28 www.tcetoday.com april 2011

tce HEALTH & SAFETY

No charge pleaseTHE pharmaceutical industry

is particularly susceptible to electrostatic problems, and there

is no doubt that static electricity costs the industry dearly in terms of production rates, yields and downtime across a wide range of operations. But when static leads to damage to plant, injury to personnel or even loss of life, the cost can be immeasurable. This article explains briefly where charge comes from and how it leads to fire and explosion hazards. With understanding comes logical solutions, so the article also explains the basics of a systematic approach to electrostatic-hazard assessment. Finally, there are some real case histories drawn from the pharmaceutical industry.

static-charge generationThe two most important ways in which unwanted electrostatic charge is acquired in industrial situations are induction and tribo-charging. It is crucial that the principles of each are understood in order to recognise where charge may be produced – an essential precursor to preventing it leading to fires and explosions.

tribo-chargingWhenever two different materials contact one another, electrons will cross the interface, making one negatively charged (excess electrons) and the other positively charged. If the two materials are good conductors (such as metals), all the exchanged charge will flow back through the last point of contact when they are separated. However, if at least one of the materials is a poor conductor, this will not happen and the charge that was exchanged during the contact will be retained on separation. It is important at this stage to dispel one very common misunderstanding. Only one of the two contacting materials must be a poor conductor for both to acquire charge.

Ian Pavey outlines the dangers of static electricity in the pharmaceutical industry

Even if the good conductor is earthed, charge will still cross the interface and the poor conductor will still carry away charge when separated. The only difference is that the good conductor’s charge will be lost to earth almost instantly. All too often it is thought that earthing plant solves all electrostatic problems. The reality is that although earthing metal plant is vital, it is not the whole answer.

The magnitude of charge acquired by tribo-charging depends on various factors but the more energetic the process, the greater will be the charge generated due to increases in contact area (solids) and separation rate (solids and liquids). Table 1 illustrates this with typical charge levels acquired by powders undergoing common processes1.

induction chargingAll but the most insulating of materials can be charged to a greater or lesser degree by induction. When exposed to an electric field – such as when in the vicinity of a charged object – opposite charges within the material will tend to separate, either being attracted towards, or repelled from, the nearby charge. Any local excess of charge at a point of contact will then be conducted away, leaving a net charge.

In order to clarify this point, a typical example of how this may occur in practice is shown in Figure 1. Figure 1(a) illustrates a person (a very good conductor) near a highly charged big bag (FIBC). The separated charges are shown, as is the negative charge lost by conduction via footwear and floor. Figure 1(b) shows the person now moved away from the vicinity of the FIBC, carrying a net charge leading to a discharge to the metal door handle

material assessmentsCharge on a material is the result of the relative rates of charge generation and Figure 1: Typical example of induction charging

a b

Process Charge:Mass Ratio (µC/kg)

Sieving 10-3 to 10-5

Pouring 10-1 to 10-3

Scroll Feed Transfer 1 to 10-2

Grinding 1 to 10-1

Micronising 102 to 10-1

Pneumatic Transfer 103 to 10-1

Table 1: Typical powder charge following common processes

april 2011 www.tcetoday.com 29

CAREERS tceHEALTH & SAFETY

charge loss (eg, by conduction). For many materials, both can be enormously affected by low-level impurities and humidity. Hence, when assessing possible static problems associated with all but the simplest of materials, it is important to obtain data from the actual material for which the assessment is being carried out, rather than relying on published data for similar materials. Charge-decay time (Figure 2) and chargeability measurements may be particularly helpful here.

static dischargesSeveral types of electrostatic discharge have been identified, each with a characteristic discharge energy. Spark discharges occur between two conductors and dissipate energy, E, given by:

E=½CV2

where C is the capacitance of the conductor on which charge was stored, and V the voltage to which it was raised. Capacitance is dependent upon geometry and location, but simplistically increases with the size of the object. Assuming a potential of 10 kV (easily attainable) Table 2 shows capacitances and spark energies available from some common objects. Brush discharges occur from insulating surfaces. The energy in a brush discharge is relatively low, often below the limit of human perception. Nevertheless, brush discharges are quite capable of igniting most common flammable vapours and gases. Ignition of a dust cloud with a brush discharge has never been positively confirmed even under laboratory conditions, but it is still often considered prudent to assume it might be possible for the most sensitive of powders. Cone discharges are associated with insulating powders and granules accumulating in containers and silos, and

“

“ there is no doubt that static electricity costs the industry dearly in terms of production rates, yields and downtime

may be somewhat more energetic than brush discharges. Propagating brush discharges arise from highly charged, thin, insulating layers on a conductive substrate and can be extremely energetic. Corona discharges are low-energy discharges from sharp points, and are unlikely to ignite any but the most sensitive of gases, such as hydrogen.

systematic hazard assessmentOnce charge generation is understood, predictions based on laboratory measurements and an understanding of the process (supported by direct measurements on the plant, where possible) can be used to determine if significant charging will occur. Knowing the charging rate and the nature of the charged materials will lead to a prediction of the discharge energy that might be possible. Whether or not this could lead to an ignition will then simply depend on how it compares with the minimum ignition energy (MIE) of the flammable materials present. For many simple gases and vapours, ignition energies are readily available in published literature, though for more complex materials (including most powders) there is no alternative to a measurement of the actual material of interest. Figure 3 shows a small dust explosion initiated by a spark of known energy as part of the series of tests

Figure 2: JCI 155v6 charge-decay analyser Figure 3: Minimum-ignition energy test

Object Typical Capacitance1 (pF)

10kV Spark Energy (mJ)

Small metal items (e.g. scoop) 10 - 20 0.5 - 1

Small containers (e.g. bucket) 10 - 100 0.5 - 5

Medium containers (e.g. drum) 50 - 300 2.5 - 15

Human body 200 - 300 10 - 15

Large plant (eg reaction vessel) 100 - 1000 5 - 50

Table 2. Capacitances and spark energies for some common objects

carried out to determine the MIE of a powder in Chilworth Global’s Industrial Explosion Hazards Laboratory.

avoiding hazardous situations Once the hazard and its potential to occur are properly understood there are a number of options for avoidance, although in practice one often stands out as the most appropriate. It may be possible to avoid charging by altering the process or operating conditions in some way. Alternatively, charging may have to be accepted and charge accumulation prevented. For example, there is no excuse for allowing charge to accumulate on conductors; they can always be earthed. However, it must be remembered that although earthing metal plant is crucial and will prevent its accumulating charge, materials inside may be quite unaffected by this measure. Sometimes it must be concluded that the risk of discharge cannot be fully controlled. In this case the only options may be to remove the flammable atmosphere by the addition of an inert gas, or to install protection for personnel against the consequences of fire or explosion.

incidents from the pharmaceutical industry It was stated at the beginning of this article that the pharmaceutical industry was

30 www.tcetoday.com april 2011

tce HEALTH & SAFETY

particularly vulnerable to electrostatic hazards and problems. In the light of the foregoing discussion it is now possible to justify that statement. Plant is often necessarily operated under conditions of low humidity. For many materials this means that they are at their least conductive and, therefore, most susceptible to charge acquisition and retention. Materials are mostly organic, often chemically very active, and frequently milled to very fine particle sizes. Experience in Chilworth Global’s test laboratory is that pharmaceutical products are among the most sensitive to ignition. In common with most other processes, ever higher speeds are required to maximise plant capacity and minimise cost. Higher transfer speeds and more energetic processes generally lead to higher levels of charge. The cleanliness and low cost of plastics as containers, liners for containers, liners for plant, or plant items means that their use has been increasing over many years. These, of course, are the very materials that will lead to many of the problems discussed. Floors (and walls) are often finished with materials suitable for washing down but with little thought to their conductivity. People and objects moving around on insulating floors have no means of dissipating charge and will become charged. Many of these problems arise in other industry sectors. However, few others bring them all together in the way the pharmaceutical industry does. The consequences can be graphically illustrated with a few real examples from Chilworth Global’s archive of incident investigations. Some of the hazards are so obvious that with hindsight it is difficult to see how they could have been missed – yet they were. Others require considerable insight even when the mechanism is explained.

conclusionsThe pharmaceutical industry is particularly susceptible to electrostatic problems in general, and hazards in particular. However, a proper understanding of the relevant phenomena and appropriate physical property information allows a systematic approach to defining operational practices and plant design to minimise the risk. tce

Ian Pavey (ipavey@chilworth.co.uk) is principal electrostatics specialist with the safety consultancy Chilworth Technology

1 PD CLC/TR 50404:2003, Electrostatics – Code of practice for the avoidance of hazards due to static electricity, BSI.

example one: dust explosion – sievingOperators scooped powder into a sieve unit. The fine particles from the sieve dropped into a large, stainless-steel bin on asymmetric wheels. One day there was an explosion and fire in the bin. Fortunately the operators were unhurt and able to evacuate the room before returning to extinguish the fire. Following a detailed analysis of the incident and measurement of relevant variables an explanation was found. The bin was on insulating wheels. Although an earthing lead was provided it was not used. Simulation experiments demonstrated that the powder would be charged by sieving, leading to charge accumulation on the bin. Measurements also showed that a discharge from the bin would be more than capable of igniting the dust cloud that would have been inside. All the ingredients were there but why did the incident actually occur this time? It turned out that as the bin filled, the centre of gravity moved until it rocked forward about its larger central wheels. When placed in just the right (wrong!) position this caused it to contact the main body of the sieve, leading to a spark. To avoid dust in the room as a whole, the bin had been covered with a plastic sheet. This caused a concentrated dust cloud to escape where the sheet was draped over the sieve outlet – exactly where the spark occurred on this one occasion. The solution, of course, was to ensure that the earth lead provided was used every time.

example two: reactor-charging explosionA 4,500 l vessel had been washed with acetone and left to dry overnight. The next day powder was manually tipped into the vessel from drums via an open access point. As drum number six was added there was an explosion and a fireball enveloped the operators, resulting in serious injuries. The investigation found the powder to have an ignition energy of between 1 and 5 mJ. Trials showed that the morning after an acetone wash, the concentration of acetone vapour could be about 50% of the lower explosible limit. Clearly the acetone on its own would not ignite, but it could contribute to a sensitive hybrid flammable atmosphere of vapour and powder. It was found that the operators’ footwear and gloves were both insulating. This identified two possible isolated conductors – the drums and

the operators. Simulation experiments showed that 10 kV could be attained by the drum during emptying. Given its capacitance this meant that a 10 mJ spark could have been produced from the drum (or indeed from the operator), and it was concluded that one of these was the source of ignition. In this case a number of solutions were possible. The vessel could be inerted and double valves or flap valves used. However, probably the most important recommendations, whether or not the others were implemented, was to provide dissipative footwear and gloves for the operators and an earth clip for the drum.

example three: electrostatic discharges over tolueneAn operator observed blue flashes in a stirred vessel containing toluene. Although under nitrogen, this observation was, to say the least, disconcerting. Chilworth Global constructed a special probe to carry out measurements on site during a trial run in which operational parameters could be varied. It was found that below a certain temperature the level of charge rose dramatically, falling away again as the temperature was raised. This temperature was coincident with the temperature at which solute started to come out of solution. It is well known that an insulating liquid with a dispersed second phase (solid or liquid) charges very much more highly than the liquid alone. The problem was solved by ensuring that the temperature never dropped below the experimentally found transition temperature.

example four: vacuum-dryer incidentsA vacuum dryer had suffered several product decompositions at the end of the drying cycle, leading to over-pressurisation of the vessel and, in some cases, loss of containment. After a wide-ranging investigation, Chilworth Global showed that the key to this problem was a hitherto unreported implication of Paschen’s Law – that discharges across a charged powder surface might be initiated by increasing pressure from hard vacuum, which for this powder was able to initiate decomposition even in the absence of air. The solution was to limit the level of vacuum used, such that breaking the vacuum at the end of the drying cycle no longer initiated discharges.

Explosion examples