Bitumen Storage Tanks and Heated Storage

The reason is a heat transfer argument, not a housekeeping preference. When a coil is submerged, liquid bitumen carries heat away from the metal by convection, and the surface of the tube sits only modestly above the bulk product temperature. The moment that coil breaks the surface, the liquid is replaced by vapour and fume. The heat transfer coefficient collapses by an order of magnitude, the metal wall climbs towards the temperature of whatever is inside it — thermal oil, saturated steam or an electric element sheath — and the thin residual film of bitumen still clinging to the tube is cooked against that surface.

What comes out of that is not softened bitumen. It is coke: a hard, carbonised, insulating deposit welded to the coil. Three consequences follow, and all three outlast the incident that caused them.

• The batch is compromised. Carbon flakes shed off the coil into the product and keep shedding for months. They appear as black specks and hard lumps in the finished mix, they blind line strainers and they foul metering equipment. Draining the tank does not remove them, because the deposit is on the coil, not in the liquid.
• The tank gets progressively worse at heating. Coke is a thermal insulator. Once a layer forms, the same medium temperature delivers less heat to the product, so the natural reaction is to raise the medium temperature — which raises the film temperature at the tube wall and makes more coke. This is a self-reinforcing failure and it is why a tank that has been dry-fired once tends to be dry-fired again.
• It is a fire risk. A carbonised deposit on hot metal in a vapour space containing bitumen fume is a credible ignition scenario, and the fume itself becomes richer as the surface temperature climbs. Bitumen is classified as a Class IIIB combustible liquid under NFPA 30 because its flash point sits at or above 93.3 °C — a typical export paving grade flashes at a minimum of 250 °C by ASTM D92, and ASTM D946 requires at least 232 °C — but a hot uncovered metal surface does not care about the bulk flash point of the tank contents.

How to make the rule real rather than aspirational

• Define a minimum operating level for every tank, measured as a stated depth of product above the top pass of the coil or the crown of the flue tube, and write it on the tank card. The number depends on the geometry of that tank, so it belongs in the tank’s own operating procedure rather than in a generic guide.
• Interlock the heat source to level. Electric immersion heaters must lose power on low level. Steam and thermal oil valves should close on low level. A burner on a direct-fired tank must trip on low level. An interlock that can be defeated with a jumper is not an interlock.
• Never energise heat to get a nearly empty tank moving. This is the single most common route to a coked coil: the tank is low, the product is stiff, and somebody turns up the medium temperature to make the last few tonnes pump.
• Isolate and cool the heating circuit before draining for cleaning, and keep it isolated until the tank is refilled above minimum level.
• On restart from cold, raise the medium temperature slowly. A coil embedded in solid bitumen is nearly as bad as a dry coil, because solid product cannot convect heat away. Melt from a low medium temperature and give it hours, not a higher setpoint.

The commercial version of this rule is simpler still: the unpumpable heel at the bottom of a bitumen tank is not waste, it is the margin that keeps the coil wet. Plan your stock turnover so you are never tempted to spend it.

Film temperature is the number that matters

A tank gauge reads bulk product temperature. Binder degradation does not begin in the bulk — it begins in the thin film against the heat transfer surface, where the temperature is set by the medium, the surface fouling and the local flow. A tank sitting comfortably at 155 °C can be quietly manufacturing coke on a coil fed with hot oil at, to take an illustrative figure, 280 °C. When you compare heating methods, compare the film temperature each one can reach with the controls out of action, not the setpoint each one is supposed to hold.

Saturated steam coils

The great virtue of saturated steam is that it is self-limiting. Saturation temperature is fixed by pressure, so the coil metal physically cannot exceed the saturation temperature corresponding to the steam pressure you supply. Cap the pressure and you have capped the film temperature, permanently, with a mechanical device rather than a control loop.

The cost is the water. A steam system needs a boiler, feedwater treatment, traps and a condensate return, and it introduces the one substance that must never enter a hot bitumen tank. A leaking steam coil is the most dangerous single failure in bitumen storage, because it feeds water continuously into the bottom of the vessel under a head of hot product. Steam also loses its advantage as you push it: raising the pressure to get a hotter medium raises the film temperature with it.

Thermal oil (hot oil) coils

Thermal oil is the most common arrangement in modern bitumen storage, and for good reasons. It delivers high temperature at low pressure, one heater can serve several tanks plus the pumps, lines and loading arms, and a coil leak introduces heat transfer fluid rather than water — a contamination problem, but not a steam ejection.

The corresponding weakness is that nothing about a thermal oil system is self-limiting. Commercial hydrocarbon heat transfer fluids are rated for bulk temperatures far above anything a bitumen tank needs, and the only authoritative figures are the maximum bulk temperature and the maximum film temperature printed on the fluid data sheet. Nothing but the operator’s setpoint stands between that rated capability and the coil wall. Two disciplines follow. First, set the hot oil supply temperature to the lowest value that actually holds the tank, and resist raising it to speed up a melt. Second, treat the fluid as a consumable: heat transfer fluids degrade with time and overheating, and a degraded fluid has a lower flash point, forms deposits that foul the coil and reduces heat transfer — which pushes film temperature up further. Sample and analyse the fluid on a schedule against the supplier’s limits; ASTM D5372 is the published guide for evaluating hydrocarbon heat transfer fluids. The fluid data sheet states both a maximum bulk temperature and a maximum film temperature, and the film figure is the one that governs a bitumen coil.

Electric immersion elements

Electric heating suits small tanks, mobile units and tank containers. There is no boiler, no medium, no traced heating loop, and control can be precise. It is also the method with the least thermal inertia protecting you from a mistake.

Watt density is the whole design question. An element sheath can run far above product temperature, and heater manufacturers publish substantially lower watt densities for bitumen and asphalt service than for water or light oils, precisely because a stagnant, high-viscosity, thermally degradable liquid cannot strip heat from the sheath quickly. Specify elements rated for bitumen service; fit an over-temperature cut-out on the sheath that is independent of the control thermostat; and hard-wire a low-level interlock that removes power. A dry-fired element destroys itself and cokes the surrounding product in minutes.

Direct-fired flue tubes

A burner firing into a tube inside the tank is cheap, self-contained and common on small yard tanks and drum melters. It also produces by far the highest film temperature of any method, concentrated at the crown of the tube nearest the burner. Where it is used, three things matter: the tube crown must sit well below the minimum operating level with real margin, the burner must trip on low level, and a rising flue gas temperature should be alarmed. That last point is the most useful diagnostic in the whole arrangement — when coke insulates the inside of the tank-side tube wall, less heat goes into the product and more goes up the stack, so flue temperature climbing at a constant firing rate is early evidence that the tube is fouling.

Cutback bitumen must never be put into a direct-fired tank, a flue-tube melter or any vessel where a flame, a burner or an exposed hot element can reach the vapour space. A cutback is bitumen dissolved in a petroleum solvent, and it is flash-tested by Tag open cup — ASTM D3143 is the method written for cutback asphalt and ASTM D1310 the general-purpose equivalent — rather than by Cleveland open cup, because the solvent flashes close to ambient temperature. ASTM D2027 sets a Tag open cup minimum of only 38 °C for MC-30 and MC-70, so the vapour above a cutback in a yard tank on a warm day can already be ignitable. The prohibition on open-flame heating is absolute: warm a cutback only by indirect means, only as far as the application demands, and only against the measured flash point on that batch’s certificate rather than the specification floor. The same applies to a tank that has previously held a cutback and has not been proved solvent-free. Full handling limits are on the cutback bitumen page.

External heat exchanger with a circulation loop

Moving the heat transfer surface outside the tank solves several problems at once. The exchanger can be isolated, opened and cleaned without a tank entry; the surface area and flow are known rather than assumed; and the circulation the loop provides is exactly what a modified binder tank requires. The price is mechanical complexity: pumps rated for the viscosity, traced and insulated pipework end to end, and a discipline about draining or keeping the loop hot, because one cold metre of untraced line is where the transfer stops.

Insulation is part of the heating system, not a finishing item

An uninsulated hot tank is a radiator, and the fuel burnt holding it at temperature is a real operating cost that belongs in any bulk-versus-drums comparison. Mineral wool under metal cladding is the standard commercial arrangement, and thicknesses in the range of roughly 75 to 100 mm on the shell are common commercial practice rather than a code requirement — the correct figure comes from a heat loss calculation for your climate, tank size and holding temperature. Two areas are routinely skipped and both cost you: the roof, where heat loss creates the cool surface on which a crust forms, and valves, manways and nozzles, which become the cold spots that set solid and are then heated with a torch. Insulation is also a personnel protection measure; ASTM C1055 is the published guide for heated system surface conditions that produce contact burn injuries, and cladding surfaces are commonly designed to stay at or below about 60 °C for brief contact.

The evidence is written into the specification you already buy against

Every paving specification includes an ageing test, and the conditions of those tests are storage temperatures. The thin film oven test, ASTM D1754, holds a 3.2 mm film at 163 °C for 5 hours. A typical Middle East export specification for a 60/70 paving grade then allows a maximum mass loss of 0.2 % and a maximum 20 % drop in penetration measured by ASTM D5 on the residue — both of those are specification lines agreed in the contract, not requirements of the test method. ASTM D946 controls the same property from the other direction, as a minimum retained penetration after the thin film oven test rather than a maximum drop. The two are expressed differently and are not numerically interchangeable, so read whichever one your purchase specification actually invokes and do not assume the other applies. The rolling thin film oven test, ASTM D2872, does the equivalent at 163 °C for 85 minutes with continuous air flow, and IS 73:2013 caps the viscosity ratio at 60 °C after RTFOT at 4.0. Those limits exist because five hours at ordinary storage temperature measurably hardens bitumen. Nobody disputes the mechanism; the specifications are built around it.

The honest qualifier

A tank is not a thin film, and it would be misleading to read TFOT limits as a storage clock. Oxidation needs oxygen, and in a tank oxygen reaches only the surface and the vapour space. The surface-to-volume ratio of a 3.2 mm laboratory film is thousands of times worse than that of a full storage tank, so a tank ages far more slowly per hour than the oven test does. What a tank has instead is time — days and weeks rather than hours — along with a hot coil surface, an open vapour space, a fresh charge of air every time it breathes, and repeated top-ups that keep the average temperature high.

As a general chemical rule of thumb, and not as a standard requirement, reaction rate roughly doubles for each 10 °C rise in temperature. That is why the difference between holding a tank at 145 °C and holding it at 165 °C is not a 14 % difference in ageing. It is closer to a multiple, and it accrues every hour the tank is switched on.

What changes, and what to do about it

The direction of travel is always the same: penetration falls, softening point rises, ductility falls and viscosity rises. Binder that arrived at 66 dmm can leave a long-held tank materially harder, and if your acceptance limit is 60 dmm you have a commercial problem created entirely on your own site. Four disciplines control it.

• Drop the setpoint when you are not producing. There is no reason to hold a tank at 165 °C over a weekend or a shutdown. Below 150 °C is the normal practice for anything not going out that shift.
• Turn stock over. Work first-in-first-out and avoid an ancient heel that never leaves the tank and quietly contaminates every subsequent fill with heavily aged material.
• Minimise heat cycles. Each melt, hold and cool is a separate ageing event. This is the strongest technical argument for taking bulk rather than drums where you can receive it.
• Re-test before release after a long hold. If a tank has been held hot for an extended period, through a shutdown or across a season, take a fresh sample and run penetration (ASTM D5) and softening point (ASTM D36) before you release it to production. The supplier’s Certificate of Analysis describes the cargo as loaded. Only a tank sample describes the tank as it stands today.

The scope of this section is deliberately narrow: what the tank does to binder that is already hot. How long each product form keeps in stock — sealed drums, emulsion, cutback and bulk hot holding — is a different question with different limits, and it is set out on the bitumen shelf life and storage page rather than repeated here.

Thermal stratification happens in every tank

With the coils at the bottom and no circulation, heat moves upward slowly and unevenly. The result is a temperature profile rather than a temperature: hottest immediately above the coil bank, cooler through the middle, and coolest against the shell and at the free surface under the roof, where a stiff crust can form and then float or hang. That profile has three practical consequences. Product drawn from the outlet may be significantly cooler or hotter than the control thermometer suggests. A single fixed temperature reading is not the tank temperature and cannot be used for a quantity calculation. And the layer nearest the coil sits at the highest temperature in the vessel for the longest time, so it ages fastest.

Circulation fixes all three. The usual arrangements are a pump-around loop that draws from the tank and returns through a nozzle positioned to sweep the bottom, a side-entry or top-entry mechanical agitator, or the external heat exchanger loop described above doing both jobs at once. Running the loop for a period before drawing product, rather than continuously, is common practice on unmodified paving grades where the only concern is temperature uniformity.

Modified binders separate chemically, and that is a specification failure

Polymer modified bitumen is a dispersion, not a solution. A styrenic block copolymer swollen with the maltene fraction of the bitumen is less dense than the bitumen phase around it, so in a static hot tank the polymer-rich phase migrates upward over hours and days. The tank ends up polymer-rich at the top and polymer-lean at the bottom. Neither layer is the product you bought: the top is over-modified and too viscous, the bottom has lost the elastic recovery that justified the price. Crumb rubber modified binder behaves worse still, because rubber particles settle out physically and can build a layer on the bottom plate.

The test that measures it

ASTM D7173 is the published practice. A sample is sealed in a tube, held vertically at 163 °C for 48 hours, cooled, and the tube is cut into three; the top and bottom sections are then tested and the difference reported — usually softening point by ASTM D36, sometimes the DSR parameter by ASTM D7175. EN 13399 is the European equivalent, conditioning an aluminium tube at 180 °C for 3 days. Note carefully that these are practices: they produce a difference, not a pass or fail. The acceptance limit belongs to the purchase specification, and IS 15462 for polymer and rubber modified bitumen sets a maximum difference in softening point of 3 °C between top and bottom sections.

What that means for your tank

• A modified binder tank needs circulation or agitation, on the regime the supplier’s data sheet specifies, and it needs a stated maximum storage period. Treat both as specification lines rather than suggestions — a separation result outside limits is a rejection, and it is a rejection you caused after delivery.
• Agitation is not permission to run hotter. Modified binders are usually stored above the range used for unmodified grades because they are more viscous, which means the ageing clock runs faster at the same time as heat and shear are degrading the polymer itself. Follow the data sheet temperature, not the temperature that makes pumping easiest.
• Sample from more than one level before releasing a modified binder that has been held, and compare. If the top and bottom disagree, the tank has already separated and circulating it afterwards may not fully recover the dispersion.
• Bitumen emulsions must never enter a heated tank at all. They are water-continuous and they break irreversibly when heated hard; they have their own storage regime, their own settlement tests and their own tankage.

The mechanism

Water flashing to steam at atmospheric pressure expands by roughly 1,700 times. Water trapped at the bottom of a tank beneath a head of hot bitumen cannot boil freely: it is held above 100 °C by the pressure of the product above it until it nucleates, and then the phase change is effectively instantaneous. The expanding steam drives the tank contents upward as a foam and out through the hatch, the vent, the dip point or whatever opening is available. A foamover can discharge a large fraction of a tank in seconds, and bitumen at 150 °C or above adheres to skin and keeps transferring heat after contact.

Note the density argument, because it explains why the hazard hides at the bottom. Cold paving bitumen typically sits at a specific gravity of roughly 1.01 to 1.06 at 25 °C — a customary range for the grade rather than a limit set by any standard, measured by ASTM D70 — so water floats on cold product. Hot bitumen at storage temperature is significantly less dense than that, so water sinks through hot bitumen and collects on the bottom plate — directly on top of the coil bank, which is the one place in the tank guaranteed to heat it.

Where the water comes from

• A leaking steam coil or a leaking steam-traced line. The most dangerous failure of all, because it feeds water continuously under head rather than as a single charge. The operational tells are a condensate return flow that does not match the steam supply, and a tank level that rises with no receipts.
• Condensation in an empty or part-empty tank. A cooling tank breathes moist air in through the vent; it condenses on the shell and runs to the bottom. A tank that has stood empty through a wet season is a wet tank until it has been proved otherwise.
• Rainwater ingress through a corroded roof plate, a perished hatch gasket, an open dip point or a failed vent screen. Roof and hatch maintenance is a safety item on a bitumen tank, not a cosmetic one.
• Wet product on delivery. A typical export specification caps water at 0.2 vol %, measured by ASTM D95, for exactly this reason. That cap is a specification line agreed between buyer and seller rather than a requirement of any binder standard, so check that it is actually written into your contract and then check the result on the Certificate of Analysis before the cargo loads.
• Water on drum heads and bag surfaces tipped into a melter with the product.
• Steam used to clear a line or blow a hose, left to condense in the pipework and then pushed into the tank on the next transfer.
• Wash water or fire water left in the bottom after cleaning or after an incident.

The controls that actually work

• Prove the bottom is dry before any hot receipt. Run the bottom drain or water draw-off to waste, and dip with water-finding paste on the bob — manual gauging practice under API MPMS Chapter 3 includes determining free water on the tank bottom. Do this every time a tank has been empty, cold, opened or worked on.
• Fill slowly on the first receipt. Introduce hot product at a reduced rate until the bottom is well covered. Any residual moisture then flashes in small quantity against a shallow head rather than under a deep one.
• Never discharge hot bitumen into a tank holding standing water, and never charge cold or wet material into a hot tank.
• Pressure test coils after any cleaning, repair or suspicion, and record the test. A coil that has been corroding from the inside is next season’s foamover.
• Treat the warning signs as an emergency. Unexplained level rise, crackling or popping on heating, foaming at the surface, or product venting steam means stop heating, stop pumping, clear the area and let it stand. Do not attempt to pump a tank you suspect of holding water.

The vapour space

Hydrogen sulphide (H₂S) evolves from hot bitumen and accumulates in the headspace of storage tanks, tankers and tank containers, where it can reach concentrations far above occupational limits. The ACGIH threshold limit value is 1 ppm as an 8-hour TWA with a 5 ppm short-term exposure limit, and the NIOSH IDLH value is 100 ppm. H₂S deadens the sense of smell at higher concentrations, so odour is not a warning system. Gas test before any tank-top work, approach from upwind, and never put your head over an open hatch. Any entry into the tank itself is a permit-required confined space operation — OSHA 29 CFR 1910.146 in the United States, and the equivalent regulation elsewhere.

Fire and burns

A bitumen tank fire is smothered, not hosed. A water jet into hot bitumen produces the same steam eruption as trapped water and will spread burning product across the bund. Use gently applied foam, dry powder or carbon dioxide, close the tank to starve the fire of air, and make sure the coil isolation valves and the burner trip can be reached from outside the bund. For contact burns, flood the area with clean cool water for a prolonged period and get medical help; do not attempt to peel adhered bitumen from skin. Full face shield, heat-resistant gauntlets and no open cuffs at any coupling being broken hot.

• Bitumen fume safety — what comes off hot binder, why hydrogen sulphide collects in tank vapour spaces, and the controls that actually work