In recent years, inorganic transparent barrier layers such as aluminium oxide or silicon oxide deposited onto polymer films have emerged as an attractive alternative to polymer based transparent barrier layers for flexible food packaging materials. For this application, barrier properties against water vapour and oxygen are critical. Aluminium oxide coatings can provide good barrier levels at thicknesses in the nanometre range compared to several micrometres for polymer-based barrier layer. These ceramic barrier coatings are now being produced on a large scale using industrial high speed vacuum deposition techniques, here, reactive evaporation on a 'boat-type' roll-to-roll metalliser. For the thin barrier layer to be useful in its final packaging application, it needs to be protected. This can be either via lamination or via an additional top coat. This study reports on acrylate topcoats, but also undercoats on aluminium oxide coated biaxially oriented polypropylene films. The effect of the acrylate layer on barrier levels and surface topography and roughness was investigated. The acrylate was found to smooth the substrate surface and improve barrier properties. Furthermore, the activation energy for water vapour and oxygen permeation was determined in order to investigate barrier mechanisms. The oxide coated film was, additionally, converted via adhesive lamination, which also provided improvement in barrier levels.

lntroductionWe present data showing, for the first time, the effect of N 2 0 flow rate on the thickness and interfacial nitrogen concentration of oxynitride films grown in a conventional furnace at one atmosphere. Our analysis, based on the concentration of nitric oxide (NO), can explain these observations as well as the variation in the published data on the kinetics of film growth. This understanding should facilitate the application of these films as tunnel dielectrics and gate dielectrics to improve device reliability (1,2,). Fig.1 shows our kinetic data with replotted results from three other groups (3,4,5). The time dependence ranges from near saturation (3,5) to an almost linear behavior (4). Our data is intermediate between these two extremes. The vertical displacement is due to differences in the oxide thickness before N 2 0 exposure which does not affect the tendency to thickness saturation. Observing the variation in this data, we decided to examine the chemistry of the process more closely. Experimental ProcedureIn this w rk, we focused on dielectrics equal to or greater than 1001 and selected p g l e growth temperature of 950°C. A thermal oxide, 75 thick, was grown on 125 mm wafers in N d 0 2 at 950OC. This initial oxide was exposed to N2O for times U to two hours. The N20 flow rate was varied from 4 to 1 f S L M . The ox nitride growth runs were camed out with all five boats h l l y loaded. A single test wafer at the center of the load was used for thickness and nitrogen measurement. Loading and flow rate variations produce changes in gas residence time which was found to be a. critical parameter. Thickness was measured with an ellipsometer with fixed wavelength (632.8nm) and refractive index (1.465). The nitrogen profile was measured by SIMS using Cs sputtering and CsN+ detection. Experimental Results and DiscussionFig. 2 shows the time dependence of oxynitride thickness for flow rates of 4 SLM and 14 SLM. The higher flow rate shows a greater tendenc to thickness saturation.This effect is shown more fully in 6f. 3 where a fixed time of 35 minutes was used for oxynitri ation of 7.5 nm oxide. SIMS was used to determine the interfacial nitrogen concentration (Nint) ,as shown in Fig. 4. It has been conjectured that nitric oxide (NO) is the critical species responsible for the interfacial nitrogen concentration (6).We show that this conjecture is consistent with published kinetic data and demonstrate below that the dependence of thickness and Nint on flow rate can be ex lained by variation in the concentration of NO at the w d r with gas residence time.Heat-up of the incoming N20 by the 950°C furnace causes decomposition by the following two reactions (7). N20+ N2+0 U11[21 Remaining atomic oxygen will ra idly recombine into 0 2 . At YOO°C-lOOOoC approximately &% of the incoming N20 decomposes according to reaction [ 11 and 2% decomposes according to [ 21 (8). Extrapolation of published kinetic data from 750°C to 950°C indicates that the half-life of the N20 decomposition by [l] and [2] is approximately ...

No abstract