Conference-ICSFS-14-Nanostructured Multimetal Granular Thin Films : How to Control Chaos

A simple method for the fabrication of films with an high density of nanometric metal grains is presented and discussed. The method is based on the successive evaporation of different metals (Au, Al and In), with different melting points: after each evaporation, rapid thermal annealing treatments are used to induce agglomeration of the deposited material. The nanograin agglomeration (in particular of the Al film) resulted strongly dependent on the previous nanograin (gold) distribution and concentration. Statistics of gold based nanograin films, deposited on silicon dioxide, are presented for different gold thickness and annealing parameters. Results on Al induced agglomeration of a successively evaporated thin film are shown for different Au nanoparticle distribution and concentration. A final indium deposition, with suitable annealing temperature and time, produces films with a very high density of metal nanoparticles (more than 2000 nanograins/μm, average radius between 5 and 10 nm). [DOI: 10.1380/ejssnt.2008.503]


I. INTRODUCTION
This work describes a fabrication technique for granular metal films, with an high density of nanometric grains.In the past, films based on nanogranular layers have been fabricated by using both semiconductors [1,2] and metals [3].Several works have been dedicated to the study of the tunnel driven electrical conduction, in a direction perpendicular to the plane of the film [4][5][6].In these works a layer of nanocrystals, embedded in an insulating material, is placed between two electrodes, used for current-voltage measurements.The principal target of these works was to investigate the possibility of employing these granular materials in the field of electronic memories, with important applications for the modern electronic industry and semiconductor market.For this important application the intergrain conduction in the film plane (grain cross-talk) was considered an unwwanted effect, and the nanograin density was purposely maintained low in order to decrease the cross talk.
Conversely, if the density of nanocrystals is sufficiently high, the hopping driven (Mott) conduction in the plane of the film can become very important, with very peculiar characteristics [7][8][9][10][11]: the main target of the present work is to describe a method for the fabrication of island films with very high density of grains with nanometric dimensions (smaller than 10 nm).Despite their chaotic nature, granular films conducting in the plane of the film have important applications, for example as active layers in sensing devices based on catalysis; recently big efforts have been dedicated to granular materials for dye sensitized solar cells [12].
Percolation theory can be used to model the hopping driven electrical nanogranular film conduction [13].This opens new possibilities for the realization of active devices based on random path conduction [14]: electrical conduction of networks near the percolation threshold can be changed by means of a control gate.
The fabrication method presented in this work is based on successive evaporations of different metals (gold, aluminium and indium, AuAlIn) and Rapid Thermal Annealing (RTA) treatments.Grain density and dimension distributions are investigated, as a function of metal thickness deposition and RTA parameters.In particular in Section II SEM and AFM images of gold granular films are shown, and statistics on the agglomeration of gold nanograins are reported and discussed.It will be shown than an high density of small Au nanograins can be obtained by starting from thin Au films.In order to increase the nanograin density, Al agglomeration on Au nanograin is investigated.In Section III it is shown that the Al nanocrystal formation is strongly dependent on the previously deposited Au grains distributions and concentration.A very high density of grains is obtained by a deposition and annealing of a third material (In), chosen with melting point very low with respect to the one of Al and Au.

II. GOLD GRANULAR FILMS
Statistical distributions of gold nanograins, on silicon dioxide, have been investigated for different process parameters (initial film thickness, RTA time and temperature).Au nanograin thin films have been fabricated by the following procedure.A) An uniform Au thin film has been deposited on a 2 µm thick silicon dioxide layer, on silicon substrate, by means of thermal evaporation of gold; the residual vacuum was 10 −6 mb, and the evaporation rate was about 0.5 nm/s.The thickness and evaporation rate of the as-deposited film has been measured by means of a pre-calibrated quartz microbalance.B) A post-deposition thermal treatment has been performed by means of a Rapid Thermal Annealing (RTA) processor (JPELEC100) in inert (N 2 ) gas at atmospheric pressure.The same up and down ramp characteristics have been employed: after a cleaning cycle (vacuum pumping for 2 min., then purge with nitrogen till the pressure arise to 1 bar), the temperature is increased (in nitrogen flux) with a rate of 50 • C/s, till the hold-on temperature that is then maintained for a suitable time; the cooling down is performed by means of nitrogen, at a rate of about 50-60 • C/s.In order to establish and control the grain dimensions distribution, several hold-on temperatures and times have been tested with different Au initial film thickness.The nanograin distribution, for different process parameters (initial Au film thickness, RTA temperature), has been determined by means of scanning electron microscopy (SEM, field emission gun FEG JEOL6500F) imaging and Atomic Force Microscopy (AFM PSIA100), no-contact mode.
Figure 1 reports three SEM photos, taken with the same magnification for nanograin Au films obtained with different initial thickness (respectively 6 nm, 8 nm, 10 nm).RTA parameters have been the same for the three films: hold on temperature and time of 600 • C and 1 min.It appears from the images that the initial thickness is very determinant for the final nanograin size distribution.In order to determine the statistical distributions of nanograin radius, each grain has been selected from a SEM image by means of a threshold level method.The grains which can be considered almost circular have been fitted with a circle (using the radius as fitting parameter).If the grain showed an elliptic shape, as in the case of lower grain densities (bigger grains), two circles have been used for fitting, considering that the grain itself can arise by the coalescence of two single grains.The statistical nanograin radius distribution showed a gaussian shape for all the investigated samples.Table I reports the average radius, the variance and the grain density for the three films; statistics and grain densities have been obtained by counting and measuring the radius of all the nanocrystals in an area of about 1 × 0.8 µm 2 .To be noted that the radius increases with the initial film thickness, from 9.13 nm (initial thickness 6 nm) to 32.5 nm (initial thickness 10 nm); according to this, the grain density shows a drastic decreasing from 1079 grains/µm 2 to 70 grains/µm 2 .
AFM investigation has been performed in order to confirm the SEM imaging investigation, and in particular for giving an estimation of the average nanograin thickness.Figure 2 reports two AFM images, taken in no-contact mode, of respectively the 6 nm film and the 10 nm film (same RTA parameters 600 • C, 1 min.).Nanograin thickness have been measured as vertical displacement between approximately the centre of the nanograin, and a clean zone.As expected the vertical dimension of the grains  increases with the increasing of the average grain radius, and with the initial thickness of the Au film.In particular the average thickness of nanograins for the 6 nm film resulted of 9.3 ± 2.7 nm, meanwhile for the 10 nm film the nanograin average thickness resulted of 35 ± 7 nm.In order to investigate the effect of post-deposition annealing parameters on the Au nanograin agglomeration, films with the same Au deposition thickness have been annealed with different hold-on temperatures.Some results are shown in Fig. 3, that reports SEM pictures (same magnification) of three nanograin films obtained with the same initial Au deposition (thickness of 6 nm), annealed for the same hold on time (1 min.)but different temperatures (700, 800 and 900 • C).From the SEM images it can be noted that the nanograin average dimensions and concentration do not show a strong dependency on the post deposition annealing temperatures.while the temperature has been changed between 600 and 800 • C. The table shows that for both thickness the diversities in radius average are not very strong; also the grain density is slightly affected by the annealing temperature.
In the case of thinner films (6 nm, higher nanograin density) a clear trend can be noted: the average dimensions of gold nanocrystals increases, and consequently the grain density decreases, for higher temperatures.In the case of 8 nm Au thickness, differences are not significant, and there is not a well definite trend; it appears that for low grain densities the effect of the annealing temperature, in the investigated range, is almost uneffective.It can be said that gold nanograin distributions (nanograin average radius and concentration) depend strongly on the initial film thickness, and slightly by the annealing parameters (at least in the RTA parameters investigated).

III. GOLD-ALUMINIUM AND GOLD-ALUMINIUM-INDIUM GRANULAR FILMS
With the aim of increasing the total grain concentration, gold granular films, fabricated as reported in Section II, have been used as a template for the successive fabrication of aluminium grain, that possibly would fill up the spaces between gold nanoparticles.An aluminium thin film has been deposited by thermal evaporation on the gold granular film, and the further agglomeration of nanograin has been investigated as a function of the successive RTA process parameters.The main result is that the agglomeration of the Al film is very critical, and it is strongly dependent both on the Au nanograin distribution, previously obtained, and on the second RTA parameters.In particular it has been noted that for all the RTA temperatures (from 450 to 650 • C, below the eutectic point of gold-aluminum alloy) and annealing times (from 1 min.to 20 min.)investigated there is not formation of Al nanograins, if the Au nanograin density is not sufficiently high.The SEM photo of Fig. 4(a) shows a typical result on a film obtained by depositing 6 nm of aluminium on a gold granular film, with a too low density of Au nanoparticles (average radius of 24.23 nm, and a grain density of 138 grains/µm 2 , obtained by 1 min. of RTA at 600 • C of an 8 nm Au film, see Table I).After a second RTA of 10 min.at 550 • C, the aluminium film does not show any agglomeration, but as it can be observed in the SEM image, it remains almost a continuous film between the gold nanoparticles.By using a higher Au nanograin density, a suitable hold-on temperature and time for the second RTA process can be found for Al agglomeration.The photo of IV.CONCLUSIONS AND FUTURE WORKS

FIG. 1 :
FIG. 1: SEM pictures of Au nanograin films; RTA parameters are the same (1 min.at 600 • C); films have been obtained with a different Au initial thickness Th, as specified on the images.

FIG. 2 :
FIG. 2: AFM measurements of two Au granular films, obtained by the same RTA parameters (1 min.at 600 • C), and an initial Au thickness of 6 nm (on the left) and 10 nm (on the right).

FIG. 4 :
FIG. 4: SEM pictures of AuAl films, obtained with different gold nanograin distributions and second (Al film) RTA time and temperatures (see text); the two images have a different magnification.

Fig. 4 (FIG. 5 :FIG. 6 :
FIG. 5: SEM picture of (b) an AuAl granular film, compared with (a) the Au nanograin film, used as a template for Al grain formation.a

TABLE I :
Radius means and variances for films obtained with the same annealing time (1 min.)and temperature (600 • C) but different initial gold thickness.

TABLE II :
Radius means and variaces for films obtained by the same gold evaporation (6 nm thick) and different annealing temperatures (same annealing time, 1 min.)