*5.3. NMR Spectroscopy*

The simplest indication of the nature of polyacetylene at the molecular level is the determination by CP MAS (Cross Polarization Magic Angle Spinning) 13C NMR that ≈5% of the carbon atoms have sp<sup>3</sup> hybridization instead of the predominant sp<sup>2</sup> hybridization. The sp3 features "can probably be ascribed to chain terminations, cross-links, or hydrogenated double bonds" [27]. This corresponds to one atom in 20. If we ascribe the putative sp<sup>3</sup> carbons to cyclobutane ring formation then this corresponds to two cyclobutane defects in 200 carbons or 50 C=C double bonds from one to the next in two chains connected by four sp<sup>3</sup> carbons in cyclobutane rings at each end. This is, of course, only the average structure in a distribution.

Another NMR experiment aimed directly at detection of bond alternation is the determination of the 13C–13C dipolar coupling constant of polyacetylene prepared with low levels of acetylene–<sup>13</sup> C2 [28,29]. In this case, the observation of two coupling constants for *trans*-polyacetylene indicates two distinct bond lengths. This work has led to the prevailing bond alternation value and the

individual C–C bond lengths for polyacetylene. For example, [30] compares a computed value to the bond lengths in [28]. There are, however, a few areas of concern in this work. The *cis*-polyacetylene isomer shows, as expected, only one coupling constant corresponding to a double bond length, 1.37 Å. Converting the sample of *cis*-polyacetyene to *trans*-polyacetylene is done by heating in vacuum at 160 ◦C for one hour. The resulting solid state nutation NMR spectra at 77 K show two coupling constants corresponding to 1.36 and 1.44 Å bond lengths. It is noted [28] that "The generation of approximately equal populations of singly and doubly bonded labeled carbon pairs in *trans*-(CH)x starting with only doubly bonded pairs in *cis*-(CH)x is intriguing." No explanation is given for this observation. The only obvious explanation is that half of the 13C=13C double bonds react with nearby predominantly 12C=12C double bonds to make 13C–13C single bonds from the original 13C=13C double bonds in cyclobutane rings. This is the kind of unambiguous experiment that is not done often enough. Presumably this also happens with half of the 12C=12C bonds. The result is best called *poly("ladderane").* Performing a CP-MAS 13C determination of random 13C labeled polyacetylene treated in the same fashion. Both before and after thermal conversion to the *trans* form would be interesting.

The authors further investigated the temperature dependence of the 13C–13C splitting up to 300 K and did not see an expected coalescence of the features from defect migration along long chains. They speculated that the signals that they were observing came from chains that did not contain mobile defects necessary for thermal averaging, i.e., locked-in bond alternation due to cross-linking. In another work [29], it was concluded that "nuclear spin-lattice relaxation rates for 1H and 13C in polyacetylene cannot be adequately explained in terms of either nuclear spin diffusion to a static paramagnetic defect or rapid one-dimensional diffusion of the defect itself. A model in which only a small fraction of the molecular chains contain defects was proposed. Nuclei on these chains are rapidly relaxed, whereas the remainder achieve equilibrium by nuclear spin diffusion. The dependence of the measured relaxation rates upon frequency and isotopic concentration agreed with the predictions of the model." As indicated by the above, an alternative hypothesis for this signal is that it comes from finite polyene chains.

### *5.4. Resonance Raman Spectroscopy*

Resonance Raman studies of "polyacetylene" [31–46] make use of the fact that Raman scattering under resonance conditions has a greatly enhanced contribution from species that have electronic excitation resonances near the Raman excitation frequency. If the sample is homogeneous, then the intensity of vibrational features in the resonance Raman spectrum will all increase or decrease together as the excitation frequency is changed. If, however, the sample is heterogeneous, with various components having different electronic excitation behavior and vibrational spectra, changing the excitation frequency will cause some vibrational features to increase and other decrease in intensity. In the case of "polyacetylene" containing multiple oligoene components, the electronic excitation spectrum moves to lower energy, and the strongly enhanced vibrational mode moves to lower frequency, as the chain length is increased. It is observed as expected that Raman excitation at longer wavelength results in lower frequency Raman active modes. The conclusion of most of these studies is that chain length heterogeneity is the most likely explanation for the experimental observations. Simple calculations reproduce the data, and alternative models were eliminated [43].

Finite linear conjugated polyenes have very strong Raman scattering. This is especially true when the Raman excitation is close to the strong electronic absorption bands of polyenes. The reason for the very strong Raman scattering of polyenes is that linear polyenes have a very strongly allowed electronic excitation that involves an appreciable geometry change in the excited state relative to the ground state. This geometry change is primarily along one normal mode of motion in the ground state near 1500 cm<sup>−</sup><sup>1</sup> with a smaller contribution from another mode around 1100 cm<sup>−</sup>1. The details of this are discussed in Section 6 below. Here, we concentrate on "polyacetylene" in the context of its anticipated Raman scattering. This issue was noted in [44] in the context of periodic solids and in a way relevant to this review in [45,46]. In these publications, it is noted that the Condon mechanism

for Raman scattering (also called A-term scattering) due to geometry changes in excited electronic states vanishes for periodic solids and for polyacetylene as we have defined it. This is because the one-electron excitation involved does not result in a significant geometry change for a system of effectively infinite size. The most argumentative statement of the issue is in [46], in which it is claimed that (a) since "polyacetylene" has a well-known Raman spectrum, it must be the case that something is missing that is beyond the Condon scattering mechanism; (b) a vibronic activity term is added (called B-term scattering), which (c) explains ("solving polyacetylene") by addition of the next term albeit with an unknown magnitude, and (d) that this refinement also removes the need to consider polyacetylene as a heterogeneous mixture of finite chains [31–42]. Specifically, quoting from [46], "In ref. 24 (of [46]), three samples of nearly monodisperse polyacetylene with lengths of about 200, 400, and 3800 unit cells were synthesized and their Raman spectra were obtained. The sidebands remained, and many of the earlier "polydisperse" explanations for the line shape quickly evaporated." The relevant ref. [24] is here ref. [35].

The authors of [46] have misrepresented [35] by claiming that the samples involved have very long conjugation length and ignoring the fact that in [35] itself these "monodisperse" samples are investigated as to the polydispersity of the conjugation lengths. Quoting from the abstract of [35] "After thermal isomerization, theoretical analysis of the Resonance Raman spectra using the Brivio, Mulazzi model indicate the ratio of long trans conjugated segments (N ≥ 30) to short trans conjugated segments (N ≤ 30) is significantly larger for 100,000 Dalton (3800 unit) polymer." It is the contention of this author that it is the N = 15–30 and perhaps a bit longer double bond species that have Raman spectra from the Condon mechanism, and that there is no evidence that anything that might be defined as polyacetylene makes any contribution to the Raman spectrum.

### *5.5. A Cautionary Note on Doping*

Our argument, outlined above, is that it is not possible that the nuclear probability distribution will exhibit bond alternation in any experiment. The reported bond alternation in "polyacetylene" is evidence for the presence of finite chains. These may dominate the signals even if they are not the major component of the sample itself. The "semiconductor" properties attributed to "polyacetylene" are, in this interpretation, due to the lack of extended conducting chains. The observed effect of "doping" on conductivity, interpreted in terms of band theory, may rather be due to enhancement of conduction via electron transfer between finite chains rather than population of a conduction band, as assumed in the conventional model. In fact, iodine "doping" of polymers such as polybutadiene that do not contain any conjugated double bonds results in significant conductivity [47,48]. This is attributed [47] to the presence of ionic iodine chains such as I3− and I5− in the material. In [47], the conduction is claimed to be electronic.

### **6. Electronic Spectroscopy of Finite Linear Conjugated Polyenes**

Interest in the electronic spectroscopy of linear conjugated polyenes began in the early days of the development of quantum mechanics. At that point, dealing with polyatomic molecules was based on qualitative molecular orbital and valence bond descriptions. Early treatments of butadiene with valence bond methods concluded that the lowest energy excitation would be a diradical with the same Ag symmetry in C2h as the ground state. The observation of a strongly allowed electronic excitation contradicted this view. For linear conjugated polyenes with an even number of carbons in the C2h point group, the symmetries of the non-degenerate molecular orbitals alternate in symmetry with increasing energy with a behavior similar to that of a particle in a box. Because of this, the HOMO-to-LUMO excitation is thus necessarily from an Ag ground state to an excited state of Bu electronic symmetry. This seemed to be in agreemen<sup>t</sup> with experiment in respect to a low-energy, strongly allowed transition that increased in intensity and decreased in energy as the polyene chain increased in length. There were, however, several aspects of linear conjugated polyene electronic spectroscopy and, especially, of the corresponding photophysics that attracted our attention that all was not right. The main things that we found of interest was the observation [49,50] that the intrinsic fluorescence (or radiative) lifetime of linear polyenes is much longer than expected on the basis of the integration of the absorption spectrum. The other items of interest were the series of spectroscopic papers on linear polyenes by Hauser and co-workers [51–56] and the discussion of that work by Mulliken [57]. In this discussion, Mulliken said:"A puzzling phenomenon reported by Kuhn, Hausser, and co-workers in their comparative study of absorption and fluorescence in polyene derivatives is the existence of a gap between the longest wave-length absorption band in the vibrational structure of N to V1 and the shortest wave-length band the corresponding V1 to N fluorescence spectrum. It is worth noting, however, that there seems to be still a small amount of absorption and emission at the middle of the gap. The width of the gap increases with increase in the number of conjugated double bonds. The absorption and fluorescence spectra are roughly mirror images of each other on a frequency scale. There seems to be no reason, especially in view of our theoretical analysis, to doubt that the fluorescence spectrum really is V1 to N. According to the theory, no other excited singlet level should be below V1." This is followed by an attempt to explain the observed spectral pattern in terms of an in-plane bending deformation of large amplitude. A similar explanation was posed for the long intrinsic lifetime [49,50].

The partially resolved vibronic spectra (Figure 6) of absorption and fluorescence of 1,3,5,7–octatetraene at 77 K have a gap between what appears to be the first absorption feature (the 0-0) near 32,000 cm<sup>−</sup><sup>1</sup> and the first fluorescence feature near 29,000 cm<sup>−</sup>1. This is not anticipated if there is only one low-energy excited electronic state. On the other hand, if there is another low-lying state, this absorption should begin where the fluorescence begins near 29,000 cm<sup>−</sup>1.

**Figure 6.** Fluorescence and absorption spectra of all *trans*-1,3,5,7-octatetraene in 3-methylpentane at 77K. Left, fluorescence on an arbitrary emission scale; right, absorption on an arbitrary absorbance scale [58].

In order to observe the hidden spectral features, it is necessary to use a solvent that at low temperature provides the same environment for all dissolved solute octatetraene molecules. This was found to be the case for n-octane, where apparently a centrosymmetric mixed crystal is formed. The spectra shown in Figures 7 and 8 are for polycrystalline matrices of n-octane with a very low concentration of octatetraene. The major features in these spectra are (a) the vibration-less origin (0,0) transition of the 11Ag to 11Bu electronic transition near 310 nm (32200 cm<sup>−</sup><sup>1</sup> in Figure 7), (b) the first absorption transition to the new low-energy, low-intensity transition near 348 nm (Figure 8) (28,730 cm<sup>−</sup><sup>1</sup> in Figure 7), and (c) the first emission transition from the as of ye<sup>t</sup> unknown excited electronic state to the ground electronic state of 11Ag symmetry near 352 nm in Figure 8. The line shapes reflect phonon side-band structure.

Figure 8 is an expanded view of the origin region combining with addition in the upper trace the first one-photon feature of Figure 7. The two central features of the upper trace are "false origins" due to modes of bu vibrational symmetry in the upper or the ground state. The lower trace is the two-photon excitation spectrum showing the true 0-0 at 350 nm plus phonon side bands and the

lowest molecular ag mode an in-plane bending vibration. These spectra present the classic pattern of Herzberg–Teller vibronic coupling, in which an electronic transition that is forbidden by symmetry is made allowed by a vibronic promoting mode that is non-totally symmetric and thus transiently reduces the molecular symmetry. The absence of the true origin transition in the one-photon excitation spectrum is a result of strict centrosymmetry in the n-octane crystal. The electronic symmetry of the ground electronic state is 1Ag and so the upper level must also be 1Ag. These are designated 11Ag and 21Ag respectively with the superscript 1 indicating single states. Use of n-nonane or n-heptane in place of n-octane induces significant intensity in the 0-0 transition due to the necessary loss of symmetry. The vibration-less origin transition is two-photon allowed as is observed. The strongly allowed one-photon transition at about 312 nm is ca. 10<sup>5</sup> times stronger than the first vibronic feature of the excitation spectrum in the 21Ag region. The vibronically active normal modes are expected to be of vibrational bu symmetry because of the proximity of the strong 11Bu excitation. The two-photon high resolution spectral feature when compared to the high resolution fluorescence spectrum permits determination of the frequency the promoting mode as 86 cm<sup>−</sup>1, which is an observed mode of this bu symmetry. All of the above were the contribution of Bryan E. Kohler and his students [59–64] following the present author's initial contributions [65–68]. The above argumen<sup>t</sup> as to the presence of a low lying 21Ag state in octatetraene is airtight on experimental grounds. The development of the theoretical situation showed very early [69–72] that the low-lying 21Ag state is derived from doubly excited configurations that mix with singly excited configurations of the same symmetry to become the lowest excited state. It is now possible to compute the electronic excitations of octatetraene and ge<sup>t</sup> very good agreemen<sup>t</sup> with experiment using advanced *ab initio* methods [73].

wavenumbers/1000

**Figure 7.** (**a**) The one-photon fluorescence excitation spectrum of octatetraene in n-octane matrix at 4.2 K. The arrows mark the vibronically induced transitions to the forbidden 21Ag excited state. The intense broad feature at 32,200 cm<sup>−</sup><sup>1</sup> is the vibration-less origin of the allowed electronic transition to the 11Bu excited state; (**b**) Two-photon fluorescence excitation spectrum of the same sample. All of the features are due to transitions to the 21Ag excited state. [59,67,68].

**Figure 8.** The upper trace left is the beginning of the fluorescence spectrum; the upper trace right is the beginning of the one-photon fluorescence excitation spectrum; the lower trace is the beginning of the two-photon fluorescence excitation spectrum. The two traces on the right are the same as the extreme left of Figure 7. [59,67,68].

Analysis of the pattern of the vibrational intensity of the strongly allowed absorption transition of spectra, like Figure 6, for a variety of polyenes shows that the features are due to the ca. 1500 and 1100 cm<sup>−</sup><sup>1</sup> C=C and C–C stretching modes. This is the case for both the transitions to or from the upper 21Ag electronic state and the allowed electronic absorption transition to the 11Bu state. This means that these excited states differ from the ground electronic 11Ag state by displacement along the total symmetric double bond and single bond contraction/expansion vibrational modes. The direction of the displacement is to upper states that have a reversal in their bond alternation pattern. The relative intensity of a vibration in an electronic transition is related to this displacement, which leads to finite overlap between the vibrational modes of the two states involved. These overlap integrals are called Franck–Condon factors. This displacement is the major mechanism that results in the vibrational structure and overall width of electronic spectra and in Raman activity of totally symmetric modes. This is called the A-term or Condon contribution to Raman scattering. This depends on the displacement of the potential energy minimum for the low-energy, strongly allowed electronic excitations. In a more general treatment of Raman scattering, there can also be cases where the mechanism of Raman intensity is due to the fact that some displacements of the atoms result in a change in the intensity of the electric dipole transition moment rather than a change in the energy of the potential energy surface that is required for non-zero Franck–Condon factors. This is especially important in electronic transitions that are between states and have electronic symmetries that cause the electric dipole transition moment to vanish at the equilibrium geometry, as is the case for the 11Ag to 21Agelectronic transitions of linear polyenes giving rise to the pattern of features shown above.

The relevance of the presence of a low-lying doubly excited 21Ag state in finite polyenes to properties of polyacetylene, as pointed out by Torii and Tasumi, is that polyacetylene in its ground electronic state must necessarily be an admixture of structures that have the standard pattern of bond alternation with another structure that has its bond order pattern reversed from that of the optimized structure [74]. The result is a polyacetylene with equal bond lengths. This has been investigated by Torii and Tasumi using the CASSCF (Complete Active Space Self Consistent Field) method. Their results for N = 6, dodecahexaene with an STO-3G basis set give an energy difference per CH group for the optimized geometry and that of the bond-reversed geometry of 2000 cm<sup>−</sup>1. A slight inflection in the potential hints at the evolution toward a double minimum expected for longer chains.

There have been several recent synthetic efforts at preparation of oligopolyenes with defined lengths. One of these [75] is relevant to the location of the 21Ag state as a function of chain length and thus to the argumen<sup>t</sup> above concerning its ultimate admixture with the ground state. However, these are experimental studies and therefore the excitation energies of the 21Ag state from the 11Ag ground state already includes the mixing which will push the 2Ag state up as it is repelled by the receding 11Ag ground state. It is, in fact, already known that in finite polyenes the two lowest 1Ag states are vibronically coupled [76–80].

### **7. Raman Vibrational Spectroscopy of Finite Polyenes**

Another of these synthetic efforts concentrates on the Raman spectroscopy of finite conjugated polyenes [81]. This study is relevant to our investigation in progress of the preparation of polyacetylene *in situ* in a host–guest inclusion complex. In [81] the Raman spectra of a series of di-t-butyl polyenes with *N* = 3 to *N* = 12 C=C bonds were measured and discussed. The main observations for this series of compounds from [81] are:


All of the above are observed in the photochemical elimination polymerization reaction discussed below, which proceed for gues<sup>t</sup> molecules in urea channel inclusion channels. In addition:


### **8. In Situ Synthesis of Oriented Insulated Polyacetylene**

Polyacetylene, whatever its limitations in terms of degree of polymerization, suffers from being entirely insoluble, conformationally disordered, subject to cross linking, and having a lack of macroscopic crystallinity. All of these factors make the characterization of this material problematic. We describe here the beginnings of a method for *in situ* generation of polyacetylene in a host inclusion complex. The objective of this is to force the *trans*-polyacetylene chain to be in its fully extended all s-*trans* configuration, to prevent close proximity of neighboring polyacetylene chains and to provide exclusion of oxygen. The initial approach to this objective is the use of urea inclusion compounds containing a reactive species. Urea inclusion compounds (UICs) are self-assembling crystalline structures formed from solution with inclusion of gues<sup>t</sup> hydrophobic compounds with an extended structure. The most extensive example is the series of n-alkane urea inclusion complexes. These have a macroscopic hexagonal structure with the n-alkanes being ordered in two dimensions but disordered about their axis of rotation coincident with the hexagonal c-axis. An important aspect of urea inclusion crystals for our application is that the terminal atoms of the gues<sup>t</sup> species are in contact in the complex. The urea host is not stable in the absence of the gues<sup>t</sup> species; the urea lattice grows around the gues<sup>t</sup> species. It is known that radical polymerization can be induced in UIC's, e.g., diene guests form very high molecular weight polymer poly(butadiene) [82–84]. The rigid urea tunnels allow gues<sup>t</sup> rotations, translations, and lateral diffusion along the tunnel axis [82]. The gues<sup>t</sup> molecules are generally more mobile than in single component molecular crystals, where reaction requires precise initial alignment.

The initial focus for our experimental effort aimed at the synthesis of long chain conjugated polyenes, in particular polyacetyelene in an all*-trans* extended conformation, begun with the preparation of a crystalline urea inclusion complex with E,E–1,4–diiodo–1,3–butadiene (DIBD) [85]. In Figure 9, we present the crystal structure of DIBD–urea as obtained by X-ray diffraction at 90 K viewed perpendicular and parallel to the channel axis [85]. Like all hexagonal urea inclusions, these crystals form as parallel channels in a host lattice densely packed with gues<sup>t</sup> species [82–85]. The gues<sup>t</sup> monomers, DIBD, are in end-to-end contact with each other and entrapped by the hydrogen bonded urea host. The internal diameter of the channels varies periodically along the channel from 5.5 to 5.8 Å. The separation of the parallel channels is 8.2 Å.

**Figure 9.** Representations of the commensurate, fully-ordered single-crystal DIBD–urea inclusion compound (UIC) complex as obtained by X-ray diffraction at 90 K viewed along the c (left panel) and b (right panel) crystal axes. Redrawn from structure Crystallographic Information File, cif of [85].

The DIBD–urea complex used in this study is unusual in that it is a commensurate structure, in contrast to most other UICs such as those formed by n-alkanes. DIBD–urea crystals have Raman features due to DIBD at 1600 and 1250 cm<sup>−</sup>1. The strong tetragonal urea feature at 1010 cm<sup>−</sup><sup>1</sup> is shifted to 1022 cm<sup>−</sup>1, the value observed for hexagonal urea. Irradiation with ultraviolet (UV) light results in new Raman modes near 1509 and 1125 cm<sup>−</sup><sup>1</sup> [86] (Figure 10a), nearly identical to spectra of *trans*-(CH)x prepared in solution [87]. The 254 nm radiation used in this experiment has very limited penetration into the DIBD–urea complex due to the high optical density. The change in composition of the urea channels is expected to be as shown in Figure 11.

**Figure 10.** Raman spectra with 532 nm excitation of (**a**) DIBD–UIC after irradiation at 254 nm; (**b**) *trans* –(CH)x; (**c**) crystalline DIBD; and (**d**) tetragonal urea [86]. The νn values at the top are the mode frequencies for polyacetylene fundamental transitions and their overtones.

The overall progress of this irreversible second order sequential reaction is anticipated to be as illustrated in Figure 12. Continued progress requires that there be considerable axial diffusional motion of the chains in order to take up the space in the channel that has been vacated by the loss of iodine. This type of diffusional motion has been demonstrated with n-alkanes in UICs [88,89]. The overall progress of the reaction can be monitored with loss of mass due to release of iodine.

**Figure 11.** Schematic figure showing progress of the photochemical reaction from diiodobutadiene to polyacetylene with an intermediate stage showing a dimer and a trimer. The picture is to scale showing the large loss of channel filling with loss of iodine. There is a 2:1 ratio in the number of carbons in the bottom/top panels. Dimers and trimers have been shown by UV-vis of the extracted material. Longer chains have been shown by Raman.

**Figure 12.** Irreversible sequential second order kinetics. The blue decreasing curve is for the monomer. The dimer peaks at p = 1 where f2 = f1, the trimer peaks at p = 2 where f3 = f2, etc. The number of carbons in the most frequent species is CN = 4(p + 1). The line colors differentiate the time dependence of the sequentially larger oligomers with their increasing delay.

### **9. Summary of Lessons from the Literature on Polyacetylene**

The treatment of the vibrational level structure of polyacetylene unambiguously eliminates the possibility that bond alternation can be supported in the absence of terminal double bond end effects. Even if the barriers were much larger, the system would undergo tunneling. In reading the literature on this subject, the statement that the potential energy has a double minimum seems often to be the equivalent that the bond lengths will alternate. This is a fallacy. It is refreshing that all of the experimental methods used to establish bond alternation fail, in some cases in rather spectacular fashion.

In the case of X-ray diffraction the nature of the samples is essentially fibers and the periodic variation of internuclear separation is a relatively fine detail to pick out reliably from such data and that further blurred by disorder.

In the case of the NMR experiments, one wonders why the experiment was done. It could not be expected to succeed in its aim as designed, and it did not. It showed instead that polyacetylene undergoes cross-linking or some other form of conversion of double bonds to single bonds. This is also noted in the Raman studies of [35]. The better NMR experiment might be to use 50% random 13C labeling.

It appears that the original interpretation of the extensive resonance Raman studies in terms of chain heterogeneity is sound with upper limits of chain length on the order of 40 double bonds.

It is our contention and that of several other workers that the Raman spectra that are observed in what is called "polyacetylene" is due to finite chains. This interpretation is consistent with observations

that the most strongly enhanced feature in the Raman spectrum vary with the Raman excitation wavelength in an expected way with lower energy excitation resulting in stronger lower frequency vibrations due to preferential enhancement by longer chains. Examination of the available published data shows that there is no reliable evidence from diffraction, NMR, or IR data for bond alternation that can be demonstrated to be uncompromised by finite chains or finite conjugation segments or other ambiguities.

Our outlook is as follows. The methodology of *in situ* synthesis of oriented insulated polyacetylene has recently taken a step forward in the preparation of urea inclusion complexes with a different gues<sup>t</sup> species that has the standard hexagonal lattice and morphology. This is expected to result in much faster reaction kinetics due to slightly looser packing, ease of orientation of the crystal channel axis, and the capability of making polyacetylene chains that are the length of the crystal. This is currently ca. 1.5 cm.

**Acknowledgments:** The author thanks Steluta A. Dinca, Damian G. Allis, Michael D. Moskowitz, Michael B. Sponsler, Mark Hollingsworth, and Luke Daemon for on-going access for discussions on specialist points.

**Conflicts of Interest:** The authors declare no conflict of interest.
