Cluster mechanisms

From the earliest days of radiation chemistry it has been known that acetylene polymerizes to a cuprene-like ( alprene ) solid (5, 6,25,28). The characteristics of the polymerization—e.g., lack of effect of temperature, doso rate, and pressure on polymer yield and negligible effect of radical scavengers—led Lind (24) to postulate an ion cluster mechanism. 

A striking observation that lacks a satisfactory explanation is the existence of magic numbers, i.e. the fact that in a distribution of clusters some species with a certain number of carbon atoms are much more abundant than others. The exact clustering mechanisms are not completely understood, and, as noted by Rohifing et al.(IO), the origin of the observed distribution of clusters may depend upon instrumental factors. Accounting for this fact, however, there still seems to be a preference for clusters with certain numbers of atoms which cannot be explained solely as due to the experimental conditions. 

The chalcogenide precursors possess many talents. Apart from forming the chalcogenide ions, they also form complexes with metal ions. As noted at the beginning of this section, and ignoring the distinction between ion-by-ion and hydroxide cluster mechanisms treated previonsly, CD processes can be divided according to two basic mechanisms participation of free snlphide ions (the... 

Probably the least-known aspect of the CD process is what determines the nucle-ation on the substrate and the subsequent film growth. In considering this aspect, we will treat the ion-by-ion and hydroxide cluster mechanisms separately, although there will be many features in common. The principles discussed should be the same for both the free chalcogenide and the complex-decomposition mechanisms. 

An expected difference between ion-by-ion and hydroxide (or any other cluster) mechanisms is that in the latter, since colloids from the solution stick to the snbstrate snrface, the crystal size is not expected to change greatly with film thickness (it may increase to some extent, since the colloids themselves can grow via an ion-by-ion process on the crystals). For ion-by-ion growth, it is likely that crystal growth occnrs on nncleii already present on the substrate, and therefore crystal size can increase with increasing deposition. 

For the cluster mechanism, while growth and termination can be similarly explained, the induction period is less obvious. The hydroxide cluster can start to adsorb on the substrate immediately after immersion of the substrate in the deposition solution, yet experiments have shown that film growth often does not occur for some time. While the reason for this is not clear, it may be connected with the... 

It will be obvious that the cluster mechanism of deposition is unlikely to lead to an oriented film, since the clusters would have to align themselves with the substrate lattice, either on adsorption or subsequently. Therefore an epitaxial film is highly suggestive of an ion-by-ion growth, which is more likely to be directed by the substrate. 

This effect is shown in Eigure 2.9 for CdSe films deposited from baths containing Cd complexed with NTA (nitrilotriacetate) and Na2SeS03 as a Se source. The nanocrystal size, measured by both XRD and TEM, varied from ca. 3 nm up to 20 nm with increase in temperature and/or change in mechanism from a cluster mechanism to an ion-by-ion deposition. The optical bandgap shifts from 1.8 eV (for bulk, zincblende CdSe) to ca. 2.4 eV for the smallest nanocrystals (ca. 3 nm). 

The main difference between the two mechanisms as they relate to crystal size (discussed in Sec. 2.6) is that the cluster mechanism is three dimensional while the ion-by-ion one is mainly two dimensional. Crystal size in the former is limited largely by the amount of reactant per nucleus The more nuclei, the smaller the final crystal size, since the same concentration of reactants is divided over more nuclei. Temperature affects this by stabilizing (kinetically) smaller nuclei as temperature is lowered, thus increasing the number of nuclei at lower temperature. 

Fig. 2.9 Transmission spectra of CD CdSe films deposited at various temperatures from CdS04/NTA/Na2SeS03 baths. All samples deposited by hydroxide cluster mechanism except 80°C HC (high complex), which proceeded via the ion-by-ion mechanism. The effective bandgap can be approximated by the wavelength (photon energy) a little shorter (higher) than the absorption onset. A second absorption knee, ca. 0.4 eV to higher energy of the initial onset, seen clearly in the 41 °C and 80°C samples, is due to a transition from the spin-orbit valence level to the lowest conduction level and is commonly observed in these films.

When reading the literature, in many (probably most) cases it is not clear whether the deposition proceeds by an ion-by-ion process. The reason is that, unless another mechanism is specifically discussed, it is often assumed that the deposition proceeds via the ion-by-ion mechanism. If the exact deposition parameters are known, which mechanism is operative can, in most cases, be calculated. Two criteria have often been cited in the literature as proof of deposition via the ion-by-ion mechanism. One is epitaxial deposition of the CD film. (Epitaxy refers to growth of one material on another in such a way as to result in coherence between the lattice of the substrate and the deposit. Often—although not necessarily—the lattice of the deposit is aligned in the same direction as that of the substrate.) This is based on the expectation that a cluster mechanism will not result in an epitaxial film for this to occur, clusters of maybe thousands of atoms would need to be able to rearrange themselves on the substrate. Some examples of epitaxial growth are given in Sections 3.4.2 and 4.I.5.2. 

More recently, there have been a small number of studies that provide strong evidence for the ion-by-ion mechanism. It must be pointed out, however, that while it is not very difficult to distinguish between an ion-by-ion and cluster mechanism in most cases, it is much more difficult to distinguish between a simple ion-by-ion and a complex-decomposition ion-by-ion mechanism. Therefore most investigations that conclude an ion-by-ion mechanism is operative, while usually assuming the simple ion-by-ion process, do not distinguish between the simple and complex-decomposition pathways. 

One investigation has shown a clear-cut boundary in the crystal size of films (CdSe, CdS, and, to a lesser extent, PbSe), depending on whether the deposition occurred via an ion-by-ion or a cluster mechanism [15]. The solution conditions... 

If the Cd is adsorbed on the substrate (either directly or indirectly through a hydroxide linkage) or on previously deposited CdS, then the same reaction would occur. If the CdS so formed remained bound to the substrate (it is assumed that CdS generated on previously deposited CdS would remain bound), the result would be film growth by an ion-by-ion, complex-decomposition mechanism. As with the cluster mechanism, it is difficult to distinguish experimentally between the complex-decomposition mechanism and the free-anion pathway. 

The general shape of the growth of CD films as a function of time is often similar for the cluster mechanism as for the ion-by-ion mechanism (Fig. 2.4). Figure 3.4 shows an actual example of CdSe deposition from a solution (containing nitrilo-... 

While the pH of the deposition solution (based on the cluster mechanism) has been found to increase by as much as 0.8 during the deposition (see Sec. 3.3.2), this increase was found to be considerably greater, up to 2.2 pH units, under illumination [92]. This could provide some clues about the mechanism of the CdSe formation under illumination. A possible pathway that could account for... 

In the other study [155], ammonia-complexed Hg(N03)2 was mixed with the selenosulphate solution. As for the corresponding HgS deposition, a white precipitate formed on addition of ammonia to the HgCNOs) [Eq. (4.9)]. This precipitate dissolved partly in the excess ammonia used, due to formation of various am-mine complexes, and completely when the selenosulphate solution was added, due to additional formation of selenosulphate (and maybe sulphite from the excess sulphite in the selenosulphite solution) complexes. It is likely that mixed ammine-selenosulphate/sulphite complexes were formed. The deposition was carried out on polyester substrates (the transparencies used in overhead projectors) at 10°C. Deposition occurred over several hours to a terminal thickness of ca. 250 nm. Bulk precipitation occurred in parallel with the deposition, suggesting that the cluster mechanism was dominant. 

---