The revolution continues! Warfare on the ground and sea will be heavily affected by megawatt-class lasers in the next two decades or so. And, as we'll see, there are transformative applications in industry, energy generation, transportation and remote sensing. Beams on the ground Mobile Tactical High Energy Laser, based on unsuccessful DF reactants Ground warfare would also be affected by a laser revolution. So far, megawatt-scale lasers have been described as efficient weapons that can reach out into space, burn down massed missile strikes, take out aircraft from extreme range and even catch re-entering ICBM warheads. They’re terrifying too: 532 nm is definitely not eye-safe, and with megawatts involved, mere reflections are enough to destroy retinas from kilometers away. In fact, they’d be starting fires everywhere. A successful beam defence of a city or forest may leave it a burnt-down wasteland if many precautions aren’t taken. But does this mean that they are supreme weapons effective in all domains? Let’s work it out in ground warfare.  MTU 1975, sporting a 30 kW CO2 laser A laser weapon as described previously, with a 1 MW output at 532 nm, focused by a 1m diameter mirror, actually fares poorly against armored targets. At the tightest practical spot diameter of 1 cm, it produces an intensity of 12,730 MW/m^2 or 1273 kW/cm^2, and it can maintain this focus out to a range of 10 km. It’s enough to bore through 19 cm/s of steel, 21 cm/s of silicon carbide, 46 cm/s of depleted uranium and 273 cm/s of high density plastic. These are the typical materials of tank armor. Damaged M1150 Assault Breacher (Based on M1A1 Abrams tank) revealing inner armor layers The heaviest US tank in service is the M1A2 Abram SEP v2, weighing over 66 tons. Its strongest armor is in the turret cheeks, adding up to around 3.8 + 10.1: 13.9 cm of steel, sandwiched between 80 cm of NERA elements made of plastic or rubber. It is sloped to give a line-of-sight thickness of 18.3 cm of steel plus 105 cm of plastic. Leopard 2A4 turret armor estimation The nearly as-heavy Challenger 3 TD has sloped turret cheeks protected by 45 cm of steel and 72 cm of NERA plastic. A T-80BVM has 36 cm of steel with a 12 cm quartz core while Russia’s newest T-14 Armata is thought to have its hull front protected by 65 cm of combined steel and NERA plates. The details of these armor schemes are secret, but as rough estimates this is sufficient.  A 1 MW laser beam focused onto a 1 cm spot could bore through the strongest armor of the heaviest tanks in about 1.02 - 2.6 seconds. Weaker front armor, like the M1A2 Abram’s 6.7 cm thick steel turret ring or the 8.5 cm thick lower hull plate of the Challenger 3 would only stand up to a 1 MW beam for 0.35 - 0.4 seconds. Thankfully, they are smaller targets that can be easy to hide with proper use of terrain. T-80 in hull-down position That incredible performance is however a ‘laboratory conditions’ result where the beam can be accurately held onto a +/- 0.5 cm point, over a target that isn’t moving, bouncing around on its suspension, nor rotating its turret in response to blaring Laser Warning System alarms. In realistic conditions, the smallest movements would move the laser to a new cm-sized spot of armor, restarting the drilling from zero.  It also ignores the aspect ratio limitations of the hole created by the laser. A meter deep but 1 cm wide hole has an aspect ratio of 100. A much more realistic result of laser drilling in unprepared outdoor conditions is a 10:1 ratio, meaning that the hole will widen to 10 cm before it reaches a depth of 1 meter. So, the actual amount of material that needs to be vaporized and removed to drill through 1 meter of armor layers is 10^2: 100x, so the actual drilling time is around 102 - 260 seconds. Combined with the armor moving underneath the beam, we have to conclude that a 1 MW laser weapon cannot reliably defeat the best armor of a tank, and it would take very prolonged firing to deal damage unless it gets very lucky to cross paths with a frontal weak-spot. During that time, the tank will either be moving to cover or firing back. The laser turret itself would have to be stationary to maintain sufficient accuracy, making it even more vulnerable to counter-attack. At far engagement ranges of 10 km+, the laser will be firing first. The tank may respond by firing its longest range weapon, which in the case of the T-90M’s 125mm cannon is the 3UBK21 Sprinter ATGM with 12 km range. The older gun-launched ATGM 9М119 Refleks has a shorter 5km max range  At that distance, the laser has over 30 seconds to find, track and delete incoming ATGMs travelling around Mach 1. It takes several minutes for the tank to be defeated, unless it knows to reverse away from the beam while wiggling its turret and swerving with its hull until it dips behind solid cover. It can assist its retreat with smoke launchers.   At medium ranges of 2 - 4 km, the tank and the laser turret would probably find and fire on each other at the same time. The laser takes several minutes to reliably destroy a tank from the front, while the tank can fire a high velocity projectile through its main gun. Modern APFSDS rounds are made of tungsten and travel at 1500-1800 m/s, so there is no chance for a laser to do anything about it. A Challenger 2 managed a tank kill at 10.5 km, supplanting the previous 4.7 km record by a Challenger 1 Worse, the laser turret is a big piece of fragile optical equipment held up high for a better field of view - protecting it the amount of damage needed to destroy it is nearly impossible. At these ranges, a conventional MBT has a strong advantage over lasers. At shorter ranges, it is a question of luck. Does the laser find the tank in a compromising position, unaware or unable to react, and will it quickly bore a channel through its armor? Or does the tank find the laser first and disable it with whatever ammunition is loaded in its main gun or secondary weapons?  A laser encountering a tank’s side armor is a very different situation. Modern MBTs are protected by layers worth less than 5 cm of steel, which can be penetrated in a quarter of a second. The hull cannot be rotated as easily as the turret, and drilling a hole with an aspect ratio of 1:5 (1 cm wide and 5 cm deep) is no problem for a laser. A tank caught out of position by such a beam would be defenceless.  Even so, the damage dealt by a laser is much less than what would happen if a tank is struck by a kinetic weapon or HEAT jet. The result of a laser drilling a hole through armor is… a hole. X-ray hitcam from 'Gunner HEAT PC' showing the post-penetration damage from an APFSDS round striking an M60A3 turret There are few ‘post-penetration’ effects to devastate the tank’s interior: no spalling, no hypervelocity fragments nor internal explosions. The beam, once inside, creates an incapacitating shockwave and adds heat that can directly burn occupants in the beam’s path, ignite fuel or set off ammunition reserves, but it has to directly touch these elements to cause damage. Continuous fire for repeated penetrations is needed to incapacitate a tank, like poking a teddy bear with a needle. In that scenario, it is not particularly better than a large autocannon with armor piercing rounds. Fearless M2 Bradley IFV engaging a T-90M with a 25mm autocannon A Bradley IFV could aim a 25 mm Bushmaster loaded with APDS rounds at the side of MBTs and easily penetrate at ranges up to 2 km. A Swedish CV9040 IFV with its 40mm autocannon can penetrate MBT side armor from over 6 km, if it could aim at targets that far.   To summarize, tanks are resistant enough to megawatt lasers that they cannot be used as a main weapon against them. In situations where they are effective, they do not perform better than smaller, cheaper existing weapons. Big lasers actually make tanks more important on the battlefield. By serving as efficient air defences out to extreme ranges, they can take out one of the biggest threats to tanks: air power. Multi-layered air defence combining lasers and missiles. Today's drone swarms have been reported as the 'death of the tank', while footage of tank columns being destroyed by Hellfire ATGMs and cluster bombs during the Iraq War highlighted how vulnerable tanks could be to air power. Protection from these threats, exaggerated as they may be, can only restore the threat that tanks pose on the battlefield.  Against lighter vehicles, the 1 MW laser is a menace. MT-LBs, BTR-80s, Strykers or Warriors would melt against such beams. The M3 Bradley has about 40 to 60 mm of aluminium protecting it; this would last about 0.05s under fire. Heavier vehicles like the Puma IFV, T-15 Armata or Namer APC would not survive for long either. The presence of megawatt lasers might push them much further back behind the front-lines, which in turn makes deploying troops more complicated. But is this any different than their vulnerability to existing threats, ranging from heavy autocannons to MBT main guns? It is more realistic that powerful lasers only have a minor effect on how lighter ground vehicles are deployed. Lockheed Martin's DEIMOS 50 kW demonstrator. The consequences for infantry are more drastic. A megawatt beam that can slice through steel would have no trouble with human bodies… although the damage would be less ‘slicing’ than ‘instant steam explosion’. That’s not more or less lethal than other heavy weapons on the battlefield (infantry don’t survive 20mm cannon fire either), and while it can fire continuously without worrying about ammunition reserves, it is more easily stopped by hard cover and deals less damage when it gets through.  What will change is that large lasers are very good at countering the tools that infantry use to become a threat to vehicles, buildings or aircraft. Infantry normally poses a threat to tanks, like these Milan ATGM operators ATGMs can be detected and lased unless fired from very short ranges. Mortar rounds, rockets, RPGs, possibly howitzers would be intercepted too. Drones swarms are eliminated in seconds. General Atomics 100 kW-class laser that fits inside a standard container All of these don’t even need 1 MW of beam power to neutralize, as today’s Iron Beam and DE M-SHORAD programs suggest. Without these tools, infantry is limited to fighting other infantry. Against mechanized forces, they must lay minefields or attempt very short-range ambushes, which is a risky tactic even if it does succeed. Hopefully the troops are issued with safety glasses so they are not immediately blinded during the assault. Finally, there’s artillery. The “king of the battlefield” is likely to lose its throne. A typical 155mm shell is protected by a 1-2 cm thick steel case, which a 1 MW laser can get through in 0.05-0.1 seconds, and it probably takes even less time to ignite the explosive filler inside. 155mm M549 rocket-assisted artillery shell Radar arrays that quickly detect rounds fired into the air exist already, and they do so at ranges of 30 to 50 km. If a laser turret takes around 1 second to switch to its target and a negligible amount of time to destroy it, then it can defend itself from a combined fire of 60 rounds per minute. That’s six 2S19M2s, ten Caesar 155mms or twelve 2A36 Giatsint-Bs, all firing ineffectively.  French Caesar 6x6 155mm artillery However, modern SPGs like the AS-90 or PzH 2000 can also shoot bursts of three rounds in less than 10 seconds. If these rounds are fired from long range and they take 30 seconds to land, then one 1 MW turret can absorb ten bursts. From short range, with 10 seconds of air time, a laser turret may defeat burst fire from three artillery pieces, only for rounds from a fourth piece to go through. Numerous SPGs would then be able to defeat laser defenders within their firing range. Of course, this implies that the artillery pieces have give up their long range capabilities to driver much closer to their target than they'd like.   An even simpler approach is to use a multiple rocket launch system. The American M270 MLRS can launch Mach 2.5 supersonic rockets. At their maximum range of 150 km (using ER GMLRS), they would take over 3 minutes to reach their target, taking a trajectory that takes them high into the sky, and therefore a laser turret can intercept 180+ of them. An M270 only has 12 tubes, so it needs to fire from within 12 x 1s x 860 m/s = 10.3 km to overwhelm the laser. The Russian Tornado-G has 40 tubes that fire slightly slower 122mm rockets, able to overwhelm a 1 MW turret from 24 km. If those rockets can split into multiple sub-munitions, then they can overwhelm laser defences from even further away, or with fewer launched rockets. The strongest counter to laser turrets What’s interesting is that defeating a laser turret using ground forces is much easier than by air strike. A laser could be overwhelmed by a single rocket launcher at short range. An asset worth $20m would be defeated by placing rocket trucks worth $1m at risk. For comparison, an air attack that requires 90 missiles and 12 F-35s in the best case, up to 300 missiles and 38 F-35s in worse conditions, would cost $18-60m if it succeeds and $978-3100m if it fails. Megawatt-scale lasers means advancing on the ground is safer and cheaper than trying to win a war from the skies. So, in the near future, ground wars would need to be won in order for air warfare can be conducted, the opposite of today’s reality of air superiority being a requirement for free ground operations.   How do ground forces adapt to megawatt lasers? Main Battle Tanks with their heavy armor would surge in importance. Their armor scheme could be easily updated to protect against laser fire from the sides by the addition of lightweight ablative shields made of graphite. This would greatly improve their durability: a laser intensity that burns through 19 cm/s of steel in ideal conditions only penetrates graphite at a rate of 9.7 cm/s. Graphite is four times less dense than steel, so per kilogram loaded on the tank, it is about eight times more effective than steel. Indonesian Leopard 2 with AMAP package that drastically thickens side armor Each armored group on the battlefield would be equipped with a mobile laser turret, to protect against air attack, ATGMs, drones and most threats that prevent them from running rampant.  This mobile laser turret would probably use an existing MBT chassis to support its multi-ton weight. Power could be derived from existing tank engines, but with additional cooling to handle the resultant waste heat. Rolls-Royce IPTMS demonstrator using a 300 kW M250 gas turbine to deliver power and cooling for high-power lasers.  Let’s imagine a 1 MW system sitting on a Leopard 2 chassis. We remove the turret, main gun and ammunition load, taking out around 15 tons. The V12 diesel engine, which provides 1500 horsepower, has an additional link to its gearbox that drives a generator (perhaps weighing 200 kg at 5 kW/kg) to produce electricity. The tank’s cooling system already handles about 1.6 MW of waste heat from the engine, so it needs to be expanded by 63% to also handle waste heat from the laser weapon.  On top of the chassis will be a tall rotating structure that includes a 2 ton laser generator (2 MW input at 1 kW/kg), a 1.5m ball turret containing a 1m mirror held under armored shutters, a fire control radar, a 100 kWh battery (around 1 ton) and various other equipment that might add another 2 tons, for a total of 6 tons. The 1K11 Stiletto, showing how a laser turret can use shutters. Sufficient armor protection to survive machine guns and shrapnel might double this total. In appearance and role, it most resembles the Leopard 2 Marksman SPAAG, but with a 1.5 m wide ball in the center instead of cannons on the turret sides. It would be just as mobile as a regular Leopard 2, and be able to fire on battery power for a full 3 minutes. This might sound short, but recall that in situations where the laser turret faced missile swarms over the course of several minutes, it spent most of its time switching between targets  and only milliseconds actually destroying them. So 3 minutes of firing would be worth 30+ minutes of operation, depending on how tough the targets are. The onboard generator can then recharge the battery in 6 minutes by running the engine while stopped, or at a slower rate while moving at reduced speed. 1K17 Szhatie Soviet laser tank. It was incredibly expensive due to multiple lasers with ruby lasing rods. If a stationary 1 MW turret costs $10m, then a 'Laser Leopard' variant is likely twice as expensive at $20m (second-hand Leopard 2A4s to provide the chassis are already under $4m).  We can also imagine a 'Laser Marder', based on the old Marder IFV platforms or similar lighter platforms, supporting the 'Laser Leopard'. Marder 1A5A1 IFV They’d have 250 kW lasers and 0.5m mirrors in ball turrets that replace their current ATGM and 20mm autocannon armament. It might only cost $10m. We calculated previously that at short ranges (under 20 km) the 1 MW laser is overkill for most targets and spends most of its time switching between targets. Additional smaller lasers would maximize the number of targets intercepted.  Naturally, a counter-laser vehicle could also be introduced. Directed Energy M-SHORAD vehicles in testing. Their ~0.5m turrets are a good visual reference. It specializes in destroying laser mirrors. In Part I, we found that a 100 kW laser focused by a 0.5m mirror can destroy the mirror (LIDT 10 kW/cm^2) of a laser turret from a distance of 22 km. A light vehicle can use ground cover and concealment to drive much closer, perhaps to within 10 km. From a hidden position at that distance, it can fire a 20 kW beam focused by a 0.5m mirror to achieve the same destructive effect. US Army HEL-TVD 100 kW laser by Dynetics. By 2045 a 20 kW laser weapon system might be small and compact enough to fit on a small off-road vehicle alongside its generator, battery and cooling system, like Raytheon is already attempting with its HEL system on a Polaris MRZR ATV. It would be sensible to add these systems to future reconnaissance vehicles too, so that they can find targets for air strikes and take out laser defences from their vantage point. Raytheon's "laser dune buggy' When large lasers are in play, infantry can do much less than before. They need to be given tools to make them a threat to laser turrets, so they can clear a space for their ATGMs, mortars and other such heavy weapons to become effective again. For example, they could gain small lasers of their own. Today’s infantry is already trained to approach within 5 km of tanks to attack them, and normally have to creep to within 3 km of them to use Stugna-Ps or Javelins. If they can do the same with a 30 cm telescope (it doesn’t need to be a fully rotating turret), then they can produce a spot size of 1 cm and destroy enemy mirrors using only an 8 kW beam. The real question is whether such a laser can be broken down into man-portable pieces that infantry can use within such short distances. Picture a laser designator, but much larger Alternatively, infantry can try taking out laser mirrors with a well-timed shot of an anti-materiel rifle when the shutters are open, or get upgraded to very high velocity ATGMs like the defunct MGM-166 LOSAT or the smaller Compact Kinetic Energy Missile which weighed only 45 kg. A future CKEM would shoot a hypersonic dart across 10 km and destroy any target, laser or not. CKEM fired from fixed tubes on the roof of a Humvee Artillery can see some modifications too. The number of rounds fired in burst mode becomes their essential performance metric, so an autoloader that can fire ten rounds in ten seconds without melting the barrel might be developed. MLRSs might equip rockets that try to maximize speed over range, so that fewer are needed. But as mentioned before, large 1 MW laser turrets can be backed up by smaller mini-turrets that multiply the number of projectiles they can intercept from short range. If Laser Marders are present alongside Laser Leopards, then the number of artillery rounds or rockets intercepted goes up linearly without also costing much more. Cluster munitions would however increase the number of targets even faster. Surface Action HELIOS: 60 kW+ from a Spectral Beam Combined Fiber Laser  Powerful lasers are already close to deployment at sea. The US Navy’s HELCAP doesn’t need much improvement to jump from 300 kW to 1 MW. We’ve also pointed out the techniques they can use to overcome the beaming challenges at sea, to shoot through atmospheric turbulence, moisture, fog and clouds. In the featureless terrain of the open ocean, lasers are sure to give their best performance.  The 1 MW green laser focused by a 1 m diameter mirror would already be very effective. It can defend ships from hundreds of subsonic missiles and dozens of supersonic missiles, per turret. It also destroys boats and UAVs out to the horizon, despite only costing as much as a single Phalanx CIWS turret today (assuming it doesn’t get its own independent radar). A ‘bolt-on’ solution (power, cooling, radar included) that can be placed anywhere would be double or triple the cost of the laser alone, but it would be more than worthwhile as a massive air defence upgrade to most of today’s large ships. Yet, you’ll find the largest ships cost billions of dollars each. It is therefore justified to consider a proportionally larger laser weapon for mounting aboard Navy ships by 2045. Let’s consider a 5 MW laser focused by a 3 meter diameter mirror. It would cost at least $60 million, or more if it is difficult to construct mirrors of that size. An Arleigh Burke class destroyer has 16 MW of electrical power generation capacity in the Flight III upgrade. Adding a 5 MW laser, which needs 10 MW of electrical input, would require adding a lot more electrical power generating capability and significant battery storage; 1000 kWh to fire continuously for 6 minutes.   Where the previous 1 MW laser was able to produce a spot intensity of 32 MW/m^2 or 3.2 kW/cm^2 at 200 km, the 5 MW mega-laser extends it to 987 km. It is dangerous to satellites at up to 4000 km altitude, instantly lethal to any aircraft it sees and a great shield against ICBM threats on top of being an extremely effective telescope and active scanner in its own right. The line-of-sight limitation however remains. This is the HMS Daring If it is mounted on a mast 21 meters high, a position occupied by the Sampson radar aboard the Type 45 destroyer, then it can fire on sea-skimming missiles flying at 10 meter altitude from a distance of 28 km. If the target is a high-flying aircraft, perhaps at 18 km altitude, then the laser can reach it from 500 km away. A laser won’t be the sole weapon installed. Navy ships are packed with missiles that are not limited by line-of-sight or distance to the visual horizon. Missiles shoot what cannot be seen, while lasers focus on everything that is within line-of-sight. This can include other ships. Lasers defeating all manner of threats at sea, even at lower power level. Navy ships are packed with missiles that are not limited by line-of-sight or the horizon. With such a laser on board, a ship can reserve its missiles for anything it cannot directly see, and fire with the laser on anything it can see. This can include other ships.  Let’s consider ship vs ship combat. As in the air warfare scenario, the side with the laser has a massive advantage, so there would be immense pressure to add these defences to ships as well. Ships can defend themselves from many missiles fired at them, especially if they have both a high-mounted (16m mast) long-range laser and smaller mini-turrets to maximize the number of projectiles intercepted. If we assume a ship has three turrets available (one 5 MW mega-laser and two 1 MW lasers) and target switching time is 1 second, then it can take out missiles as soon as they appear over the horizon.  The fastest anti-ship attacks are delivered from essentially short-ranged ballistic missiles. China is building up an arsenal of these with missiles like the DF-21, which crosses over 1500 km while climbing to 1200 km altitude before diving down at a terminal velocity of Mach 10. It’s not known if the DF-21’s warhead deploys decoys while outside the atmosphere, or if it carries less re-entry shielding than full ICBM warheads, but even if we assume the worst case scenario, a 5+1+1 MW laser defence system can take out over 120 of them after they hit the atmosphere on their way down.  Sea-skimming missiles fare better as they spend less time exposed to laser fire. The P-1000 Vulkan is a monstrous 11.7m long missile that can reach Mach 3 while flying only 50 meters over the sea. It pops up over the horizon from 40 km away, giving defenders less than 39 seconds to respond. In that time, 78 missiles will be destroyed. The Mach 3 Indian BrahMos ramjet missile flies lower: 3 meters above the waves, apparently, meaning it only appears to the lasers from 20.5 km. That reduces the number taken out to 40. Laser defences with megawatt-scale beams and multiple turrets can take out enormous missile salvos. They can absorb full salvos from 4-5 other ships, even if they’re the largest and most advanced sea-skimming missiles. So, in an equal engagement, no missiles get through. However, from a distance of 25 km, ships would start to become directly visible to each other.  A warship design optimized for laser weapons A 5 MW beam focused by a 3m mirror can produce a 1.7cm wide (1.5x diffraction limit) spot at 25 km, enough to cut through 34 cm/s of steel or 151 cm/s of aluminium alloy. If the beam is intentionally widened to 38 cm, then the drill rate is 1 cm/s through aluminium alloy. That is enough to open Domino’s Extra-Large Pizza sized holes in any ship’s hull (5-10mm of aluminium) every second. Ships cannot instantly rotate their hull nor quickly exit the laser fire range, so they are likely to leave any direct engagement with hundreds of such holes. While there are on-going efforts today to counter directed energy weapons, it is unlikely ships will emerge unscathed from such engagements. Ideas for very long range bombardment using railguns or coilguns could be revived, as hypersonic inert projectiles delivered at a high rate of fire from behind the horizon would be an excellent counter to lasers. The same electrical output installed for laser weapons could be used to power these guns. Eventually, we might see the return of battleship design aspects: multiple large guns, relying on thick (ablative) armor to survive laser strikes and less concerned by air or missile attack... but that's further into the future than 2045. Pan Spatial's USS Teton futuristic battleship. Even better than heavy armor is diving underwater… or better still, attack while remaining underwater. Submarines would be largely unaffected by powerful lasers. Even a few meters of seawater is enough to render megawatt beams useless, meaning they’ll remain well protected. Their main means of attack are torpedoes and missiles launched from underwater. Torpedoes would be as effective as usual. Submarine-launched missiles like the Exocet SM39 or UGM-84 Harpoon must emerge from the water, forcing them to face laser defences, but a submarine has the advantage of being able to launch them very close to their target. MBDA SM40 Exocet with 120 km range For example, they could be launched from a distance of 5 km, greatly reducing the time laser turrets get to fire at them. SLBMs used in an anti-ship role face a similar challenge that regular ballistic missiles do, but they too can be launched from closer distances. The same goes for SLCMs. It all implies that submarines (either large crewed designs or new XLUUV drones) could take on the primary anti-ship role, while regular surface ships focus on anti-air and anti-sub operations. Anduril's Ghost Shark XLUUV There are further uses for megawatt lasers, other than sinking ships and felling missiles. Laser light can serve as a power transmitter to any craft equipped with photovoltaic panels. A PV panel specifically tuned to the wavelength of the laser can convert the light it receives into electricity with triple the efficiency of a regular silicon solar panel, up to 86% efficiency. Electrically powered aircraft, like an observation drone, could install laser-PV panels on the undersides of their wings and get recharged by ship beams at sea, allowing them to stay aloft indefinitely. In a further future, those drones could also carry aloft relay mirrors to extend the ship beams far over the horizon. Simpler, cheaper flat mirrors won’t cost much, making them acceptably expendable, but they would get around the direct line-of-sight requirement for laser fire.  And, while megawatt lasers generally are very harmful to the presence and effectiveness of air forces, they can at least permit them to focus on offensive missions, while laser turrets take on the role of defence. The majority of an Aircraft Carrier Group’s air wing’s time is spent on defensive patrols. Instead, all those E-2 Hawkeyes and F/A-18 buddy tankers could be re-dedicated to serving strike missions. The same can be said for a Navy ship’s missile loadout: it can be mainly offensive instead of nearly entirely defensive, as today. Instead of an Arleigh Burke-class destroyer dedicating over 60% of its 96 VLS tubes to defensive SM-2s, SM-6s and Evolved Sea Sparrow missiles, they can be filled nearly entirely with LRASMs or the upcoming Naval Strike Missile.     The only real obstacle to these lasers at sea by 2045 is how long it takes to modify and upgrade ships, let alone design and launch new classes that make full use of new laser weapons!  Gone with the photon So far, we have looked exclusively at the military domain. The laser revolution extends far beyond that.  PowerLight laser power beaming demonstration What we’d like to see is megawatt-class lasers assisting with transportation, by using them to transmit beamed power. However, this is unlikely to happen in the next 20 years. Shipping and aircraft require too much power while cars and trucks are too numerous to dedicate a beam to each vehicle.  Ships are large and the most efficient way to move things around. "ONE Innovation": 24,136 TEU, driven by 62 MW MAN B&W engines to 22 kts The gigantic Triple E-class container ships are 399m long and 58.6m wide, so their top surface area is roughly 23,381 m^2. If 75% of that area is covered in 20% efficient solar panels, it could generate 3.5 MW of electricity. That’s not enough to match the output of its 2x31 MW diesel engines, and it’s only that much in perfect weather conditions, during the day and without clouds. If those panels are replaced by photovoltaic versions tuned specifically to receive a single laser wavelength, they can reach an efficiency of 68% or better. A direct free-space optical power link Even so, they’d need a laser intensity of over 5200 W/m^2 to produce the required 62 MW of power. That’s a pretty dangerous beam to be shooting down at the ocean among commercial shipping lanes: an uncooled surface in the path of that beam would reach a temperature of 550 Kelvin. Aircraft have it much worse. There are several ongoing attempts at aircraft electrification, from hybrid electric/thermal propulsion to ‘turbo-electric’ architectures to pure electric flight. NASA next-generation turboelectric aircraft design While electric motors are approaching the point where they could replace turbine engines at the same power density level (5-15 kW/kg), electric flight is hampered by the insufficient energy density of batteries (40x lower than carbon fuels) and the weight of generators or fuel cells. Lasers can get theoretically rid of these obstacles by delivering power from a remote source. All the aircraft needs to do is install large photovoltaic panels on its underside to catch laser beams, convert them into electricity and use it as needed, without storage. There is a reason why the MOTH2 UAV is so thin and lightweight A beamed-power airplane coated in these panels on its underside could feed electrical power directly into electric engines to propel itself, as long as it is within reach of ground-based lasers, allowing it to cross indefinite distances without expending fuel. It is likely to still have an onboard turbine-generator to produce additional power during take-off, or to take over during emergencies. If you don’t want megawatt laser beams wandering over airports near cities, then requiring that electric beam-riding aircraft also have fuel reserves for independent take-off and landings would be fair.  However, the way aircraft are currently designed does not suit the needs of beamed power very well. If you removed the fuel from a 737-800 until it weighed around 50 tons, for example, and placed it at cruising altitude with a lift-to-drag ratio of 10, then you’d need roughly 50 kN of thrust (which is similar to the 29.4 x 2 kN maximum cruise thrust mentioned here). Power = Thrust x Velocity, so the plane would need a net engine output of 11 MW at 800 km/h. The plane’s wing and fuselage area add up to roughly 230 m^2, meaning if it was all coated with PV panels, they’d need to pump out 47.8 kW/m^2. Considering 70% efficiency as realistic, that’s an incoming laser intensity of 68 kW/m^2! Far too high. If that beam wandered over a regular airplane, it would produce a temperature over 1046 K, enough to melt its aluminium surfaces.   Alternatively, go all-in with intense beams to drive a laser-powered turbofan A flying wing design might double this lift-to-drag ratio in ideal flight conditions, but that would still require dangerous beam intensities. Instead, lasers could be used to power small drones. Whether it’s carrying packages or circling above to transmit radio waves or monitor the ground, it is much easier to deliver the electricity they need using beams of modest intensity. Attempts to realize this are ongoing, although there is competition between using lasers to deliver this energy or microwaves. Laser power transmission exists, but needs to be scaled up 100,000x for humans Megawatt lasers allow these drones to scale up to the size of light aircraft (the Daher TBM 960 has a 633 kW engine) or medium helicopters (the Leonardo AW09 with 5-8 seats uses a 750 kW turbine), more than capable enough of carrying people. They just need to be equipped with photovoltaic panels of around 20 x 20 meters in size. Cheap yet powerful lasers could make laser rocket launches practical. If it takes 1 MW to launch 1 kg to space, and that MW only costs $10,000, then a $100m laser launch facility can start delivering ten ton payloads into orbit. Depiction of a laser launch facility using many small lasers to minimize costs A laser launch may last 10 minutes, and the equipment would be around 50% efficient overall, so 1.2 GJ or 333 kWh of electrical energy would need to be expended per launch per kg. That’s $54 at the average electricity prices in the US, but it can be as low as $23 in India! Add in the $10 cost of propellant (mass ratio 2.6 rocket at $6/kg) and other running costs, and you space launch within $100/kg. That’s an order of magnitude lower than the launch costs of chemical rockets (even the Falcon 9 and the upcoming Starship) and promising enough to warrant serious efforts.  Which suits the next possibility very well. Megawatt lasers could also open a path to Space Based Solar Power return to being a serious contender for solving the planet’s energy problems. NASA/Boeing 100 kWe laser-beaming power satellite proposal Lasers with the right wavelength and assisted with cloud-boring techniques have the advantage of being unperturbed by the atmosphere, and easily focused across very long distances using relatively small focusing mirrors. Chinese plans for this energy source do consider laser beaming, as opposed to typical microwave-based Western designs. A 2m wide mirror focusing 532 nm beams can produce a spot that’s 24 m in diameter… from geostationary orbit. If 10x solar intensity is acceptable to beam down from space, with an automatic interruption for birds or planes flying through it, then the satellite could deliver 4.5 MW of power all day (3.2 MW of electricity after conversion inefficiency). This should be compared to the kilometer-scale transmitters envisioned for microwave-beaming power satellites that also needed receivers hundreds of meters wide (while also acting as a massive radio jammer for mobile phone or Wifi frequencies), which in turn justified photovoltaic panels growing to the Gigawatt scale. 1970s DOE/NASA solar power satellite studies resulted in gigantic designs That makes each power station a huge investment requiring billions of dollars each in multiple launches. A MW-scale laser-beaming power satellite could fit inside a single Falcon 9 launch and start operating as soon as it reaches its intended orbit. If that launch is also done using lasers, then it can become cheaper still.  There are further possibilities with using tightly focused beams to deliver power to spacecraft with electric thrusters across cislunar (10,000 km) distances. Leik Myrabo designs for power beaming technologies  A laser-electric drive could rapidly spiral in and out of Low Earth Orbit using small receiver panels (1-5 kW/m^2) instead of titanic solar panels (200 W/m^2) that are 5-25x larger and heavier. Since electric thrusters have many times the specific impulse of chemical rockets, or even upcoming solid-core nuclear thermal rockets, travel between the Earth and the Moon could be done that much more efficiently. Finally, we can see large lasers being added to the roster of sensors we use to detect and track aircraft and ships, or weather patterns, land erosion and deforestation. Flying laser sensors are cutting edge technology They can act as giant LIDARs, which would act as much higher resolution RADAR that with very long range or excellent resolution, that detects radio-transparent objects or illuminate surfaces for other sensors to work on. Such lasers could be mounted on aircraft or observation satellites. Space based laser detection of submarines We haven't forgotten laser cutting of metals, but it seems that industry is limited by other factors and therefore more powerful lasers won't help much. A caution All that’s been written so far is based on rough calculations, publicly available data and projections decades into the future. Some aspects might not have been covered enough, such as the important role of sensors and electronic warfare in air defence scenarios, or the omnipresent danger of blinding people with large lasers. There will be details that will turn out wrong, either in their optimism or pessimism, but hopefully they’ll be near misses and the general picture remains true (plus or minus a few years). If the sort of future described here starts to take form, then hard questions need to be posed about how nations defend themselves, what role nuclear weapons will have, or whether current spending plans are wise…  Either way, there will be a laser revolution and like most revolutions, it will shake things up.

Imagine you could take a train ride to space. Tracks that slope up into the sky, higher and higher, until you reach a plateau above the planet where it’s a straight line up to orbital velocity. That’s what’s possible with a Lofstrom Loop. But sending you into orbit is just one of the things it can do! The mechanics of a Lofstrom Launch Loop (presented here) are simple and straightforward but it is the implementation of each of its parts that is difficult.  Let’s start with what we have today: rocket launch. A see-through SLS rocket revealing how much of it is just propellant liquids. Modern rocket engines are very effective at converting the chemical energy stored in their fuel and oxidizer into heat. Large expansion nozzles then do a decent job at turning that heat energy into thrust to accelerate a spacecraft. However, most of a launch vehicle today is propellant, not payload. And nearly all of that propellant is spent accelerating the rest of the propellant. A vehicle like the Falcon 9 FT is 4% payload by mass. In other words, it is wasting 96% of its energy on liftoff. Ideally you want to spend energy on accelerating only the payload and nothing else. You could do this by putting the payload on a track and pushing it along with magnets, like a train. A design like the Maglev, which can be thought of as an electric motor unrolled into a long line, can reach theoretically unlimited velocities with great efficiency.  Japan's 603 km/h Maglev train. However, there’s the problem of drag. A train on the ground has to push air out of its way. That’s the main source of energy losses and the reason why it has a top speed. It’s a problem that cannot be solved with just More Power: at high enough speeds, the air doesn’t get out of the way fast enough and instead compresses in front of the train. Aerodynamic heating is dangerous - at Mach 2.5, it becomes dangerous for aluminium… at orbital velocity (equivalent to Mach 25), it is enough to vaporize every material in existence. There are ways around that issue, like enclosing the train in a vacuum tube or equipping it with an enormous heatshield. These are difficult and expensive options.  StarTram envisaged a shuttle accelerated within an enclosed vacuum tube from the ground. It’s best to get rid of the drag problem altogether. A train raised out of the atmosphere can accelerate to any speed on a long enough track. But how do you lift tracks up to an altitude of 80 km or higher and keep them there? You can’t attach them to balloons. There’s no propeller or wing that can generate lift up there either. An 80 km tall pillar of steel won’t work either; it would have to be shaped more like a pyramid and weigh several billions of tons. That's due to the specific strength limits of construction materials. The solution is a dynamic support structure; held up not by the strength of its materials but by the momentum of a high velocity rotor.    Dynamic Support Structures The Space Fountain is a simple early concept for a dynamically supported structure. This is the key component of the Lofstrom Loop. Dynamic support is the only way to build things over 80 km tall without mobilizing mountains of resources. That’s because we do not have materials that are strong enough. Conventional structures like a pillar have a limit: a maximum height beyond which the materials it is made of cannot be stacked any further without collapsing under their own weight.   That height limit depends on the ratio of their mechanical strength to their density, divided by the acceleration of gravity. A traditional building material like bricks has a compressive strength of 20 MegaPascals and a density of 2000 kg/m^3. A pillar of bricks can only go up to a height of around 1000 meters before the lowest layer breaks. Construction steel is better, with a strength of 350 MPa and a density of 8000 kg/m^3. It can be stacked to an altitude of 4460 m. That’s still very far from 80 kilometers of altitude.  Wikipedia's table of specific tensile strengths. To go further, we need to use shapes that get wider at the base, like a pyramid, or use materials that are extremely strong but also very light. Neither is a good option. A gigantic 80 km tall pyramid, large enough to support a train track at its peak, would need several hundreds of billions of tons of steel. A structure made of a super-material like carbon nanotubes would be much slimmer, but also cost trillions to make because the material is so expensive.  14 cm long carbon nanotubes are considered 'ultra-long'. We would need several thousands of km of them. A radically different structure is necessary, and that’s where the dynamic support structure comes in. It’s actually easy to understand. You’ve worked with one before without realizing it. Have you ever played with a water hose? Water in fountains following parabolic arcs. If you turn the tap all the way open and point the nozzle up, you can make tall arcs of water. Push your hand into the stream and you will feel a force lifting it up and away. That force is due to the momentum of the moving water resisting your attempt to bend its trajectory.  Your hand takes the place of this 'deflection plate'. A dynamic structure relies on that same principle to raise a platform up to the sky. Instead of water, we’ll use a rotor, which is a metal cable going around in loops inside a tube. A basic dynamic support structure section. This is possible to build in a vacuum. Inside an atmosphere, we have to protect those moving masses from drag.  The air in the tube is removed, creating a vacuum that reduces drag forces on the rotor to zero. Magnets surrounding the tube can also levitate the metal cable, so it does not have to slide against anything. Cross-section of the rotor and vacuum sheath. With no drag and no friction, we can accelerate the metal cable to incredible velocities and then not have to worry about it experiencing resistance or slowing down afterwards. Those incredible velocities can give a lightweight rotor a lot of momentum, which in turn means it takes an enormous amount of force to bend it. A dynamic support structure makes use of that enormous force to carry a load.  Sketch of a bridge supported by moving masses. To do that, we point the moving mass of iron inside its vacuum tube upwards. It will try to form an arc. Place a load at the top of that arc and it will flatten down until the load’s weight matches the forces resisting the bending of the rotor. That flat section is perfect for a train ride all the way to space. We call this sort of arrangement a Lofstrom Loop after its inventor (first presented in 1985).  Increasing the load a rotor can support or the altitude it can reach is simply a case of making the rotor loop go faster. It can far surpass the performance of any solid material without becoming much heavier or more expensive. There are disadvantages of course. Dynamic structures require power. This may be acceptable for a space launch system when the electricity is replacing thousands of tons of propellant. It’s not suitable for a building or tower that normally stands for a century without maintenance. And even if costs were ignored, it is not reasonable to expect no loss-of-power or blackout events over the course of an entire century!  They are not intrinsically self-correcting either. A regular structure made of solid materials withstands deformation from all directions. A loop will resist being bent in one direction but is fine with flopping over to the side. This exacerbates another problem: dynamic structures need active control. Tacoma Narrows bridge is an infamous example of a structure lost to uncontrolled vibrations. An iron rotor spinning inside a tube of magnets is not passively stable and will hit the walls unless its drift is corrected. The most dangerous event to look out for are certain vibrations that will self-amplify, creating waves in the rotor that grow until they strike the tube walls.  And while dynamic structures are scalable, so they can be built and demonstrated using small loops first, they only have advantages over regular steel and concrete when reaching for extreme heights. Not 1 km, but several tens of kilometers above the ground. We rarely need to build that tall, so dynamic structures won’t often be practical to use.  Lofstrom Loop As stated above, the basic concept for how the launch loop operates is simple to understand. The implementation is the hard part. Note that there are few technical details on how a Launch Loop would actually work outside of Lofstrom's publications, which include an earlier version and a later paper. In this post are included illustrations from both papers, but only numbers from the latter. The Orions Arm depiction of a Launch Loop. Lofstrom’s ‘Launch Loop’ uses an iron rotor travelling at 14,000 m/s. It’s a hollow tube only 5 cm in diameter with walls just 2.5mm thick. A meter length section of it would mass just under 3 kg. A vacuum sheath protects the rotor, with the vacuum maintained by pumps spaced every 10 km. Another cross-section of the loop's sheath.  Ferromagnetic levitation is achieved using a combination of permanent magnets and electromagnets. 40 megawatts of electricity are required to maintain their field strength. The currents running through these magnets generate heat, which needs to be dissipated using radiator fins along the sides.  In total, the iron rotor weighs 15,600 tons. It runs a complete loop in around 5 minutes. The iron rotor could be a ribbon, as presented in the earlier work by Lofstrom. Additional equipment includes pumps to maintain vacuum conditions along the track, position sensors and high frequency electronic controls that can adjust the magnetic field strength to compensate for vibrations, magnetic instability, buckling and so on.  Track weight. A lot of optimization is possible. On average, the track mass is 7.1 kg/m. The weight of the track is less than the vertical force from the dynamic structure, so it can hold itself at any altitude without external support. Still, it is attached to the ground with stabilization cables that compensate for the push and pull of winds or other perturbations. A little bit of surplus lifting force pulls the stabilization cables taut and allows the track to carry the additional weight of a train on top without bending.  Lofstrom adds three more components to complete the launch loop: an elevator, deflection stations and a rotor motor. Pulley elevator at 80 km altitude. The elevator carries a payload (a train) vertically up to the top of the track. It is a simple set of pulleys connected to the flat part at the top of the loop. Lightweight, yet sized to lift several tons at Mach 1 so that the 80 km climb takes less than 4 minutes. An electric motor on the ground moves the elevator. 600 tons per hour costs 130 MW of electricity from the motor.  Deflection stations turn the loop around at either end. One version of Lofstrom's deflection stations. They have to take the momentum of the hypervelocity rotor and turn it around 180 degrees… a difficult task! To manage it, they use a 28 km wide semicircle of magnets. 100 megawatts are consumed here to produce strong enough magnetic fields to bend the iron rotor. Each station has 5000 tons of mass and has to be well anchored into the ground. They form the most significant pieces of surface infrastructure. The final piece is a large electric linear motor, 10 km long. It constantly pulls on the iron rotor to increase its velocity, which is necessary to compensate for the small drag losses the rotor incurs as it runs around the loop. A linear motor as an 'unwrapped' rotary motor. This comes from both magnetic field discontinuities slowing down the rotor and aerodynamic drag from residual gases inside the vacuum sheath. 60 MW of motor power are needed to overcome it all. As we will see in the next section, launching off the loop also takes away from the rotor’s momentum. The main task of the motor is actually to make up for the momentum lost with each launch. Lofstrom proposes an additional 300 MW of motor power to maintain a high launch rate of 48 payloads to Low Earth Orbit per day, with each payload massing 5 tons. Lofstrom’s design slopes off the ground at 9 to 20 degrees to reach an altitude of 80 km. It creates a 2000 km long ‘launch track’ outside of most of the atmosphere. That length of track is used to accelerate payloads into space.  Deploying the structure requires the motor and deflection stations to be online. The iron rotor is slowly brought up to speed - a process which may take weeks. As it rises off the surface, stabilizing cables are added until the whole track sits at 80 km altitude.  Deployment can be sped up by providing more power to the linear motors.  To summarize: the iron rotor travels at nearly twice orbital velocity, pulling itself and all attached structures off the ground. Magnets turn it around at each end of the loop. The rotor is held, levitating, inside a vacuum tube while a linear motor keeps it up to speed. It’s more accurate to imagine this structure as a ‘flying rope’ than something solid you can walk on.  Altogether, the launch loop consumes 500 MW of power and delivers 240 tons to LEO per day. It has the capacity to send off payloads much more frequently, but it would require proportionally more power. For example, using 17,000 MW, the launch loop can send off 9600 tons per day to the Moon!  If no launches are scheduled, it continues to consume 200 MW of power. Theoretically, rotor has enough momentum that it’ll continue gliding along for a long time, but an actual prolonged loss of power would end the active control that suppresses wobbles, vibrations and other shakes to keep it safe.  The Launch So how does a launch actually get done? The key element here is the magnetic eddy currents that the fast-moving iron rotor can induce in a nearby magnet. If we place a row of magnets next to the launch track, it will be both repelled and dragged along by these eddy currents. This is similar in principle to the electrodynamic suspension felt by a Maglev train passing over a track of magnets. A 10 meter long ‘magnet rack’ passing over the Lofstrom Loop will feel 50 kN of lift force and 150 kN of drag force. This is enough to hold a 5 ton payload up against gravity while pulling it along at up to 3g of acceleration.  As the payload accelerates, three things happen: the iron rotor slows down (a momentum transfer), the relative velocity between the payload and rotor decreases and the rotor gets deflected downwards (equal and opposite reaction to payload weight). If left unattended, the payload will gain speed while losing lift, until it crashes into the rotor, destroying the Launch Loop.  Instead, the Launch Loop must compensate for each effect. The linear motor on the ground can gradually bring the iron rotor back up to its initial velocity in between launches. The magnet rack has to continually get closer to the iron rotor to keep increasing the lift force as the relative velocity decreases. Magnets closer to each other feel stronger forces!  The payload’s passage pushes down on the loop. The acceleration track falls by 1.2 m/s when a 5 ton weight passes over it. The passage of payloads over the track make it look like they're surfing a giant wave. Lofstrom proposes to release counterweights, timed exactly, to match that push and prevent the track from actually moving. 5% of the track mass would be dedicated to counterweights fired by solenoid coils. Simultaneously, the entire launch loop must release tension at the deflection stations to allow the track to stretch appropriately.  A 5 ton payload accelerated at 3g down the whole 2000 km length of the acceleration track removes 735 GJ of kinetic energy from the iron rotor, slowing it down by 3.6 m/s at the start of acceleration and 14.3 m/s by the end. Now, the payload design. We have 5 tons to play with. Lofstrom places a pressurized passenger capsule on top of the 10 meter long magnet rack. Let’s call this the Space Train. In case of a launch abort or other failure, the Space Train simply falls off the acceleration track. It may be travelling at extreme velocity, so it needs a heatshield. Also, wings and a parachute to land on the ground (or more likely, splash down). In the back, a rocket engine is necessary to circularize the Space Train’s orbit with an apogee kick. The deltaV required to enter a circular Low Earth Orbit at 300 km altitude, after leaving the Launch Loop, is a mere 65 m/s. So the apogee kick engine can be very small.  The Space Train can also have a ‘cargo’ configuration. This may hold 2-3 tons of useful cargo if it needs to be recovered, or nearly the whole 5 tons if it’s just given an unpressurized aerodynamic shell on top of the magnet rack.  Launch performance is adjustable. Use the full or partial length of the track, use the full or partial strength of the magnet rack, to reach the desired final velocity. At full 3g acceleration, the Space Train can reach 7450 m/s after accelerating for 943 km. That same velocity can be reached by using the full 2000 km of the track at a milder 1.41g. It’s enough to send the payload to a near-Low Earth Orbit trajectory with 300 km altitude.  At 2.49g and using the full 2000 km, the payload can be accelerated to 9875 m/s. That is enough to reach a Geostationary orbit. The apogee kick deltaV necessary to stay in GEO increases to 1490 m/s.  The maximum performance is 10,547 m/s. Passengers would have to experience an uncomfortable 3g for six minutes and a half. But this puts the Space Train on a trajectory to the Moon! DeltaV to circularize is 833 m/s.   Here is a performance table: M288 is a hypothetical space station in a High Earth Orbit. Note the option to shoot off payloads at 10g. This is not for passengers, but for insensitive stuff like food or rocket fuel.  The main advantage of the Launch Loop is its ability to send useful payloads into any orbit around the Earth or Moon using only electricity. Higher launch rates make the process more efficient. At the 500 MW scale, 40% of the electricity is spent maintaining the rotor instead of accelerating payloads. At 17 GW, that load falls to 1.2%. A launch to Low Earth Orbit (7451 m/s) adds 27.8 MJ of kinetic energy to each kilogram of the Space Train. At the 500 MW scale, 180 MJ of electricity is consumed on average for each kg launched. So the efficiency of the launch loop is roughly 15%. A 10.5 km/s launch to the Moon represents 55.1 MJ/kg of kinetic energy. At the 17 GW scale, the launch loop spends 153 MJ for each kg sent to the Moon. Its efficiency is a much better 36%.  As of February 2023, the average residential electricity rate in the U.S. is about 23 cents per kilowatt-hour. That translates to 6.39 cents per MJ. That means a Launch Loop can expect an energy cost of $11.5/kg at 500 MW, to $3.5/kg at 17 GW. That's much lower than current rocket launch prices exceeding $2350/kg for a Falcon 9 Heavy, or even the optimistic $100/kg goal for the SpaceX Starship. The differences become even more stark when comparing between a Launch loop and a rocket headed beyond LEO.  Of course, these figures do not include the cost of delivering electricity to some remote structure out over the ocean, other running costs like maintenance, or the overheads from paying back the construction of the whole thing. Still, they are a useful reference for how cheap a launch loop could make space travel. For the environmentally conscious, the fact that all this electricity can be generated by solar panels or nuclear reactors on the ground instead of burning rocket fuels in the upper atmosphere is a major benefit.  Getting Loopy Lofstrom’s design for a Launch Loop is a massive structure of over 2000 km in length that needs to be placed near the equator. It has two slopes rising to 80 km altitude that need a large flight exclusion zone around them to prevent run-ins with aircraft. They also make it vulnerable to strong weather, like a hurricane. Preferably, it would be installed somewhere that doesn’t have to face such winds frequently.   Here is a map of suitable locations: The yellow bars show how large the Loop looks on a map. The squiggly tracks are the paths taken by violent storms. All the locations are intentionally far from inhabited areas. That’s because if a Launch Loop fails, it can fail spectacularly.  The structure can be severed by a collision. A payload crashing into the acceleration track, an airplane crashing into its slopes, a meteor striking the wrong spot, a ship running into the deflection stations… these scenarios can be made less likely but never completely ruled out. What happens afterwards depends on where the loop is cut. A cut in the ascending slope would disgorge the iron rotor at 14 km/s. It is travelling much faster than orbital velocity, in both forward and return directions. The iron would therefore sail out into space, escaping Earth’s gravity entirely, only catching the unlucky satellites passing overhead. The hollow vacuum tube that held the iron rotor would start to fall downstream of the cut. Since it would only mass around 5 kg per meter, it drifts to the surface harmlessly. A cut in the high-altitude acceleration track would have a similar outcome. The structure might recover from such an event, by rebuilding the cut section, inserting a new iron rotor and raising it back into place. A cut in the descending slope is more dangerous. It would point the 14 km/s iron rotor down at the ground. This turns it into a gigantic hypervelocity shotgun that continues firing for around five minutes. A zone extending for hundreds of kilometers east of the cut could be riddled with shrapnel. The deflection station at the foot of the slope might also get hit. That would make recovery much more difficult.  A very unlikely but still possible event is a cut in the short section between the foot of the slopes and the deflection station would also result in a hypervelocity shotgun, but the damage would be contained to a relatively small area near the surface. 1500 TJ of kinetic energy would be released at that spot, deposited at a rate of 5 TW, so it would resemble a miniature nuke.  Not all damage causes a catastrophic failure.  Despite its impressive scale, the acceleration track is very thin and so narrow it would not be visible to the naked eye most of the time. That makes hitting it with anything rather unlikely. And if it does get hit, chances are the strike will create a hole in the vacuum tube rather than a full cut. The hole will fill with air. If it is a small hole or an opening at high altitude, the vacuum pumps will compensate for the inflow of air and maintain a good level of vacuum until repairs arrive. A large hole, or an opening created in the thicker air at low altitude, will fill with air quickly. A 14 km/s iron rotor encountering air will start to experience drag, which doesn’t slow it down by much, but does create heat. If the rotor heats up to 1000 K, the Curie temperature of iron, it will lose its magnetic properties and crash into the tube walls. Thankfully, this takes a very long time to happen and a sea-level air breach would not be ignored for long.  This brings us to thermal limits. The biggest cause of iron rotor heating is the launch of a payload. A full 2000 km launch with 3g acceleration and 5 ton payload raises the rotor temperature by 84 Kelvin. At 900 K, the rotor radiates away enough heat through blackbody radiation (the only way to lose heat in a vacuum tube) to accommodate 80 successive launches of 5 ton payloads per hour. The permanent magnets and other structures surrounding the rotor would have to be shielded from this heat to enable such a high launch rate.  Smaller payloads cause a smaller temperature rise. The thermal limits create two additional failure modes: magnetic failure in the iron rotor from a heat spike above 1000 K, or overheating if the launch rate exceeds the sheath’s cooling capabilities, disabling the magnets and breaking the loop.  Beyond Lofstrom A depiction of a deflection station, by Katie Byrne for this video. Some aspects of Lofstrom’s design proposal can be modified. Larger Space Trains launched less frequently seem more practical than a handful of passengers accelerated every few minutes. The use of loaded counter-weights seems unnecessary when the stabilization cables could be tensioned for the same effect. High temperature superconductors to replace some or all of the permanent magnets would be an expensive option today, but would be a plausible option by the time we actually consider building a Launch Loop. The vertical elevator that places a Space Train atop the acceleration track seems dangerous and impractical when the Space Train could climb the West slope instead. As the incline is about 310 km long, and the majority of it is at stratospheric height or higher, so a Space Train could climb it at supersonic speed in less than 10 minutes. It would also be reasonable to install an actual set of rails to support the Space Train when it is not being dragged along on its magnet rack. A launch loop with a more substantial track structure. They can be relied upon at low speeds, or the Space Train can slow down to rest on them if the launch is aborted - Lofstrom’s design has ditching into the ocean as the only other option.  We can further expand on Lofstrom’s design.  A shorter track can be envisaged. Maybe because LEO is the only destination you want, an acceleration track of around 950 km long is suitable. Perhaps because the Launch Loop is something new and not yet rated for human spaceflight, meaning it is reserved for cargo launch until it has proven its safety. That cargo could survive higher accelerations on the shorter track. At 10g, a track length of only 283 km is needed to reach LEO, 567 km to reach the Moon.  Perhaps your destination is not somewhere beyond Earth. A short dynamic structure could serve as a launcher for intercontinental flights. London to JFK in 1 hour at Mach 5. It doesn’t even have to reach 80 km altitude. Merely pushing a passenger plane to 10-20 km altitude is enough to eliminate sonic boom issues entirely. It could obtain almost all the speed it needs from the launch loop, eliminating the need for complex and heavy hypersonic engines, instead covering most of the flight distance by gliding at high average velocity. The only propulsion needed is landing or divert engines, which can be small, subsonic and loaded with a minimum amount of fuel. If a hypersonic glider is too much to consider, then a launch loop placing a merely supersonic plane directly into its most efficient cruise speed and altitude is already a huge benefit. Imagine a Concorde, but affordable and without sonic booms.  The low speed takeoff, the climb to altitude and the high thrust acceleration to supersonic flight all cost a significant load of fuel, which has knock-on effects on the size of the wings needed to lift that fuel, the unnecessary drag incurred during the rest of the flight and so on. All that could be skipped. Furthermore, passenger flights are much more frequent than space launches, which works to the launch loop’s favour as it only becomes better with higher launch rates. The location of a plane-launching loop is flexible as the plane can maneuver to its destination and lands on its own. The small low-altitude dynamic structure wouldn’t need a 14 km/s iron rotor inside, but a much slower rotor that doesn’t have a ‘hypervelocity shotgun’ failure mode. It may be safe enough to place near cities… Or, you could imagine a longer track. One which lets a payload reach velocities that directly place it on a trajectory to Mars, Venus or beyond. It might need to deliver 15.5 km/s to its payloads, so it’d have to be 4081 km long if limited to 3g acceleration. A 4000 km launch loop anchored between two Atlantic islands.  A multi-interplanetary launch loop would need at least one end to be mobile so that it may align itself with the target planet’s inclination; a job perhaps for a floating platform that sinks itself into place before launch.  Size of a 4000 km launch loop stationed at Cape Canaveral. The track could be extended even further to shorten the trips to other planets. The only limit would be slowing down at the destination. Aerobraking can do wonders in this regard.  Alternatively, we could exploit the launch loop’s ability to shoot off many payloads in a short span of time before it reaches its thermal limits. Lofstrom’s design could manage 15 x 5 tons in 6 minutes, followed by another 15 x 5 tons after half an hour of cooling. In two hours, that’s a total of 300 tons. A thicker rotor with greater heat capacity could manage even more. These many small payloads would all be travelling in roughly the same direction with small differences in speed and a separation between them of a few hundred kilometers. Spacecraft launched in multiple pieces, connecting in space, is a often-proposed concept. But here, it would be done while already underway to Mars.  They would have weeks or months to group up, connect and form a single large spaceship. That large spaceship would then be used to brake into the destination.  For example, a futuristic 5500 km Mars Launch Loop could send off a hundred 5 ton payloads in a short burst. They leave the loop at 18 km/s, putting them on a 45 day trajectory to Mars. Loop-launched craft aimed at Mars could slow down with intense aerobraking, or if going really fast, with powerful rocket engines.  The majority of those payloads are blocks of propellant, the rest are 5 ton nuclear engines, 5 ton crew compartments and so on. Assembled, they form a 500 ton nuclear spaceship. The rocket engines could have a specific impulse of 1200s. So, with a mass ratio of 5 (100 tons dry, 400 tons of propellant), the onboard deltaV capacity is 18.9 km/s. This is enough to slow down into a Mars orbit without aerobraking.   By the way, those propellant blocks could also be fusion fuel pellets, making the Launch Loop an excellent way to set up a Fusion Highway.  The Launch Loop would replace the solar sails setting up the pellet 'highways'. And the Launch Loop is not restricted to Earth. It could be installed on the Moon, Mars or anywhere else. A launch loop anchored to a distant planet. Illustration from a page of JayRock's Runaway to the Stars. An extraplanetary launch loop wouldn’t even have to rise to a higher altitude if it is installed on an airless body. A Lunar Launch Loop could, for example, sit flat on the ground. This makes the launch loop the equivalent of a gentle, easily scalable mass driver with less extreme pulsed power requirements than a railgun or coilgun.  A launch loop delivering a payload to a skyhook. Also a JayRock piece.  It may even be paired up with an orbital tether, so it would not need circularization engines. The tether could handle inclination changes, so the launch loop facility would not need to move itself to reach different destinations. In the far future, if Launch Loop technology is accepted despite its faults and developed into a mature, reliable service, then we could have one loop launch a Space Train perfectly onto the track of another launch loop. The Space Train would grip onto the returning iron rotor to slow down at the destination. The entire journey could be completed without using a drop of rocket fuel, at incredible speed. It may or may not be more difficult than lining up a rendezvous with the end of a rotating tether.

Reset your expectations of solar sails. They are a fast and free way to travel to any point in the Solar System, as many times as you want, any time of the year. Solar sails can carry passengers and they have a nearly unlimited number of uses. You just have to... think big! Solar Sails Art by Stepan Polivanov. Sunlight is abundant. The energy needs of any activity can be met by simply collecting enough of it. All of human civilization, for example, could be powered by the perfectly using the sunlight passing through an 86 km square near Earth. There’s more than enough sunlight to power space travel. How convenient that we can just harvest what we need to get anywhere in the solar system! The usual approach is to use photovoltaic devices. They convert sunlight into electricity, which can then power electrical devices directly. They’ve had a long history: the first solar panels in space were launched with Vanguard 1 in 1958. Those panels produced 1 Watt in total. Today, the largest solar-powered installation in orbit is the International Space Station. Its eight arrays could generate up to 240 kW when new. Accumulated damage from micrometeorites and radiation will necessitate their gradual replacement though. The latest development is the large-scale use of electric thrusters. The DART mission used the NEXT-C thruster at a 3.5 kW power level, and thousands of Hall-effect thrusters are deployed with the SpaceX Starlink constellation. However, solar panels are not the only way to make use of solar power. Solar sails directly convert sunlight into thrust. Their efficiency usually exceeds 90% and they’re very resistant to degradation. A spacecraft equipped with these sails can accelerate endlessly without requiring a drop of propellant. How is this possible? How much acceleration can be achieved? How far can we push this technology? Let’s dive into these questions in the following sections. Light and Speed  Space sailing! By Jett Furr. The way a solar sail works is pretty simple. Light has no mass but it has momentum. When it strikes a surface, it generates radiation pressure proportionate to its intensity. If the surface is reflective, that pressure can nearly be doubled.  An equation describes radiation pressure: Radiation Pressure = Light Intensity x (1 + Reflectivity) / C Radiation Pressure is in Pascals (Pa). Light Intensity is in W/m^2. Reflectivity is a figure between 0 and 1. C is the speed of light in m/s. In the vacuum near Earth’s orbit, the intensity of sunlight is 1361 W/m^2. A surface with a reflectivity of 0.9 (90% of sunlight is reflected) would feel a radiation pressure of about 1361 x (1 + 0.9) / (3 x 10^8) = 8.6 x 10^-6 Pa or 8.6 microPascals.  You can convert radiation pressure into thrust if you know the area of the reflective surface. Radiation Thrust = Radiation Pressure x Area Radiation Thrust is in Newtons (N). Radiation Pressure is in Pascals (Pa). Area is in m^2.  A square solar sail that’s 10m on each side would have an area of 100 m^2. Following the previous example, it is feeling a pressure of 8.6 microPascals. The total thrust it generates is 8.6 x 10^-6 x 100 = 8.6 x 10^-4 or 0.86 milliNewtons. That’s not a lot of thrust; it’s equivalent to the weight of 3 grains of rice.  From thrust, we can work out acceleration if we know the mass of the sail. The mass of a sail is related to its area: usually a sail with 10x the area is also 10x heavier. We’ll therefore use the area density. Solar Sail Acceleration = Radiation Thrust / (Area Density x Area) Solar Sail Acceleration is in m/s^2. Radiation Thrust is in Newtons (N). Area Density is in kg/m^2. Area is in m^2. A typical area density for today’s solar sails is 10 grams per square meter, or 0.01 kg/m^2, as found on the IKAROS sail. A 100 m^2 sail would therefore mass 1 kg. The acceleration we can expect from 0.86 milliNewtons of thrust is 0.00086 m/s^2, or 0.086 milligee. It would take this sail nearly two minutes to gain 1 m/s. In more familiar terms: its “0-100 km/h” time is 9 hours! If we combine all the previous equations, we can write: Sail Acceleration = Light Intensity x (1 + Reflectivity) / (C x Area Density) Notice how ‘Area’ disappears from the equation. A bigger sail will not accelerate faster. Improving reflectivity does not affect acceleration much as the difference between a good 90% reflectivity and an incredible 99.999% reflectivity is small. That leaves two factors that really affect acceleration: Light Intensity and Area Density. Light Intensity varies between 64 MW/m^2 at the surface of the Sun to 0.873 W/m^2 near Pluto. Increasing that figure means getting closer to the Sun, but that might not always be where you’re headed. Area Density is the only contributor to sail acceleration that we can really affect.  In the example we used, we ended up with a solar sail that accelerates about four times slower than a snail because it only had access to Earth-levels of sunlight and had a 0.01 kg/m^2 mass per area. If we attach a payload to that sail, it will accelerate even more slowly. Angles relevant to a solar sail. What’s more, conventional solar sails can only reflect sunlight back in the direction it came from or off to the side. For a spacecraft trying to spiral out of its current orbit then into a new one, this is a problem. It will have to angle its solar sail to produce sideways thrust to increase its orbital velocity. The efficient angle is 35° to the Sun. At this angle, the sail’s ‘effective thrust’ is actually 82% of maximum as it is not fully facing the Sun. The result is a space vehicle that takes months to years to accomplish the most minor of maneuvers, let alone interplanetary transfers that require deltaV changes in the thousands of m/s. This is currently acceptable for small probes that can fit inside rocket upper stages as secondary payloads.  But can we do better? Less is More Lightsail 2's deployment. The material that today’s solar sails are made of is aluminized Mylar supported by extensible trusses. It is only a few micrometers thick (7.5 micrometers in the case of IKAROS) and manages to reflect 90% of sunlight. 2 micrometer thick aluminized PEN being spooled and handled manually The main struggle is getting a reflective surface that can be packaged like origami into a very small volume, survive the vibrations of a rocket launch, then reliably deploy out in space into a rigid structure. Solar sail deployment Also, since solar sails so far have been of modest size, the mass of secondary equipment like control actuators or deployment mechanisms take up a proportionally larger fraction of total mass. Simply making larger sails will improve their area density. They also need features to prevent tears from ripping through the entire sail, or to keep edges of the sail material from curling inwards.  The first step to improving area density is to take current materials and improve them. The DLR/ESA 60m sail is an example of a realistic near-term goal. Reducing Mylar sheet thickness to just 3 micrometers could reduce area density to 4.2 g/m^2. Add in the mass of the aluminium coating and the graphite booms and you achieve 5.3 g/m^2. The Encounter sail is even larger, and projected Mylar sheets of just 1 micrometer thickness, for an overall area density of 1.9 g/m^2.  Another avenue to reduce area density is to cut away solid support structures in favour of other ways to keep the sail to its intended shape. Inflatable booms can be lighter than graphite rods. Rotating sails can be lighter still - they can use centripetal forces to both extend the sail and to keep it under tension. These result in circular sails, or even ‘heliogyros’ where the sail is divided into ‘blades’ kept under tension like a helicopter’s rotor. Controlling their motion is more complicated though. A heliogyro sail catching up to Halley's Comet. If we put together the thinner materials and innovative configurations, we are likely to achieve sails with 1 g/m^2 area density. This would grant them accelerations on the order of 0.86 milligee, likely to be 0.5 milligee with payload included. That’s enough to get a spacecraft from geostationary orbit to Earth escape in about a week. But we can do even better. A better material than Mylar can be considered. Carbon fibers supporting an aluminium layer can achieve 0.5 g/m^2 or better. At some point, we can imagine removing the supporting layer entirely and relying instead on the metal reflector alone. This layer can be very thin: as low as 0.14 micrometres in this proposed example. It is a way to achieve sails with an area density of 0.33 g/m^2, rising to 0.45 g/m^2 when including support booms.    Nanotechnology can come to our aid.  Sub-wavelength metasurfaces allow solar sails to remain reflective while having gaps smaller than the wavelengths of sunlight. It is similar to how the metal screen in your microwave oven’s door reflects the 2.45 GHz (12.2 cm) wavelengths inside but lets visual wavelengths (0.4 to 0.7 micrometers) through. A solar sail can use the same concept to become mostly empty space while remaining reflective. The shortest wavelength we want to capture would be 0.4 micrometers, so we could use a quarter-wavelength grid spacing of 0.1 micrometers (so it would be 75% empty). That would reduce the sail mass four-fold, down to 0.1 g/m^2.  More advanced sail configurations. Theoretically, we can shave down the reflective surface down to the absolute minimum thickness that can interact with sunlight. This seems to be 10 nanometers for Aluminium. A sail of that thickness would have an area density of 0.027 g/m^2. If we further apply the sub-wavelength gaps, we could bring it down to the level of 0.01 g/m^2. That would enable accelerations on the order of 87 milligee. But it comes with many downsides. This sort of ultra-thin surface is very weak (tensile strength of 0.225 Pa) so while it would survive the radiation pressure of the Sun, it might not handle any other sort of disturbance. It would also become vulnerable to the ambient space environment: solar wind, interplanetary dust, radiation damage and so on. Radiation damage to solar sail surfaces. Better performance can be achieved if we abandon reflective sails and use diffraction instead. The many ways a diffractive sail can manipulate light. Diffractive sails let sunlight pass through them while bending the angle at which it emerges, which produces a force at the opposite angle. If they bend sunlight to the right, they feel recoil from the left. Diffraction gratings made with very thin transparent materials like silica, or reflective strips made of aluminium spaced by about a micrometer, can do the job. There are few details on how such a grating could be constructed, but they are often directly compared to conventional sails in terms of area density.   Diffractive sail thrust. The most interesting feature is that control of their optical properties (switching the angle of the grating, for example) allows them to redirect that sunlight in any direction, and therefore control the direction of their thrust. Conventional sails can only accelerate away from the Sun or at an angle away from it. Refractive sails would work more like an engine in that they can choose with direction to accelerate. They can do this with over 90% of the energy they gather, as described here. This allows the sail to always point at the Sun for maximum efficiency while thrusting sideways or even towards it. Advanced control would even allow them to discard control thrusters or control gyroscopes for further mass savings. A 'photon thruster' sail. Another interesting configuration for solar sails is the ‘photon thruster’. It’s essentially a set of two mirrors working together: one large collector and a smaller mobile mirror to redirect the light in any direction. The advantages are a much wider range of maneuvers that become possible, and since the collector mirror can always face the Sun directly, there is no loss of efficiency from angling the thrust. The downside is increased complexity and the additional weight of the mobile mirror. There are even more speculative types of solar sail, but their applications are limited.  ‘Black sails’, for example, absorb sunlight. They do not reflect or redirect the sunlight, so they only accelerate away from the Sun (try using the equations above with R=0). It’s useful for well timed exploration of the outer solar system or interstellar missions, but not much else. The upside is that they can use the thinnest and strongest materials available: carbon nanomaterials. Their very low density works in their favour. Aerographite, for example, has a density of just 0.18 kg/m^3. A 1 mm layer of it is opaque and could create a solar sail with an area density of 0.18 g/m^2. It is unclear whether it remains opaque at 0.1 mm thickness, which translates to 0.018 g/m^2. The potential acceleration would be as high as 0.25 m/s^2. A 3mm wide graphene light sail under testing At the material limit, we have a single layer of graphene. It’s just one atom thick (0.345 nanometers) and its area density would be 0.00077 g/m^2. However, pure graphene only absorbs 2% of sunlight while letting the rest through. That limits its acceleration to 0.12 m/s^2 near Earth. It has been suggested that adding a layer of Molybdenum disulfide (MoS2) can increase the absorption to 37% of sunlight. The area density of a graphene/MoS2 bilayer would increase to 0.004 g/m^2 allowing its potential acceleration to become 0.4 m/s^2. It’s also possible to create black sails with area density below the material limit, by using dusty plasmas. These sails would magnetically confine a plasma that can hold charged particles within itself. Those particles would absorb sunlight. Since plasma sails can grow to immense sizes of several tens of kilometers in diameter, with very little mass within them, they effectively constitute solar sails with area density as little as 0.001 g/m^2 or less. Few studies have focused on dusty plasmas or their characteristics, so there is no data on their sunlight-absorbing properties. We therefore can’t guess at their potential acceleration. The Sails Let’s design four solar sails to use as a reference for the rest of this post. They’re meant for increasing payloads and more demanding missions with each technology level. Note that we’re adding 20% to the area density to account for all the other equipment a sailcraft needs to navigate. Characteristic Acceleration is the acceleration the sail manages when receiving full illumination at Earth sunlight levels (1361 W/m^2).  Modern Sail Area Density: 10 g/m^2 Sunlight Interaction: 190% Area: 100 m^2 Sail mass: 1.2 kg Payload: 1 kg Total Mass: 2.2 kg Characteristic Acceleration: 0.00039 m/s^2 This Modern Sail is of the type we have already launched into space. 190% efficiency is thanks to 90% reflection nearly doubling the momentum it gains from sunlight. It’s made of Aluminized Mylar several micrometers thick, held up by 10 m wide square rigging, propelling a microsat-sized payload. Due to its tiny total mass, several of these can be sent into orbit by the smallest of launchers. Advanced Heliogyro Sail Area Density: 2 g/m^2 Sunlight Interaction: 190% efficiency thanks to reflection Area: 50,000 m^2 Sail mass: 120 kg Payload: 100 kg Total Mass: 220 kg Characteristic Acceleration: 0.0019 m/s^2 This sail is made of six blades rotating around a central truss. Thanks to its Mylar layer being thinned to 1 micrometer, its area density is much improved over the Modern Sail. It’s small enough to fit into the secondary payloads of a regular rocket like a Falcon 9 on a geostationary transfer orbit mission.  Nanofilm Sail Area Density: 0.1 g/m^2 Sunlight Interaction: 180 % Area: 4,000,000 m^2 Sail mass: 480 kg Payload: 1000 kg Total Mass: 1480 kg Characteristic Acceleration: 0.022 m/s^2 This sail uses a reflective surface that’s just a tenth of a micrometer thick, further lightened with sub-wavelength holes. Carbon nanotube struts hold it up into a 2000 m wide square. Since some shorter wavelengths pass through its holes, it has a reduced sunlight interaction efficiency. At this scale, it is matching the useful payloads of deep space probes while requiring only a fraction of the initial mass. Launching it up from the ground is a dubious proposition though. Diffractive Meta-Sail Area Density: 0.01 g/m^2 Sunlight Interaction: 90% Area: 2,500,000,000 m^2 Sail mass: 30,000 kg Payload: 10,000 kg Total Mass: 40,000 kg Characteristic Acceleration: 0.26 m/s^2 This sail is 50 km wide. Despite its immense size, it’s still only the equivalent of the payload of a single heavy lift rocket today. The sail surface is a nanostructured diffraction grating about 10 nanometers thick, actively controlled by piezoelectric actuators to be able to give complex shapes to the sunlight that passes through it. Its sunlight interaction efficiency is 90% as it does not reflect the light but merely redirects it. This is a sail large enough to transport humans on interplanetary missions. However, it is certainly too fragile to be launched atop a rocket or even folded before deployment. The whole structure would have to be assembled in orbit.  Solar trajectories The most important aspect of solar sail travel is the sail’s acceleration relative to the local gravity.  A sail with very low acceleration spiralling out of LEO to escape velocity In Low Earth Orbit, local gravity is quite similar to the 1g found on the surface. Solar sail acceleration would be much much smaller than the local gravity. The Modern Sail would have an acceleration of 0.039 milligees, for example.  In this environment, solar sails can only make very gradual spiralling trajectories as they attempt to raise or lower their orbit. A useful approximation for how long this takes is to divide the difference between the initial and destination orbits’ velocities by the sail’s acceleration.  Let’s try it. The orbital velocity at 1000 km altitude is 7350 m/s. The orbit velocity at 35,800 km (geostationary) altitude is 3075 m/s. The difference is 4275 m/s. The Modern Sail would take 4275/0.00039 = 10,961,538 seconds or about 4.16 months to complete the maneuver. Realistically, it will take a while longer as we have not accounted for the time spent in Earth’s shadow, the inefficiencies from accelerating sideways to the sunlight and the effects of atmospheric drag, which can be significant even far above the atmosphere. The Advanced Heliogyro Sail with its 0.0019 m/s^2 acceleration would do it in 26 days.  This approximation holds true so long as sail acceleration is far below local gravity.  In the following table, epsilon E is the ratio between the spacecraft’s acceleration and local gravity. The third column is the ratio between the deltaV needed to reach escape velocity, and the initial orbital velocity. And in the fourth column, we have the number of orbits needed to complete the maneuver.  When E is very low (10^-5), we have to expend practically as much deltaV as our current orbital velocity to escape into deep space. In this specific example, your deltaV requirement is 95.75% of your current orbital velocity. It also takes an incredible number of orbits to complete this sort of maneuver (over 50,000!). As E increases, you need less deltaV and less orbits to reach escape velocity. When E is a hundredth (10^-2) of local gravity, you only need to spend 76% of current orbital velocity to escape and it takes just 51 orbits to do this. While deltaV savings don’t mean much to a solar sail, they do translate into additional time savings.  However, you will notice that solar sails with the performance levels we find interesting will achieve accelerations comparable to or exceeding local gravity. Consider what happens as we move away from Earth. At an altitude of 100,000 km, the local gravity is 0.035 m/s^2. At 400,000 km, or 1 lunar distance, it is only 0.0024 m/s^2. Compare these values to the accelerations that advanced solar sail designs can achieve. The Advanced Heliogyro Sail manages 0.79 times the local gravity at 1 lunar distance. That’s a ‘thrust to weight’ ratio we typically find on fighter jets! In other words, high performance sails can pull off acrobatic maneuvers, albeit in slow motion from our perspective. Throughout the rest of the solar system, we have this useful chart: At 1 AU from the Sun, solar gravity is 0.0059 m/s^2. At 5.2 AU (Jupiter’s orbit), it falls to 0.00022 m/s^2. Note that the Sun’s gravity and the intensity of sunlight follow the same inverse square scaling with distance (D^-2). Intensity of sunlight determines a sail’s acceleration. Therefore, a solar sail will always maintain the same ratio of acceleration to gravity, regardless of how far it is from the Sun. We call this the sail’s beta B or ‘lightness ratio’. For example, if a sail manages an acceleration of 0.001 m/s^2 at 1 AU, which is six times lower than solar gravity, then it will achieve 0.000037 m/s^2 near Jupiter, which is also six times lower than solar gravity over there.  Small lightness ratios leads to very long travel times. A ratio of 0.015, which is an acceleration of around 0.00008 m/s^2 near Earth, would mean an annoying long transfer time of 1647 days to Mars. That’s 4.5 years. Increasing the lightness ratio dramatically shortens travel times. It makes solar sails act more like powerful rockets. A lightness ratio of 0.1 is enough to take on fast trajectories to Mars. 176 days is about 6 months, which is shorter than the usual 9 month Hohmann transfer chemical rockets can manage. And unlike regular rockets, a solar sail can make its way from Earth to Mars or back without having to wait for any transfer windows. Another study shows that a Mars mission can take less than 120 days using sails with a lightness ratio as low as 0.2. However, that involved meeting Mars with a velocity of 5-12 km/s. An Advanced Heliogyro sail would make up that velocity gap in 70-170 days. The other option, as the study suggests, is aerocapture into Mars orbit.  We can reasonably assume that sails in that performance category can complete trips to the Outer Planets at least as fast as minimum energy Hohmann transfers: 2.7 years to Jupiter, 6 years to Saturn. Solar sails are relatively quick even with low performance levels. But if we increase performance to a lightness ratio of 0.5, we would enable a solar sail to directly move away from the Sun and accelerate onto a solar escape trajectory that eventually resembles a straight line into interstellar space.  We have this trajectory for a sail with a lightness ratio of 0.6, where it takes 2.8 years to reach Jupiter: The same paper finds an 11.5 year trajectory to Saturn with a lightness ratio of 0.5: And we know solar sails can do even better. At the upper end of their performance, they will act like advanced propulsion systems, such as nuclear-electric thrusters or fusion rockets. Their lightness ratio exceeds 1, so they can effectively ‘cancel out’ the Sun’s gravity and pick up straight line trajectories to wherever they want. Metaphorically, they’re unleashed to go wherever they please.  The trajectories possible with different lightness ratios. If a solar sail with a lightness ratio of exactly one exits Earth’ sphere of influence, it will start to travel in a straight line with Earth’s orbital velocity and tangent to Earth’s orbit at that point. If it is timed correctly, it will drift at 29.78 km/s straight towards its target, whether it be Mars, Jupiter or beyond. That is enough to reach those planets in 88 days and 243 days respectively. With a lightness ratio exceeding one, a solar sail can take that initial orbital velocity and add to it. It can shorten the trip by many months. The Nanofilm Sail has a lightness ratio of 3.73. It can keep accelerating towards a target like Jupiter, crossing the 4.2 AU minimum distance to that planet in 103 days. At the end of the trip, it is travelling at 85.8 km/s. The Diffractive Meta-Sail has a lightness ratio of 44. It completes that same trip to Jupiter in 37 days, reaching a velocity of 253.1 km/s by the end of it. Those same sails can reach Saturn in 203 days and 66 days respectively. A solar sail cannot accelerate into the Sun. Of course, just zipping past your destination isn’t always desirable. A sail can just turn itself edge-on to the Sun to stop accelerating away from it and let gravity take over. Solar gravity is however very weak. The sail performance doesn’t really matter in that case: the trip always takes about 630 days (1.72 years). For Saturn it is 1660 days (4.5 years). Caution! These are only rough calculations that have to be taken with a grain of salt, as they are basically assuming the sail makes a straight line for its target and stops at the end with near-zero velocity. A realistic trajectory, such as one that tries to reduce the relative velocity during the encounter with Jupiter (13 km/s) or Saturn (9.7 km/s) will be different and might take a bit longer. Still, they are much faster than minimum energy Hohmann transfers, and they can be launched any time during half the year where Earth and the target planet are on the same side of the solar system. Now, there are tricks to make those trips shorter.  From "Multiple-satellite-aided capture trajectories at Jupiter using the Laplace resonance" Jupiter and Saturn have large moons that can help capture a solar sail into an orbit around the gas giants. Gravity assists can change the velocity of a passing spacecraft by a maximum of twice the planet or moon’s orbital velocity. The Moon, for example, orbits Earth at about 1 km/s, so theoretically it could grant a velocity change of up to 2 km/s. Lunar Gravity Assists have been calculated to provide 1.5 km/s of deltaV savings for capturing asteroids into Earth orbit from a single pass. Earth's Moon used for gravity assists and slingshots. We can therefore expect proportional deltaV savings from other large moons. Jupiter has Io, which orbits at 17.3 km/s. A Nanofilm Sail that only has to slow down to 17.3 km/s before reaching Jupiter could shorten its trip time to 312 days (-50%). If Enceladus and its 12.6 km/s orbital velocity can be used to capture into Saturn, then the Nanofilm Sail’s trip is reduced by 51% to 814 days. Again, gravity assists are much more complicated than this but we do get a rough estimate for what sort of benefits we get from them. A solar sail at Saturn Returning from the gas giants is not easy. It takes a very long time to maneuver far from the Sun, and even in the best case afterwards, only solar gravity can pull the sails inwards. For example, the Nanofilm sail has an acceleration of 0.000813 m/s^2 near Jupiter. It would take it around 185 days to cancel out Jupiter’s orbit velocity and start falling straight into the Sun. From a distance of 5.2 AU, it would ‘fall’ for 740 days until it is at a 1 AU distance from the Sun. Intercepting Earth is done near the Sun, using maneuvers that take up to 25 additional days. The total mission time is therefore 950 days or 2.6 years. That’s only slightly faster than a Hohmann transfer. Using a better sail can shorten the 185 day and 25 day portions of the trip, but does nothing to the 740 day freefall period. Inner solar system missions don’t have to cover as much distance as missions to the gas giants, but they do have to contend with targets that have much higher orbital velocities: Venus at 35 km/s and Mercury at 47 km/s. Neither of them has a useful moon for gravity assists either. Capture into those planetary systems is therefore quite challenging. A Mercury sample return mission. For low performance solar sails, we expect to see slow spiralling trajectories. Thankfully, their acceleration increases as they get closer to the Sun. On the way to Venus, a solar sail’s acceleration nearly doubles. The average acceleration is 145% of its characteristic acceleration. Near Mercury, a solar sail is zipping around at 6.7 times its characteristic acceleration. A sail’s average acceleration is 383% its Earth-normal acceleration on the way to Mercury.  We have here data on Venus transfers with a sail of characteristic acceleration of 0.1, 1 or 2 mm/s^2. That’s lightness ratios of 0.017, 0.17 and 0.34. We see that increasing performance brings the transfer time closer to the Hohmann minimum-energy trajectory that takes 146 days to Venus. Mercury remains a challenging target for low performance sails. One way around that is a flyby of Venus to shorten the travel time to the innermost planet. We can see significant benefits for sails with lightness ratio below 0.1. Venus flyby on the way to Mercury. Another paper studies interplanetary trajectories for a diffractive sail. The range of characteristic accelerations is 0.1 to 1 mm/s^2. Surprisingly, it takes about the same amount of time to reach Venus as it does Mercury; about 200 days.  To Venus: To Mercury: High performance sails with excellent lightness ratios have similar troubles with the inner planets as they do returning from the outer planets. It’s the ‘cannot fall inwards faster than gravity’ problem. Even if they cancel out all of their orbital velocity, they can only be pulled inwards by the relatively weak solar gravity.  At Earth’s orbit of 1 AU, solar gravity is 0.0059 m/s^2. Near Venus’ 0.72 AU, it is 0.0113 m/s^2. Mercury is the closest to the Sun at 0.39 AU and experiences 0.0387 m/s^2 of solar gravity.  We can use a table to calculate the position of an object falling into the Sun every 1 hour. It should take about 42 days to fall from Earth to Venus, accumulating about 26 km/s of velocity by the end of it. Falling from Earth to mercury takes 57 days, reaching 52.5 km/s. Falling all the way into the Sun takes 65 days, which matches theoretical results. The direct trajectory that gives the minimum possible travel time from Earth to Venus or Mercury for a high performance solar sail is the following: cancel out Earth’s orbital velocity (29.78 km/s), fall to the desired altitude, cancel out the accumulated fall velocity, accelerate to the target’s orbital velocity.  For a Nanofilm sail travelling to Venus, the Earth maneuver takes 15.7 days. Falling towards Venus takes 42 days. Cancelling out the 26 km/s inwards velocity and gaining the planet’s 35 km/s orbital velocity take together (26,000+35,000)/(0.022 x 1.91): 16.8 days. The total trip is 58.5 days before optimizations. A high-tech Diffractive Meta-sail travelling to Mercury would cancel out Earth’s orbital velocity in 1.3 days. It then falls for 57 days. Removing the 52.5 km/s inwards velocity and catching up to Mercury orbiting at 45 km/s can be done in one intense 97.5 km/s maneuver. It is made more manageable by the 3.83x boost to sail acceleration nearer the Sun, so it only takes  (97,500)/(0.26 x 3.83): 1.1 days. The total trip duration is 59.4 days, with 96% of the time spent just falling towards the Sun. Returning from Venus or Mercury to Earth is like to be even faster.  'Sundiver' maneuvers with a sail You may have noticed how much faster sails get when they approach the Sun. The fact is exploited fully with the SunDiver maneuver (also called a solar photonic assist), where sails get to within 0.1 AU of the Sun for a massive 100x increase to their characteristic acceleration. Even a sail with a modest lightness ratio of 0.3 can reach Neptune in 933 days thanks to a single SunDiver maneuver. It greatly speeds up travel to the far reaches of the Solar System. Combined with the Oberth effect from swinging around so quickly near the Sun, you can get sails with modest characteristic acceleration to incredible velocities exceeding 400 km/s. Characteristics of an aerographite spherical sail At the upper limit, using carbon nanomaterials like Aerographite, we can push solar sails to 2% of the speed of light. Of course, there is no way to stop these sails afterwards. They can fly past planetary targets or simply head out into interstellar space.  'Pole sitter' statite Finally, high performance solar sails, especially designs with a lightness ratio equal or exceeding one, have the notable ability to just… stop. They can hover in any place. This turns them into ‘statites’; stationary satellites. These are useful if you want to place an observatory or telescope in an exact position or lay down a fusion highway.  A solar sail following a 'displaced orbit' A high performance sail could also enter ‘short orbits’ that only partially cancel out gravity near a planet, or follow complex trajectories (neither lines nor orbits) that take them between arbitrary points in space, like sitting exactly halfway between the Moon and Earth or above a specific crater on a spinning asteroid. Since they need no propellant, they can keep this up indefinitely.  Hybrid Braking So far we have talked about solar sails exclusively using sunlight for propulsion. This doesn’t have to be the case. ‘Hybrid sails’ use a mix of propulsion systems. OKEANOS combines a solar sail with thin film photovoltaic cells. For example, the payload of a solar sail can maneuver independently. Several sailing missions propose dropping a capsule during a flyby of Mars so that it can aerobrake to the surface. This saves on having the solar sail fight the planet’s gravity all the way down to a low orbit and then back out to escape velocity. A payload could also detach and use its own thrusters to go down to a planet, then rendezvous with the sail to return to interplanetary space. This can be the case with a lander with in-situ refuelling that lands nearly empty then returns to orbit nearly empty again.  A solar sail aerocapture into Saturn. The large area of a solar sail also makes it very useful for aerocapture. The Modern Sail has an average mass per area of 0.022 kg/m^2. The Mars 2020 mission that landed the Perseverance rover used an aeroshell with a mass per area of about 95 kg/m^2. This means the solar sail can slow down 4320x faster in Mars’ thin atmosphere. Aerocapture into Mars orbit with a sail. Or more realistically: it can brake from interplanetary velocities into a Mars orbit using the thin gases present high above Mars. The same is possible with the more advanced sails in the upper atmospheres of the gas giants. Whether the sails can survive the maneuver without being damaged is another question.   A simple hybrid sail. A “hybrid sail” is a more interesting design where the solar sail serves double-duty as part of another propulsion system. The obvious use case is a solar collector for a photovoltaic system. The huge reflective surfaces of a solar sail interact with large amounts of solar power but only convert a tiny fraction of it into useful kinetic energy. The Nanofilm Sail and its 4,000,000 m^2 surface area collects 5.44 GW of sunlight near Earth. If only a tiny fraction of that power is converted into electricity and fed into an ion thruster, we would get useful thrust levels throughout the solar system. It is most useful when braking towards the outer planets like Jupiter or Saturn. In that situation, the solar sail is still handling hundreds of megawatts of power but it still takes months to slow down into a capture orbit. If the Nanofilm Sail sacrificed half of its payload to a 400 kW solar-electric propulsion system weighing 200 kg (including a 9000s Isp thruster) and had 488 kg of propellant, it would be able to provide 9 Newtons of thrust anywhere. It would be enough to slow down by 20 km/s within 34.5 days. That’s an average of 0.0067 m/s^2 or 8.3x the photon-only acceleration near Jupiter. The major benefit is that when combined with gravity assists from a moon like Io, the sail only needs to slow down to (20+17.3): 37.3 km/s. That saves a massive amount of time. The usual 343 day trip that’s been reduced to 235 days with the gravity assist is further reduced to 159 days with the electric thruster. An even greater benefit is to be had near Saturn. At that distance, the sail is still capturing 59.3 MW of power but the photon-only acceleration falls to just 0.00024 m/s^2. The electric thruster would provide 28x that much acceleration. And it would mean the sail only needs to slow down to (20+12.6):32.6 km/s, which shortens the 748 day trip to 364 days! Similar calculations can be made for the other sails. There may be trade-offs where adding an engine and propellant to the payload lowers the sail’s characteristic acceleration but shortens the overall trip. You only need to find the right balance. This tends to favour high power density (W/kg) propulsion systems with very high specific impulse. However, do be wary of the fact that many propulsion systems cannot easily be scaled down. A Modern Sail with 1 kg of payload might find room for cold-gas micro-thrusters but probably not for an electromagnetic nozzle. The sails themselves are very fragile so an unshielded nuclear reactor is not a good idea either.   The ultimate hybrid sail is probably one where the sail material serves as a photonic thruster near the Sun, and then a second stage is used for braking for which the sail material is consumed as propellant.  A plasma thruster - VASIMR - could theoretically use aluminium gas as propellant. Imagine a plasma thruster that can turn solar sail material into a cloud of ions, ready to be shot out of a nozzle by magnetic fields. Apply this idea to the Diffractive Meta-Sail design. The sail alone represents 75% of the total mass of the vehicle. Its massive 2,500,000,000 m^2 area can capture 3.4 TW of sunlight near Earth, decreasing to 125.6 GW near Jupiter and 9.23 GW near Uranus. Instead of redirecting 90% of that sunlight into empty space to produce photon thrust, it could focus it instead onto an advanced photovoltaic receiver behind the sail. A 200 MW propulsion system (30% efficiency) would never need more than 666 MW of Sunlight, or 7% of the sail area to function. We can therefore consume 93% of the sail material as propellant and still have enough area to power the plasma thruster. The overall mass ratio is 3.3. The electric propulsion system would occupy half the payload capacity (5 tons) if we assume a fantastic power density of 40 kW/kg, achievable with fully superconducting designs. If the plasma thruster has a specific impulse of 12,000s, we’d get an average acceleration of 0.13 m/s^2 and a total of 140.5 km/s of deltaV.  In other words, a Diffractive Meta-Sail that swaps out half its payload for a high-Isp propulsion system could shorten its 112 day trip to Jupiter to less than 44 days. It would use its solar sail to accelerate all the way out to 2.1 AU, reaching a peak velocity of 210.8 km/s, before braking, then switching to its electric thruster for the last 12.5 days. The final maneuver is a capture around Io at 17.3 km/s.  Could we have just relied on the electric thruster alone? Probably not. The solar sail provided a ‘free’ 263.8 km/s of deltaV - more if we count the climb out of Earth’s gravity. An electric-only mission would have to provide a full (263.8+140.5): 404.3 km/s of deltaV to complete the same mission. Even with 12,000s Isp, that entails a mass ratio of 31, which is totally impractical.   The fission sail concept uses a very thin layer of uranium. Fission events are started by antimatter or neutrons from a nuclear reactor. There are even more options for combining propulsion systems with a solar sail. Solar thermal thrusters would provide a quick burst of thrust but not much deltaV. The backside of a solar sail could be coated with radioactive materials, turning it into a two-sided fission sail. The electric sail rides the solar wind instead of sunlight. The sail could be electrically charged, turning it into an electric sail that rides the solar wind. That resource is available everywhere in the solar system and doesn't have the same drop-off as sunlight. We can imagine a craft that starts as a solar sail, then consumes its sail materials and spins them back out as a web of charged elements, creating a much larger electric sail.   Riding more than sunlight A massive 'pinwheel' helio-gyro with human-sized payload capacity. In previous ToughSF posts, we frequently mention how well space infrastructure pairs with various propulsion systems. The same goes for solar sails. A simple example would be space tugs. They would move solar sails out of the deep gravity wells of planets, saving them the weeks of spiralling outwards to deep space. Or, they could ‘catch’ sails coming in from interplanetary trajectories, which would be especially useful for settlements around Jupiter or Saturn waiting for cargo deliveries.  There is no reason why a solar sail couldn't also work as a laser sail (with low intensity beams) Another well-described piece of space infrastructure is beamed power. Laser beams especially work very well with solar sails. Most sails would be able to receive a large range of laser wavelengths owing to the fact that they are supposed to reflect most solar wavelengths from UV to infrared. Lasers could boost solar sails away from Earth, or help them brake into capture orbits around the gas giants. The beam energy can be used to directly provide thrust via radiation pressure, or indirectly by being directed in hybrid sails that have photovoltaic cells to power electric thrusters (and similar designs). We’ve mentioned already how reducing the braking deltaV can drastically shorten trip times.  A Skyhook is a low orbit rotating tether. Rotating tethers are another classic example. They can fling solar sails into higher orbits or catch them on the way down. If a solar sail is too fragile to survive the g-forces from swinging around a tether, it could detach its payload and the two would go separate ways: the payload making a short trip down a rotating tether and the sail performing long spirals to park in another orbit.  One unique pairing between tethers and sails is the ‘self-winding tether’. In this previous post, we explained how tethers are two-way momentum exchange devices. They absorb momentum from incoming payloads, store it, and then spend momentum to launch payloads to higher orbits. An imbalance between incoming and outgoing momentum has to be compensated by using rocket engines or electrodynamic propulsion. But what if there wasn’t an ambient magnetic field to exploit or enough propellant to keep up with the demand for momentum? The solution would be a solar sail. It could carry a dumb weight up to a high altitude then ‘drop it’ down to the tether. The tether absorbs its momentum. The sail then spirals down to the tether’s orbit, picks up the weight again and carries it back up for another drop. This process can be repeated indefinitely, with nothing expended except free sunlight. Solar sails can also enable mass stream propulsion. Basically, many sails in a row, all working together like a train, carry a number of masses each. The sails head out on a long loop that takes them far away from where another spaceship is sitting. They then accelerate back in, picking up free velocity from sunlight. At the last moment, they all drop their masses and divert to the side. The masses continue in a stream up to a high velocity rendezvous with the spaceship. If the spaceship has a sort of receiver/nozzle like a pusher plate, magnetic nozzle or other such device, it can ‘ride’ the mass stream without having to spend any energy or propellant of its own. The mass stream effectively concentrates the energy gathered by solar sails over a long period into short bursts. A fleet of solar sails, perhaps bringing a mass stream up to speed. Mass stream propulsion has been described in previous posts. Essentially, it strips a spaceship down to just a payload and a nozzle. The nozzle converts the mass stream into thrust. Since the spaceship does not have propellant, reactors, radiators or any such equipment, it is very lightweight. Thrust comes externally without being limited by mass ratios, and it can be used to both accelerate and brake. Combined, we get a way to accelerate hard up to incredible velocities for rapid trips around the solar system. The high velocity masses would be consumed in a magnetic nozzle like this one. If sails use the SunDiver maneuver, we get sails up to 400 km/s and beyond. They would release (or become, with sacrificial micro-sails) mass streams that drift to a rendezvous with a mass-stream-rider spaceship, which would also accelerate to those velocities. It’s enough for an Earth to Saturn trip in less than 37 days. Braking would be handled by pre-positioned mass streams orbiting the destination. Best of all, the energy to drive all these mass streams is gathered for free from the Sun. There would be no need for giant railguns positioned at both the departure and destination points! Another feature of solar sails can be exploited to maximum effect. Sails with a lightness ratio better than one can position themselves anywhere as statites. This means that rotating tethers or laser beams and relay mirrors can be placed wherever they are needed. A statite could release a string of masses at the right place to act as a slow mass stream that a fast spaceship can brake into. There’s a huge range of creative uses for equipment or mass that can be moved freely to any point in the Solar System.  Sailing prospects So, where and how could Solar Sails be used? We’ve clearly shown how solar sails are not necessarily slow or restricted to small payloads. They can get to places quickly and are very flexible in their uses. Solar sailing should become commonplace in the future. Solar sails are most likely to be used to send low-cost probes to the rocky planets first. Then they would find use as ‘asteroid hoppers’ travelling from asteroid to asteroid as long their control electronics keep working. ‘Moon Tour’ missions where they survey the many bodies orbiting Jupiter and Saturn would come later. These sorts of missions make use of sails’ zero-propellant travel capability. What happens next is more speculative.  The 'Lunar Flashlight' uses the sail as large light source to explore the Moon. Nuclear propulsion in the future may be fast, but it costs fuels like uranium or tritium. Solar sailing is nearly free once the sail is built. There will always be a niche for ‘free’. Sails could be used to return precious metals from asteroids back to Earth. They can redirect whole asteroids, given enough time - for either Earth's defense or other operations. Only a 'free' propulsion system could make use of large masses of otherwise low value resources, like icy comets They could also shuttle supplies from Earth to Mars on a predictable schedule, all year round. With the SunDiver maneuver and aerocapture, they could be the first missions to actually enter orbit around Uranus or Neptune.  Mining with sails. Once we master solar sail technology and produce better lightness ratio designs, the options expand greatly. Uncrewed missions could use gigantic solar sails to carry payloads anywhere in the solar system. Passengers would use mass streams accelerated by solar sails to get to distant locations within days or weeks. Hybrid sails would bridge the gap for payloads that don’t need to get to places so quickly, or could be used when transport infrastructure is not available.  A sail on a 'non-Keplerian' orbit. Beamed power infrastructure to keep colonies alive around icy moons could rely on mirror relays held in position by statites. We might even see huge sail ‘motherships’ that have an onboard laser to move around a fleet of smaller dedicated laser sails. Meanwhile, statites sitting over the Sun’s poles read solar activity and warn everyone else of approaching solar storms within minutes.  The possibilities are numerous. Don’t dismiss solar sailing!

Rotating tethers can reach incredible velocities when they are built out of high strength materials. With some design features, they can greatly surpass the exhaust velocities of chemical or even nuclear rockets. They can become propulsion systems with impressive performance... and might look like the classic 'saucer' spaceship. How would they work? What performance could they achieve? Rotating Tethers Cover art by Mack Szbtaba. Rotating tethers are a fascinating topic that have been treated in depth by previous posts on ToughSF, such as using them to extract energy from planetary motion or make space travel much shorter.  Two SpaceX Starships in a 1500m tether formation spun to generate artificial gravity. In summary, a tether made of high strength-to-weight ratio material can withstand enormous forces while remaining lightweight. If spun in a circle, usually many kilometers wide, it can support a load on one end as long as it is supported by a counter-weight on the opposite side. The tip velocity achievable before the tether breaks from centrifugal force will reach several kilometers per second. It can be boosted even further if the tether is tapered: wider at the base and thinner towards the tip. With this technique, tethers made of mass-produced materials like Kevlar can cover a significant fraction of orbital velocity, making it good enough to be used to build a skyhook. Skyhook principle of operation. The important factor here is how heavy of a tether we need to handle a certain payload mass spinning at a certain velocity. First we need to work out the characteristic velocity of a tether, which depends on its material properties: tensile strength and density. Characteristic velocity = (2 * Tensile Strength / Density)^0.5 Characteristic velocity in m/s Tensile Strength in Pascals Density in kg/m^3 For Kevlar, the values we have are 3,620,000,000 Pa and 1,440 kg/m^3. Kevlar’s characteristic velocity is 2242 m/s Then we need to find the ratio between the tether’s tip velocity and the characteristic velocity, which we’ll simply call the Velocity Ratio VR. VR = Tip Velocity / Characteristic Velocity If our tether is spinning at 3300 m/s, then the VR is 3300/2442 = 1.351 Finally we get to the Tether Mass Ratio. It is the ratio between the tether mass and the payload mass it can handle.  Tether Mass Ratio (TMR) = 1.772 * VR * e^(VR^2) A tether with a VR of 1.351 will have a Tether Mass Ratio of 1.772 * 1.351 * e^(1.351^2) = 14.85. It means that a 1485 kg Kevlar tether can handle a 100 kg payload at its tip while spinning at 3300 m/s.  The HASTOL concept relied on 3250 m/s tethers.  The Tether Mass Ratio is square-exponential. It climbs extremely rapidly with increasing VR. Doubling the tip velocity to 6600 m/s, for example, raises the Tether Mass Ratio of a Kevlar tether to 7122. Now a 712.2 ton tether is needed for the same 100 kg payload; a nearly 48x increase. As a consequence of this scaling relationship, large rotating tethers are optimized for velocities only slightly above their material’s characteristic velocity. Then some safety margin has to be added on top. It is not practical to have a 10 ton capsule matched with a tether of several thousand tons. Hundreds of launches would be needed to justify the presence of the tether.  Large tethers also have some additional complications limiting their performance, such as the need to add multiple redundancy against micro-meteorite strikes and shielding against solar radiation that would otherwise degrade their materials. All of these measures cut into the mass actually dedicated to supporting a payload. Hoytether multiple redundant tether lines. But that is not the only way to use tethers. We can design them for an entirely different role. Higher Velocities It is possible to imagine much smaller tethers, perhaps a few meters across, spinning at much higher velocities. They would be completely enclosed in a protective container. The idea of a smaller, faster tether launching objects is not new. In fact, it is being worked on at full scale by alternative launch companies like SpinLaunch today. The idea is that we can increase tether velocity to many kilometers per second, then release small masses from the tether tips. This can be water or dust grains or whatever can flow down the tether’s length. Their release generates recoil in the opposite direction: that’s thrust. Momentum is lost with each release, though it can be regenerated by an electric motor that spins the tether.  Counter-rotating tethers ejecting water for propulsion. If we mount a tether like this on a spacecraft, it can be used as a rocket engine as propellant exiting in one direction and thrust produced in the opposite direction. As long as two counter-rotating tethers are used, there is no torque. Essentially, they become an electric thruster with an ‘exhaust velocity’ equal to the tether tip velocity.  There are many advantages. The tethers can use nearly any propellant they can pipe to their tips. Whether it is dust gathered from an asteroid’s surface, nitrogen scooped up from the edge of Earth’s atmosphere or water derived from a lunar mining operation, it can all go in the propellant tanks with minimal processing. That means there is no need to haul a chemical factory with you to every landing site in the Solar System. An orbital gas scoop. The tether itself should be practically frictionless and have nearly 100% efficiency. It operates mechanically (no electric currents or coolant flows) so it should produce negligible heat even at extreme power outputs, which are in turn limited only by its RPM.  A frictionless magnetic bearing is necessary to enable high efficiency rotating tethers. A tether rocket compares favourably in many ways to existing technology like Hall effect thrusters or MPD thrusters. They do not have to pay the energy penalty to ionize their propellant, nor do they have the pulsed energy storage concerns of mass drivers (railguns, coilguns). Further advantages will be described later in this post.   These tethers can be spun to very high velocities at the expense of impressive mass ratios. The g-forces exerted at their tips would be immense, but it is acceptable as their payloads won’t be fragile spacecraft. Also, since they are on a much smaller scale, it becomes much more affordable to build them out of the best materials available. For example, Toray’s polyacrylonitrile fiber T1100G with a characteristic velocity of 2,796 m/s or new UHMWPE fibres (Dyneema) being tested to a characteristic velocity of 2900 m/s.  These may seem like tiny gains over the characteristic velocity of widely available Kevlar, but remember that the Tether Mass Ratio is square-exponential. Small improvements lead to huge decreases in tether mass. Here is a table of the performance we can get: All of these materials make it possible to achieve tether tip velocities exceeding the best performance of chemical rockets (460s Isp or 4512 m/s) with a moderate mass ratio. Kevlar struggles when going faster than that. T1100G or UHMWPE can get us 7500 m/s exhaust velocity with a Tether Mass Ratio in the thousands. An exhaust velocity exceeding that of nuclear thermal rockets (1000s Isp or 9810 m/s) is achieved with T1100G at TMR 2.27 million and UHMWPE at TMR 0.89 million. A Tether Mass Ratio in the millions sounds extreme but consider it in these terms: a tether of 1 ton mass would be handling 1 gram of propellant at its tip. If it is 1 meter in radius, and the tip velocity is 10,000 m/s, then it makes a complete rotation 1591 times a second 95,460 RPM). It is not so extreme: commercial hard-drive disks spin at 7200 RPM and ultracentrifuges manage 100,000 RPM. We could compare at them to uranium gas centrifuges spinning at 90,000 RPM.  Rows of uranium gas ultracentrifuges. If this 1m long tether releases a 1 gram drop of water every time it completes a rotation, it will have a mass flow rate of 1.59 kg per second. Thrust is propellant flow rate times exhaust velocity, so multiplying that figure by 10,000 m/s gives us a thrust of 15.9 kN. Thrust power is equal to half the thrust times exhaust velocity, which in this case is 0.5 * 15,900 * 10,000 = 79.5 MegaWatts! Let’s try to design two realistic Hypervelocity Tether Rockets, one with T1100G aiming for an exhaust velocity of 6000 m/s which is ideal for travel between the Earth and Moon, and another using slightly more advanced UHMWPE aiming for 10,000 m/s which is better for interplanetary travel. The g-forces at the tether tips will exceed 1,000,000g, which is troublesome as there would have to be some moving part that controls the flow of propellant that can open and close thousands of times a second. A piezoelectric poppet valve that can open and close 2000 times a second. Putting as many components as possible on the external container (control electronics, magnetic actuators) rather than on the moving tip could help. Lunar Tether Rocket The Toray T1100G material is selected because you can order spools of it right now. The individual fibres have a tensile strength of 7000 MPa and a density of 1790 kg/m^3. With its characteristic velocity, 6000 m/s tip velocity means a Tether Mass Ratio of 380. Why 6000 m/s? Because it allows a rocket to make the 8400m m/s deltaV trip from Low Earth Orbit to Low Lunar Orbit and back with a propellant mass ratio of 4 (that’s 3 kg of propellant for each 1 kg of empty rocket). That is modest for an upper stage of a launch vehicle, let alone a lunar transfer stage. The tether here can have a length of 3.67 m. It would rotate at 15,607 RPM. If it aims to shoot off 10 grams of water with each rotation, then it will have a mass flow rate of 2.6 kg/s. The tether itself will mass 3.8 kg but we can bump that up to 5.7 kg to add a 50% safety margin. A counter-weight doubles that value to 11.4 kg. It will feel 60 Newtons of recoil with each release, which seems like it can easily be handled by a suspension mechanism. To counter torque effects, we must add a second tether rotating in the opposite direction, which adds another 11.4 kg for a total of 22.8 kg. Average thrust from both tethers is 31.2 kN. Thrust power is 93.6 MW.   This power can be delivered by a high power density megawatt-scale electric motor. An example of this today would be the H3X HPDM-3000 that manages 2.8 MW of output with a power density of 12.7 kW/kg. It is already meant to be stacked in multiple units. 93.6 MW of power would need to be delivered by 7370 kg of these electric motors. The motors are 94% efficient, so there’s 5.97 MW of waste heat to consider. The motors operate at 60°C, so 4282 m^2 of double-sided radiator panels are needed to handle their waste heat. This may need 4282 kg of 1 kg/m^2 radiator panels based on carbon fibre heat pipe technology.   In total, this propulsion system masses 11,675 kg. If we add a 10% mass margin for equipment like water pumps, tether container walls, coolant pipes, we arrive at a total mass of 12,843 kg. The tethers are by far the smallest component, representing only 0.178% of the mass total. Toray T1100G Tether Rocket Performance Tip velocity = 6000 m/s Total Mass = 12,843 kg  Thrust = 31.2 kN Thrust-to-weight ratio = 0.247 Average power density = 7.3 kW/kg If you add a power supply, propellant tanks, structural components and a payload, you get the rough draft of an Earth-Moon spaceship. The Hypervelocity Tether Rocket here far exceeds the performance of most electric propulsion systems you could slot into its place on such a spaceship. Aerojet Rocketdyne’s Hall thrusters struggle to reach 0.26 kW/kg. NASA’s more advanced electric thrusters aim for up to 4 kW/kg, but at a reduced efficiency of 60 to 85%. They are superior in terms of specific impulse, but that is not particularly needed in cis-lunar space.  Interplanetary Tether Rocket Now we look at a 10,000 m/s UHMWPE tether. It will be more advanced but still within the realm of ‘near future technology’. Tether Mass Ratio is 891,437.  The tether is short: 0.95 m in radius. It spins at 100,000 RPM. The amount of propellant released with each rotation is 1 gram. That means a tether mass of 891.4 kg and a mass flow rate of 1.67 kg/s. With counter-weights and a second counter-rotating tether, the tether assembly adds up to 3566 kg. We bump this up to 5349 kg for a 50% safety margin. The average thrust produced from the two tethers is 33.4 kN. Thrust power is 167 MW.  Fully superconducting electric motors can reach astounding kW/kg values At this power level, it is sensible to switch superconducting devices. NASA’s 2035 goals for turboelectric propulsion on aircraft uses high temperature superconductors to achieve 40 kW/kg at 99.99% efficiency. The electric motor mass would only need to be 4175 kg. The waste heat produced at 65 Kelvin would be 16.7 kW. A superconducting design. A 201 kW Stirling cryocooler of 300 W/kg, would raise the temperature to 300 Kelvin (30% of Carnot efficiency) and 670 kg of equipment. The radiators to handle the final heat load (16.7 + 201 * 0.7 = 157.4 kW) add another 171 kg.  In total, this propulsion system masses 10,365 kg. If we add a 10% mass margin as before, we arrive at a total mass of 11,401 kg.  UHMWPE Tether Rocket Performance Tip velocity = 10,000 m/s Total Mass = 11,401 kg  Thrust = 33.4 kN Thrust-to-weight ratio = 0.298 Average power density = 14.65 kW/kg This design has even higher performance and better specific impulse. It is well suited for missions to Mars. Its performance is somewhat comparable to a solid-core nuclear thermal rocket using liquid hydrogen, as it has the same exhaust velocity but it does not need bulky cryogenic propellant tanks or a full electrolyzing ISRU plant to refuel it. If solar or beamed power is available, it could do away with nuclear technology altogether and still achieve comparable performance.  Neither of these designs are optimized. There could be further performance gains to be had from selecting a better tip velocity or cooling solution. For example, the propellant water could first be used to cool the electric motors to save on the mass of radiators needed. Or, we could employ several tethers to multiply the thrust the engine could produce without having to also increase RPM or tip velocity.   Staging tethers on tethers Rockets get around the problem of exponential mass ratio by using staging. Tethers can employ the same strategy. Instead of placing a payload on the tip of a tether, another smaller tether can be attached. Each tether would spin independently of each other, and at the right moment, their tip velocities would add up.  Here is an example with Kevlar: We want a tip velocity of 10,000 m/s. As we calculated previously, this would require an impractical tether with a Tether Mass Ratio of over 139.1 million. If we instead break it down into tethers of 5,000 m/s velocity, and stage them tip-to-tip, we would obtain stages with a mass ratio of 240. Two stages would add their tip velocities to 10,000 m/s and multiply their mass ratios to 240 x 240 = 57,600. This is obviously much lower than one huge tether.  There is very little literature available on this idea. The closest concept is the Tillotson Two-Tier Tether, as depicted here. There will be challenges to designing a two-stage tether for use as a rocket. There’s the issue of transferring propellant between the tethers, which could be very troublesome if you want solid particles as propellant. Designing a rotating joint that can work smoothly when under high g-forces can’t be easy. Then there’s the difficulty of restoring momentum to the second-stage tether. A second-stage tether also needs its own counter-weight, which could double the overall mass ratio. But, if all these challenges can be solved, then we would get much more impressive tether rockets. Here is a table for two-stage performance: The same material selection as in the previous section is given a second stage so that the total Tether Mass Ratio for both stages reaches 500, 50,000 and then 500,000. The final ratio is doubled to account for the second stage tether’s counterweight. In this arrangement, even Kevlar exceeds 11 km/s tip velocity. UHWPE manages 13.1 km/s with a final tether ratio of 1 million.  Let’s update the two tether rocket designs with staged tethers: Toray T1100G Two-Stage Tether Rocket Performance Tip velocity = 7430 m/s Total Mass = 12,843 kg  Thrust = 25.2 kN Thrust-to-weight ratio = 0.2 Average power density = 7.3 kW/kg We maintained the 380 final tether mass ratio from the Toray 1100G tether rocket. However, with two stages, we get an exhaust velocity of 7.43 km/s. Thrust power from the electric motor is identical so the thrust-to-weight ratio has to fall to 0.2. UHMWPE Two-Stage Tether Rocket Performance Tip velocity = 10,000 m/s Total Mass = 6095 kg  Thrust = 33.4 kN Thrust-to-weight ratio = 0.56 Average power density = 27.4 kW/kg The UHMWPE tether rocket aims for the same tip velocity, but with two stages the final Tether Mass Ratio (x2) can fall from 891,437 to just 7128. The tether assembly is reduced from 5349 kg to 42.7 kg, raising the overall thrust-to-weight ratio and average power density significantly.  Note that for both of these designs, we are only calculating the mass of the engine - the part that converts electrical power to thrust. A complete spaceship would have to include an electrical generator, be it an onboard reactor, solar panels or a laser-photovoltaic receiver. In a realistic study, you will find that high engine power densities means the average power density of the propulsion module of a spaceship approaches that of the power generating section alone. The overall performance of a spaceship won’t improve much if you have a terrible power generator (0.2 kW/kg solar panels) but excellent engines (20 kW/kg).  Solar-electric spacecraft with football fields of photovoltaic panels might not benefit much. Two-stage tether tip velocities means we obtain a propulsion system that can make shorter interplanetary trips. 1200 seconds of specific impulse means that a spaceship that’s 75% water (a mass ratio of 4) has 16.3 km/s of deltaV. It can start in Low Earth Orbit and arrive in Low Mars Orbit in 88 days, or complete a trip to Io’s orbit around Jupiter in 1.73 years instead of the usual Hohmann transfer of 2.73 years. This is without the assistance of aerobraking and with the ability to quickly load up on propellant at the destination for the return trip.  A relatively quick trip from Earth to Jupiter. Theoretically, a third tether stage is possible. It would push the potential performance of tether rockets well into the domain of electric thrusters (2016s Isp with UHMWPE) while retaining the upper hand in thrust-to-weight and power density. However, the problems mentioned above would all be exacerbated.  Carbon extraordinaire So far we have restricted ourselves to materials available in bulk today. Better materials exist; we only need to learn how to manufacture them in large quantities. The most promising of these are carbon nanomaterials: nanotubes and graphene. Carbon nanotubes are being grown right now, up to lengths of 50 centimeters. Graphene flakes are regularly added to epoxy resins and nanocomposite materials to enhance their strength. In the future, we could see them being produced in much larger quantities, enough to use for tethers.  In order of difficulty of manufacture, we have multi-walled carbon nanotubes, single-walled carbon nanotubes and then graphene. Here are their ‘perfect’ properties: The characteristic velocity of these materials can exceed 10 km/s. When used in a tether with a Tether Mass Ratio (TMR) of 100, they can achieve tip velocities approaching 20 km/s. In a TMR 10,000 tether, they approach 30 km/s and they can push beyond 60 km/s with a TMR of 1 million. That’s better than what most electric thrusters are capable of today. Of course, it is unlikely we will be able to form tethers of several meters in length with zero defects, errors or safety margins using these materials in the near future. The strength of a single perfect fibre is reduced when it has to be bundled with many other fibres, bringing down the ‘engineering strength’ to about half of the maximum with no other factors involved. Even at their weakest, carbon nanotubes far surpass other materials. If we assume that a half of the theoretical maximum could be achieved in bulk quantities, the tip velocities we would actually achieve would be reduced by 42%. Then, we could apply staging. A two-stage hypervelocity tether rocket with specific impulse of 2000 to 4800s seems achievable with these materials. The overall power density of the rocket is difficult to estimate because access to carbon nanomaterials would also affect the weight of components like electric motors or radiator panels. The final design could easily exceed 100 kW/kg. It does mean that the performance of the power generating source becomes critical to good overall performance. Even a nuclear reactor with radiators and a turbine that we consider excellent today at 10 kW/kg would become a performance bottleneck when paired with a 100 kW/kg carbon nanotube tether rocket.  Mechanical Rocketry What’s it like to use hypervelocity tether rocket engines? The radiators are tapered to fit inside the reactor's shadow shield, with the water tanks serving as extra shielding.  They can simply be mounted on spacecraft and used to travel by throwing propellant out. It would look rather weird: they have no nozzles, only need small propellant tanks and their most distinguishing feature might look like a wheel... or if the tethers are placed internally, the whole spaceship might be configured like a disk. Not aliens, a spaceship with equatorial tether-rockets (and fancy lighting)! Meaning, your diamond hard science fiction can have fully justified 'flying saucers' roaming the Solar System. The tethers can thrust in different directions by selecting a different firing port for their exhaust. A disk-shaped spaceship with firing ports along its rim can accelerate in any direction. It just has to take care not to aim its exhaust at nearby objects.  Docking might have to be done entirely using secondary propulsion (RCS thrusters). Water can drill holes through asteroids, space stations and other spacecraft when shot out at 10 km/s. Over long distances, it would disperse into harmless mist but at short distances it would be dangerous. Dust or other solid particle propellant would not disperse and would remain dangerous forever. Their use in the Outer Solar System or between asteroids might be justified by the vast distances involved, but not in cluttered low planetary orbits, especially if exhaust velocity is less than escape velocity (the dust would circle back around).  Spaceship pilots might need to pay attention to how long it takes for their tethers to reach operational RPM. Thrust would not be instantaneous, which makes delicate or urgent manoeuvres troublesome.  Thrust levels can be adjusted by firing more or less frequently. Theoretically, the tether can be spun down to a lower tip velocity to allow for more propellant to be fired with each rotation. The potential thrust would increase exponentially as the tether velocity is decreased. However, the other critical component in a tether rocket is the electric motor. Its output is tied to its RPM, so spinning slower might also mean less watts from the motor. The solution to this is a gearbox… but the practical details of building a MW-scale 100,000 RPM set of gears are best left to people in the future. It should be noted that electric motor power does not have to exactly match the thrust power of a tether rocket. The spinning mass of a tether can be considered a type of flywheel, so it can store energy. Energy can be accumulated gradually by a small motor (which enables some mass savings), then released quickly from the tether. This is most useful for spacecraft that aim to raise their orbit via multiple short burns at the periapsis of their orbit. It maximizes the contribution of the Oberth effect and was used by Rocketlab’s Photon stage for the CAPSTONE lunar mission. It’s possible to rely on rotating energy storage alone for propulsion. An asteroid mining spacecraft could land on a target, hollow it out for raw materials, build flywheels-tethers out of the leftovers and spin them up before leaving. Those tethers would then eject pieces of asteroid dust for propulsion until their energy ran out. RAMA proposed this architecture but with a different way of converting stored energy into thrust (using catapult sling arms). In fact, asteroid mining is one of the best applications of tether rockets. The ability to use any propellant, the decent exhaust velocity (for an electric rocket) and the ability to store energy then release it quickly combine to make tether rockets ideal for asteroid hopping spacecraft. The deltaV for travelling between asteroids can be very low, which suits the tether rocket perfectly. An asteroid mining spaceship. Perhaps the ring sections could be tether-rockets... Sunlight may be too weak to keep a powerful motor running continuously in the asteroid belt, so slowly accumulating energy into a flywheel is a good option to have.  Being able to use asteroid dust as propellant means the mining ships can hop to very ‘dry’ targets without worrying about the availability of water to refuel themselves. The tether itself could be made of locally sourced materials, such as glass or basalt fibres that exhibit ‘good-enough’ characteristic velocities of 1.5 km/s to 2 km/s. Glass fibre tethers would be larger and heavier than carbon nanotubes, but that’s actually an advantage if they double as energy storage flywheels. Manufacturing basalt fibres. This creates a ‘low performance’ niche for tether rockets. They could excel here as well as they do in the ‘high performance’ role with super-materials and extreme tip velocities. Other Applications Beyond simple use as rockets, hypervelocity tethers can have a variety of further applications. Drilling and excavation A high pressure water drill. A series of high velocity impacts concentrated onto a small area can serve as an efficient drill. Water or dust at 10 km/s can overcome the mechanical strength of practically any material, so what the target is made of does not matter. The impacts can be tuned to bore a hole through a target, or create shockwaves that fracture it into smaller pieces for easy excavation. One idea is to have the spinning tether first serve as a rocket to bring a spaceship close to an asteroid, then become part of mining equipment to dig into the asteroid’s surface and expose the dense core potentially loaded with precious metals. Just make sure to anchor the tether well! Mass Streams 'Pellet beam' propulsion. A tether could launch those pellets. The hypervelocity tether can be used as a mass driver to shoot a series of projectiles to propel other spacecraft. This is known as mass stream propulsion. The spacecraft riding these mass streams only need a device to catch the projectiles - it can be as simple as an ablative pusher plate or as complex as a magnetic nozzle that drops solid targets into the path of the mass streams and pushes off the resulting plasma explosions. Either way, the riders are unburdened by propellant, reactors or radiators, so they can have fantastic acceleration. Mass drivers are usually fixed structures that do not have to worry about their weight, so the tethers can aim for extreme mass ratios. A two-stage T1100G tether with a TMR of 100,000 per stage would have a tip velocity of 17.5 km/s. Spacecraft riding these mass streams could achieve a good fraction of this velocity, perhaps 16 km/s. More mass streams headed in the opposite direction would be waiting for them at their destination for braking. Together, they enable fast interplanetary travel.   Railguns or coilguns could also be used as mass drivers, but they are usually much less efficient and take up a lot more room than tethers.  Stealth Drive Dark, non-radiating and doesn't even leave a trail of hydrogen behind it. You might imagine that a hypervelocity tether would make for a good weapon. It could drill through any target and its firing rate would allow for enough shots to ensure hits at long range. However, this is unlikely. Hypervelocity tethers have no barrel, so they are inaccurate. It would be difficult to put them in a turret. Their large rotating mass means they act like a gyroscope that resists turning. The way the tether mass scales with projectile mass means that only the smallest projectiles are possible. That removes the option of using ‘smart’ guided projectiles with sensors and RCS thrusters to track a target as these may have a minimum mass of several hundred grams.  Worse, they would be extremely vulnerable to battle damage. A small cut on the tether might lead to it completely disintegrating… inside your spaceship.    So spinning tethers are a bad weapon. Does that mean they have no military use? There is one final advantage that comes into play. The exhaust of a tether rocket can be cryogenically cold. The entire launch process does not release any heat. Even the electric motor can be of a superconducting design bathed in liquid helium at <4 Kelvin. So long as you have access to electrical power, the tether rocket can be a completely stealthy propulsion system.