[Skip to content]

TWI
Search our Site
.

Coating quality in arc spraying - getting it right

TWI Bulletin, May/June 1992

 

Mohammed Amin
Mohammed Amin

Since 1966 when Amin joined TWI he has earned a reputation as an authority on synergic control in MIG welding. The results of his work are incorporated in MIG welding power sources the world over.

He has also studied the arc and metal transfer behaviour of various materials used in the MIG process and has originated a Patent for a shielding gas mixture for optimum weld quality. More recently his attention has been directed to the arc spray process and TWI's arc spraying equipment.

More must be learnt about the process variables of arc spraying if its popularity as a reclamation and corrosion resistant coating method is to continue to increase. Mohammed Amin surveys existing literature and examines the effects of spray particle characteristics on coating quality.


Arc spraying involves striking an arc between two consumable wire electrodes. As the wires melt they are fed continuously into the arc where the molten droplets generated are broken into finer particles and projected at high velocity on to the target component by a jet of compressed air (or gas) fed through the arc spray gun (Fig.1). The arc spraying process is a major consumer of solid wires of the type common to MIG and TIG welding.

Fig.1. A spray beam of 'atomised' particles
Fig.1. A spray beam of 'atomised' particles

Arc spraying has been used for a number of years for depositing corrosion resistant coatings, in particular zinc and aluminium, on to large outdoor steel structures such as bridges and ships. There are also many engineering applications where arc sprayed coatings are used, e.g. reclamation of worn or over-machined components.

Assuming specified preparation of the substrate, the coating quality is determined mainly by the spray particle characteristics of size, temperature and velocity. In turn, these characteristics depend on the spray process parameters including standoff distance, wire feed speed (or current), arc voltage and atomising air pressure. The effect of substrate material to be coated on the particle characteristics is considered to be insignificant. The dependence of the particle characteristics on the process parameters and effects of the particle characteristics on coating quality are reviewed here.

Particle characteristics

The spray particle characteristics - size, temperature and velocity - are related to the quality of the coating in terms of:

  • Bond strength - both between substrate and coating (i.e. adhesion), and between the layers themselves (ie. cohesion);
  • Porosity (or coating density);
  • Surface finish (or surface roughness);
  • Oxide content;
  • Residual stresses.

The dependence of coating quality on particle characteristics and process parameters is summarised in Tables 1 and 2, respectively.

Table 1 Summary of particle characteristics for optimum coating quality

Coating quality Requirement Particle characteristics
Size Velocity Temperature
Bond strength High Large High High
Porosity Low Small High High
Roughness Low Small High High
Oxidation Low Large Low Low
Residual stresses Low Small Low Low

Table 2 Spraying parameters for optimum coating quality

Coating quality Requirement Spraying parameter
Standoff Wire speed (current) Voltage Atomising air pressure
Bond strength High Optimised
For maximum particle velocity
Optimised
High - good for large droplets but high heat input
Low - good for low heat input but small droplets
High (30-40V)
Good for high heat input but high levels of porosity
Optimised
Low - good for large droplets with high heat content but low particle velocity and low impact
High - high velocity and high impact but small particles with low heat content
Porosity Low Optimised
For maximum particle velocity
Low Low (say less than 30V) Good for small droplets but must operate within stable voltage regime (about 26-29V) High
Good for small, void filling particles and high particle impact
Roughness Low Optimised
For maximum particle velocity
Low Low (say less than 30V) Good for small droplets but must operate within stable voltage regime (about 26-29V) High
Oxidation Low Short
Minimum flight time; less opportunity for oxidation
High
Large droplets, hence minimum surface area to volume ratio but low particle velocity and low impact
High (30-40V)
Good for large particles but high levels of porosity
Optimised
Low - good for large particles presenting small surface for oxidation, but low particle velocity and more time for oxidation
High - good for high particle velocity and less time for oxidation but small particles presenting large surface for oxidation
Residual stress Low Long Low
Good for minimum heat input but poor bond strength
Low (say less than 30V)
Good for minimum heat input but must operate within stable voltage regime (about 26-29V)
High

Particle size

Fig.2. Typical relationships between the particle volume and standoff distance for two wire feed speeds, using 1.6mm diameter 1%C, 1%Cr steel wires, 28V arc voltage and 0.48MPa air pressure
Fig.2. Typical relationships between the particle volume and standoff distance for two wire feed speeds, using 1.6mm diameter 1%C, 1%Cr steel wires, 28V arc voltage and 0.48MPa air pressure

Standoff distance 
Particle size varies considerably within an arc spray beam, with diameters ranging from 0.1µm (metal vapour) to greater than l mm. In the proximity of the arc, the particles are relatively large, about 0.3-l mm and irregular in shape. As the particles are projected away from the arc, they are then divided and sub-divided into smaller particles by impingement of the atomising air jet, and take on a spherical shape due to surface tension. Typical relationships between the droplet volume and the standoff distance are shown in Fig.2. That is:

  • For a given wire feed speed, the particle volume (or size) is large at a short standoff distance and small at a long distance;
  • The particle volume decreases rapidly as the standoff distance increases up to 50mm, beyond this distance, the volume decreases only gradually;
  • Consequently, the number of particles increases with the standoff distance.
Wire material

Particles detached from aluminium wires have been found to be smaller than those from stainless steel wires. Furthermore, a given degree of atomisation of aluminium particles occurs at a shorter standoff distance than stainless steel.

Wire feed speed (or current)
The particle size has been found to increase with an increase in the wire feed speed (and therefore current) at a given standoff distance, such that:

  1. At a given standoff distance up to 150mm, the particle size is large at a high wire feed speed and small at a low wire feed speed.

  2. For the two wire feed speeds (or currents at 100 and 50A), the particle size difference is large at a short standoff distance and small at a long distance. Typically, the droplet volume is about three times larger for 100A, compared with that for 50A, at 20mm standoff distance, and the volume is practically the same for both the currents at 150mm.

  3. That is, the effect of the wire feed speed on the particle size is strongly dependent on the standoff distance.
Fig.3. Effects of arc voltage, atomising gas pressure and a small coaxial nozzle on particle size distribution for type 420 stainless steel wires
Fig.3. Effects of arc voltage, atomising gas pressure and a small coaxial nozzle on particle size distribution for type 420 stainless steel wires
Arc voltage
Particle size increases with an increase in arc voltage. Furthermore, the particle size distribution is a function of the arc voltage. For a high voltage, the average particle size increases, and the particle size distribution is much greater (Fig.3).

Typically, for stainless steel, the modal droplet size is 20µm at 27V (0.345MPa atomising air pressure) whereas it is more than 20 times larger, 430µm, at 40V (0.138MPa atomising air pressures). In addition, the spray beam becomes less directional, i.e. the spray beam angle becomes wider and fumes are greater.

Atomising air pressure
The average particle size decreases as the atomising gas pressure increases. A particular benefit of high air pressure is a reduction in the incidence of large globules of molten metal at the wire tips.

Fig.4 Typical relationship between the measured temperature of the wire ends and arc power, using 1.6mm diameter mild steel wires, 0.4MPa air pressure and an open nozzle system
Fig.4 Typical relationship between the measured temperature of the wire ends and arc power, using 1.6mm diameter mild steel wires, 0.4MPa air pressure and an open nozzle system

Wire diameter
A smaller diameter wire produces smaller particles. Use of wire less than 1.6mm diameter is uncommon, because of the difficulty in feeding at high speeds without a specially designed wire feeder.

Nozzle size
Qualitative observations indicate that an atomising gas nozzle with a smaller diameter produces smaller particles.

Particle temperature

Arc power (current X voltage)

The temperatures of the particles and the ends of the electrode wires are related to the arc power (Fig.4). For example, the temperature is about 2100K at 1.3kW and rises to 2250K at 4.5kW arc power. For a given current, a higher arc voltage increases the temperature (and therefore heat content) of the spray particles. 

Fig.5 Relationship between the measured particle temperature and standoff distance using 1.6mm diameter mild steel wires, 150A arc current, 28V arc voltage, 0.4MPa air pressure and an open nozzle system
Fig.5 Relationship between the measured particle temperature and standoff distance using 1.6mm diameter mild steel wires, 150A arc current, 28V arc voltage, 0.4MPa air pressure and an open nozzle system

Standoff distance
Although the maximum particle temperatures occur in the proximity of the arc, little fall-off in temperature is found until the particles are a significant distance from the spraying gun (Fig.5). Typically, particle temperatures of about 2300K in the proximity of the arc remain substantially constant up to a distance of 0.2m, probably because of the exothermic reaction of iron with oxygen in the atomising air. Beyond this point particle temperature decreases as the heat loss due to radiation and convection becomes significant.

At a given standoff distance, there is a distribution in particle temperature across the spray beam cross-section. The particles at the periphery of the spray beam are cooler than those located towards the centre.

Wire feed speed
The temperatures of the particles just after their detachment from the wires (T i) and just before their deposition on the substrate (T c) are related to the rationalised wire feed speed parameter V +/(V + + V -) (Fig.6 and 7).

Fig. 6. Dependence of T i and T c on the rationalised wire feed speed parameter, using 1.6mm diameter aluminium wires, 30V arc voltage, 8 m/min wire feed speed on positive polarity and air atomisation
Fig. 6. Dependence of T i and T c on the rationalised wire feed speed parameter, using 1.6mm diameter aluminium wires, 30V arc voltage, 8 m/min wire feed speed on positive polarity and air atomisation
V + is the wire feed speed of the wire at positive polarity.
V - is the wire feed speed of the wire at negative polarity.
T i is the temperature of the particles at the fusion stage.

T c is the temperature of the particles at the adhesion stage.

Fig.7. Dependence of T i and T c on the rationalised wire feed speed parameter using 1.6mm diameter stainless steel wires, 30V arc voltage, 6 m/min wire feed speed on positive polarity and air atomisation
Fig.7. Dependence of T i and T c on the rationalised wire feed speed parameter using 1.6mm diameter stainless steel wires, 30V arc voltage, 6 m/min wire feed speed on positive polarity and air atomisation

The temperature rise from T i and T c occurs in the arc region. In both cases, T i and T c vary between the melting point and the vaporisation point of the materials and both T i and T c are maximum at about 1550 and 2170°C for aluminium (Fig.6) and at 1930 and 2500°C for stainless steel (Fig.7) when the value of the rationalised wire feed speed parameter is between 0.4 and 0.5, i.e. when both the wires are fed into the arc at approximately the same speed.

Particle heat loss
Particles lose heat by radiation and convection during their projection from the arc to the substrate. The rate of heat loss by convection is determined by the relative velocity between the atomising air and particles, and it is between two and three times greater than the heat loss by radiation. The rate of heat loss by radiation does not vary substantially with the standoff distance. However, the total heat loss by convection and radiation is small because the particle projection time is small, less than about 1msec.

Particle velocity

The particle velocity changes because of momentum imparted to the spray particles by the drag action of the atomising gas. The main features of the velocity (Fig.8) are:
  • In the proximity of the arc, the particle velocity is low, about 10 m/sec, and the atomising air velocity is high, about 500 m/sec.

  • As the particles are projected away from the arc (i.e. the standoff distance increases), the average particle velocity increases, whilst the atomising gas velocity decreases.

  • The particles attain a maximum velocity at the point where the particle velocity equals the decreasing air velocity. Typically, the maximum particle velocity is about 50 m/sec (Fig.8).

  • Beyond the standoff distance for maximum velocity, the average particle velocity exceeds the atomising air velocity and begins to decrease.

  • Beyond the standoff distance for maximum velocity, the difference between the air and particle velocity is small; typically the difference is about 25-70 m/sec (Fig.9).

  • Smaller particles, corresponding to the lower wire feed speed, are accelerated more rapidly by the atomising air stream (Fig.9).

  • Larger particles, corresponding to the higher wire feed speed, lose momentum less rapidly than the smaller particles in the region of decelerating atomising air (Fig.9).

  • At a given standoff distance, there is a distribution in particle velocity across the spray beam, with the particles at the periphery slower than those at the centre of the beam.
Fig.8. Relationship between the measured particle velocity and standoff distance, using 1.6mm diameter mild steel wires, 0.4MPa air pressure and an open nozzle system
Fig.8. Relationship between the measured particle velocity and standoff distance, using 1.6mm diameter mild steel wires, 0.4MPa air pressure and an open nozzle system
Fig.9. Dependence of the air velocity, particle velocity and their relative velocities on standoff distance, predicted by a model, for 1.6mm diameter, 1%C, 1.7%Cr, 1.5%Mn steel wire, 31V arc voltage and 0.41MPa air pressure
Fig.9. Dependence of the air velocity, particle velocity and their relative velocities on standoff distance, predicted by a model, for 1.6mm diameter, 1%C, 1.7%Cr, 1.5%Mn steel wire, 31V arc voltage and 0.41MPa air pressure

Coating quality

Effects of particle size, temperature and velocity on coating quality in terms of bond strength, porosity, surface roughness, oxide content and residual stresses are summarised in Tables 1 and 2, and described in the following sections.

In general, the conditions which favour high bond strengths are low oxide content, high particle temperature and high impact energy. These conditions are met by particles of larger size, because the surface area per unit volume is lower than that for the small particles. Therefore, the cooling rate of the large particles during projection by radiation and convection is reduced. Similarly, oxide levels, which are primarily caused by surface oxidation of the molten droplets in the atomising gas, are reduced.

The operating parameters which can generate large particles have been identified as:

  • Short standoff distance;
  • High wire feed speed or high current;
  • High arc voltage;
  • Low atomising air pressure.

These parameters should be used to set up arc spray conditions to maximise the bond strength. However, there are limits to which these parameters can be extended without detriment to other coating characteristics. For example, a short standoff distance generates large particles, but it will produce a coarse, rough surface finish.

Therefore, maximum bond strength through generation of large particles requires careful balance and optimisation of the arc spraying parameters if porosity, poor finish, low particle velocities and high stress levels are to be avoided.

The temperature of the spray particles has a considerable effect on the bond strength of a coating (Fig. 10 and 11). The bond strength F reaches a maximum value when the particle temperature (T i or T c) is also a maximum. Typically, the maximum bond strength is about 7.5MPa when TT i is maximum at 1525°C or T c is maximum at 2150°C for the aluminium coatings. Similarly, the maximum bond strength is about 7MPa when T i is maximum at 1875°C or T c is maximum at 2525°C for the stainless steel.

Fig.10. Effect of particle temperature on bond strength, using 1.6mm diameter aluminium wires, 30V arc voltage, 8 m/min wire feed speed on positive polarity and air atomisation
Fig.10. Effect of particle temperature on bond strength, using 1.6mm diameter aluminium wires, 30V arc voltage, 8 m/min wire feed speed on positive polarity and air atomisation
Fig.11. Effect of particle temperature on bond strength, using 1.6mm diameter stainless steel wires, 30V arc voltage, 6 m/min wire feed speed on positive polarity and air atomisation
Fig.11. Effect of particle temperature on bond strength, using 1.6mm diameter stainless steel wires, 30V arc voltage, 6 m/min wire feed speed on positive polarity and air atomisation

Maximum particle temperature is achieved by using as high a heat input as the substrate can withstand before distortion becomes a problem. For both aluminium (Fig. 12) and stainless steel (Fig. 13) the bond strength is directly proportional to the heat input Q per unit mass of the spray particles.

Fig.12. Relationship between bond strength and heat input per unit mass of spray particles, using 1.6mm diameter aluminium wires and air atomisation
Fig.12. Relationship between bond strength and heat input per unit mass of spray particles, using 1.6mm diameter aluminium wires and air atomisation
Fig.13. Relationship between bond strength and heat input per unit mass of spray particles using 1.6mm diameter stainless steel wires and air atomisation
Fig.13. Relationship between bond strength and heat input per unit mass of spray particles using 1.6mm diameter stainless steel wires and air atomisation

In principle the heat input can be increased by increasing both the current and the arc voltage. However, as reported previously, a high arc voltage produces considerable porosity in the coating.

Furthermore, maximum particle temperatures occur when the wire feed speeds from the negative and positive terminals are approximately equal, i.e. the rationalised wire feed speed parameter (V +/(V + + V -) should lie between 0.4 and 0.5. A value of 0.5 is usual for commercial spraying equipment.

Porosity

Porosity occurs because voids between the solidifying particles remain unfilled and become embedded in the coating. With an increase in the particle velocity, the particles strike the substrate with greater impact and the molten metal is driven firmly into the voids on the substrate surface or those of the coating layer deposited previously. Porosity is minimised by using conditions which generate small droplets at high velocity, i.e. :

  • Low wire feed speed or low current (less than about 150A);
  • Low arc voltage (less than about 3OV);
  • High atomising air pressure (greater than about 0.4MPa).
However, as with bond strength, it is not possible to apply these rules without regard to their effect on other coating properties. For example, increasing the standoff distance up to about 40-50mm, reduces the particle size (Fig.2), as well as increasing the particle velocity (Fig.8 and 9). Consequently, both effects reduce porosity. But an excessive increase in the standoff distance, typically about 150mm, reduces the particle size further. It also has the undesired effect of reducing the particle velocity. With this reduction, the droplets would not be driven firmly into the voids. Therefore, a transition occurs in the porosity levels from low values of 2-3% to high porosity of 15% around the standoff distance of 150mm (Fig. 14). In addition, a lower velocity causes a lower bond strength. Therefore, standoff distance must be restricted to below around 150mm.

Fig.14. Relationship between porosity and standoff distance for Al 5%Si wires on to a mild steel substrate, using 150A current, 27.5 arc voltage and 0.45MPa atomising air pressure
Fig.14. Relationship between porosity and standoff distance for Al 5%Si wires on to a mild steel substrate, using 150A current, 27.5 arc voltage and 0.45MPa atomising air pressure

Similarly, the arc voltage must be restricted to a narrow range of typically 26.5-29V to achieve minimum porosity (Fig. 15). However, a limit to the maximum atomising air pressure has not been reported (Fig. 16).

Fig.15. Relationship between coating porosity and arc voltage for Al 5%Si wires on to a mild steel substrate, using 150A current, 0.45 MPa atomising air pressure and a 150mm standoff distance
Fig.15. Relationship between coating porosity and arc voltage for Al 5%Si wires on to a mild steel substrate, using 150A current, 0.45 MPa atomising air pressure and a 150mm standoff distance
Fig.16. Relationship between coating porosity and air pressure for Al 5%Si wires on to a mild steel substrate, using 150A current, 27.5V arc voltage and a 150mm standoff distance
Fig.16. Relationship between coating porosity and air pressure for Al 5%Si wires on to a mild steel substrate, using 150A current, 27.5V arc voltage and a 150mm standoff distance

Surface roughness

Smooth coatings are produced when small particles impinge on the substrate with maximum impact. The spraying conditions which provide these characteristics are:

  • Long standoff distance (up to about 150mm);
  • Low wire feed speed or low current (less than about 200A);
  • High atomising air pressure (greater than about 0.4Mpa);
  • Low arc voltage (less than about 3OV).

Particle size decreases with an increase in the standoff distance (Fig.2), typically from about 0.3mm diameter at 15mm to 0.1mm diameter at 100mm, but the standoff distance can only be increased to about 150mm without a rapid increase in porosity (Fig. 14). To achieve a smooth coating surface, the air pressure should be maintained as high as possible while maintaining stable arc operation. Arc voltage cannot be used to reduce surface roughness significantly because the voltage levels are restricted to a narrow operating range, typically 26.5-29V, with 27.5V being the optimum voltage, by the need to achieve minimum porosity, typically 2-3 area % (Fig. 15).

Operating conditions producing minimum surface roughness, i.e. a long standoff distance up to 150mm (Fig. 17) and high air pressure, say 0.45MPa (Fig. 18) also result in the lowest porosity levels, typically 2-3 area % (Fig.14 and 16).

Fig.17. Relationship between coating surface roughness and standoff distance for Al5%Si wires on to a mild steel substrate, using 150A current, 27.5V arc voltage and 0.45MPa atomising pressure
Fig.17. Relationship between coating surface roughness and standoff distance for Al5%Si wires on to a mild steel substrate, using 150A current, 27.5V arc voltage and 0.45MPa atomising pressure
Fig.18. Relationship between coating surface roughness and atomising air pressure for Al5%Si wires on to a mild steel substrate, using 150A current, 27.5V arc voltage and a 150mm standoff distance
Fig.18. Relationship between coating surface roughness and atomising air pressure for Al5%Si wires on to a mild steel substrate, using 150A current, 27.5V arc voltage and a 150mm standoff distance

Oxidation

Oxidation of the surfaces of the spray particles occurs during flight and on any exposed surfaces after striking the substrate. The higher the level of oxidation, the lower is the bond strength between the coating and the substrate, and the interlayer bond strength. The conditions which produce minimum surface oxidation are those generating particles with the minimum ratio of surface area to volume (large particles), and those with the minimum time-of-flight, i.e.:

  • Short standoff distance (less than about 50mm);
  • High wire feed speed or high current (greater than about 150A);
  • High arc voltage.

As before, strict adherence to these conditions would be detrimental to other coating characteristics, e.g. large particles generate a rough surface finish, a high current results in a high heat input which causes high stress levels, and a high arc voltage, typically greater than 29V, generates considerable porosity. Therefore a careful balance of the spraying parameters is required to optimise the coating characteristics.

Residual stresses

Residual stress levels should be kept to a minimum, because high stress levels may overwhelm the coating to substrate bonds (tensile strength 10-20MPa) resulting in the coating separating from the substrate.

Residual stresses increase as the coating thickness increases, making thick coatings of large flat areas difficult to produce. For steel coatings, the stress rises sharply for the initial 0.1 mm coating thickness and then it rises very gradually. Typically, for a low carbon steel coating, the stress is about 100 MPa in a 2mm thick coating. However, these residual stresses may be of benefit during reclamation of cylindrical components.

Residual stresses are kept to a minimum by spray particles of small size with a low heat content. These characteristics can be achieved by using:

  • Long standoff distance (greater than about 100mm);
  • Low arc voltage (less than about 3OV);
  • Low wire feed speed or low current (less than about 150A);
  • High atomising air pressure (greater than about 0.4MPa).

Discussion

Arc spraying is most widely used for high deposition rate coating of outdoor steel structures with aluminium or zinc for corrosion resistance applications. The level of porosity is not generally a major concern, since the coatings are either sacrificial or 'self-healing' in operation. The most important coating characteristic is bond strength. The adhesion of the sprayed coating to the substrate is usually assisted by minimising the level of oxidation and the residual stress in the coating, to prevent it from peeling off the substrate.

It can be seen from Table 2 that the optimum spraying parameters for minimum levels of oxidation and residual stress do not match in terms of the standoff distance, ideally short for low oxidation and long for low residual stresses; or the arc voltage, ideally high for low oxidation and low for minimum residual stresses. In practice, the optimum conditions for bond strength (Table 2) will also produce the optimum balance between the spraying conditions for low oxidation and residual stresses.

Whilst bond strengths are still very important for engineering applications, such as reclamation of worn components or overmachined parts, the levels of porosity are more important for engineering applications than for corrosion resistant applications. Generally, the lower the porosity, the longer the coating will last in service under conditions of wear, and the fewer the defects in the machine finished surface. However, a minimum level of porosity is often of benefit, particularly for load bearing surfaces, where oil can be absorbed into any surface breaking voids.

The spraying parameters for minimum porosity are not wholly consistent with those for maximum bond strength, mainly because the principal particle characteristic requirement for minimum porosity, small droplet size, is opposite to the large droplet size (maximum heat input and minimum oxidation) for maximum bond strength.

It is particularly interesting that the optimum spraying parameters for minimum porosity are identical to those for minimum surface roughness. This is not surprising perhaps, since the smaller the particle size, the more likely they are to fill the surface voids on impact and the smoother the surface finish. This suggests that a visual inspection of the surface may be an excellent qualitative guide to the coating density. It also suggests that it would be possible to derive both empirical and theoretical relationships between the surface roughness and the porosity of the coating.

Conclusions

  • Spray coating quality, mainly bond strength and porosity, is related to the three fundamental particle characteristics of size, temperature and velocity. In turn, these characteristics are determined by the spraying parameters including standoff distance, wire feed speed (or current), arc voltage and atomising air pressure.

  • Maximum bond strength between the substrate and coating (typically 7-8MPa) is achieved with particles of large size (typically 0.5mm diameter for stainless steel), at high temperature (typically 2100°C for aluminium and2500°C for stainless steel) and with high velocity (typically 50 m/sec).

  • Minimum porosity (about 3 area % or less) is achieved with small (substantially less than 0.5mm diameter), void filling particles with high kinetic and thermal energy, i.e. high particle temperature and high velocity.

  • To achieve spray particle characteristics for high bond strength and low porosity requires a complex optimisation of all spraying parameters.

  • A qualitative indication of the coating density may be obtained by visual inspection of the surface finish. High density coatings are characterised by a smooth surface finish.

Acknowledgements

Useful discussions with M D F Harvey and Dr W Lucas are gratefully acknowledged.