• 沒有找到結果。

Phase transition from the dot array to the stripe pattern without

Part II: The study of the surface structures and phase transitions

Chapter 3 Results and Discussions of the self-assemble layers of

3.8 Phase transition from the dot array to the stripe pattern without

In special conditions, the stripe pattern can directly form from the dot array without crossing the metastable phases, for instance in Figure 3-27 with the two serial STM images depended on working potentials.

Figure 3-27. The two serial STM images dependent on the applied potential. (a)-(c) represent the behavior of the stripe pattern on step edge. (d)-(f) reflect the effect of the stripe pattern on the domain boundary of the dot array. Each corresponding potential is given on the right lower corner of each figure. (a)-(c): Ubias = 100 mV and It = 0.50 nA. (d)-(f): Ubias = 400 mV and It = 0.1 nA. The sizes of (a)-(f) are 57.62×57.62 nm2.

The first serial images from (a) to (b) indicate the preferential stripe pattern formation on the step edge marked by the white arrow. The second serial images from (d) to (e) show the stripe pattern stacking along the domain boundary indicated by the black arrow. The white and the black lines of Figure 3-27 represent the growth orientations of the stripe row on the step edge and the domain boundary. These sites are favorable sites for the viologens adsorption to form the stripe pattern. Such stripe pattern is not the result from phase transition of the metastable phases, but formed from the radical cationic viologen molecules of the solution species. It is apparent that one stripe pattern row of Figure 3-27(b) absorbs on the step edge without following the orientations of the step edges of Figure 3-27(a) which shows many acute angle edges. The step edge and the domain boundary have common characters including favorable reaction and adsorption sites, reduction of the motion, and vacancies so that

( ) a ( ) b ( ) c

( ) d ( ) e ( ) f

0 mV -115 mV

0 mV -120 mV

-120 mV

-140 mV

the viologens can absorb and form the stripe pattern rows directly on these postions.160-162 As the potential up changes to -120 mV and -140 mV, this stripe pattern row disappears and instead forms the metastable phases following neighboring metastable phases as shown in Figure 3-27(c) and (f). This shows that the applied potentials can overcome the bonding hindrance of the π − π stacking interaction of the stripe pattern formation. Furthermore, there is an interesting phenomenon in the black ellipse region of Figure 3-27(e) and (f). A short stripe row at the step edge can still exit on the surface. The reason is that the short stripe row locates on the kink site, and hence it is not easy to change. At more negative potential, this row also will follow the big domain orientation.

The serial STM images of Figure 3-28 show the effect of the domain size of the stripe pattern when decreasing the potential. In Figure 3-28(a), the enclosed black ellipse indicates the initial absorption position from the solution species at the domain boundary and then the stripe rows start to grow until attaching the next boundary as shown in Figure 3-28(b). The black and the white arrows of Figure 3-28 represent the larger and the smaller domains of the stripe pattern, respectively. Here, the smaller domain has the two stripe pattern rows and the larger domain includes the five stripe rows. The smaller domain will disappear at more negative potential, but the larger domain is maintained. That reflects the larger domain with the stronger intermolecular interaction then the smaller domain. pattern” is formed at more negative potential. Phase transition process from the stripe pattern to the closed stacking stripe pattern is shown in Figure 3-29(a) to (b). Type A of the metastable phase can also be found on the surface of Figure 3-29(a). The reason is mentioned in section 3.7. Figure 3-29(b) shows that many defects and dislocations are produced on the surface after phase transition due to the surface relaxation.

( ) a ( ) b ( ) c

-140 mV

-100 mV -120 mV

Figure 3-29. The stripe pattern of (a) transfers to the closed stacking stripe pattern. (c) and (d) show the stability of the dimer stripe pattern row. (a): Ubias = 400 mV, It = 0.1 nA, and at -170 mV. (b): Ubias = 400 mV, It = 0.1 nA, and at -255 mV. (c): Ubias = 432 mV, It = 1.0 nA, and at -215 mV. (d): Ubias = 432 mV, It = 1 nA, and at -290 mV. The sizes of (a)-(d) are 57.62×57.62 nm2.

The whole surface filled with the dimer stripe pattern row phase is shown in Figure 3-29(c). The dimer stripe pattern row almost does not occur in phase transition as shown from Figure 3-29(c) to (d). This case is very rare to observe because the dimer stripe pattern row configuration is that the alkyl chains have to insert in the spaces between neighboring bipyridiums. Once the dimer stripe pattern row is formed, such stable and rigid feature stays until the hydrogen evolution reaction. A dimer stripe pattern row is also similar to the closed stacking stripe pattern according to their structure models of Figures 3-26(d) and 3-31(a), respectively. The opportunity to occur phase transition from the dimer stripe pattern row to the stripe pattern is near

( ) a ( ) b

( ) c ( ) d

some defects as indicated by the white and black arrow regions of Figure 3-29(d).

The surface structure of the closed stacking stripe pattern is shown in Figure 3-30.

The white arrow represents the direction along [100] and the black arrow is along the major orientation of the closed stacking stripe pattern arrangement. The angle δ between the white and black arrows is a 50 degree.

Figure 3-30. (a) and (b) are the STM images of the closed stacking stripe pattern at -290 mV. (c) and (d) show the relations of the closed stacking stripe pattern domains.

(a): Ubias = 466 mV, It = 0.1 nA. (b): Ubias = 8 mV, It = 1.5 nA. (c): Ubias = 400 mV, It

= 0.1 0.1 nA, and -290 mV. (d): The sizes of (a) and (b) are 7.02×7.02 nm2 and that of (c) and (d) are 57.62×57.62 nm2.

Compared to the stripe pattern, the transverse distance between neighboring stripe rows is closer, and the biprydinium core orientations between neighboring rows are slightly rotation relation in order to avoid great overlapping of the alkyl chains resulting from the alkyl chains directly absorbed on top of the alkyl chains. Further, for the closed stacking stripe pattern, a bilayer configuration cannot be observed, unlike the stripe pattern with the topmost and the underlying layers. Figure 3-30(c)

( ) a ( ) b

5

o

90

o

( ) c ( ) d

and (d) show the domain orientation relations of the closed stacking stripe pattern.

The rotation relations of a 5 degree and a 90 degree are the same as that of the stripe pattern.

Based on the above observations, a structure model is drawn in Figure 3-31(a). The distance between neighboring molecules along the longitude direction is a 0.44 nm which is a 0.40 between the stripe pattern and a 0.54 nm for the metastable phases.

The distance from the stripe pattern to closed stacking stripe pattern is changed to 0.44 nm from 0.40 nm. That effect means that the alkyl chains are not in the end to end of the alkyl chains configuration replaced by inserting into the spaces between the alkyl chains of the neighboring closed stacking stripe pattern rows. Not only the π − stacking interaction plays an important role for the closed stacking stripe pattern formation but also the hydrogen bonding of N-H. Figure 3-31(b) indicates phase transition mechanism from the stripe pattern to the closed stripe pattern in a side view.

The bilayer formation structure of the stripe pattern is in the upper panel of Figure 3-31(b) and the monolayer structure of the closed stacking stripe pattern is in the lower panel of Figure 3-31(b). The closed stacking stripe pattern can only be formed at more negative potential compared to the stripe pattern. The working potential influences the phase formation. The electrode at more negative potential enhances the interaction between viologen molecules and the anion layer.163 Therefore, the increasing electrostatic force drives the topmost layer of the stripe pattern to move down to form the monolayer feature. Due to such vertical motion, the structure of the stripe pattern greatly rearranges in its geometrical stacking and hence the closed stacking stripe pattern generates so many vacant defects and dislocations as seen in Figures 3-32, 3-29(b), and 3-30(d).

Figure 3-31. (a) shows a structure model of the closed stacking stripe pattern. (b) represents the mechanism from the stripe pattern to the closed stacking stripe pattern in side view.

Figure 3-32. The vacant and dislocation defects appear on the closed stacking stripe pattern (Ubias = 435 mV, It = 0.1 nA, and 57.62×57.62 nm2).

( ) a

( ) b

δ

3.10 Chloride Desorption and Dimer phase

When the chlorine anions start to desorbs from the Cu(100) surface, the morphology of the closed stacking stripe pattern is destroyed to a cluster-like formation at around -400 mV as shown in Figure 3-33(a). The cluster-like formation is expected to be a dimer formation resulting in an ordered-dimer phase formation. In this potential region, no ordered structure is found and the Cu(100)-1×1 structure is identified by removing viologen molecules on the surface in Figure 3-33(b). Viologen molecules absorb on the pure Cu(100)-1× 1 surface. Figure 3-33(c) shows the mechanism of the chloride desorption. At sufficiently negative potential, the electrode surface is accumulating many extra electrons which make negatively charged chlorine desorb from the surface. Here, the interaction between the copper atoms and the chlorine anions is overcompensated by electrostatic repulsive force. After the chloride desorption, the HER occurs if no viologen molecules are on the Cl/Cu(100) surface.

However, when adding viologen molecules on the Cl/Cu(100) surface, the HER is blocked due to the presence of the dimer phase (Figure 3-24(d)) on the Cu(100)-1×1 surface which reduces the charge transport probability of the hydrogen cations from the solution. From Figure 3-33(d), the dimer phase occupies a large area on the bare Cu(100)-1×1 surface. That phenomenon is also reflected in the CV curve resulting in the smooth and broad window at the region Ⅳ. Only the two domain orientations are observed and the angle between the two domains is a 90 degree.

: Chlorine anions

The surface structure of the dimer phase is shown in Figure 3-34(a). The intermolecular π − π interactions between the radical viologen molecules and the hydrogen bonding of N-H between the nitrogen atoms and carboxylic acid groups play an important role to form an ordered dimer phase.164 It is noticed that the dimer phase is established by the hydrogen bonding of N-H between neighboring dimer rows; otherwise the domain size of the dimer phase would be small and many domains with many orientations would appear. Figure 3-34(b) shows the underlying pure Cu(100)-1×1 surface structure after removing the surface molecules. The structure model of the dimer phase is drawn in Figure 3-34(c).

Figure 3-34. The surface morphology and the atomic structure of mono-cationic V-(C7-COOH)2 molecules in (a) of Ubias = 466 mV, It = 0.1 nA and (b) Ubias = 2 mV, It

= 15 nA in 7.09×7.09 nm2 at -405 mV. (c) is the structure model of the dimer phase.

(d) illustrates the schematic solution/solid interface. (e) describes the retarding HER effect.

( ) a ( ) b

( ) c

( ) d ( ) e

The orientations of the bipyridinium cores do not perfectly matched the [011]

direction due to the intermolecular interaction of the hydrogen bonding. After a 90 degree rotation, corresponding positions of viologen molecules almost do not change on the two-fold symmetry Cu(100)-1×1 surface. The chloride desorption apparently destroys the ordered phase (closed stacking stripe pattern) to form a disordered phase.

Unlike the chloride desorption, the HER does not break the surface configuration because viologen molecules rearrange into the dimer phase on the pure Cu(100)-1 × 1 surface by the stronger electrostatic attraction force between the copper electrode and viologen molecules as shown in Figure 3-34(d). The stable dimer phase exits on the pure Cu(100)-1 × 1 surface resulting in retarding the HER due to favorable HER sites occupied by the viologens as indicated by the arrow in Figure 3-34(e). The reaction formula of the HER is 2H + 2e ⟶ H for the acid solution case.

In a special case, the viologen solution is directly injected into the cell on the Cl/Cu(100) surface within the HER and then form a dimer phase as shown in Figure 3-35. Such experiment sometimes resulted in many small domains, but a 90 degree rotation relation is still maintained due to the two-fold symmetry of the Cu(100)-1×1 surface. The rotation relation means that the positions of viologen molecules prefer to absorb on the four-fold site. Once viologen molecules absorb on the Cu(100)-1 × 1 surface at very negative potential, viologen molecules are trapped on the surface by electrostatic attraction force without much diffusion resulting in the small domains.

The charge transport and rearrangement occur locally only on the pure Cu(100)-1 × 1 surface.

Figure 3-35. Viologens exposure onto the Cl/Cu(100) surface at the hydrogen evolution reaction region. (a): Ubias =400 mV, It = 0.1 nA, and at -425 mV. (b): Ubias

=400 mV, It = 0.1 nA, and at -400 mV. The sizes of (a) and (b) are 57.62×57.62 nm2.

( ) a ( ) b

3.11 Determining the orientations of the alkyl chains for all phases

The bipyridinium cores show the high density of electron states and hence the positions of the bipyridinium cores are apparent to determine for all phases. The orientations of the alkyl chains are also clear seen from the STM images only for the dicationic viologens as shown in Figures 3-8(a) and 3-13(a). Therefore, all orientations for the dot array and the oblique row phase are confirmed. However, the alkyl chains of the phases for the radical viologens are hard to observe due to different heights. The heights of the topmost level of the alkyl chains are lower than that of the bipyridiniums, unlike the dicationic viologens as illustrating in Figures 3-36(a) for the dicationic viologens case and (b) for the radical viologen case in a side view due to different stacking configurations between the “face on” of the dicationic viologens and the “face to face” (π − π stacking) of the radical viologens.

Figure 3-36. (a) is the “face on” configuration of the dicationic viologens. (b) describes the “face to face” configuration of the radical viologens.

In order to determine the orientations of the alkyl chains of the radical viologens, different orientation cases are considered as shown in Figure 3-37. As the alkyl chains are parallel to the bipryidiniums for the metastable phases, it is hard to form the greater overlapping between the alkyl chains of the neighboring molecules, and hence the orientations should have the angle different from the (a) and (b) as shown in (i).

For the stripe pattern of (c), a closed stacking stripe pattern of (d), and the dimer phase of (e) and (f) conditions, the orientations of the alkyl chains are not parallel to that of the bipryidiniums. There are no enough spaces for neighboring molecules in the π − π stacking to form the stripe pattern and a closed stacking stripe pattern. It is also hard to form the dimer phase due to the larger overlapping of the alkyl chains.

Therefore, the orientations of the alkyl chains are near parallel to that of the bipryidiniums as indicated by (ii) and (iii).

( ) a

( ) b

Figure 3-37. The various orientations of the alkyl chains correspond to the metastable phases of (a) and (b), a stripe pattern (c), the closed stacking stripe pattern (d), and the dimer phases of (e) and (f). All image sizes are 7.09×7.09 nm2 and the scanning parameters are the same as the above sections.

3.12 Summary

In summary, the seven phases are found including the dot array, an oblique row phase, the metastable phases, a stripe pattern, the closed stacking stripe pattern, and the dimer phases for a 0.10 mM carboxylated viologens within 10 mM HCl on a Cu(100)-1×1 surface. Among these phases, only the dimer phase on the pure Cu(100)-1×1 surface, and the other phases are on the Cl/Cu(100)-c(2×2) surface. The phases of the dot array and the oblique row phase are formed in the dicationic state and the other phases are in the radical state of the viologens. The carboxylated viologen molecules at the dicationic state prefer the face-on configuration on the surface and form the dot array and the oblique row phase due to electrostatic interaction between the dications of viologen molecules and the chlorine anion layer.

The well-ordered network configurations of the dot array and the oblique row phase are braced with the hydrogen bonds of N-H and O-H, respectively. Further, carboxylic

( ) a ( ) b

( ) d ( ) c

( ) e ( ) f

or

(i)

(ii) (iii)

acid groups favorably interact with the bipyridinium cores rather than neighboring carboxylic acid groups. Therefore, the observable rate of the dot array is greater than that of the oblique row phase.

After one electron charge transfer to the dicationic state, phase transition occurs from the dot array to the metastable phases including type A and B in the face to face configuration. The cluster-like phase is called type A and the stripe-like phase is called type B. There are various orientations of type A in the range from 0 to 23 degree due to the stiff stacking with long alkyl chains. Type B combines at least four molecules in a π − π stacking formation. Type A and B can frequently exchange at -130 mV. Such exchange process is ready to occur phase transition from type B (C) of the metastable phases to the stripe phase.

Because the formation of type B of the metastable phase is more similar to the stripe pattern than type A, type B can immediately transfer to the stripe pattern, not type A. Even type A can remain on the surface at more negative potentials. Further, the four kinds of the stripe patterns are observed. These four kinds represent the makes positively charged viologens strongly attracted on the surface. Because of the enhanced electrostatic attraction force between viologen molecules and the electrode, the bilayer of the stripe pattern transfers to the monolayer of the closed stacking stripe pattern. During this phase transition, the viologen coverage does not change.

Therefore, surface density of viologen molecules increases and then neighboring stripe rows are closer and concomitantly, and hence many defects appear in the closed stacking stripe pattern. For the closed stacking stripe pattern, the major interactions are the π − π stacking between neighboring bipyridiniums and the hydrogen bonding of N-H. The closed stacking stripe pattern is stable on the surface until the chlorine desorption which occurs at more negative potential (around -380 mV).

During the chlorine desorption, the anion template is removed which breaks up the formation of the closed stacking stripe pattern and leads to the formation of the separate dimers. After that, viologen molecules in the radical cation state still can stay on the surface by electrostatic attraction forces with the copper surface.

As the anion layer of the chlorine completely desorbs from the Cu(100)-1 × 1 surface, the disordered dimer phase starts to form an ordered dimer phase along the [011] direction on the Cu(100)-1×1 surface. Viologen molecules rearrange to the ordered dimer phase on the pure Cu(100)-1 × 1 surface which makes the HER

retarded compared to the case of the pure Cl/Cu(100)-c(2 × 2) surface without the molecules. At more negative potential, the HER occurs and the cathode current is increasing exponentially. The rapid reaction results in a desorption of the molecules from the surface or in a greatly enhanced mobility which makes it hard to observe any phases on the Cu(100)-1 × 1 surface.

The intermolecular interactions are the hydrogen bonds of N-H for a dot array phase, the metastable phases, a stripe pattern (bilayer formation), the closed stacking stripe pattern, and a dimer phase to construct the network formation, besides the oblique row phase comes from the hydrogen bonding of O-H. Because the hydrogen bonding of N-H is more stable than that of O-H, the oblique row phase tends to transfer to the dot array phase resulting in the smaller occupied area and the lower observed probability compared to the dot array phase.

Chapter 4 The bilayer formation of carboxylic viologens and the influence of the chlorine and the bromine anion layers on the growth behavior of the thin film

In this chapter, the phenomenon of the bilayer formation and the influences of the chlorine and bromine anion layer on the growth behavior of viologen thin films are discussed. The bilayer formation is confirmed by a dilute viologen solution on the Cl/Cu(100) surface, 0.1 mM viologen solution on a HOPG surface, and 1.0 mM viologen solution on the Cl/Cu(100) surface. The choice of a chlorine or bromine anion layer can result in a different behavior of phase transitions by comparison of the 1.0 mM viologen solution on the Cl/Cu(100) surface and the 0.1 mM viologen about 60 ml HCl is injected into the cell to reduce the viologen concentration (0.1 mM) at 0 mV. After reduction the concentration of the cell, the dot array also is found on

In this chapter, the phenomenon of the bilayer formation and the influences of the chlorine and bromine anion layer on the growth behavior of viologen thin films are discussed. The bilayer formation is confirmed by a dilute viologen solution on the Cl/Cu(100) surface, 0.1 mM viologen solution on a HOPG surface, and 1.0 mM viologen solution on the Cl/Cu(100) surface. The choice of a chlorine or bromine anion layer can result in a different behavior of phase transitions by comparison of the 1.0 mM viologen solution on the Cl/Cu(100) surface and the 0.1 mM viologen about 60 ml HCl is injected into the cell to reduce the viologen concentration (0.1 mM) at 0 mV. After reduction the concentration of the cell, the dot array also is found on