A 20 tonne grab crane carries 26 tonnes of steel and 61 tonnes of CO2e before it scoops anything. A 2 tonne cleanroom crane carries 1.65 tonnes of steel and 5.3 tonnes of carbon, over half of it in the stainless sheet. Four sub-types, four different carbon problems, and a customs line that catches the sheet but not the machine.
Four special crane sub-types share a bridge, a hoist and a control cabinet, and almost nothing else. A 20 tonne grab crane carries 26 tonnes of steel before it scoops anything, and accounts for about 176 tonnes of CO2e over twenty years on the EU grid. A 2 tonne cleanroom crane carries 1.65 tonnes of steel and about 6 tonnes of operating carbon. One is an electricity problem with a steel bill attached. The other is a materials problem that will never repay a variable frequency drive.
Two grab crane quotations landed on my desk in September, same 20 tonne rating, same 22.5 metre span. One machine weighed 18.5 tonnes. The other weighed 24.5. The heavier one was about 40,000 dollars cheaper, because it used thicker plate in a lower grade and skipped the box-section detail that costs a day of welding. Six tonnes of extra plate is roughly 14 tonnes of CO2e before the crane lifts a bucket. Nothing in the comparison sheet had a column for it, and the buyer had no reason to ask.
That is the whole argument for looking at this properly. The mass is on the drawing. The carbon follows the mass, plus a handful of materials that carry far more carbon per kilogram than steel does. What follows is the mass data by sub-type, the material splits, the grab arithmetic nobody puts in a quotation, the coating question that is not really about paint, and the customs boundary that catches the stainless sheet while letting the finished crane walk through.
Because they are sized by different things. A hook crane is sized by the load and the span. A grab crane is sized by the load, the span and the grab. A precast stacker is sized by the load, the span and the concrete element it has to hold square while it travels. A cleanroom crane is sized by the surface finish and the cleanability, and the structure follows along. An Ex crane is sized by the enclosure, and everything downstream of the enclosure grows with it.
The numbers below are planning figures from our own quotation and shipping records, not a published survey. Machine mass excludes the grab where the grab is a separate item. Carbon is at the worldsteel 2025 global average of 1.92 tonnes CO2e per tonne of crude steel for the mixed case, and at 2.34 for the BF-BOF plate we normally buy, which is the number used throughout this article.
| Sub-type | Typical rating | Machine mass | kg per kg of capacity | Embodied carbon | t CO2e per t of capacity |
|---|---|---|---|---|---|
| Grab crane, double girder | 20 t | 18.5 t + 7.6 t grab | 1.31 | 61 t | 3.06 |
| Precast stacker | 50 t | 62 t | 1.24 | 145 t | 2.90 |
| Pickling line crane | 10 t | 13.5 t | 1.35 | 32 t | 3.16 |
| Cleanroom crane | 2 t | 1.65 t | 0.83 | 5.3 t | 2.65 |
| Hazardous area (Ex) crane | 10 t | 17.8 t | 1.78 | 42 t | 4.19 |
| Reference: single girder hook crane | 10 t | 6.5 t | 0.65 | 15 t | 1.52 |
So the range is not a rounding difference. A special crane carries between 1.7 and 2.8 times the embodied carbon of a standard hook crane per tonne of rated capacity, and the grab crane carries two machines' worth of steel because the grab is a load that rides to work with you every cycle.
If you are buying any of these, the specification page on special cranes lists the standard build for each sub-type. What it cannot tell you is which of the four you actually need, because that is a process question, not a catalogue question.
Nobody quotes grab dead weight as a percentage of capacity, and yet it is the single number that decides throughput, hoist sizing, and the carbon of every lift. The published sizing data is consistent across suppliers: a mechanical grab runs 40 to 55 percent of its payload in dead weight, and an electro-hydraulic grab runs 60 to 90 percent, because you are now lifting a motor, a pump, a tank and hydraulic oil alongside the material.
| Crane rating | Grab type | Grab dead weight | Usable payload | Dead weight as % of payload | Grab embodied carbon |
|---|---|---|---|---|---|
| 5 t | Single rope mechanical clamshell | 1.4 t | 3.6 t | 39% | 3.3 t |
| 10 t | Four rope clamshell | 2.8 t | 7.2 t | 39% | 6.6 t |
| 16 t | Four rope clamshell, heavy duty | 5.2 t | 10.8 t | 48% | 12.2 t |
| 20 t | Electro-hydraulic clamshell | 7.6 t | 12.4 t | 61% | 17.8 t |
| 32 t | Electro-hydraulic clamshell | 12.5 t | 19.5 t | 64% | 29.3 t |
| 50 t | Electro-hydraulic orange peel | 21.0 t | 29.0 t | 72% | 49.1 t |
| 80 t | Electro-hydraulic orange peel | 34.0 t | 46.0 t | 74% | 79.6 t |
The 80 tonne line is the one that stops conversation. A 34 tonne grab is 80 tonnes of CO2e, more than the crane that carries it, and it is a wear item. Grab shells in coal and iron ore get relined every one to three years and the body itself is replaced on a five to ten year cycle. Over a twenty year life you buy the grab two or three times and the crane once.
Fill factor matters more than people expect. Clamshell grabs fill to 85 to 90 percent of nominal volume. Orange peel grabs fill to 70 to 80 percent. If you match the grab to a material with the wrong bulk density, say a coal grab at 0.85 t/m3 dropped into an iron ore job at 1.8 t/m3, you lift half the tonnage per cycle while still paying the full dead weight in energy and the full grab mass in carbon. Same bucket, same crane, less than half the throughput.
On a 20 tonne grab crane the grab is 38 percent of the mass moving on every single cycle. It is not 38 percent of the electricity, because hoisting efficiency and duty patterns are not linear, but a planning figure of 35 percent is close enough for a budget.
Our 20 tonne A7 grab crane draws about 42,800 kWh a year at 5,000 powered hours. Thirty-five percent of that is 15,000 kWh a year spent lifting the bucket itself, which is 62 tonnes of CO2e over twenty years on the EU grid at 206 g/kWh, or 161 tonnes on the Chinese grid at 537 g/kWh. Set that against the grab's own embodied carbon of 17.8 tonnes and the payback is about six years in Europe and about two years in China.
Which is a slightly uncomfortable conclusion. On a grab crane, the grab's manufacturing carbon is paid back by the electricity you spend carrying it around. That is the opposite of what most carbon questionnaires assume, and it is why the first thing to fix on a grab crane is the grab-to-material match, not the motor efficiency class.
Not the structure. On a 2 tonne cleanroom crane the carbon steel is roughly 700 kilograms, and the stainless is roughly 500. The stainless carries more carbon than the steel does, because stainless is not steel with a nicer surface. It is an alloy with 18 percent chromium and 8 percent nickel, and the nickel and chrome units arrive through a supply chain that is far more carbon intensive per kilogram.
| Material group | Mass in a 2 t cleanroom crane | Factor used (t CO2e/t) | Carbon | Share |
|---|---|---|---|---|
| Carbon steel structure | 700 kg | 2.34 | 1.64 t | 31% |
| Stainless 304 cladding, rails, fasteners | 500 kg | 5.57 | 2.79 t | 53% |
| Hoist and electricals | 300 kg | 2.63 | 0.79 t | 15% |
| Aluminium components | 60 kg | 6.60 | 0.40 t | 8% |
| PU wheels, seals, PFPE grease | 40 kg | 3.00 | 0.12 t | 2% |
| Total | 1,600 kg | blended 3.31 | 5.3 t | 100% |
Five hundred kilograms of stainless carries 53 percent of the carbon. The stainless factor of 5.57 is not an assumption I invented for this article: it is the default value that the UK government published on 16 October 2026 for CBAM reporting on CN 7219 31 00, which is cold rolled stainless flat product over 600 mm wide. The neighbouring stainless headings are all in the same band.
| CN code | Material | CBAM default value | Used for |
|---|---|---|---|
| 7219 31 00 | Cold rolled stainless flat, over 600 mm | 5.571 t CO2e/t | Cladding, panels |
| 7220 11 00 | Hot rolled stainless narrow strip | 5.508 t CO2e/t | Edge trim, small pressings |
| 7222 11 11 | Stainless bars and rods, hot rolled | 6.466 t CO2e/t | Pins, shafts, fasteners |
| 7222 40 10 | Stainless angles, shapes, sections | 6.491 t CO2e/t | Cleanroom track supports |
| 7223 00 19 | Stainless wire in coils | 6.515 t CO2e/t | Rope, mesh guards |
| 7308 90 98 | Steel structures and parts | 6.374 t CO2e/t | Brackets, walkways |
| 7326 90 98 | Other steel articles | 3.439 t CO2e/t | Misc brackets, guards |
| Reference | Global average crude steel | 1.92 t CO2e/t | worldsteel 2025 |
Which means the relevant question is not "can we use stainless" but "where does stainless actually do work". In a cleanroom the answer is the surfaces that shed particles, the rails the wheels run on, and the fasteners that have to survive wipe-down. Cladding a whole girder in 304 because the buyer likes the look of it adds about 1.5 tonnes of CO2e to a 2 tonne machine, and no cleaning benefit at all above the 1.2 metre line where nobody wipes.
The particle number explains why cleanroom buyers are conservative here. An ISO 5 space allows 3,520 particles of 0.5 micrometre or larger per cubic metre, against roughly 35 million in ordinary city air, and ISO 14644-3:2019 sets out the tests that prove it, including filter leak, air flow direction and recovery. A crane that sheds paint flakes, grease or zinc dust fails those tests, and the crane is a tiny fraction of the room's energy and a large fraction of its particle risk. That is why the stainless is there. It is not there to be a carbon decision, and it should not be defended as one.
Not in the way people think. A pickling line is a C4 or C5 environment under ISO 12944-2, and the coating system that survives it is a 220 to 260 micrometre multi-coat build with a zinc-rich primer, a glass flake or high-build epoxy intermediate, and an aliphatic polyurethane topcoat. The paint itself is maybe 65 kilograms on a 10 tonne crane, which is a fraction of a tonne of CO2e. The blasting and the surface preparation cost more carbon than the paint does.
| Corrosivity class | Where it applies | Total dry film | Coating design life | Full repaints in 20 years |
|---|---|---|---|---|
| C2 | Heated workshop, warehouse | 100 microns | 10 to 15 years | 1 |
| C3 | General production hall | 160 microns | 8 to 12 years | 2 |
| C4 | Steel mill, foundry, coastal | 220 microns | 6 to 10 years | 2 to 3 |
| C5 | Pickling line, chemical plant | 260 microns | 4 to 8 years | 2 to 5 |
So the paint is 0.3 tonnes of CO2e and the repaint programme is a decision about whether the structure reaches twenty years at all. Girder flange section loss above 15 percent means the coating has already failed and the member needs weld repair before you repaint it. That is the real carbon item, and it is measured in tonnes of plate, not kilograms of paint.
Which is why I keep telling buyers the same thing on pickling jobs: put the control cabinet outside the acid mist zone, use IP65 field hardware, and accept a slightly uglier coating schedule on the walkway. The electricals fail before the paint does, and replacing a cabinet is a carbon event that nobody put in the budget.
More than the certificate. Flameproof equipment is heavy by design. An Ex d motor sits inside a cast iron enclosure that has to contain an internal explosion without deforming, with flame paths machined on every joint. That is not a cosmetic difference: the same output at the same speed is a bigger, heavier machine. Control enclosures grow the same way, and spark-resistant bronze hooks and blocks are larger than the carbon steel parts they replace for the same safe working load.
| Item | Standard build | Ex build | Carbon effect |
|---|---|---|---|
| Hoist motor and brake | 11 kW, aluminium frame, 95 kg | 15 kW Ex d, cast iron, 210 kg | +0.27 t |
| Control enclosure | Sheet steel IP55, 60 kg | Ex e / Ex d, 180 kg | +0.28 t |
| Hook block | Forged steel, 70 kg | Spark resistant bronze, 120 kg | +0.32 t |
| Load chain | Grade 80 alloy, 10 t rated | Stainless, derated, next size up | +0.15 t |
| Structure and panels | 13.5 t total machine | 17.8 t total machine | +10.1 t |
| Total, 10 t capacity | 42 t CO2e equivalent build | 42 t CO2e | +32% mass |
Bronze is the quiet one. A copper alloy is roughly 88 percent copper, and copper sits between 1.1 and 8.5 tonnes of CO2e per tonne depending on ore grade, energy source and route, against 2.34 for plate steel. Per kilogram of safe working load, a bronze hook block is around three times the carbon of the forged steel part it replaces. There is no way around it, because the whole point of the bronze is that it will not spark.
The stainless chain derating is the second order effect. Stainless has lower tensile strength than alloy steel, so a stainless chain of the same dimensions carries less, and the machine has to be built one size up to hold the same rating. That one step up propagates: bigger chain wheel, bigger motor, bigger brake, heavier frame.
Whichever one is actually running. I have seen buyers ask about IE4 motors on a cleanroom crane that runs 300 hours a year, and ignore the fill factor on a grab crane that runs 5,000. The two decisions are not the same size.
| Action | Sub-type it helps | Carbon saved | Cost | Payback |
|---|---|---|---|---|
| Match grab volume to bulk density | Grab | 15 to 25% of hoisting energy | A new grab, or nothing if specified right | 1 to 2 years on throughput alone |
| Cut lift height to the process minimum | All | 5 to 15% of hoisting energy | Engineering hours | Immediate |
| Stop the standby load between shifts | Precast stacker, pickling | 60 to 90% of standby, 2 to 6 MWh a year | USD 400 to 1,200 | Under 1 year |
| Restrict stainless to functional surfaces | Cleanroom | Up to 1.5 t per machine | Negative, it is cheaper | Immediate |
| Specify the C5 coating and keep it | Pickling | Avoids a girder replacement, 32 t | Included in the machine price | One avoided failure |
| Regenerative drive on the hoist | Grab, heavy utility | 8 to 15% of hoisting energy | USD 14,000 to 25,000 | 4 to 8 years on a grab duty |
| IE4 motor upgrade for carbon reasons | All | 1.5 efficiency points, about 0.09 t a year | USD 40 to 80 per kilogram of extra copper and aluminium | Never on carbon grounds |
The last line is not a mistake. Moving a 22 kW motor from IE3 to IE4 saves 1.5 percentage points, which is about 440 kWh a year at high duty, which is 0.09 tonnes of CO2e. The extra copper and aluminium in the IE4 rotor is 40 to 80 kilograms of CO2e. On carbon alone the payback is five to ten years, which is the design life. Buy IE4 because the electricity bill matters, or because a tender requires it, and say so honestly.
This is the table I would put in front of a buyer who has one page and two minutes. Operating carbon assumes 206 g/kWh for the EU grid in 2024 and 537 g/kWh for the Chinese grid, both published averages, and the same planning duty per sub-type used in our energy work.
| Sub-type | Embodied | Operating, EU grid | Operating, China grid | Embodied share, EU | Where the money goes |
|---|---|---|---|---|---|
| Grab, 20 t A7 | 61 t | 176 t | 459 t | 26% | Electricity, then grabs |
| Precast stacker, 50 t A5 | 145 t | 103 t | 269 t | 58% | Steel, once |
| Pickling, 10 t A7 | 32 t | 71 t | 185 t | 31% | Steel, then electricity |
| Cleanroom, 2 t A3 | 5.3 t | 6.2 t | 16 t | 46% | Stainless, once |
| Ex, 10 t A5 | 42 t | 84 t | 219 t | 33% | Cast iron and bronze |
The precast stacker is the surprise. It is the heaviest machine and it spends most of its life standing still, so 58 percent of its twenty year carbon has already been spent when it arrives on site. Everything that reduces its mass or its standby load is worth doing, and the mass is fixed at the design stage by the load path, the lift height and the travel length. After delivery you can only work on standby.
The grab crane is the opposite. Two thirds of the carbon happens during the twenty years, which means the operating decisions dominate, and the operating decisions are made by whoever loads the bucket. That is not a purchasing decision at all, it is an operations decision. I have watched plants spend six months selecting a motor class and then load a coal grab at half its rated volume for two shifts a day.
Not the crane. HS 8426 sits outside Annex I of Regulation (EU) 2023/956, exactly like 8425 for hoists, so a complete special crane imported into the EU does not trigger a CBAM declaration. What sits inside Annex I is most of what the crane is made of when you buy it separately.
| Item on a special crane | CN code | Inside Annex I | Default value | Exposure on one machine |
|---|---|---|---|---|
| Complete crane | 8426 | No | n/a | Zero |
| Stainless cladding, 500 kg | 7219 31 00 | Yes | 5.571 | 2.8 t CO2e, about EUR 229 |
| Grab body, 7.6 t | 8431 41 00 | No | n/a | Zero |
| Runway brackets, 1.2 t | 7308 90 98 | Yes | 6.374 | 7.6 t CO2e, about EUR 629 |
| Bolts and nuts, 300 kg | 7318 15 88 | Yes | 6.077 | 1.8 t CO2e, about EUR 150 |
| Misc steel articles, 400 kg | 7326 90 98 | Yes | 3.439 | 1.4 t CO2e, about EUR 113 |
Certificate prices for 2026 were published quarterly by the Commission: 75.36 euro in the first quarter, 75.28 in the second, and 82.32 in the third, which is the figure used above. The fourth quarter price is due on 4 January 2027 and from 2027 the price is published weekly. The first declaration and surrender for the 2026 import year is due by 30 September 2027.
The threshold is 50 tonnes of CBAM goods per importer per calendar year, across all CN codes combined. A single special crane's accessories rarely reach that, but a plant building programme will. A precast yard that buys its runway brackets, crane rails and holding-down bolts locally is buying domestic goods and has no exposure at all, which is the honest answer for most of the buyers who ask me about this.
One more thing, because it comes up in every second conversation. Splitting a crane into loose steel and declaring it under 7308 does not remove it from the crane heading. General Interpretative Rule 2(a) puts unassembled or disassembled goods in the same heading as the assembled article, and the Court of Justice confirmed that reading in Case C-107/22 in April 2023 for goods that need only bolting, riveting or welding to assemble. What changes is the freight profile and the packing cost, not the classification. The parts a buyer purchases separately are a different matter, and those really are 7308 transactions.
Ten lines. Each one is answerable by any serious supplier in a day, and any supplier who cannot answer them has not calculated anything.
| # | What to ask | Why it matters |
|---|---|---|
| 1 | Total machine mass, by main assembly, and the grab mass separately | Everything else derives from this |
| 2 | Steel grade and mill origin for plate over 10 mm | BF-BOF against EAF is 2.34 against 0.69 |
| 3 | Mass of stainless, aluminium, bronze and copper, by CN code | These carry two to four times the factor of steel |
| 4 | Grab volume, dead weight and the bulk density it was sized for | Wrong density is half the throughput at the same carbon |
| 5 | Replacement interval and mass for the wear items: grab shell, rope, chain, wheels | Recurring carbon is the part buyers forget |
| 6 | Coating system, total dry film and design life to ISO 12944-2 | Decides whether the structure reaches 20 years |
| 7 | Installed power by motion, and whether a regenerative unit is fitted | Basis of every operating estimate |
| 8 | Standby load in watts, itemised | On stackers and pickling cranes it is the largest controllable item |
| 9 | Duty class to ISO 4301 and the design life basis | A3 against A7 is a fortyfold difference in cycles |
| 10 | Which supplied items are CBAM goods if bought separately | Tells the buyer where the declaration risk sits |
Line five is the one that gets the blank look, and line nine is the one that changes the answer. Everything on a special crane is downstream of the duty class, and the duty class is set by the process, not by the buyer's preference. If the process is genuinely A7, buy an A7 machine and stop trying to save carbon on the motor. If somebody assumed A7 because the plant runs three shifts, go and count the cycles first. That single measurement is worth more than every efficiency upgrade in this article.
The special crane range covers grab, precast, pickling, cleanroom and Ex builds from 0.5 to 100 tonnes. If you send us the mass and material breakdown of a machine you already own, we will put it through the same numbers and tell you which of the five lines above is worth arguing about first.
Yes, per tonne of capacity, and by a wide margin. A 10 tonne single girder hook crane carries about 15 tonnes of CO2e, which is 1.5 tonnes per tonne of capacity. The five sub-types in this article run from 2.65 to 4.19 tonnes per tonne of capacity, so between 1.7 and 2.8 times more. The reason is not exotic materials on most of them. It is that the machine is sized by the process as well as the load, and the grab, the stacker mast or the Ex enclosure adds mass that a hook crane never carries. You can see the standard builds on the special crane page.
It does, and this is the single mistake behind most grab sizing failures. Rated capacity covers grab dead weight, payload and rigging. A 20 tonne crane fitted with a 7.6 tonne electro-hydraulic grab lifts 12.4 tonnes of material, not 20. On an 80 tonne crane with a 34 tonne orange peel grab the payload is 46 tonnes. Buyers who quote grab volume without subtracting dead weight end up with a machine that cannot reach its stated throughput, and the fix is a bigger crane, which is another 20 to 30 tonnes of steel.
No, and doing so is the most common carbon mistake on this sub-type. Stainless carries 5.5 to 6.5 tonnes of CO2e per tonne against 1.92 for average crude steel, so cladding a whole girder in 304 adds about 1.5 tonnes of CO2e to a 2 tonne machine for no cleaning benefit above the height anyone actually wipes. Put the stainless where it does work: running surfaces, wipe-down panels below about 1.2 metres, and fasteners that see cleaning chemicals. The rest can be a good coating.
Not to the complete crane. HS 8426 is outside Annex I of Regulation (EU) 2023/956, the same as 8425 for hoists, so an imported special crane triggers no declaration. The materials inside it are a different matter: stainless sheet under 7219 and 7222, structural brackets under 7308, bolts under 7318 and general steel articles under 7326 are all CBAM goods with default values between 3.4 and 6.5 tonnes of CO2e per tonne. The threshold is 50 tonnes per importer per year across all CBAM goods, and most single-machine buyers never reach it.
The Ex crane, at about 4.2 tonnes of CO2e per tonne of capacity, because flameproof enclosures, cast iron motor frames and bronze hook blocks are all heavier than the parts they replace, and the stainless chain derating pushes the whole hoist up one size. The grab crane is second at 3.1 when the grab is counted as part of the machine, which it should be, since it ships with the crane and wears out faster than it.
Not directly. The paint on a 10 tonne pickling crane is about 65 kilograms, which is a fraction of a tonne of CO2e, and the blasting costs more carbon than the paint does. What the coating decides is whether the structure reaches twenty years. Girder flange section loss above 15 percent means the member needs weld repair before repainting, and at that point you are talking about tonnes of plate rather than kilograms of paint. So specify the C5 system for the C5 environment, then protect it, and treat the repaint interval as a structural decision.
Related articles:
Hoist Carbon Footprint: What a 5 Tonne Hoist Costs Over 30 Years
Combination Crane Track Carbon: Why the Lightest System Carries the Most
Gantry Crane Runway Carbon: What Two Rails Cost Before the Crane Arrives
Special Crane Energy Efficiency: Four Sub-Types, Four Different Problems
Special Crane Standards: ATEX, ISO 14644 and ASME B30.20 Explained
Special Cranes: Grab, Precast, Pickling, Cleanroom and Ex Builds