Conductive Polymer Functionalization by Click Chemistry

Conductive Polymer Functionalization by Click Chemistry

(Parte 2 de 3)

Scheme 3. Fluorophore Synthesis Scheme 4. Schematic of the Film Reaction of 3 with 1, Where R Substitutes 1a a The gray area indicates where the film has been exposed to the reaction mixture and thus also that there are unreacted azides around this area.

Figure 1. Fluorescence microscopy of the (a) clicked surface (14) and

(b) the reference prepared without CuSO4 under otherwise equivalent conditions. The drop of the reaction mixture did not cover the lower right corner in (a), and thus this part of the film has not been functionalized.The imagesare recordedusing equal lightingand camera settings.

Figure 2. AFM topography image of the clicked surface of 14.

Figure 3. N (1s) high-resolution peak for PEDOT-N3 (3) and the product triazole (21).

Table 1. XPS Results of PEDOT-N3 (3) and the Triazoles 21 and 2 (All Numbers in atom %) a Inclusive 3% tosylate from reoxidation based on XPS analysis of pure PEDOT films showing tosylate to EDOT ratios of approximately 1:3 in oxidized conductive films. b Reference sample, PEDOT-N3 without any further treatment. Quantification in parentheses refer to analysis without electron charge compensation. c Reference sample, PEDOT-N3 exposed to equivalent reaction conditions except that CuSO4 was omitted.

4324 Daugaard et al. Macromolecules, Vol. 41, No. 12, 2008

considerably smaller than the reduction in conductivity, since the film thickness is increased by 75% after the click reaction, which in its own effect gives a reduction in conductivity.

Mechanical stirring is not possible with the polymer film, and all mixing of reagents is thus dependent on diffusion. Therefore,the necessarycatalystconcentrationwithinreasonable time frames was investigated. UV/vis spectroscopy of the films produced with a catalyst amount of 5, 20, and 100% catalyst relativeto the alkyne (1) is presentedin Figure 4 and compared to a thin layer of 1 spin-coated on a glass slide. The UV/vis spectrum shows that the 20% and 100% samples have the same three characteristic peaks as 1, although they are blue-shifted by 5-10 nm. The strong absorption peaks supports the assumption that more than a surface layer of fluorescein is present after the reaction.

The UV/vis absorption of the three fluorescein peaks reaches a maximumalreadyat 20% catalyst,and conductingthe reaction with 100% catalyst made no significant difference. Two explanations for this result seem feasible. Either 20% is sufficient, given the diffusion, to make the reaction proceed to completionwithin reasonabletime or the concentrationof azides is lower than assumed and all reactive sites can be initialized with 20% of the alkynes. The concentration of azides on the surface estimated based on the film thickness, film size, and the density of PEDOT that gives approximately 0.1-0.2 mg/ cm2 of azide polymer. In addition to this, access to all the azides in the film must be limited by sterical hindrance and diffusion into the polymer film, and it is expected that some of the sites may be inaccessible.

Using 20% catalyst loading the reaction time was monitored using UV/vis as shown in Figure 5. The reaction proceeds mainly within the first 20 h, and then the increase in intensity gradually decreases over time. This rate decrease is believed to be due to steric hindrance in the bulk of the polymer film as the loadingincreasesand less free sites are availablefor reaction. It would demand unreasonable long reaction times to obtain complete reaction of all azides, and thus the loading obtained after 20 h has been found sufficiently high.

Film thickness measurements by profilometry corroborate these results (see Figure 6) by showing a slow progression toward larger film thickness for reactions occurring over several days.

The results from polymer films of pure EDOT-N3 show a very high density of accessible reactive sites. Many applications may benefit from a lower density of sites with controllable average spacing, e.g., uses targeting the immobilization of large biomolecules as sensors or cellular stimulants. In order to be able to load the surface with a lower amount of azides, the copolymerizationof EDOT and EDOT-N3 was investigated.

As mentioned earlier, EDOT-N3 was less reactive than EDOT and thus polymerized without pyridine. In the copolymer formulation mixture it was necessary to include some pyridine in order to limit the EDOT polymerization since that would otherwise be too fast. It was decided to simply mix the reaction mixtures for homopolymerization of the two monomers in the ratio desired of the copolymer. This is a tradeoff as polymerization has to be slowed down sufficiently to control the EDOT reactivity without inhibiting the polymerization of EDOT-N3 completely. The reaction was performed as shown in Scheme

5, and as with the homopolymer the product is insoluble and cannot be characterized using NMR or SEC.

Figure 4. UV/vis spectroscopy of the fluorophore (1) loading on

PEDOT-N3 (3) as effect of catalyst concentration using a constant reaction time of 20 h.

Figure 5. UV/vis spectroscopy of the fluorophore (1) loading on

PEDOT-N3 (3) as a function of reaction time using a constant catalyst loading of 20%.

Figure 6. Increased thickness of the triazole functional polymer (14) as a function of reaction time using a constant catalyst loading of 20%.

Figure 7. Increase in thickness by the click reaction as a function of EDOT-N3 content in the copolymer.

Figure 8. UV/vis absorptions per nm film thickness of the clicked copolymers, 15-20.

Macromolecules, Vol. 41, No. 12, 2008 Conductive Polymer Functionalization 4325

Film thickness analysis of the copolymer films by profilo- metry revealedthat increasingEDOT-N3 content and decreasing pyridine content led to thicker polymer films. An explanation could be that more EDOT-N3 remains after spin-coating and subsequentheatingof the monomersas EDOT-N3 is less volatile or simply that lower pyridine concentrationgives less inhibition.

If a specific film thickness was desired, it would have to be optimized for every ratio of the monomers in the copolymer. Since our interest in the copolymers are merely in creating a controllable density of reactive groups in the polymer, this has not been investigated further. Functionalization of the copolymers with 1 using equivalent reaction conditions was performed. As mentioned, the copolymer films are of different thickness, and thus the increased thickness by the click reaction was investigated as shown in Figure 7.

The increase in thickness was found to be dependent on the

EDOT-N3 content,which indicatesthat the reactionis conducted on the bulk film. This has been corroborated with fluorescence spectroscopy performed on the backside of the films which shows fluorescence.The degree of loading was also investigated using UV/vis spectroscopy as shown in Figure 8. To eliminate the effects of different film thicknesses, the results were recalculated and the relative absorbance per nm was plotted.

The corrected data clearly show that an increasing amount of reactive groups can be obtained using the copolymer approach. Interestingly, an upper degree of loading seems to be reached with 60% EDOT-N3.

The effect on conductivity with the varied loading of the copolymer was simultaneously investigated, as shown in Figure 9. As mentioned above the functional polymers have a lower conductivity than pure PEDOT, and it is clear that the conductivity decreases with decreasing contents of EDOT. Consequently, the gain in functionality is balanced by a loss in conductivity.

To show that the method can be used as a general method for PEDOT functionalization, 2,2,3,3,3-pentafluoropropylpent- 4-ynoate, 10, and an alkyne-functionalized MPEG 5000, 12, were prepared. Both alkynes were produced through an ester synthesis using DCC in 8% and 87% yield, respectively. The fluorous and MPEG-functionalized polymers, 21 and 2, have no special UV/vis absorptions that could be used to confirm the reaction. However, the changes in the surface energy properties of the PEDOT-N3 after reaction are substantial, as can be seen from the water contact angle measurements shown in Figure 10.

It is known that introduction of fluorous compounds on a surface will decrease the surface energy and make the surface more hydrophobic.36 By attaching the fluorous 10 to the conductive polymer, there was a distinct increase in both the advancing and receding contact angle by ∼20°, which clearly shows that the reaction has proceeded and a more hydrophobic surface has been formed. The contact angle hysteresis is ∼50°, indicating substantial chemical heterogeneity at the surface. It wouldnormallybe expectedto obtaina higheradvancingcontact angle on a fluorous surface like 21, but the effects from the tosylate ions that were introduced into the polymer through the reoxidation are believed to affect the surface oppositely. Covalentattachmentof poly(ethyleneglycol)entitiesto PEDOT-

N3 is expected to increase its surface energy, resulting in lower water contact angles. Immobilization of MPEG 5000 on the surface in 2 yields the expected lowering of the advancing contact angle by 20°-25°, depending on using either H2Oo r DMF as solvent.The recedingcontactangleshowsno significant development in either solvent compared to the untreated

PEDOT-N3. This is again consistent with a chemically heterogeneous surface, where the PEDOT/tosylate base polymer presentsmore hydrophilicfunctionalitiesamidstthe PEG chains.

Profilometry measurements on these samples showed that the reaction performed in DMF increased the thickness by ∼130 nm, corresponding to 60% for the fluorous hydrocarbon, 21, and 60 nm corresponding to 30% for the MPEG functionalized polymer prepared in DMF, 2. 2 prepared in H2O had approximatelythe same thicknessin the productas in the starting material.

XPS analysis of the clicked surfaces corroborates the results found by contact angle and by thickness goniometry and profilometry, as shown in Table 1. As mentioned earlier, the quantifications are deviating due to degradation of the azides in the sample during the analysis. However, the relative differences between PEDOT-N3 (3) and the respective products can still be used for qualitative assessments. Reaction of 3 with

10 to give 21 introduced 10.6% fluorine to the surface layer, and the high-resolution nitrogen peak shows no residual azide. Thereby the reaction is confirmed, though the extent of reaction cannot be conclusively determined. The results with MPEG

Scheme 5. Copolymerization of EDOT-N3 (2) and EDOTa a Pyridine was added with EDOT to inhibit the polymerization.

Figure 9. Conductivity of the clicked copolymers, 15-20, compared to PEDOT.

Figure 10. Contact angles of H2O with the fluorous (21) and MPEG- functionalized triazole (2, reaction in DMF or aqueous environment) compared to the starting material PEDOT-N3 (3).

4326 Daugaard et al. Macromolecules, Vol. 41, No. 12, 2008 coupling to give 2 also correlate well with the contact angle measurements,where there is clearly a higher loading of MPEG with DMF as solvent compared to H2O. The high-resolution peak of nitrogen shows in both cases residual azide nitrogen

(not shown), which would be expected since diffusion of the long MPEG chains of 12 into the PEDOT-N3 is much more difficult than with the smaller 10. Regarding the MPEG experiments, the reaction is much more likely to occur only in the upper surfaceor perhapsonly on the surface.However,there is an increased loading in the case of DMF compared to water. One may speculate that this is due to water being inefficient in swelling and wetting the PEDOT-N3 surface. Access to reactive groups would be limited to the azides on the actual surface; thus, after reaction of about one layer of MPEG the reaction would stop. The higher loading observed with MPEG reactions in DMF could be an effect of better wetting and swelling of the surface, which would give access to an increased number of reactive sites.

Conclusion

We have developed the synthesis of a new azide monomer,

EDOT-N3, and demonstrated that it can be polymerized to

PEDOT-N3, which can be used as a precursor to obtain conductive polymers with different functionalities. A click reaction with a fluorescein derivative has been performed on the surface, and the reaction conditions have been optimized for other applications. It is possible to copolymerize EDOT-N3 and EDOT in different ratios and to functionalize these after polymerization. Through the copolymer it is possible to control the number of reactive sites on the surface, in case a lower loading is desirable. The copolymers have also been shown to be a good way to obtain functional conductive polymers while minimizing the loss in conductivity.

By coupling the fluorous alkyne and the MPEG alkyne to the surface, we have shown the versatility of the method. We believe that this method can be applied for other alkynes with different functional groups. The approach is well suited for new sensor devices, and we are currently working on the development of different systems.

Acknowledgment. The Danish Research Council for Technology and Production Sciences (through the framework program “Designand Processingof Polymersfor MicrofluidicApplications”, grant 26-04-0074) is thanked for financial support.

References and Notes

(1) Dai, L. M.; Soundarrajan, P.; Kim, T. Pure Appl. Chem. 2002, 74, 1753–1772. (2) Mateiu, R.; Lillemose, M.; Hansen, T. S.; Boisen, A.; Geschke, O.

Microelectron. Eng. 2007, 84, 1270–1273. (3) Goetzberger,A.; Hebling, C.; Schock, H. W. Mater. Sci. Eng., R 2003, 40, 1–46. (4) Hung, L. S.; Chen, C. H. Mater. Sci. Eng., R 2002, 39, 143–2.

(Parte 2 de 3)

Comentários