1 Introduction

The efforts described in this paper are related to those of Professor Dr. Hiroshi Nishihara to whom this special issue is dedicated. While Dr. Nishihara has contributed to a number of efforts in the area of metal-containing polymers, he pioneered in the synthesis of molecular wires that allow conductance of electrical current. This paper is in dedication of the effects.

The search for non-metal electrically conductive materials has been occurring for over 30 years. The importance of this effort was recognized by the awarding of the Nobel Prize in 2000 to MacDiarmid, Heeger, and Shirakawa for their work with polyacetylene [1–7]. Conductivity is dependent on the micro- or fine structure of the fibrils, doping agent, extent, and technique and aging of the sample as well as other factors so it is complex [8–14].

The transformation of these materials from being semiconductors to conductors occurs through exposure of the polymeric materials to dopants that allow the materials to increase their conductivity 10–109 fold. The structural requirement for successful doping is that the polymer chains possess what is referred to in molecular orbital terms as whole chain delocalization and in valence bond theory as whole or entire chain resonance such that the polymer chain acts as a conduit for electrical charge to transverse along the polymer chain.

We have been involved in the search for semiconductors that contain metals within coordination and condensation polymers for about 30 years focusing on polymer chains that can exhibit whole chain resonance with many of these being semiconductors [15–24, 49].

We originally reported the synthesis of titanocene-containing polymers in 1971 [25]. The topic of metallocene-containing polymers was recently reviewed [26]. Some of these polymers were found to spontaneously form fibers [27]; were able to moderate laser emission when present in samples at levels of ppm [30]; and act as plant growth hormones carriers [28, 29]. We also found that some of these polymers are potent anticancer agents [32, 33]. In 1972 we reported the synthesis of titanocene polyamines of the form studied in the present investigation employing the solution and interfacial techniques [33, 34].

Recently we began doping some of metal-containing polymers. One of these polymers was derived from the condensation reaction of titanocene dichloride and 2-nitro-1,4-phenylenediamine (Fig. 1). We recently reported on the successful doping of this product where surface exposure to iodine vapors resulted in the increase in conductivity to 104 [24]. This was the initial report of a somewhat traditional condensation polymer being successfully doped to significantly increase its conductivity. More recently, we reported on the bulk doping of this polymer where doping occurs through a simple addition of the iodine to the titanocene polymer that resulted in a similar increase in the polymer’s conductivity [49]. Here we describe the effort to increase this phenomenon to other similar polymers to begin the process of investigating the structural window where such doping is effective at increasing electrical conductivity.

Fig. 1
figure 1

Synthesis of polymer from reaction of titanocene dichloride and 2-nitro-1,4-phenylenediamine where R is simple chain extension

Full size image

2 Experimental Section

2.1 Synthesis

Titanocene dichloride (CAS # 1271-19-8), dibutyltin dichloride (683-181-1) and 2-nitro-1,4-phenylenediamine (5307-14-2) were used as received from Aldrich Chemical Company; zirconocene dichloride (126-72-5), hafnocene dichloride (12116-66-4) from Alfa Inorganics; iodine (7553-56-2) from Fisher Scientific; and N-methyl-1,4-phenylenediamine (5395-70-0) from Ivy Fine Chemicals.

2-Nitro-1,4-phenylenediamine (or other Lewis base; 3.00 mmoL) was dissolved in 30 mL of water. Sodium hydroxide (6.0 mmoL) was also dissolved in the water. Titanocene dichloride (or other metal-containing dihalide; 3.00 mmoL) was dissolved in 30 mL chloroform. The aqueous phase was first added to the Kimex emulsifying jar. Stirring (about 18,000 rpm) was begun and the chloroform phase was quickly added (3–4 s addition time) and stirring was continued for an additional 15 s. The precipitate was collected using suction filtration and washed with water and chloroform to remove unreacted materials. The solid was washed into a glass Petri dish and allowed to dry under room conditions of temperature and pressure.

2.2 Electrical Measurements

A standard procedure was employed to obtain the electrical measurements [1, 8–12, 35–43]. Briefly, a known amount of sample and iodine were mixed together with grinding to form a powder. The 2-nitro-1,4-phenylene diamine and titanocene dichloride polymer was finely ground to a powder employing a mortar and pestle in the usual fashion and viewed under an optical microscope (Olympus CH30 microscope) to determine an average size of the particles. The particles were irregular spheres with an average diameter of 1.02 × 10−6 (±0.04) for (n = 20) meters. The power was pressed into pellets using a Carver Laboratory Press. These pellets were utilized for the conductivity measurements. The thickness and diameter of the sample pellet were measured using a micrometer screw gauge. Pellet thickness (0.59-mm) and diameter (13.0-mm) were held constant for the current studies. A GenRad 1650 B impedance bridge was employed for the electrical measurements. The sample holder consists of an enclosed cage which houses the copper electrodes. Pellets were placed between the copper electrodes and an alternating current applied. The sample holder was connected to the impedance bridge which measures the various electrical properties. The impedance bridge was also connected to a 1311-A audio oscillator which allowed the frequency to be varied. The conductivity measurements were done for the following frequencies: 0.4, 1, 2, 5, and 10 kHz. Measurements and calculations were made in the usual manner [1, 8–12]. A blanket of nitrogen was present during the doping and measurement procedures to maintain an oxygen and moisture free surrounding. The assembly was calibrated employing materials with known conductivities.

Sample heating was carried out by placing the disks into a preheated oven set at 50 °C for the specified time.

2.3 Spectroscopy Measurements

Infrared spectra were obtained employing a JASCO FT/IR 4100. MALDI MS were obtained employing a high resolution electron impact positive ion matrix assisted laser desorption ionization time of flight, HR MALDI-TOF, mass spectrometry employing a Voyager-DE STR BioSpectrometer, Applied Biosystems, Foster City, CA. The standard settings were used with a linear mode of operation and an accelerating voltage of 25,000 V; grid voltage 90 % and an acquisition mass range of 100–1,000 Da. Fifty to 200 shots were typically taken for each spectrum. Molecular weight was determined employing a Brice-Phoenix Universal Light Scattering Photometer.

3 Results and Discussion

3.1 General

We have begun investigating the results of polymer doping. While there are a number of different doping procedures employed in the production of conducting polymers, the most widely employed is the exposure of the polymeric material, often in a compacted disk, to elemental iodine vapors either as a vapor or as part of a mixture as we have done in the present study [1, 8–12, 35–43]. We studied the effects of doping of over 50 products. Of these, we found one polymer whose electrical conductive significantly increased when doped with iodine. The material was the product from reaction between titanocene dichloride and 2-nitro-1,4-phenylenediamine (Fig. 1). The titanium atom in titanocene dichloride has a vacant orbital that can accept electrons allowing resonance to occur through it [44]. Thus the polymer can exhibit whole-chain resonance so its structure is consistent with the general structural criteria for conductivity. Here we will focus on studies related to determining the extent of the structures that allow increased conductivity when doped.

Table 1 contains the molecular weight of the polyamines investigated in the present study.

Table 1 Molecular weight and chain length for the metallocene polyamines involved in the present study
Full size table

3.2 Variation of Dissipation Factor and Dielectric Constant for Doped and Undoped Materials

Since the emphasis is on conductivity we will focus on this parameter. Even so, it is of interest to briefly describe the various measured values that are employed in determining conductivity and to compare these values for the doped and undoped materials. We will use results from the titanocene-2-nitro-1,4-phenylenediamine product to illustrate these values since these values are typical of other products described in this paper that showed increased conductivity with doping.

The dielectric constants, K, for the undoped samples vary from 0.4 to 1.9 indicative of materials with a relatively low dielectric constant. The dielectric constants for the doped products have increased varying from 8 to 38. For comparison, dielectric constants (60 HZ) for typical polymers vary from about 2 for polytetrafluoroethylene to 8.4 for poly(vinylidene fluoride) with values for nylons, polycarbonates, and polyesters generally in the range of 3–5 [13, 14]. Values of K for the highly polar Group IVB polyoximes [18, 23, 24] varied from 2 to 10 and 5 to 120 for a series of palladium polyamines [15]. Both the polyoxime and polyamine samples tested can exhibit whole-chain resonance in their backbones.

Dissipation factors, D, vary from 0.4 to 1.4 for the undoped samples and from 0.4 to 2.7 for the doped samples. Thus, the dissipation factors do not widely vary when the sample is doped. Values for D for the Group IVB polyoximes [18, 23, 24] vary from 2.6 to 10 and for the palladium polyamines [15] from 0.2 to 28. Values for commercial polymers are as follows: nylon 66: 0.02; polyethylene: 0.0002; polypropylene: 0.0003; polytetrafluoroethylene: 0.0003; and poly(vinyl chloride): 0.01 [13, 14]. In general, more polar molecules have higher D values.

The D and K values are consistent with the major changes found between the doped and undoped samples being contained in the K term. The K term is largely a measure of the polarity of the material so that the larger values found for the doped samples are consistent with the materials becoming more polar as they are doped. This trend is found for all of the samples reported on here.

In general, changes in conductivity with applied frequency are usual and the particular trend varies with the particular material [13, 14, 35, 40].

3.3 Experimental Variables

We carried out two studies to establish general conditions for the remaining studies. These studies involved evaluation of the amount of iodine used for doping and sample heating effectiveness.

3.3.1 Effect of Amount of Iodine

In this study, specific amounts of iodine are mixed with the polymer. Figure 2 contains the conductivity for four amounts of iodine added, from 3 to 15 %, for the product from titanocene dichloride and 2-nitro-1,4-phenylenediamine. Unlike doping silicon to produce semiconductors where the amount of doping agent is generally less than 1 %, the amounts of iodine employed to promote polyacetylene from a semiconductor to a conductor are in the range of 3–15 % [1–8, 36, 41]. Thus, it was this range that was studied. Bulk conductivity generally increased as the amount of iodine increased increasing generally over 102 fold for all frequencies for the 10 and 15 % doped samples compared to the undoped sample. The greatest increase was found for the highest applied frequency with conductivity increasing about 103 fold converting the material from being a semiconductor to being a near-conductor. The changes from addition of 10 % compared to 15 % iodine are small. Ten percentage doping was employed for the remaining studies.

Fig. 2
figure 2

Plot of log bulk conductivity as a function of frequency for the product from titanocene dichloride and 2-nitro-1,4-phenylenediamine containing varying amounts of iodine from 0 to 15 %

Full size image

3.3.2 Effect of Heating

In the first study, the iodine was simply present as part of a mixture brought together with mixing [24, 49]. More effective mixing might occur when the sample is heated since the iodine is volatilized and should be more completely dispersed within the sample. Further, heating may help in the formation of an active compound allowing for better conductivity. Figures 3, 4 contain results for different heating times.

Fig. 3
figure 3

Conductivity of the product from titanocene dichloride and 2-nitro-1,4-phenylenediamine containing varying iodine amounts heated for 10 s as a function of applied frequency

Full size image
Fig. 4
figure 4

Conductivity for the product from titanocene dichloride and 2-nitro-1,4-phenylenediamine mixed with varying iodine amounts heated for 480 s as a function of applied frequency

Full size image

A similar pattern to that found for the unheated samples is found for the samples heated for 10 s (compare Figs. 2, 3 unheated and heated, respectively). Conductivity increases as the concentration of iodine is increased with the difference between the conductivity of the undoped and doped in the range of 102–103 fold. Similar results are found for samples heated for 20 and 30 s heating. But by 60 s heating, there is some conductivity loss and after heating for 480 s (Fig. 4), all of the samples approximate the conductivity of the undoped sample.

In general, as the heating times increase the conductivity of the doped samples decrease eventually returning to the conductivity of the undoped sample. We believe this is due to the heating preferentially removing surface iodine so that while more effective mixing may be occurring internally, it is not experimentally observed because of the removal of the surface iodine. As noted before, the measurements are made using two flat electrodes so that surface phenomena is critical to the overall measurement and conductivity may well be different within the bulk of the sample. The suggestion that the decrease is due to evaporation of surface iodine is consistent with a decrease in metallic coloring on the surfaces of the sample disks as heating is increased.

We attempted to look at only the surface for the presence/absence of the R–F moiety. Our infrared and mass spectral analyses for post heating of the bulk and surface showed the persistence of the R–F moiety. Even so, our surface samples may have included material not precisely at the surface so that our spectral analyses does not allow us to confidently draw conclusions regarding the role that the R–F presence or absence plays in the conductivity.

3.4 Results of Varying the Nature of the Metallocene Metal for Polymers Derived from 2-Nitro-1,4-Phenylenediamine

While there are many metals that can be incorporated, we chose to look at the analogous products from tin since we have lots of experience with organotin-containing materials and the corresponding metallocenes but derived from zirconocene dichloride and hafnocene dichloride since these metals are in the same family as titanium and should offer similar tendencies.

Here we will compare results for the analogous polymers derived from 2-nitro-1,4-phenylenediamine except where the metal moiety is zirconocene or hafnocene. Results for the titanocene analogous titanocene product were given in Sect. 3.3 where we established that the product from titanocene dichloride and 2-nitro-1,4-phenylenediamine can be doped under a variety of conditions. For the additional studies we employed samples doped with 10 % iodine.

Electrical properties for 2-nitro-1,4-phenylenediamine and zirconocene dichloride and hafnocene dichloride products containing 10 % by weight of iodine were studied as a function of applied frequency and different heating times. Results appear in Figs. 5, 6. The initial plot in the figures is for a doped sample that is not heated. Conductivity increases on the order of 103–104 are found for both the zirconocene and hafnocene polymers. In general, the conductivity decreases with heating time possibly due to the vaporization of iodine from the surface of the pellet as already noted. Finally, after heating for 480 s for the zirconocene product and 60 s for the hafnocene the conductivity approaches the undoped sample.

Fig. 5
figure 5

Conductivity of the 2-nitro-1,4-phenylenediamine zirconocene dichloride polymer containing 10 % iodine heated for different times

Full size image
Fig. 6
figure 6

Conductivity for the product of 2-nitro-1,4-phenylenediamine and hafnocene dichloride containing 10 % iodine heated for different times

Full size image

3.5 Conductivity of the Product of 2-Nitro-1,4-Phenylenediamine and Dibutyltin Dichloride Product

Tetravalent tin such as present in dibutyltin dichloride has vacant orbitals that are often involved in back bonding with carbonyls [44]. Thus, it fits our criteria of allowing delocalization of electrons through it. Electrical properties for the 2-nitro-1,4-phenylenediamine dibutyltin dichloride polymer were studied (Fig. 7). There is little or no change in conductivity between the doped and undoped samples.

Fig. 7
figure 7

Conductivity of the 2-nitro-1,4-phenylenediamine and dibutyltin dichloride product containing 10 % iodine heated at different times

Full size image

3.6 Effect on Conductivity of Titanocene Products Formed from Substituted 1,4-Phenylenediamines

Another structural window involves varying the nature of substituents on the 1,4-phenylenediamines. There are commercially available a wide range of substituted 1,4-phenylenediamines. In the present study we looked at only several products with one containing an electron donating substituent, a second containing an electron withdrawing substituent, and finally one containing no substituent.

The first product was derived from 2-methoxy-1,4-phenylenediamine which contains the electron donating O–CH3 moiety. The results for 2-methoxy-1,4-phenylenediamine and titanocene dichloride are given in Fig. 8. There is no observable change in the conductivity between the doped and undoped samples.

Fig. 8
figure 8

Conductivity plot for 2-methoxy-1,4-phenylenediamine titanocene dichloride polymer containing 10 % iodine and different heating times

Full size image

Electrical properties for the 1,4-phenylenediamine and titanocene dichloride product are given in Fig. 9. As in the case with the product from 2-methoxy-1,4-phenylenediamine, there appears to be little change in the conductivity of the product when it was doped.

Fig. 9
figure 9

Conductivity plot for 1,4-phenylenediamine titanocene dichloride polymer containing 10 % iodine and heating at different intervals

Full size image

Figure 10 contains conductivity results for the product of 2,5-dichloro-1,4-phenylenediamine which contains electron withdrawing chlorine groups. Here, there is a small, generally about tenfold, increase in conductivity as the sample is doped.

Fig. 10
figure 10

Conductivity plot for the 2,5-dichloro-1,4-phenylenediamine and titanocene dichloride product containing 10 % iodine for different heating times

Full size image

In summary, little or no change was found for the conductivity of doped and undoped samples of other products whether the substituents were electron donating (such as 2-methoxy-1,4-phenylenediamine) or electron withdrawing (such as 2,5-dichloro-1,4-phenylenediamine) or non-substituted (1,4-phenylenediamine). Thus, there may be something special about the nitro group and its ability to allow for an increased conductivity when doped for this set of polymers. Future studies might consider dinitrodiamine products but caution should be taken since some aromatic dinitro compounds are known to be thermally unstable.

3.7 Effect of Substitution on Nitrogen

We also looked at a number of metallocene-polymers derived from diamines that were structurally similar to 2-nitro-1,4-phenylenediamine except not containing the “nitro” functional group. We found one product that exhibited similar doping behavior but not containing the nitro group. The electrical properties for the N-methyl-1,4-phenylenediamine and titanocene dichloride product appear in Fig. 11. The results are similar to those found for the nitrophenylenediamine products. As in the cases with the nitrophenylenediamine products, the greatest increase is found for the unheated sample with increases in the range of 106 throughout the frequency range. The conductivity decreases approaching that for the undoped sample at longer heating times. It is currently not known why this particular product gave such increases but suggests that another structural window to be investigated is the modification on the amine such as employing N,N′-dimethyl-1,4-phenylenediamine along with other substituted amines. For this product, doping moves the conductivity of the product from a semiconductor to the lower range of conductivity.

Fig. 11
figure 11

Conductivity of the product of N-methyl-1,4-phenylenediamine and titanocene dichloride containing 10 % iodine heated at different time intervals

Full size image

The results are also consistent with the idea that the presence of the nitro is not necessary for enhanced conductivity to occur through doping with iodine. Thus, increased conductivity through doping is a wider phenomenon that will need further study.

3.8 Atmospheric Stability of 2-Nitro-1,4-Phenylenediamine and Titanocene Dichloride Doped Products

Many doped samples are unstable either to moisture or oxygen in the atmosphere [1–11]. Figure 12 contains a study following conductivity over a 21 h period. The room was maintained at 25 °C with 25 % humidity.

Fig. 12
figure 12

Conductivity of the 2-nitro-1,4-phenylenediamine and titanocene dichloride product in an open atmosphere for different times

Full size image

There appears to be little change in conductivity over the 21 h test period (Fig. 12).

3.9 Structural Change Characterization

In doping, while it is assumed that some chemical reaction involves activation of the increased conduction from doping, the nature of the reaction site is not usually given. This site may be sufficiently stable to allow standard detection methods. To attempt to gain some knowledge concerning the identity of the activated site or species, infrared spectroscopy and MALDI TOF MS spectroscopy were carried out on doped and undoped samples for products that responded with increased conductivity to doping.

Comparisons of the doped and undoped materials were made over the infrared range of 4,000–400 wave numbers (all band assignments are given in wave numbers). There were few differences (Figs. 13, 14 for the titanocene-2-nitro-1,4-phenylenediamine product). There was a modest difference in the Ti–N band occurring at 496 for the undoped polymer and 473 for the doped material. Further, there is a new band found at 664 for the doped sample tentatively assigned to the formation of the C–I moiety [45, 46]. It is possible that activation occurs through the formation of the C–I moiety.

Fig. 13
figure 13

Infrared spectrum for the undoped product of 2-nitro-1,4-phenylenediamine and titanocene dichloride

Full size image
Fig. 14
figure 14

Infrared spectrum for the product of 2-nitro -1,4-phenylenediamine and titanocene dichloride doped with iodine

Full size image

For the zirconocene product a new band is found at 666 and for the hafnocene product a new band is found at 669, both tentatively assigned to the formation of a C–I moiety. The product of N-methyl-1,4-phenylenediamine has a new band at 664 also assigned to the formation of a C–I product.

MALDI MS was also performed on the undoped and doped samples and results compared (Figs. 15, 16 for the titanocene-2-nitro-1,4-phenylenediamine product). Three new ion fragments were found for the doped material. The ion fragment at 254 Da is assigned to molecular iodine, I2. (All of the spectra show this band and it will not be mentioned again.) The ion fragment at 232 Da is assigned to HNPhNHI and the ion fragment at 323 is assigned to TiNHPhNO2NHI both consistent with the formation of an aromatic C–I moiety. For the zirconocene product additional ion fragments containing iodine are found about 216 (HNPhI), 231 (HNPhNHI), and 324 (ZrNHPhNHI). For the hafnocene product new bands are found again at 216 and 231 and an additional band at 301 (HNPhNO2NHI, Na). For the titanocene product with N-methyl-1,4-phenylenediamine new bands are found at 151 (NaI) and 227 (PhI, Na).

Fig. 15
figure 15

MALDI MS for the undoped product of titanocene dichloride and 2-nitro-1,4-phenylenediamine

Full size image
Fig. 16
figure 16

MALDI MS for doped product of titanocene dichloride and 2-nitro-1,4-phenylenediamine containing 10 % iodine

Full size image

For doped samples, MALDI MS showed the formation of iodine containing species, namely an iodine-containing phenylene, and infrared spectroscopy showed the formation of a new C–I bond.

Figure 17 shows possible mechanisms for the formation of the iodine-containing phenylene ring employing electrophilic, nucleophilic and free radical attacks. It is not known if this is the initial step that assists in the increase in the conductivity or if this is simply one of the steps involved in the doping process resulting in the formation of an iodine-phenylene product.

Fig. 17
figure 17

Possible mechanistic schemes for the substitution of iodine to the phenylene ring

Full size image

4 Future Directions

The discovery of somewhat traditional condensation polymers that can be doped to improve their conductivity is potentially important. Such polymers may offer their own different and potentially valuable electrical profiles that can be employed in fashions different from the current often vinyl-derived polymers. Combination of different polymers may be employed as a series of switches that allow more than the present off and on behavior. Further, the discovery of a condensation polymer that behaves in a manner similar to vinyl-derived polymers with respect to doping is significant and may open doors to a number of other condensation polymers that behave similarly including the non-metal containing condensation polymers.

Several areas where additional work is needed are obvious. The current work is simply the beginning of describing the structural window associated with the successful doping of similar condensation polymers.

As noted above, the current study is only the start of describing the structural window where successful doping can occur. Studies involving additional products need to be carried out including investigation of other metallocenes such as vanadocene. Additional nitro- and non-nitro-containing phenylenediamines need to be studied. Along with the 1,4-phenylenediamine, similar products except derived from the 1,3-phenylenediamine should be investigated. Further, additional N-substituted aromatic diamines should be considered.

Since it is believed that the formation of the iodo- substituted 2-nitro-1,4-phenylenediamine may be involved in the increased conductivity, the corresponding polymer, that is from iodio-2-nitro-1,4-phenylenediamine, could be synthesized and its electrical properties measured. If it has a significantly larger conductivity than the corresponding non-iodio product then this would be (tentative) evidence that formation of this structure is important in increasing its conductivity. If there is no increased conductivity for this product then it is evident that other activation is needed and may or may not involve the formation of this compound.

Non-metal compounds might also be tested. These might initially include aromatic polyamines such as formed from the reaction between terephthaloyl chloride and 2-nitro-1,4-phenylenediamine. This polymer should offer chain electron delocalization.

The products might be examined employing additional doping agents such as arsenic pentafluoride (oxidizing agent), bromine (oxidizing agent), and metallic sodium (reducing agent). None of these are as easy to handle as iodine.

Conductivity is believed to occur from the transfer of free radical sites along polymer chains and their “hopping” or transfer from one chain to another. It is known for conductive polymers that conductivity varies dramatically for isotropic systems in comparison to directionally heterogeneous systems. Thus, for aligned polyacetylene chains conductivity is about 106 times greater in the direction of the alignment compared with that found at right angles to this alignment [47, 48]. For the present system, pressing of the polymer into pellets may cause alignment against the direction of applied pressure resulting in the chains being aligned along the face of the pellet and not at right angles to the face of the pellet. Since bulk conductivity is measured through the face of the pellet, any favorable enhancement due to the chain alignment is opposed. In the future, conductivity should be measured with electrodes connected to the sides of the pellets and not through the faces of the pellets. Application of a strong magnetic field to pre-pressed material may also encourage preferential alignment. This polling is regularly employed to align materials. Thus, the conductivity changes that are occurring within the sample may be much larger than measured.