Read the Case Study titled “Barton Solvents – Static Spark Ignites Explosion inside Flammable Liquid Storage Tank”
tutor4helpyou
j.reppermund-britton-static_accumulators.pdf
Paper presented at SCHC Spring 2009 Meeting Houston, Texas April 7-8, 2009
1
Hazards of “Static Accumulating” Flammable Liquids
James Reppermund (Presenter) Consulting Engineer, Howell, NJ
and Laurence G. Britton PhD CEng CPhys
Consulting Scientist, Charleston WV
Abstract In June 2008, The U.S Chemical Safety Board issued its report 2007-06-I-KS on its
investigation of an explosion and fire at Barton Solvents in Valley Center, Kansas. The
recommendations issued with this report included the need for improved MSDS
communication about the hazards associated with a particular class of flammable
liquids. This paper attempts to explain what happened, why it happened and offers
suggestions as to what statements can be added to future MSDSs for these products to
alert the MSDS readers of the unusual hazards of these products
Introduction In June 2008 the U.S. Chemical Safety Board (CSB) issued Case Study No. 2007-06-I-
KS describing a tank explosion at Barton Solvents. The recommendations included
improved communications for MSDS preparers. In brief:
1. Warn of liquids that are both “static accumulators” and can form ignitable vapor-
air mixtures inside storage tanks
2. Warn that bonding and grounding may not be enough
3. Give specific examples of additional precautions needed.
4. Include conductivity testing data so that companies can apply published guidance
Paper presented at SCHC Spring 2009 Meeting Houston, Texas April 7-8, 2009
2
What Happened at Barton Solvents In July of 2007, Barton Solvents experienced a catastrophic fire at its Valley Center,
Kansas facility. The fire destroyed the tank farm and caused the evacuation of
approximately 6000 local residents. This incident occurred during multi-stage unloading
of a multi-compartmented tank truck of VM&P (Varnish Makers’ and Painters’) naphtha,
a NFPA Class IB flammable liquid.
The most likely cause of ignition was considered to be a spark caused by a loose
connection between the metal float and the grounded metal tape in the storage tank’s
level gauge system. An analysis showed that the float might briefly attain a high voltage
during multi-stage loading of the tank and spark to the grounded tape. However, it could
not be ruled out that ignition might have been caused by a non-spark static discharge
from the liquid itself. A possible location for such a static discharge (brush discharge)
was from the liquid to the side of the float.
The particular grade of VM&P naphtha involved (flash point 58ºF) is one of a
comparatively small number of commercial hydrocarbon products that has both a low
conductivity and a vapor pressure that provides a persistent, easily ignitable vapor-air
mixture close to the liquid surface in closed vessels or containers. This is where ignition
must occur in cases where static discharges are produced by the charged liquid itself.
In the Barton Solvents case, the ungrounded component of the float gauge was also
located close to the liquid surface and the liquid was loaded at 77ºF (about 20ºF above its
flash point). The most easily ignitable vapor-air mixture typically occurs about half way
between the Lower Flammable Limit (LFL) and the Upper Flammable Limit (UFL). This
condition can be exhibited by many NFPA Class IB liquids and (at higher ambient
temperatures) by many Class IC liquids. Some Class IB liquids, such as most gasolines,
Paper presented at SCHC Spring 2009 Meeting Houston, Texas April 7-8, 2009
3
quickly exceed the UFL close to the liquid surface as a tank is filled, owing to the
presence of volatile “light end” components. However, many pure liquids such as
toluene, and hydrocarbon mixtures such as some VM&P naphthas, lack volatile “light
ends” and do not exceed the UFL during tank filling. They are therefore more prone to
ignite by static discharges. Toluene, for example, maintains its most easily ignitable
vapor-air mixture near the liquid surface throughout tank filling in the high-70s
Fahrenheit and many static ignitions of toluene have been reported.
What is Static Accumulation?
The liquids that are the object of the CSB recommendations are low conductivity liquids,
also known as Static Accumulating liquids.
The defining characteristic of electrical conductivity is how quickly electrical charge
moves over the surface of a material or through the body of a material. When electrical
charges can move easily, the material is defined as a conductor. When electrical charges
move very slowly, the material is defined as a non-conductor or an insulator. Solid
materials can be classified both by volume resistivity and surface resistivity, since the
movement of electrical charges across a solid surface is distinct from movement through
the bulk material. Since liquids have electrical charges distributed throughout the body of
the liquid, they are classified only by volume resistivity. It is customary to use the inverse
of resistivity, conductivity, to electrostatically classify liquids. The units of volume
conductivity are Siemens per meter. One Siemens is the conductance of a material in
which an electric current of one ampere is produced by an electrical potential of one volt.
The Siemens is the SI equivalent of the “Mho” (which, in turn, is an inverse Ohm).
Low conductivity liquids (also called non-conductive liquids or insulating liquids) have a
high resistance to the flow of electrons and will retain significant electrical charge for
seconds or even minutes. Virtually all refined, petroleum-based hydrocarbon products
Paper presented at SCHC Spring 2009 Meeting Houston, Texas April 7-8, 2009
4
have low conductivity and are considered to be “static accumulating”. Many of these also
create a persistent, easily ignitable vapor-air mixture close to the liquid surface in closed
vessels or containers filled at ordinary ambient temperatures. Examples include the
following Class IB flammable liquids:
Benzene Cyclohexane
Heptane VM&P Naphtha
Hexane Toluene
Various non-petroleum based liquids such as simple ethers, carbon disulfide and
hexamethyldisilazane fall into this category as well.
Why is Static Accumulation a Hazard?
Low conductivity liquids, also called non-conductive liquids, have a high resistance to
the flow of electrons and will retain an electrical charge for significant lengths of time.
As with insulating solids such as plastics, once these materials are charged, they will
remain charged even when in contact with grounded metal surfaces. Since the electrical
charges are unable to move quickly to ground, they can build up or accumulate in the
liquid receiver (tank, container, etc) provided there is a continuous source of charging.
This is why low conductivity liquids are also called Static Accumulating Liquids.
When a static accumulating liquid becomes charged, it can cause ungrounded conductors
that are in contact with the liquid or near to the liquid to become charged. If the charged,
ungrounded conductor becomes grounded, there can be a spark. If the spark has sufficient
energy and if the spark occurs in an ignitable vapor-air mixture, the result will be a fire.
Other hazards are more insidious and less obvious. When a charged, static accumulating
liquid is pumped into a tank the surface voltage on the liquid in the tank increases as
Paper presented at SCHC Spring 2009 Meeting Houston, Texas April 7-8, 2009
5
liquid level builds. The voltage can in some cases be sufficiently high for static
discharges to occur from the liquid itself, even if the tank is properly bonded and
grounded. Discharges known as “brush discharges” often occur to grounded projections
above the liquid surface such as the ends of filling pipes (which is one reason a slow start
is often used until dip pipes are submerged). Static discharges may also in some cases
form “streamers” at the tank wall that travel across the liquid surface. Both types of static
discharges may ignite flammable vapors in air under the right conditions, which might
only occur once during years of operation.
To avoid ignition various precautions are required to limit the accumulation of charge.
These are given in codes of practice and include such measures as restriction of flow
velocity, depending on the size of tank, filling pipe diameter and other conditions.
How do You Determine the Potential for Static Accumulation?
The propensity of a liquid to accumulate static electricity can quickly be determined by
measuring the liquid’s electrical conductivity. Instruments are commercially available
that can quickly and (relatively) inexpensively make this measurement. The instrument
selected to make these measurements must be capable of measurements in the pico
Siemens (pS) range. One pico Siemens is equal to 1 x 10-12 Siemens. Typical laboratory
conductivity meters only measure in the micro Siemens range, roughly 6 orders of
magnitude larger than what is needed. While there is no longer a single ASTM Standard
Test Method that is applicable to the testing of all liquids, including high conductivity
liquids such as alcohols, ASTM D2624 Standard Test Methods for Electrical
Conductivity of Aviation and Distillate Fuels addresses the conductivity range up to 2000
pS/m, which includes the low conductivity liquids discussed in this paper. However,
instruments are commercially available that measure conductivities over very wide
ranges, based on other standardized test procedures. Hence it is possible to determine the
conductivity of almost any liquid for MSDS reporting purposes.
Paper presented at SCHC Spring 2009 Meeting Houston, Texas April 7-8, 2009
6
Liquids having a measured conductivity of 100 pS/m or less are considered to be “Static
Accumulating” liquids.
Liquids having a measured conductivity greater than 100 pS/m are considered to be either
“Semi-Conductive” liquids or “Conductive Liquids”. The hazards of such liquids are
specific to the handling conditions and will be mentioned only briefly in this paper.
The demarcation of 100 pS/m as given above is arguably conservative for hydrocarbons
(at least in the case of tank filling), and the Petroleum Industry in particular uses a lower
demarcation of 50 pS/m. However, the conductivity of “Static Accumulating” liquids is
highly dependent on temperature and purity. The conductivity of a liquid handled in a
chemical plant on a cold day might be only one-third of that measured in the laboratory.
Also, at a given temperature, it is common for samples of the “same” liquid to have quite
different conductivities depending on the source of the liquid. A “pure” liquid such as n-
heptane has virtually no intrinsic conductivity and what is measured is the effect of trace
contaminants. Different samples of n-heptane could have conductivities varying by at
least two orders of magnitude. Another complication is that the rate at which a charged
liquid loses its charge depends on its dielectric constant. This typically ranges from about
2 for hydrocarbons to about 4 for other “Static Accumulating” liquids such as simple
ethers. Hence, for general reporting purposes such as MSDS, the higher demarcation of
100 pS/m should be used.
Hazards of Suspended Water Droplets & New Hypothesis for “Water Slug”
Hazards
According to CSB, the Barton tank likely contained sediment plus water. It was an air-
breathing tank so water (condensed from humid, ambient air) would gradually
accumulate in the tank bottom over time. The Barton tank volume (~15000 gallons)
Paper presented at SCHC Spring 2009 Meeting Houston, Texas April 7-8, 2009
7
exceeded 50 cubic meters and places it in the “large” tank size category defined in
CENELEC 50404.
It is well known that static charging can be greatly increased when mixtures of oil and
water are pumped together, especially when subjected to high shear such as during
passage through a component such as a partly closed valve that produces small water
droplets having a large interfacial area relative to the continuous oil phase. CENELEC
50404 (2003) warns of this hazard in Chapter 5.4.4.2.1 with respect to “medium” tanks
(1-50 cubic meters) where it states: “For two-phase flow or if water bottoms could be
stirred up in the tank, the filling velocity should be restricted to 1 m/s”. Here the usual
“two-phase flow” warning has been extended to address suspension of water droplets
derived from water bottoms already in the receiving tank.
Most air-breathing tanks are likely to contain water bottoms that could be stirred up
(provided water is insoluble in the lading) and the authors are unaware of what (if any)
procedures are used to limit its accumulation. It is not uncommon for a “side-bottom-
entry” fill pipe to double as the outlet pipe, in which case the limiting factor would be
entrainment of settled water bottoms during tank emptying. The inlet/outlet pipe is often
located close to the tank floor.
CENELEC’s “1 m/s” flow velocity restriction with respect to water bottoms in “medium”
tanks (1-50 cubic meters) is not currently provided in other codes such as NFPA 77.
Also, CENELEC does not apply the 1 m/s flow rate restriction to tanks larger than 50
cubic meters, which are designated as “large” tanks. The reason behind the selected
volume cut-off is likely based on the maximum capacity of single compartment tank
trucks (<50 cubic meters and typically about 26 cubic meters). The 50 cubic meter cut-
off allows ready differentiation between tank trucks and rail cars, which have a larger
capacity (typically about 89 cubic meters for single compartment cars). Greater flow
velocities are allowed for rail cars than for tank trucks. The reader should refer to
Paper presented at SCHC Spring 2009 Meeting Houston, Texas April 7-8, 2009
8
CENELEC 50404 and NFPA 77 for specific recommendations. It is important to
recognize the practicalities involved in “tank size” definitions. Although the Barton tank
strictly exceeded 50 cubic meters capacity, the geometry of vertical storage tanks results
in faster accumulation of liquid level and larger surface voltages than would apply to a
rail car of equal capacity, all other factors being equal. It is prudent to apply significant
latitude when considering what flow rate restrictions should be applied to vertical storage
tanks.
The discussion has so far focused on hazards caused by suspension of small water
droplets. We now propose a new hazard scenario in which large “slugs” of water derived
from tank water bottoms pose a spark ignition hazard, even where all tank components
are properly bonded and grounded. This hazard is different than that of “increased static”
caused by suspension of small water droplets; it is potentially far more severe and has not
previously been recognized.
If a tank has significant water bottoms and is side filled at the bottom, as was the Barton
tank, it is possible that large water “slugs” will be launched into the liquid and convected
to the surface. Electrostatic charging of water slugs may occur via a variety of
mechanisms once they are adrift in the charged oil. Collision with grounded tank
components and break-up of slugs, particularly in regions of high electric fields, is in
many ways a more plausible spark ignition scenario than the much-studied “supertanker
water washing” explosion scenario advanced in the early 1970s (Britton 1999 pages 217-
218).
The hypothesized water slugs will at this point be “charged ungrounded conductors” that
may spark to the tank wall. The minimum voltage for vapor ignition via sparking is less
than about 10 kV, depending on the size of the slug. Such voltages are commonly
exceeded when filling medium sized storage tanks. Hence, all that is needed is for a slug
of sufficient size (capacitance) and voltage to attain the correct trajectory through the
Paper presented at SCHC Spring 2009 Meeting Houston, Texas April 7-8, 2009
9
liquid. Conversely, the minimum voltage for ignition via a brush discharge is at least 25
kV (see later). This 25 kV threshold applies only to negative charging with an optimized
electrode and vapor-air mixture above the liquid surface; in all other cases the threshold
voltage is greater. It follows that increased static due to suspended water droplets should
be far less of a hazard than the formation of large water slugs. Note that the specific
gravities of some common Class IB liquids are not very different from that of water;
while heptane has a specific gravity of about 0.7, that of typical VM&P is about 0.8 and
toluene is about 0.9.
This “water slug” hypothesis might explain some atmospheric tank explosions that did
not involve high flow velocities or other adverse conditions, such as pumping oil-water
mixtures or the location of a microfilter close to the tank. If large water slugs can float
around in a receiving tank, even slow filling velocities might not exclude the possibility
of spark ignition; indeed, slower velocities might favor the creation of larger slugs.
However, the use of decreased flow rates will reduce the liquid voltage in the tank and
hence the ignition frequency.
As a practical matter, it would be helpful to gather information on the accumulation of
water bottoms in air breathing storage tanks. If there is general consensus that the
problem needs to be addressed, we hope that the matter will be taken up by an
appropriate safety organization.
The “water slug” hypothesis is not currently recognized in codes of practice. A suitable
warning statement would need to address water bottoms directly, such as “Do not load
liquid into tank containing water bottoms that could be stirred up”. As noted above, there
is no “safe” flow velocity associated with the hypothetical scenario.
Paper presented at SCHC Spring 2009 Meeting Houston, Texas April 7-8, 2009
10
What Can Be Done About Static Accumulation?
Static accumulating liquids can become charged by numerous different operations
including but not limited to:
• Spraying
• Air entrainment
• High velocity flow or agitation
• Two phase flow or mixing
• Settling of an entrained solids or immiscible second phase
• Passing through a micro-filter
Static accumulation in low conductivity liquids cannot be prevented although it can be
reduced by reduced velocity and the addition of conductive liquids or of conductivity
enhancing materials.
Part-per-million levels of antistatic additive can be used to raise the conductivity of a
nonconductive liquid to above 100 pS/m depending on the needs of the customer. This
eliminates the need to use the phrase “Static Accumulating Liquid”. However, static
accumulation cannot be prevented under all conditions.
The phrase “Static Accumulating Liquid” (or “Static Accumulator”) must be confined to
those liquids that may accumulate hazardous levels of static charge when pumped into
properly grounded metal tanks or containers. The purpose of the warning is to identify
those liquids that may accumulate sufficient surface voltage for a so-called “brush
discharge” to occur. This is generally associated with liquid conductivities less than 100
pS/m and usually much less than this value. However, as discussed above, conductivity is
not constant and for communication purposes, a safety factor is needed to account for
batch-to-batch variation plus the effect of low ambient temperatures.
Paper presented at SCHC Spring 2009 Meeting Houston, Texas April 7-8, 2009
11
Codes of practice on “Static Electricity” warn of precautions needed to fill tanks safely,
such as limiting the flow velocity and using dip pipes. The codes also warn of special
hazards such as nonconductive tanks, plastic lined tanks, entrainment of air or water,
passage through static generators such as micro-filters and partly blocked strainers, and
suspension of water bottoms in a tank. The user should follow these recommendations.
However, the recommendations are not consistent in different codes and also vary with
the type and size of the tank. It is impossible to summarize all of these on a MSDS and
reference should be made to codes such as NFPA 77 and CENELEC CLC/TR 50404
(which has far more explicit information on tank filling precautions). In the next year it is
expected that a new IEC document will be issued that will update and replace the
CENELEC document.
It must be emphasized that various common process operations such as two phase mixing
and spraying can accumulate static electricity at much higher conductivities than 100
pS/m. For various mixing operations, it is common practice to increase the liquid
conductivity to several thousand pS/m to avoid static problems, such as by adding a
suitable conductive liquid to a nonconductive hydrocarbon. Even alcohols and ketones,
which typically have conductivities of 1 million pS/m or more, can accumulate hazardous
static on ungrounded spray nozzles such as in painting applications.
Even where all other precautions are taken, an ungrounded person may be the cause of a
static spark, independent of any electrical properties of the liquid. Hence general warning
statements about static ignition should be given separately along with boilerplate
warnings about open flames and the like.
What Should be on the MSDS to warn of Static Accumulation
Paper presented at SCHC Spring 2009 Meeting Houston, Texas April 7-8, 2009
12
1. Include a representative “standard” conductivity measurement (25ºC) on the
MSDS for all static accumulating liquids and identify them as both “static
accumulating liquid” and “low conductivity liquid (non-conductive liquid)”.
Warn that the value may change with temperature and purity, including how the
liquid is stored and handled.
2. Include where possible a representative conductivity value for all liquids, so that
various code practices can be applied. The conductivity should at a minimum be
given for liquids with conductivity up to 2000 pS/m, which is within the
capabilities of various commercial instruments and includes the demarcation of
1000 pS/m used by the Petroleum Industry for “high conductivity liquid”. Since a
higher demarcation of 10,000 pS/m is widely used in the Chemical Industry
(especially for operations such as liquid-solid mixing), it would be prudent to
have a 20,000 pS/m capability, as available in some commercial instruments.
Ideally, the instrument should be capable of wide range determination from less
than 1 pS/m to about 10,000,000 pS/m, so conductivities can be given not only for
“static accumulating liquids” but also for commonly used conductive solvents
such as many esters, alcohols and ketones. Note that some liquids have
intermediate conductivity (between 50 and 1000 pS/m, or between 100 and
10,000 pS/m, depending on the code of practice referred to) and are described as
“medium conductivity” or “semi conductive”. These require special consideration
in various Codes of Practice. It can be seen that a conductivity value is more
useful than a description that varies with the Code of Practice referred to.
3. Suggested warning statements for “static accumulating liquids” include:
• “This liquid may accumulate static electricity when filling properly
grounded containers.”
• “Bonding and grounding may be insufficient to remove static electricity.”
• “Static electricity accumulation may be significantly increased by the
presence of small quantities of water or other contaminants.”
Paper presented at SCHC Spring 2009 Meeting Houston, Texas April 7-8, 2009
13
• “Restrict flow velocity according to (CITE APPLICABLE CODE)”
• “Refer to Codes of Practice” (insert applicable code, such as CLC/TR
50404 in the EU, for guidance). We are awaiting the issuance of an IEC
“Static Electricity” code in late 2009 or early 2010 that is internationally
recognized and should greatly help the MSDS preparer.
To address the ignitability issue, MSDS preparers must consider the criterion of a vapor
pressure that provides “a persistent, easily ignitable vapor-air mixture close to the liquid
surface” in closed vessels or containers. This is simple in the case of pure liquids because
the vapor pressure at different temperatures can be simply related to the known
flammable limits. For mixtures, it is more complex. However, for a first pass the criterion
may be applied to NFPA Class IB and IC liquids as discussed by the CSB, with
exceptions made where applicable. Some Class IB liquids such as gasoline and light
naphthas might be excluded while under cold weather assumptions some borderline Class
IA liquids might be included. Figure 1 below shows the vapor pressure curve of a typical
hydrocarbon Static Accumulating Liquid (Toluene in this case), indicating the
temperatures where the Lower Flammable Limit will occur and where the Upper
Flammable Limit will occur.
Paper presented at SCHC Spring 2009 Meeting Houston, Texas April 7-8, 2009
14
0
10
20
30
40
50
60
-10 0 10 20 30 40 50
Figure 1: Toluene Flammability Limits at 1 atm Temperature Limits of Flammability (TLF)
and Most Easily Ignitable Temperature
V ap
or P
re ss
ur e
(m m
H g)
Equilibrium Temperature (C)
UFL = 7.1 vol% (54 mmHg)
Upper TLF = 38 C
LFL = 1.1 vol% (8.36 mmHg) Lower TLF = 3.2 C
Most Easily Ignitable ~26C (Britton 1999)
A suggested warning statement is:
• “This liquid may form an ignitable vapor-air mixture in closed tanks or
containers”
• “Additional advice on handling and processing low conductivity liquids
can be found in the following documents –
Paper presented at SCHC Spring 2009 Meeting Houston, Texas April 7-8, 2009
15
o NFPA 77 – Recommended Practice on Static Electricity, National Fire Protection Association
o RP-2003 – Protection Against Ignitions Arising Out of Static, Lightning, and Stray Currents, American Petroleum Institute
o TR- 50404 – Code of Practice for the Avoidance of Hazards Due to Static Electricity, CENELEC, European Committee for
Electrotechnical Standardization”
o Generation and Control of Static Electricity in Coatings Operations, National Paint and Coatings Association
o Britton, L.G., “Avoiding Static Ignition Hazards in Chemical Operations”, AIChE-CCPS (1999)
In some cases tank inerting might be considered. This is described in:
o NFPA 69 – Standard on Explosion Protection Systems, National Fire Protection Association.
Plate 1 shows a roughly two-inch long “positive brush” discharge from a negatively
charged diesel oil surface to a grounded electrode (Britton, L.G., and T. Williams, “Some
Characteristics of Liquid-to-Metal Discharges involving a Charged Oil”, J. Electrostatics,
13 (1982) pp. 185-207). The picture was taken using a high gain image intensifier so does
not show the liquid surface or the electrode. The upper electrode was a ½-inch steel
sphere, intended to represent a probe such as the end of a thermowell above electrically
charged liquid in a tank. Discharges of this type were able to ignite mixtures of propane
or butane in air at liquid voltages above -25 kV.
Paper presented at SCHC Spring 2009 Meeting Houston, Texas April 7-8, 2009
16
Plate 1
Two-inch Long, Incendiary Brush Discharge from Negatively Charged
Oil to ½-inch Grounded Spherical Electrode
