 700 0 $ 600. s a .-3; 500 in c Q, 400 in 0,c..-v)8 300 20c 1oc c 0 20 40 60 80 100 120 140 Fig. 2 Variation of composite sheet resistance (Q/O) with conducting polymer overlayer thickness (6) for polypyrrole-quartz samples 1-5 (6 values calculated from TG data) I 1.20 1.25 1.30 1.35 energy/MeV Fig. 3 RBS spectra of polypyrrole-quartz composites of various coating thicknesses: (-..-) sample 5; (---) sample 4; (--.-) sample 2; (-..-) sample 1; and (-) bare quartz fabric polymer overlayers.Moreover, it is a welcome confirmation that the assumptions inherent in the calculation of the 6 values from the TG data using eqn (I) are, indeed, justified. Computer simulations of the silicon edge shifts enable us to calculate approximate overlayer thicknesses from the RBS data. These values are presented in Table 1. We wish to emphasize the close agreement between 6 values of the polyp- yrrole-quartz samples determined independently using the RBS and TG techniques. However, the 6 values calculated from the TG and RBS data for the polyaniline-quartz com-posite (sample 6) are substantially different (368+_ 8 nm and 28 f3 nm, respectively). Close examination of this particular sample using an optical microscope showed that loosely bound macroscopic aggregates of bulk polyaniline powder are randomly distributed throughout the quartz fibres. Obvi- ously, the presence of this powder affects the TG results for the sample and results in a considerable overestimate of 6.On the other hand, this bulk powder is not detected by the RBS technique, which only 'sees' the polyaniline coating that adheres directly to the quartz fibre. This surface-polymerized overlayer is clearly very thin and appears to be uniform when examined by SEM. Compared with the polypyrrole deposition process, we have generally found it much more difficult to eliminate the extraneous bulk-polymerized powder in order to prepare solely submicronic coatings of polyaniline on quartz or textile substrates.The use of NaVO, rather than (NH4)2S208 as a chemical oxidant for the aniline polymeriz- ation should be beneficial in this regard5. The conductivity of a single filament of a polypyrrole-quartz composite (sample 3 in Table I) was calculated using the relation o=l/RA (2) where R is the resistance in ohms (calculated from the current and voltage readings), A is the cross-sectional area of the filament, and 1 is the length of the filament between the inner two silver epoxy contacts. Assuming that the single filament is uniformly coated with polypyrrole (see Fig. 1) then A is given by A =2na6 (3) where a and 6 have their previously defined meanings. For sample 3, 6 2575 nm (from Table I), 1=0.80 cm (from the sample geometry), and we measure R to be ca.1.1 MR. Taking into account the various experimental uncertainties, we calcu- late the conductivity of the polypyrrole coating to be 35 f7 i2-l cm-'. This value is slightly higher than chloride- doped bulk polypyrrole powder prepared using FeCl, as the chemical oxidant', but it is similar to the conductivity of polypyrrole toluene-p-sulphonate powder prepared using iron(@ toluene-p-sulphonate in methanol or aqueous media (23-46 K cm -1).1491 As we have already noted, EDAX (and XPS4) studies indicate that aromatic sulphonate anions are incorporated as dopant anions in the polypyrrole-quartz composites in prefer- ence to chloride anions derived from the FeC1, oxidant.6 This observation explains the higher conductivity of the polypyr- role coating and its improved air ~tability.~ We wish to emphasize that the conductivity measurements in the present work were made on an 8 month old polypyrrole-quartz composite which had been stored under ambient conditions.Thus, it is likely that the conductivity of the freshly prepared sample would have been somewhat higher than 35 0-l cm-'.16 We examined the morphology of the polypyrrole-quartz and polyaniline-quartz composites by SEM. The thinnest polypyrrole (and polyaniline) overlayer (sample 1 in Table 1) seemed to be smooth and featureless (see Fig.4). However, the thicker polypyrrole coatings become less uniform, with some distinctly globular features reminiscent of bulk polypyr- role being observed (see Figs 5 and 6).We have reported similar features on polypyrrole-polyester textile com- posites6 Recently, D. F. Evans et al. observed similar morpho- logical changes in electrochemically synthesized polypyrrole and polythiophene films of increasing thickness by STM' We believe that, in the present work, this change in mor- phology can be rationalized by either one or both of the following two hypotheses. First, the substrate surface could be playing an important role in influencing the order and morphology of the conducting polymer overlayer. Obviously, this effect would be reduced for increasing coating thicknesses and for sufficiently large values of 6. The overlayer mor-phology would eventually be expected to revert to that of conventional, chemically synthesized polypyrrole.Secondly, it is possible that, for the thickly coated samples, the textile substrate surface area is insufficient to fully accommodate the Fig. 4 SEM image of a PolyPyrrole-quartz composite (sample 1; 6= 47 f4 nm by RBS) Fig. 5 SEM image of a PolYPYrrole-quartz composite (sample 4;6 = 94+ 5 nm by RBS) Fig. 6 SEM image of a polypyrrole-quartz composite (sample 5; 6 = 120k5 nm by RBS) amount of deposited polypyrrole. If this were the case, we might expect that some globular solution-polymerized pyrrole would be deposited in addition to and on top of the surface- polymerized pyrrole. Finally, the high resolution of the STM enables us to examine the morphology of these suPPosedlY smooth, feature- less PolYPYrrole overlayers in much greater detail- A typical STM constant current-height image of sample 3 is shown in J.MATER. CHEM., 1991, VOL. 1 Fig. 7. The polypyrrole coating appears to be made up of partially ordered aggregates of nanoparticulates (average diameter 5-10 nm). We have observed similar nanomorphol- ogies on other polypyrrole and polyaniline textile composites (polyester and nylon substrates), sterically stabilized polypyr- role and polyaniline colloids, and electrochemically synthe- sized polyaniline films.6 Conclusions We have shown that both TG and RBS can be used to determine the average coating thickness (6) of very thin (45-130 nm) polypyrrole overlayers on quartz fibres. The results obtained from these two independent methods are in close agreement.Furthermore, the latter technique provides direct evidence of the remarkable uniformity of these overlayers. The polyaniline-quartz sample we examined contained deposits of bulk polyaniline powder, which invalidated the measurement of the coating thickness 6 by thermogravimetry. However, the RBS technique enabled us to determine 6 to be ca.28nm. This overlayer was also smooth and uniform, as evidenced by SEM studies. It is likely that the observed powdery deposits could be minimized or even eliminated by using a more appropriate chemical oxidant, such as NaVO,, for the aniline polymerization. Our scanning electron microscopy studies confirm that, in some cases, these conducting polymer coatings can be remark- ably smooth and uniform.However, it seems that thicker coatings (6x125 nm) are much less uniform, with a distinctly globular morphology being observed. This change in mor- phology could be caused by a reduced ordering effect of the substrate. Alternatively, the globular deposits could simply be solution-polymerized pyrrole. The high-resolution of the STM enables us to examine the nanomorphology of the conducting polymer overlayers. In the case of the polypyrrole-quartz composites, the polypyrrole coating is made up of partially ordered aggregates of very small particulates (5-10 nm diameter). Similar morphologies have been observed in other conducting polymer systems. Finally, we have measured the electrical resistance of an isolated 9 pm diameter filament abstracted from a polypyr- role-quartz composite. Since the coating thickness 6 of this filament is known from our TG and RBS experiments, we have been able to determine, for the first time, the conductivity of the polypyrrole overlayer to be ca.35 R-' cm-'. Fig. 7 Constant current-height STM image of a polypyrrole-quartz composite (sample 3; 6 =75 k5 nm by RBS). Applied bias voltage 1600 mV; current 0.14 nA J. MATER. CHEM., 1991, VOL. 1 In conclusion, we have established a reliable characteriz- ation methodology for conducting polymer-quartz com-posites. We believe that our results have significant implications for the polypyrrole-textile and polyaniline-tex- tile composites currently being produced on a commercial basis by Milliken Research Corporation.These latter systems are important because they have the potential to overcome the various technological problems that have previously plag- ued the commercial development of conducting polymers. S. P. A. wishes to thank SERC for the travel grant that enabled this collaborative project to be carried out. This work was funded by the U.S. Department of Energy Advanced Industrial Concepts Division. References 1 Proc. Int: Con$ Synth. Met. (ZCSM '88) ed. M. Aldissi, Synth. Met., 1989, 27-29, and references therein. 2 R. V. Gregory, W. C. Kimbrell, and H. H. Kuhn, Synth. Met., 1989, 28, 823. 3 R. V. Gregory, W. C. Kimbrell, and H. H. Kuhn, Proc. A.C.S. Div. Polym.Chem. 1989, 30, 165. 4 R. V. Gregory, W. C. Kimbrell and H. H. Kuhn, Proc. 3rd Znt. SAMPE Electron. ConJ, 1989, 570. 5 H. H. Kuhn and W. C. Kimbrell, U.S. Pat. 4803 096, 1989. 6 S. P. Armes, M. Hawley, S. Gottesfeld, J. Beery and M. Aldissi, Lungmuir, in the press. 7 J. R. Tesmer, D. M. Parkin and C. J. Maggiore, MRS Bull., 1989, 12(6), 101. 8 T. Bettencourt, Technical Products Department, J. P. Stevens Company, personal communication. 9 S. P. Armes, M. Aldissi, G. C. Idzorek, P. W. Keaton, L. J. Rowton, G. L. Stradling, M. T. Collopy and D. B. McColl, J. Coll. Znterface Sci.,1991, 141(1), 119. 10 G. R. Mitchell, F. J. Davis, R. Cywinski and A. C. Hannon, Polym. Commun. 1989, 30, 98. 11 D. E. Stilwell and S. M. Park, J. Electrochem. SOC., 1988, 135, 248 1. 12 W. K. Chu, J. W. Mayer and M. A. Nicolet, Backscattering Spectrometry, Academic Press, New York, 1978. 13 S. P. Armes, Synth. Met. 1987, 20, 367. 14 S. P. Armes, unpublished results. 15 J. A. Walker, L. R. Warren and E. F. Witucki, Am. Chem. SOC. Polym. Prep. 1987, 28(2), 256. 16 S. P. Armes and M. Aldissi, Polymer, 1990, 31, 569. 17 T. H. Chao and J. March, J. Polym. Sci.Polym. Chem., 1988, 26, 743. 18 R. Yang, D. F. Evans, L. Christensen and W. A. Hendrickson, J. Phys. Chem., 1990,94, 61 17. Paper 0/05588C; Received 12th December, 1990