Vacuum Web Coating

ZnS vs Titania for high Refractive Index coatings

**  Question  **

We currently metallise PET with aluminium. We are considering retro-fitting our coater so we can manufacture HRI products.

Firstly, what are the pros and cons of zinc suplide compared to titanium dioxide (or any other alternatives you suggest we consider).
Second, do we need a plasma treater ? We have heard conflicting information - some stating a plasma treater is essential and some stating it isn't. Please help.

**  Answer  **

Zinc Sulphide tends to be a soft porous coating but can be deposited at metallizing speeds (100's m/min).  The ZnS can be thermally evaporated from a slot source that spans the whole web width. This ease of deposition means that the cost of ZnS is much lower than for depositing titania.  The ZnS can dissociate and recombine easily and so stoichiometry is generally not an issue.  Operators may not be so happy with the material as it can smell when the system is vented because of the residual Sulphur.

Titania tends to be denser, harder and is usually deposited by slower techniques ( 1 - 10's m/min).  If you are after a hard wearing, highest quality hologram then titania will probably be the preferred material.

Many companies using the coating for holograms want to deposit the coating and emboss later.  It is possible to emboss into the polymer through the ZnS and still get a good hologram.  The ZnS is soft and deformable but it is a test of how good the adhesion is between the ZnS and substrate.  ZnS has the advantage that it does less wear on the holographic shims than the harder titania and so production costs can be further reduced.

Part of the problem of depositing titania is getting the stoichiometry right. Sputtering from a titanium target and reacting the coating with oxygen to produce titania is not neceaasrily easy.  The oxygen cannot be controlled selectively to reach the growing coating and not the sputtering cathode & so the cathode gets poisoned. The sputtering rate of titania is 20x slower than for the metal and so the process tends to avalanche to sputtering slowly from an oxide target.  There are methods of correcting for this runaway process but they all push up the deposition costs. Ideally if the coating is to be deposited by sputtering a dual cathode with an AC power supply including  arc control would be preferred.    One system built a few years ago used a sputtering source with an additional oxygen ion gun to control the oxygen input to a minimum excess.  This system operated at around 2 m/min winding speed.

It is possible to deposit titania from electron beam deposition sources but getting a consistent stoichiometry is not trivial.  Often an additional oxygen plasma is used both to densify the coating to make it more like a sputtered coating but also as a method of reducing the excess oxygen that is required to convert the metal to titania.

Uniformity can also be an issue.  Sputtering has a deposition rate fall off towards the ends of the cathode so that to get high uniformity across the whole web width requires a cathode length wider than the web width.   For electron beam sources it depends if you have a series of individual sources that have their deposition flux integrated across the width or if there is a single sweeping electron beam with a single crucible that spans the web width.

Critical to both will be the position of the pumping system in the system you are planning to convert and the method of feeding in the oxygen gas.  Ideally the pumping will be symmetric about the web centreline (including any cryopumps) if this is not the case there are additional problems if achieving uniformity.  Moisture from the webs can be a source of oxygen to the coating as well as the oxygen supplied to the plasma as the controlled gas source.  If there is asymmetric pumping then there will be a pressure gradient across the web and it then becomes difficult to deliver sufficient oxygen, an no more, to all parts of the sputtering cathode or vapour flux from the e-beam source.  Any imbalance will lead to non-stoichiometric coatings.

Plasma cleaning.  There are many different plasma cleaning sources available. Some work better than others.  The choice of power, time and gas composition can all affect the effectiveness of any plasma treatment.  In general it is better to plasma clean than not.  Most polymer film has contamination on the surface. This contamination is a source of poor adhesion. It either needs to be better bound into the polymer web or volatilised and pumped away leaving behind the polymer web surface.  If the surface is under treated there the adhesion will not be optimised. If the surface is over treated it may still have a high surface energy but it will also have a weak boundary layer on the surface caused by too much polymer chain scission by the over treatment.  Thus plasma treatment is a balance that has to be optimised for each polymer film. Changing supplier can be it is best to re-optimise as different film suppliers will use different proprietary formulations and so the surfaces may well be different.

Argon plasma can roughen the surface and sputter efficiently but there is no mechanism for converting any hydrocarbons into gaseous species that can be pumped away and so what ever is sputtered from the surface may well fall back and still be a contaminant to the surface even though it has been plasma treated.  Thus using an oxygen/argon plasma is probably the most favoured  plasma treatment gas mixture for the widest range of polymer webs.  There are exceptions and sometimes a different bonding site is required that makes oxygen less favoured.

My advice would always to be to keep the process as simple as possible.  Hence do not plasma treat if you do not have to.  Thus I would do some trials first to check if the adhesion of the coating is good enough i.e. fit-for-purpose. 

For ZnS my expectation is that the adhesion will be a problem if you do not use any plasma treatment. I would expect a plasma treatment to be a critical part of the process. 

I

if you elect to go with titania then it may be much less critical, depending on the process you adopt.  If you are using an oxygen additional plasma this may aid your adhesion, if you are using a sputtering source this will already be bombarding your substrate to some extent and will have a better adhesion than from an evaporation source.

I think this just about answers your question as far as I can.  The choices will depend on the substrate material, the system you are wanting to convert, the cost you have to produce the coating coupled to the performance of the coating.

Choice of gases for vacuum plasma treatment.

**Question**

How to determine the best plasma combination for metallised PET. Normally which combination of gas is best suited for metallised PET films and in what percentage. What is generally the best colour for plasma when seen during vacuum metallisation.

**Answer**

There is no set proportion of gases for plasma treatment. Ideally there will be both argon and oxygen present.  The oxygen can be derived from pure oxygen or from water vapour and so some systems locate the plasma treatment in the winding zone to get both a higher pressure and to make use of the water vapour extracted from the polymer as it unwinds.   The water is broken down in the plasma to provide the plasma with the oxygen.  Systems that have the plasma treatment zone in the deposition zone area usually rely more on the oxygen being provided from bottled gas.

Another consideration in using oxygen is that the excess oxygen that does not react with any surface contamination to form carbon monoxide or carbon dioxide has to be pumped away through pumps that often contain thin films of hot oil.  This can be a potential explosion risk.  This is more of a risk as the gas is compressed at the rotary pumps. Hence it is important to have something like a nitrogen purge, added to the pumping line before the backing pumps, to protect the pumps from too much oxygen. The aim is to never have an oxygen concentration greater than ambient air (20%).

So most typically there will be a gas feed into the system of argon and oxygen with the oxygen being of the order 10% - 20%.   The heavier argon provides the physical bombardment of the surface and the oxygen provides the chemical recombination that converts the hydrocarbons into volatile species that can be pumped away. Without the oxygen these would be sputtered from the surface but most likely would fall back onto the surface and recontaminate the surface.

There is a big problem in talking about the colour of plasmas. Around 8% of the population has some problem or other with colour vision. On top of this if you have ever looked at a plasma for some time your colour perception will be changed as you eye sensors become saturated with some wavelengths and de-tune the sensitivity and so when you look away all the colours in the room will be different to when you viewed them before looking at the plasma.  Another problem is people’s different description of the same colour. What is pink to one person may be lilac to another or red to someone else.  Thus the only true method of describing colour is using a scanning spectrometer.

I would normally use a combination of mass flow controllers and/or pressure monitoring along with the voltage and current information from the power supply to control the plasma and trust the instruments rather than my perception of the colour.

Deterioration of metal adhesion

1) How fast will the adhesion of metallised layer on film for Polyester film (MPET) deteriorate?

2) Process steps:
12µm laminated with 11pt IP 208, coated with NC lacquer and then gone
through the tunnels. The laminated film board coated with UV ink and then
follow by UV OPV (Over print varnish). Perform adhesion test using 3M tape.
The metallised layer was taped off from the film?
Why did it happen?

Answers.

The first thing I always advise is to check out the plane of
failure.  It is always the assumption that the failure is at the interface
but it may not be.  Assuming it is a failure at the interface then common
causes of lack of adhesion are a weak boundary layer or poor adhesion and
wetting of the deposited aluminium.

The weak boundary layer can be as a result of low molecular weight material
left on the surface before metallizing. The low molecular weight material
may be residual oligomer from the polymerisation process, or additives such
as slip agents or pre-treatments that have been applied to help film
handling.

The surface of the polymer may or may not have been treated using flame,
corona or plasma treatment to raise the surface energy.  It is preferable
for the surface to have been treated as this raises the polymer film surface
energy and will usually increase the adhesion.  If the surface has been
treated but the adhesion still poor then it may be because the treatment has
been ineffective this could be for a variety of reasons.  The treatment
could have been done a long time before metallization and the effects
degraded with time.  The treatment may not have been optimised for the
particular film used. The film may have been treated on one side only rather
than both sides and contamination from the back surface transferred onto the
front surface when the film was re-wound.  If the film was plasma treated
using only an argon atmosphere rather than a mixed argon/oxygen atmosphere
which is preferable.

If the adhesion between the metal and polymer is poor then it is also
possible for material to migrate to the interface over time and this can
progressively reduce the adhesion.

Thus I would start by confirming the true plane of failure.
Assuming it is truly a failure at the interface I would then track back
looking at the film specification and checking on any/all pre-treatments
taking particular interest in the treatment type and time between treatment
and metallization.