Medical Device Link.

MANUFACTURING

When packaging fails

Image: iStockphoto

Few colleges in Europe and Australasia offer degrees in packaging. Courses are available in pharmaceutical packaging, but training in medical device packaging is a neglected area. As a result, medical device packaging engineers have to learn their trade on the job and young engineers are likely to repeat the mistakes of their predecessors.

Many packaging problems can be avoided if the knowledge gained by others is shared. It is hoped that this article will help engineers avoid incurring some of the packaging problems that others have already experienced. All the problems and failure modes that are identified in this article were observed while the people involved believed that their packaging operations and testing were being conducted according to “validated” procedures.

Poor heat seals

The underlying cause of defective heat seals was examined in packages produced at various sites in many countries. In more than 90% of these incidents, inadequate contact pressure at the seal site led to failure. For the others, inadequate heating of the mating surfaces was the main issue.

When sterile barrier packs are found to have a poor heat seal this is naturally a major cause for concern. This defect can manifest itself in many forms. It can be a general reduction in seal quality throughout the seal area or the defect may be intermittent, and in other cases only part of the seal area is affected. The defect may be apparent to the eye when examining the unopened package, but often defects only become visible when the package is opened. However, new ultrasound scanning techniques now enable online detection of sealing defects and make possible early detection of problems.

Once a machine has been adjusted to produce properly sealed packages and validated appropriately it should continue to produce good results. Something has to change for the output quality to deteriorate. If seals become inadequate, the sealing conditions have changed or the packaging material properties are different.

The sealing problems described below occurred on various types of machine: some were sealing lids on to thermoformings, others were closing sachets and pouches on band and impulse sealing machines.

In one example, investigation of the cause of the sudden appearance of weak unacceptable seals indicated low sealing pressure. Closer inspection revealed that a sand blasting cabinet had been installed near the clean room and the air supply had been taken from the pipe supplying the clean room. As a result whenever sand blasting operations were taking place the sealing machine had an inadequate air supply, which resulted in poor seals.

Table I. General reduction in a machine’s seal quality. (click image to enlarge)

On another occasion where similar poor seals were in evidence, the company had already taken steps to ensure an adequate clean room air supply with the incorporation of an air receiver in the air line adjacent to the clean room. The static air pressure was noted as being 7 bar; however, when the sealing machine was cycled, the pressure fell to less than 2 bar and did not return to 7 bar until after completion of the sealing cycle. The cause of the low pressure was water condensing out of the air supply and almost filling the air receiver. Water prevented air storage, which stopped the receiver from acting as a reservoir for rapid supply of air to the sealing machine. Draining the receiver restored normal seal quality.

At another location similar poor sealing results were obtained following the repositioning of a blister sealing machine. It was clear that when the machine was repositioned an 8 mm bore air feed tube was used in the clean room instead of the required 20 mm. As in the previous example, there was a severe pressure drop during the important part of the sealing cycle. More causes of a general loss of seal quality are shown in Table I.

Otherwise good seals marred by voids

Often sealing problems manifest themselves as voids in parts of the seal area. These occur because of uneven sealing pressure or temperature.

On several occasions during the investigation of voids in seals that were symptomatic of local pressure loss on a blister sealing flange, the silicon rubber gasket was removed only to find that it had been covering a hole in the base tool; the rubber had no support over that hole. The usual reason for circular holes is careless placement of tool mounting holes. Slots have also been found under the seal area where a “T” shaped gasket strip was once secured, but had later been replaced by a flat gasket sheet; no attempt had been made to fill the open channels.

It is unwise to position print so that printing ink comes into contact with the hot sealing tool. This is because on each sealing cycle a small amount of ink can adhere to the tool and build up to act as an insulating layer that subsequently leads to a loss of seal in the area under the print.

Table II. Uneven quality of seal across the seal area. (click image to enlarge)

Excessive pressure when sealing the edge of a laminated pouch may lead to squeeze out of the laminating adhesive and heat seal coating. However, this is readily visible and should stimulate corrective action. Without tool cleaning the contamination can interfere with proper pressure distribution. More causes of uneven seal quality and voids are listed in Table II.

It is unfortunate that some sealing problems are intermittent and therefore can be difficult to identify. In the case mentioned above where air was stolen by a sand blaster, use of the sand blaster was intermittent, which allowed perfect seals for much of the time.

Hand fed impulse sealers

Figure 1. Illustration of "T-peel." (click image to enlarge)

Hand fed impulse sealers are a particular problem, however, improved control systems on new models now make successful validation a real possibility. The simpler designs are prone to defective seals because of incorrect parameters in the circumstances, which are listed in Table III. In addition, uneven wear of the Teflon covering the sealing wire, corrosion of the wire and general wear can lead to loss of system reliability.

Misleading results from the measurement of seal strength

By the time that validation is underway, the properties of a “good seal” will have been established for the product package combination. Doubtless one of the properties that will be defined and measured is the peel strength of the seal. Many manufacturers will find that during production they have been faced with a report that package seals are visually acceptable, but the measured values for seal strength are all too low. In these circumstances a healthy distrust of the report’s validity is appropriate. It is important for manufacturers to be aware of what they have not been told, for example:

• Has the measuring device gone out of calibration or just been recalibrated?
• Is there a new laboratory technician running the tests?
• Were the samples measured before being allowed to cool after sealing?
• Was the test environment correct (temperature/humidity)?

If the answers are “no, no, no” and “yes” in that order, it is important to check that the tests have been conducted to a properly defined and validated test procedure. For comparison of results to have meaning, a consistent test procedure is essential. The following points should be considered and each characteristic specified in an unambiguous manner:

• For each peel test, record one of the following: peak peel strength value, average strength over the sample,or, average value over the test after removal of the start and stop periods as outlined in ASTM F88.¹
• Ensure that the width tested is the actual sample width and that the effective width is not changed by peeling at other than at a 90° angle to the sample edge.
• Use a constant prespecified speed of test.
• Always use the same “peel angle”(see Figures 1 and 2).
• Use a consistent means of support (or nonsupport) of the sample’s tail.
• If a burst test is used to measure seal strength, remember that the results depend on pack geometry and material stiffness as well as seal strength.

Figure 2. Illustration of 180° peel. (click image to enlarge)

Once the specification and actual test methods have been aligned then the mismatch between visual and measured values should be eliminated (see Table IV).

Curl in blister flanges

Table III. Inconsistent results when using hand fed impulse sealers. (click image to enlarge)

Curled seals are primarily a cosmetic issue. Unless the curl is severe, package function is not likely to be affected. One constraint that limits the acceptability of curl in a double blister pack is that there must be sufficient space in the outer pack to accommodate the sealing flange of the inner. Each company must decide for itself the degree of curl that it will allow in its own package seals.

During the forming process stress is locked into thermoformed mouldings. If during later processing, part of the moulding is heated to its forming temperature, then the stress will be relieved by the area shrinking. During the sealing process, when the moulding is heated above its forming temperature, curl occurs if the lid does not shrink at the same rate as the moulding. The thermoformer can reduce the risk of curl by forming at the highest possible film temperature.

If nothing changes in the sealing process, a sudden occurrence of curl can probably be traced to the thermoforming conditions. Operators should not try to avoid curl by lowering the sealing temperature or reducing the dwell time; this action can be disastrous because it may lead to seal failure.

When investigating the sudden appearance of curl, conditions should be checked to determine if they have been unwittingly changed; a new sealing gasket with the wrong surface texture or thermal conductivity may be the cause. If the gasket colour has changed, be suspicious.

Damage to sterile barrier systems

Table IV. Some reasons why reported results may differ from what is expected. (click image to enlarge)

Even when packages have excellent seals there are some types of failure that still occur during distribution. When the damaged packages are returned for evaluation the cause may not be immediately obvious. Table V provides examples of possible seal failures.

EtO sterilisation failure

Table V. Some failures that may be found after product distribution. (click image to enlarge)

EtO sterilisation failure seems to rarely be a packaging issue. However, misguided packaging and pallet loading can cause problems. In Europe this is usually only a winter-time problem. The EtO sterilisation process requires the presence of water vapour to be effective. In the summer, humidity levels tend to be high, but in the winter when production plants are heated to provide a comfortable working environment for the human operators, this lowers the humidity. If validation takes place in the summer, allowance should be made for the extra humidification needed in the winter sterilisation process.

Operators should be warned not to shrink or stretch wrap pallet loads if these stabilisation methods were not present during the validation cycles. The film wrap may prevent proper humidification and later aeration of the pallet load.

Knowing the known

Known reasons for sterile barrier systems not to perform as intended have been outlined in this article. It is hoped that this article has been informative and adds to the archive of data that help avoid future packaging problems.