Aim of Process
To polymerise an image of the inner layer circuitry in the correct image integrity to the correct position onto the dry film etch resist.
The exposure process is the polymerisation of the oligomers/monomers in the resist chemistry. Polymerisation is accomplished by exposing the resist to a
known amount of ultra violet radiation.
The following sections detail eight critical areas that should be taken into consideration to ensure successful exposure.
There are several different types of exposure units used in the industry which range from single-sided exposure manual printers to double-sided exposure fully automated types. Different types of light reflectors are used with varying efficiency of light transfer. The height of these reflectors above or below the panel to be exposed will determine the light energy reaching the panel. This will directly impact the rate of exposure. The quality of the reflectors will determine the ENERGY DISTRIBUTION across the exposure frame.
Light will not be reflected evenly across the frame leading to resist exposure difference across the imaged panel. On some older manual exposure units the difference in light energy across the exposure frame can be as high as 40%. This does mean that we can have a difference of up to one step on a 21 Step tablet across the frame.
PHOTEC photoresists have a peak spectral absorption of about 360 manometers. Therefore, the lamp used must have a peak spectral output at the same wavelength. If there is a mismatch between output of the lamp and the absorption characteristics of the resist there may be incorrect exposure, even if the required number of millijoules is applied. In general, the type of lamp used is a high pressure mercury lamp. To obtain the required wavelength output the mercury is normally “doped” with small quantities of iron.
Lamps should be changed regularly as the efficiency of the lamp changes with age. Regulating the current that is applied to the lamp controls the light output. In addition, spectral output may also change. It is common practice to change the lamps after 1000 hours of use. A meter is normally incorporated in the machine to record the time that the lamp is used.
Some lamps are cooled with a water jacket around the lamp. This cooling does prevent heat build up within the exposure machine but it also reduces the amount of energy transmitted. In severe cases this may increase the length of exposure time to reach the desired number of millijoules.
A short, high illumination intensity is preferred to a longer, low intensity illumination. The optical density of the resist changes during exposure. Since the resists are of the surface polymerisable type, the surface layer of the resist will consume energy, resulting in reduced energy passing to the lower levels within the resist. Therefore, unless there is sufficient energy intensity, most of the energy will be consumed as it passes through the resist and the
resist at the copper interface will not be exposed sufficiently.
During normal exposure we will see an exposure difference between the resist on the surface and that at the base of the resist. The difference in exposure can be two steps on a 21 Step Density tablet. Two steps are a reduction of 50% of the light energy.
Exposure intensity is measured in milliwatts.
The actual time of exposure in seconds is millijoules.
Millijoules = milliwatts x time
Each resist requires a certain amount of energy to reach the optimum exposure state. The amount of exposure will determine the actual chemical properties of the resist. In all cases the technical data sheet of the resist should be followed.
If the exposure energy is too low, then the resist will be attacked by both the developing solution and the subsequent processing chemicals. The adhesion of the resist can be affected leading to open circuits after innerlayer etching and short circuits on electroplated circuits.
If the resist is exposed with too high an exposure energy, the image on the phototool will not be transferred in a 1:1 ratio. This means that the exposed traces will be wider than on the phototool.
Phototool quality is extremely important for high circuit density products.
Image density ( Dmax ) should be in excess of Dmax 3,5. Edge definition of the opaque areas should be sharp and well defined. Diffused image edges can lead to variation in track widths after development or incorrectly exposed edges which may cause problems in the development, electroplating or etching processes.
Background clarity of the clear areas of the phototools is essential to ensure short exposure times. A high background density may increase exposure times by up to 100%. If not corrected to compensate for high background density, poor resolution and resist side wall profile may occur.
The use of phototool emulsion protection systems extends the life of the phototool. When setting exposure times the thickness of the protective coating (normally 3-5 microns) must be considered to ensure accurate line width reproduction. This thickness increases the off-contact distance from the resist and thus makes the beam collimation of the exposure unit more important. If the exposure unit has poor beam collimation, the use of a phototool protective system will be dependent upon the permitted deviation from the normal 1:1
reproduction of the phototool.
Degree of Collimation
To obtain accurate phototool reproduction, the exposing energy (light) which is hitting the panel at right angels to the resist is required. This light should be evenly distributed across the exposure frame.
Collimated light is defined as light that is very close (90 degrees) to the panel surface. Normally the angle of declination is about 0,5 degrees. It must be noted that the higher the degree of collimation, the more dirt and scratches will have on the phototool. These defects will be reproduced on the exposed panel.
There are several factors that can influence the angle at which the light hits the panel being exposed. The light emitted by the exposure lamp passes through several different layers before it actually reaches the photoresist
These include the glass or polyester vacuum frame, phototool, phototool protective layer and the polyester cover sheet on the resist. All of these layers can give diffraction of the light.
Any air trapped between any of the layers will also cause light scatter. This air gap can lead to “off contact“ printing. The actual print out image will be diffused and lead to rejects.
A vacuum delay is necessary to allow time for all the air between the various layers to be removed. We need the closest contact possible between the phototool and the photo-resist, the vacuum delay allows the time for the air to be removed and for the phototool to pull down onto the resist surface. A typical vacuum delay is about 10 seconds but for high circuit density boards or high-resolution boards then the vacuum delay is increased up to 30 seconds in severe cases.
Exposure Unit Temperature
Heat is a key factor during exposure. Polymerisation reactions started by ultra violet light can be continued by heat energy and continue after the light energy has been stopped. This may result in over exposure of the resist even if the number of millijoules applied is correct. The actual degree of over exposure caused by the heat in the exposure unit will depend on the actual temperature and the inhibitors used in the formulation of the resist.
The exposure unit should have sufficient airflow within the unit to remove the heat emitted by the lamp and to maintain the temperature within the unit at or near room temperature.
Aim of Process
To chemically remove all the unexposed ( monomers ) resist and leave the circuit image, with a good side wall profile.
The mylar protective sheet is then removed ( this is here to help with the lamination of the resist, to provide damage to the resist, and provide oxidation of the resist. ) The cores are then put onto a horizontal conveyorised processor, which will
Spray a heated ( 28 degrees C ) solution of Sodium Carbonate ( 1% ) at the core developing the image within 100% of the first chamber, leaving the other developing chamber for providing vertical side walls.
The core is then transported onto the next chambers, which are 5 independent water rinses, which are sprayed at the dry film resist to ensure that all dry film and carbonate residues are fully rinsed off.
The core is then passed through a drying stage and then automatically unloaded for inspection.Now for the theory………
Development is the removal of unexposed portions of the negative working resist. The development stage is critical as it determines the quality of the resist remaining on the surface in terms of track profile, adhesion, etc.
As circuit density increases, the track width becomes smaller and more closely packed. As such, the development process becomes more important.
On exposure the resist is polymerised and this alters the dissolution kinetics between the exposed and unexposed resist.
The development mechanism is a diffusion controlled process, that is the developing solution penetrates the unexposed resist and partially removes the resist in the form of a colloid of binder polymer carboxylate salts. This layer must then be removed by mass transfer of developing solution on the resist surface and mechanical action by spray pressure before the next layer can be attacked.
Eight key factors must be taken into consideration to achieve the correct development action, cleanliness of the developed surface and optimum track profile.
· pH concentration of the developing solution
· Resist loading within the developing solution
· Temperature of the developing solution
· Development Time (breakpoint): removal of the resist
· Spray nozzles (type, volume of solution, spray impact)
· Antifoam type and quantity
· Water rinsing (time and temperature)
The above factors demonstrate how the mechanical aspects of the development machine and the chemical control of the process are extremely important. Equipment design is very important and often the dry film resist supplier has to use equipment that is already installed. However, the resist supplier must ensure the following points so that resist performance is not jeopardised.
· The spray from the nozzles is effective and even over the entire board surface. Development is even on the upper and lower board surfaces.
· The spay impact pressure is sufficiently high to remove unexposed resist from within fine lines and spaces and yet is not too high to break tents, particularly those over oblong holes.
· There is no puddling on the upper surfaces of the panel.
· If more than one chamber is incorporated in the development part of the machine, the space between the chambers should be clean and the conveyor mechanism should remain wet. This can be achieved by using “misting “nozzles.
· Rinsing is an integral part of the development mechanism. Ideally, the rinsing time should be 50 % of the development time.
· Drying reduces swelling of the resist that occurs in the development stage.
If wet resist passes directly to a cupric chloride etching solution, increased organic contamination may occur in the etchant.
The development mechanism is a diffusion controlled reaction, The sodium carbonate reacts with the carboxylic (acid) radicals in the unexposed resist and solubilises the resist. This carbonate resist mixture must be removed with fresh sodium carbonate before the reaction may proceed.
Dry film resist is not dissolved in the developing solution, but is held in suspension in a colloid form. Any mechanism by which this colloid is broken down can lead to scum formation on the panel surface. It may also appear as an “oily“ substance on the solution surface. These mechanisms include mechanical stress created by sheer forces in the pumps, spray nozzles, antifoams, dry film resist load in the development solution, etc.
If the colloid is destroyed then a scum will deposit in the developing solution and an oily substance may be seen floating on the surface of the developing solution.
The concentration of sodium carbonate used for development must be within the range specified for each dry film resist. If the concentration is too low the unexposed resist will not be completely removed and a scum remaining on the surface will lead to etching or electroplating problems. If the concentration is too high the resist will be attacked at the resist substrate interface leading to poor track profile, and in severe cases, lifting of the resist. This in turn will lead to electroplating deposits under the resist or track width reduction during
The dry film resist has acid radicals within its chemical structure. As the resist reacts with the sodium carbonate, the acid radical is neutralised and the pH of the developing solution will fall. If the pH of the developing solution falls below a specified pH, the developing mechanism will cease. Therefore, it is essential to replenish the developing solution by either pH control, conductivity, or area of resists developed etc.
The quantity of dry film resist within the developing solution is defined as resist loading. As resist loading increases, the dissolution kinetics is changed and therefore the speed of development changes. There is also a greater tendency for the colloid resist particles to become unstable and precipitate back onto the panel. This leads to open circuits in the case of pattern plating and short circuits in the case of innerlayer production or circuits produced by the tent and etch technique.
The specific recommendation for resist loading will depend on the type of
circuit being produced. Circuits with lines and spaces greater than 150 microns a loading of 0,4 square metres of 40 micron thick resist per litre of developing solution is recommended. As the lines and spaces become smaller, the resist loading becomes correspondingly lower. For lines and spaces of 100 microns and lower, the resist loading should be less than 0,1 square metres per litre of developing solution. (See Appendix 1 for method of analysis for resist loading in developing solutions.)
The dry film resist has acid radicals within the chemical structure. As the resist is dissolved in the sodium carbonate solution this acid radical is neutralised and the pH of the developing solution will fall. A 10-gpl Sodium Carbonate solution will have a pH of about 11,2.
Most dry film resist specifications require that the temperature of the developing solution be controlled within close limits. Normally this is the optimum temperature plus or minus two degrees Celsius.
Thermostatically controlled heaters are required to control the temperature. If the action of the pump mechanism generates heat the developing solution temperature will rise and could exceed that set on the heater thermostat. In this case cooling coils are also necessary to remove this excess heat. If persistent problems arise always check the actual temperature of the solution with a thermometer; never solely rely on the temperature indicated on the
When the developing solution temperature is too high the speed of development is greater and will lead to over development unless the total development time is reduced. It will also give rise to a poor resist side wall profile. A high temperature may lead to breakdown of the colloid and give rise to scum on the panel surface. If the temperature of the developing solution is too low then under development can occur and leave resist residues on the panel surface and also give a larger “ foot “ to the resist profile.
Development Time (Breakpoint)
The minimum development time is the time taken for the resist to be removed from the copper surface. Actual time in seconds will depend on a number of factors including concentration and temperature of the developing solution, volume of solution being sprayed onto the surface, type of nozzle, height of the nozzle above the surface, etc.
For PHOTEC dry film photoresists, a breakpoint is required when the resist is 50-60% developed through the total effective spray length of the developing machine. Note: The effective spray length may not be the overall length of the developing machine. It is essential to inspect the machine and check the actual length where the developing solution is sprayed onto the panel.
If the breakpoint is less than 50% of the total effective spray length over development will occur. This in turn will lead to a poor resist sidewall profile and undercut at the resist to copper interface. This results in under plating in the case of pattern plating or over etching when innerlayers are produced or the tent and etch technique is used.
If the breakpoint is greater than 60% then under development will occur.
Under development will lead to track width reduction in the case of pattern plating and in severe cases will lead to plating inhibition or plating adhesion failures.
In the case of etching innerlayers under development will lead to a track width increase. In severe cases this may also lead to organic contamination in the etching solution. Organic contamination in the etching solution will give small areas of copper not etched on the surface (copper spots).
Development is a diffusion-controlled mechanism. This means that the developing solution must diffuse through the unexposed resist and through the interaction of the sodium ion in the sodium carbonate solution react with the carboxylic acid groups within the resist to form a colloidal particle. These colloidal particles are then washed from the resist surface by fresh developing solution. As the colloids are removed it leaves the surface in a condition for more sodium carbonate to diffuse into the resist surface. This mechanism is repeated until all the unexposed resist is removed.
Because resist removal is a diffusion controlled mechanism, a large volume of low pressure sodium carbonate to reach the resist surface is required during the first stage of development. To achieve, spray nozzles are used in the first part of the development machine that spray a cone shaped pattern onto the surface (i.e. cone nozzles). These cone nozzles give a low pressure on the resist surface. It is recommended that each nozzle deliver in excess of 4 litres per minute of developing solution. (A typical cone shaped nozzle will have a spray impact of 0,4-0,5 Bar on the panel surface.)
To ensure the complete removal of resist at the base of fine lines and spaces a higher pressure is required so that the developing solution will reach the base of resist at the resist to copper interface. In fluid dynamic terms this is a deep recess. A nozzle that delivers a fan shaped spray pattern is used for this purpose.
The nozzle with a fan shaped spray is referred to as a high impact nozzle and will have a typical spray impact of 8-9 Bar on the panel surface. Equipment design is important to achieve the correct development of the resist across the entire panel. Nozzle height and the angle of the nozzle jet must be such that all the entire panel is sprayed at equal volume and pressure and that no overlapping of the sprayed solution occurs. Any overlapping of the sprays will reduce the effective pressure that impinges onto the surface. Prevention of development solution “puddling” on the top of the panel must be prevented, otherwise this will affect the replenishment of fresh solution on the resist surface.
To even out the differences in solution replenishment on the top and lower panel surfaces different spray pressures are used between top and bottom.
Normally a 0,2 – 0,3 Bar higher pressure on the top spray bar is used. Typical spray pressures for development are Top 1,5 Bar and Lower 1,3 Bar.
However, the exact spray pressure is equipment related and should be arrived at experimentally.
The type and quantity of any antifoam used in the development solution will determine the stability of the colloid formed. If the incorrect antifoam is used precipitation within the development solution will occur. This may lead to scum formation on the panel surface that will interfere with the subsequent plating
or etching operations. To eliminate these problems, an antifoam that is a polyethylene oxide or polypropylene oxide block copolymer based should be used.
Siloxane based antifoams are known to cause instability of the colloids in the development solution and give rise to scum or sludge. Silicone based antifoams should not be used as they may provide problems in subsequent processing operations such as electroplating. Trials should be made with all new antifoams to check for colloid stability prior to use on production.
The concentration of antifoam used should be minimised to limit the foam build-up in the development solution. Actual concentration used should be within the range recommended by the supplier.
To prevent the possibility of resists particles or large colloidal particles being re-deposited back onto the panel it is essential to filter the solution. The filter should be in line between the pump and the spray bar manifold and equipped with filter to remove particles greater than 30 microns. The filters should be changed regularly to prevent loss of developer solution volume being supplied to the spray nozzles. Pressure indicators should be fitted into the spray manifold after the filters; this will give a more accurate indication of the spray pressure.
Water rinsing after development is an integral part of the process. The type, volume and temperature of the water are important to ensure dry film resist performance. Water with a hardness of 8 – 12 on the DIN scale should be used. (See Appendix 4 for water hardness conversion.) The reason for using slightly hard water for rinsing is that the divalent cautions. Calcium and magnesium converts the soluble sodium form of the polymers present on the resist sidewalls after development into less soluble carboxylate salts. This effectively stops further development while improving both resolution and resist sidewall profile.
Soft water or softened water is not an effective rinsing medium. The pH of soft water is easily affected by drag-in of developer solution and in severe circumstances may continue the development action.
The pH of the rinse water should be less than pH 9,0 and the temperature of this rinse water should be less than 30 0C to stop development action and prevent attack on the resist surface.
We are following how a BGA (Ball Grid Array) device is created, the first operation where this type of component becomes visible is after innerlayer developing.
Some of the detail of this was provided by Enthone.