Meeting Report
Cranfield University
5th July 1995
The Group was welcomed by Sue Impey on behalf of the Surface Analysis
Group in the Support to Industry and Management Section at Cranfield.
The morning session was opened by Dave Sykes (University of
Loughborough) who discussed the reasoning behind the choice of technique. At Loughborough
the initial selection of technique may be based on general "suitability" and
will include considerations such as area to be analysed, sample conductivity, stability
and whether quantitative or qualitative elemental, chemical and/or trace analysis is
required. Inevitably, pragmatic considerations such as instrument availability, sample
throughput required and, of course, cost also play a significant role. Overall, the
analysis should be fit for purpose and should "meet the customers expectations at a
price they can afford". Dave suggested that the analyst should disregard the possible
consequences of the results, eg. job losses, litigation or bankruptcy, although these are
bound to be in the back of their mind.
As examples, Dave cited a company making Plaster of Paris cottages
which had problems with painting due to the use of excess silicone mould release agent. In
this case SSIMS is used if the requirement is show that silicones are there but XPS would
be used if the problem was "How much was there?".
A second routine dynamic SIMS problem was calibrating the acceptable
level of trace metals from three furnaces in a semiconductor manufacturing plant. Once the
levels rise above the critical value the whole batch of material must be discarded because
the processed devices will be useless. Although this result places a severe cost on the
plant, the cost of rejecting the batch of completed devices is far greater.
Dave cited a case study where only a tiny fragment from a valuable
Greek sculpture was available. In this case the sample size virtually decided that the
LIMA technique should be used. At the other extreme, a building with hundreds of 2m x 1 m
glass panels showed dark discolouration which was an indication of corrosion of the
coating on the glass. The consequences of this analysis were dire for the installer who
had not stored the glass properly.
Graham Smith (Shell Research) described how fundamental analysis of oil
seal degradation can lead to suggestions for better tests.
An engine oil is a sophisticated blend of additives which has to pass a
long list of engine manufacturers specified tests. The crankshaft oil seal is a
fluoroelastomer and must not degrade in hot engine oil and leak so it is subjected to a
7-day compatibility test where it is heated in fresh oil at 150C under tension and then a
full tensile test is performed on the seal. A loss of less than 20% of the initial tensile
strength is acceptable but more than this and the oil fails the test.
These fluoroelastomers are well suited to XPS analysis, particularly
using a monochromator, because a large chemical shift occurs for each fluorine
substitution onto the hydrocarbon backbone. SEM images of the seals after testing show a
crazy paving effect and EDX analysis clearly indicates loss of fluorine from the surface.
XRD indicates that the fluorine component has broken down and reacted with calcium
hydroxide in the filler to produce calcium fluoride. Cutting a 10 degree taper section and
analysing using small spot XPS shows that it is one part of the copolymer blend that
degrades preferentially in this manner and this has provided some additional flexibility
for developers to formulate new materials. For example, the presence of amines catalyses
the cross-linking of CHF groups and allows HF to be formed which can react with the
filler. Use of CF2 groups instead would cause less reactive F2 to be formed.
Graham pointed out that although the test is carried out using a specific fluoroelastomer,
this is no longer the favoured formulation used to make oil seals. Also, the test is run
using fresh oil which is more likely to fail the test and is unrealistic in a motoring
context. The work described here has increased understanding of the degradation process
and enabled a more relevant test procedure to be proposed.
Dick Chappell (Pilkingtons) described the AES/XPS analysis carried out
in support of glass production at Pilkingtons. XPS is used for general surface analysis
and depth profiling AES for analysis and characterisation of coatings on glass.
The bottom surface of float glass contains predominantly Si, O, Na, Ca,
Mg and some Sn from the molten tin upon which the glass is floated to form sheets. The top
surface has a lower level of Sn (from vapour deposition) and some sulphate and chloride.
Tin concentration profiles vary considerably due to glass thickness and bath conditions.
XPS produces thermal damage after extended exposure and AES gives electromigration
problems and peak shifting that can be so severe as to subject the channeltrons to
unexpectedly high count rates - causing reduced lifetimes. An additional problem is the
accurate positioning of both electron and ion beams due to local charge build up and
deflection of the beams. With high sputtering requirements it is also possible to
contaminate the electron optics with insulating material that requires mechanical removal.
During weathering Na, Ca and Mg are preferentially leached, in that
order, but under static conditions the concentrations of the leached elements build up and
Mg is often back precipitated.
There are many patented coatings for glass to achieve desired heat
transmission or reflection. Most of these are carried out on an enormous scale on whole
sheets of glass on conveyor belt systems. Some recently developed systems use sputtering
or CVD coatings technology on the float bath to coat as the glass sheet is manufactured.
Dick described a number of these coatings in which a metallic layer is incorporated as the
active layer but this has to be protected by overcoating with a stabilising oxide such as
SnO2. One particular system has an aluminium on silver layer structure. During heat
treatment, as used for toughening, the aluminium diffuses through the silver layer
clearing it of impurities as it goes. Some coatings are surprisingly sophisticated - K
glasses have an F-doped SnO2 layer with an SiCO undercoat to produce an interface which
reduces reflection.
With such coatings technology it is no surprise to find that many
problems occur because the coating is susceptible to "corrosion" under adverse
conditions. Many problems arise because the glass has not been treated with sufficient
care and according to instructions during storage and installation.
Sue Impey (Cranfield University) described work carried out within the
surface analytical facility in support of the research and development programmes in the
School of Industrial and Management Science at Cranfield
The first example was a study of the Cr6+ content of the fumes from MIG
welding. Reference Cr203 and CrO3 were studied along with samples of the fumes collected
on filter papers. Initial results indicated no Cr6+ but it was found that Cr5+ degraded in
air to Cr3+ although it was stable in vacuum. This meant developing rapid sample
collection strategies and anaerobic sample transfer.
A second example was given of a Ti:Fe alloy rocket fuel container
corrosion due to the presence of N204 oxidiser. An important aspect of this work was
definition of the oxide thickness. XPS was used with argon ion depth depth profiling but
the alloy is so reactive that it is impossible to remove the oxide completely at the
interface. The approach adopted was to ensure that sufficient oxygen was available to keep
the surface oxide in equilibrium but to monitor the onset of carbide formation as an
indicator of the position of the interface. This occurs because there is no carbon in the
oxide itself but once the titanium substrate is exposed it reacts with CO in the vacuum to
give a measurable carbide-like feature.
A third example was taken from examination of NiZn plated high strength
steel fasteners. The normal cadmium plated devices are susceptible to hydrogen
embrittlement and the NiZn alternative can also be affected depending on plating time and
pH. An XPS depth profiling study of very thin plated films showed that under the right
conditions the Ni layer forms an interfacial barrier to hydrogen between the Zn and steel.
In this study the shallow crater edge was imaged by SEM/EDX to confirm the existence of
the Ni layer at the interface.
Finally, Sue described some work on graded coatings consisting of mixed
BC and WC deposited on steel using a Ni alloy buffer layer. BC has been found to be the
most erosion resistant ceramic and is finding increasing applications in high-tech
coatings research. XPS is particularly advantageous for studying the chemistry of BC.
Len Hazell (CSMA Ltd) gave his personal view of the surface science
support requirements of UK Industry and how well this is serviced by the various surface
analytical facilities within the UK. The recession has enforced many recent changes in UK
Industry's view of "outsourcing" their analytical requirements and any business
now has a number of options to solve its surface problems. The UK Surface Analysis Forum Surface
Science Services Survey shows that the UK is unique in its provision of a roughly even
split of contract surface analytical facilities involving Universities, Industry and
Government or ex-Government agencies. CSMA Ltd is the only independent commercial
laboratory offering both electron and ion spectroscopy service in the UK.
Len listed some of the factors which influence manufacturing industry
to sub-contract work into the UK service providers network. Clearly, cost, speed of
response, quality, expertise and availability ofspecialised facilities are key parameters.
However, there are also various European or UK Government subsidy schemes designed to
foster closer ties between Industry and the Academic community for longer term projects.
There is a major difference between the UK surface analysis provision
and the "rest of the world". Elsewhere, it is illegal for Universities to become
involved in day-to-day problem solving in competition to industry and/or independent
analytical laboratories. By contrast, in the UK, Government policy has virtually
necessitated that Universities compete. Len's view is that this undermines the true value
of the industrial laboratory and the independent analysts and does not help the UK surface
science base in the long term. He also believes that the Universities do not want to be
put under such pressure and would rather be doing what they do best, ie. teaching,
fundamental underpinning of the techniques, instrument development and long term applied
scientific research. These are areas which he feels have been neglected recently in the
UK. For example, with few exceptions, all instrument development is being carried out by
the major instrument manufacturers (who are no longer UK owned) and practical technical
underpinning is limited to standardisation work at NPL and BSI initiatives (funded largely
by UK User groups).
In summary, Len indicated that problems within manufacturing industry
will not go away and the cost culture will introduce more last minute requests for support
which demand faster response. We are continually seeing problems at a smaller scale and
this trend will continue. While the accountants are still in charge of UK industry there
will be more "outsourcing" in the immediate future with some sole provider
arrangements and increasing access to networks of analytical services. However, this will
only form a stable surface analysis infrastructure in the UK if the government encourages
a clearer definition of roles and ensures adequate funding for development.
The afternoon session was dedicated to workshops.
| Leaders |
Len Hazell/Graham Beamson |
| Raporteur |
Len Hazell |
The workshop began with a discussion of where charge neutralisation is
essential in XPS and AES and what the potential benefits are. For XPS, the higher energy
resolution from monochromators is available with increased sensitivity, lower sample
damage and smaller spot sizes but the neutralisation has to cope with uniform and
differential charging. In AES, charge neutralisation is often necessary to obtain a
spectrum of any kind.
Charging problems arises on insulating materials due to an imbalance of
charge "in" compared to charge "out" of the surface. The precise
nature and level of charging depends on the incident beam energy, flux distribution,
surface inhomogeneity and roughness effects. These create local electrostatic fields which
distort the electron optics behaviour of the analyser and affect the voltage referencing
and sensitivity, making quantification unreliable.
Graham Beamson described some methods of charge neutralisation that he
had utilised on the Scienta instrument. Simply using a conventional twin anode as well as
a monochromator can help but the low electron flux available, the sample damage and the
increase in background make this approach unsatisfactory. Early types of electron flood
gun had low emission current at the low beam energies used and were particularly poor if
the sample angle was changed, as for example in a take-off-angle experiment. The latest
flood gun is fairly sophisticated and specially designed for the Scienta. It uses a pulsed
filament heating/extraction cycle to give a very monochromatic line profile beam onto the
sample. It is a powerful source so it incorporates a dog-leg to avoid infra red radiation
from the hot filament damaging the sample. It works well at all sample angles.
Discussion indicated that it was probably not necessary to go to such
lengths to obtain satisfactory charge neutralisation on other instruments and some clear
ideas were obtained on the approach to take. It appears to be most important to align the
sample normal to the flood gun in order to avoid the charge build up deflecting the low
energy electrons, typically less than 10eV. the flood gun energy should be adjusted to the
lowest possible value but such that any additional voltage shifts the spectrum to a lower
binding energy by the same amount.
Other methods such as using a grid suspended above the sample surface,
bombarding a nearby gold foil with ~ 1OOeV electrons to generate a flood of lower energy
secondaries, applying very thin in-situ metallic coatings, use of wet samples and using
argon ion bombardment to induce surface conductivity in polymers through surface damage
were explored. Bob Bulpett (Brunel) had successfully eliminated differential charging on
one oxide sample by applying an extremely thin gold coating from a SEM plasma coating
unit. The gold peaks were not intense and did not affect the analysis but confusion due to
differential charging was removed.
| Leader |
Steve Harris (British Aerospace) |
| Raporteur |
Sukanta Biswas (Cascade Scientific) |
The group consisted of companies with in house surface analysis
instrumentation, equipment manufacturers and service laboratories. Several questions
associated with QA were discussed. Surface analysis itself is quality assured from some
labs but not others. Certainly the feeling was that the customers required quality
assurance in analysis and there was a demand for QA work in surface analysis, particularly
using the mature techniques of AES, XPS and SIMS. ISO9000 is thought of as a good idea and
the QA aspect in general was considered important in promoting surface analysis.
The topic of "round robins" was discussed and thought to be a
good idea if the experiments are designed properly. Wrong choices of materials and
experiments often wastes time and money.
Current limitations in the use of surface analytical techniques in the
QA environment covered topics such as, the requirements for better databases, totally
automated analysis, reduced cost instruments. All were seen to be important to encourage
wider use of surface analysis in manufacturing.
The question of whether instruments will ever be widely installed
directly on the production line to check if a product passes or fails was thought to
depend on what information needs to be provided. Whether pictures, numbers, depth profiles
or some other result is presented was probably a very individual thing, varying from
client to client. Also, the question arises as to how we present ourselves to the
customer. Are we surface scientists, surface analysts or technicians? Ultimately, we need
to change the customers, and their managements, views of the utility of surface analysis
in order to keep surface science alive in QA applications.
| Leader |
Bob Wild (IAC, Bristol) |
| Raporteur |
Albert Carley (University College of Wales, Cardiff) |
The workshop concentrated on sputter depth profiling. The two main
problems encountered were changes in the chemical state of the surface, such as reduction
of metal oxides by ion bombardment, and preferential sputtering which leads to a build up
of one of the overlayer components at the substrate overlayer interface. It was felt that
the processes involved were not well enough understood to allow reconstruction of the true
profile from the measured one. It is important that the bottom of the eroded crater is as
smooth as possible. This is best achieved by rotating the sample during sputtering which
also leads to a better defined interface. However, few instruments have this capability.
Measuring the crater depth after profiling allows the sputter rate to
be calibrated. Techniques such as SEM and AFM (provided the scanned area is large enough)
may be useful for this.
After a brief discussion on some experimental details, such as choice
of source parameters and beam and sample geometry, Bob gave his views on the potential
problems involved in computer controlled depth profiling. The main difficulty lies in
allowing the computer to calculate signal intensities using an adequate algorithm applied
within a chosen energy window. Using integrated peak intensities rather than peak heights
accommodates changes in peak shape (chemical state) across the interface.
The final discussion concerned the need for reference materials in
depth profiling. Few exist, although several are proposed and there is a need for a cheap,
widely available material. Pilkingtons tin oxide coated K-glass was suggested as a
possible secondary standard and Dick Chappell at Pilkingtons kindly offered to supply some
pieces upon request |
