arXiv:2009.08874v2 [physics.pop-ph] 31 Oct 2020 mlingam@fit.edu orsodn uhr aav Lingam Manasvi author: Corresponding ointrs yJl atri 07( 2007 “tech- in termed Tarter appositely Jill - by (ETI) nosignatures” intelligence logical o tpeo-yo pee aebe netknto undertaken been have spheres Stapledon-Dyson for 2019a the in 2011 underscored ( its increasingly century and potentiality 21st been recognized, the has was however, importance technosignatures field, non-radio this of minus- of remains inception very sampled space ( the cule parameter hitherto, conducted of searches fraction numerous the of spite ra- at mostly electro- onward, artificial ( 1960s wavelengths for dio the quest from the signals by magnetic dominated been has opsdo wrso bet otpteeeg of energy the tap to objects are ( of star megastructures the swarms are these of category - composed this spheres in technosignatures striking. Stapledon-Dyson known more it best commensurately tech- humans, ad- The be resultant than more would their technology considerably that nosignatures their anticipated are conventionally of ETIs is terms their these of in As output vanced energy star. radiative host entire the harnessing 2020 Gray 2001 Tarter eavlal erc( to proven metric has valuable Kardashev a Nikolai be late the by pioneered h erhfrsgaue fetaersra techno- extraterrestrial of signatures for search The yee sn L 2020 using Typeset 3, November version Draft ncasfigpopcieEI,teKrahvscale Kardashev the ETIs, prospective classifying In ; ). atre l 2010 al. et Tarter rgte l 2014 al. et Wright oeta pesta ol eraial yselregns n det and engines, stellar by stellar realizable the be accelerating could and that thrust speeds generating potential of purpose the with a.I a epsil odrv onso hi bnac,btth but a abundance, which high-spe their engines, for on such of bounds searches distribution derive commensal spatiotemporal to the possible conduct regarding be to ast may Hence, utilized It speeds. Way. such be produce contrast may to In unlikely very stars conditions. are Way optimal Milky under the in attainable be perhaps might 1 tpeo 1937 Stapledon eateto eopc,PyisadSaeSine,Flori Sciences, Space and Physics Aerospace, of Department tla nie r yohszdmgsrcue htetate extract that megastructures hypothesized are engines Stellar osrit nteaudneof abundance the on Constraints .Tp IEI,frisac,aecpbeof capable are instance, for ETIs, II Type ). ; odne l 2017 al. et Worden yo 1966 Dyson A T 1. E rk 1965 Drake X 2 INTRODUCTION twocolumn nttt o hoyadCmuain avr University, Harvard Computation, and Theory for Institute adse 1964 Kardashev ; ; ; rgte l 2018 al. et Wright ikvc2018 Cirkovic ; yo 1960 Dyson ria 1983 Freitas ; hlvki&Sgn1966 Sagan & Shklovskii tl nAASTeX63 in style ; ak ta.2020 al. et Lacki .Svrlsearches Several ). aav Lingam Manasvi ; ; ; igm&Loeb & Lingam rduye al. et Bradbury atr2007 Tarter Cirkovi´c 2015 ´ .Snethe Since ). ;in ); ABSTRACT - ) ,2 1, ; ; 0 . 01 n baa Loeb Abraham and c aIsiueo ehooy ebun L391 USA 32901, FL Melbourne Technology, of Institute da okn ute aki ie h ulttv oinof notion qualitative in the appears time, engines in stellar back further Looking a elydt eetafato fterdainback ( radiation the star of host fraction the mirror a toward gigantic reflect a the to wherein deployed with engine was stellar credited a widely for is design Shkadov gener- first the typically Leonid upon and thrust. work draw useful ate which extract to engines, energy stellar star’s are category ilar ( in reviewed as date, tla nie aebe xlrdi eea te pub- other several in ( explored lications been have engines Stellar quote: following the from seen as monograph: wide-ranging characteristically his in nario Zwicky 2021 tla nie nteMlyWay Milky the in engines stellar nte ru fmgsrcue eogn oasim- a to belonging megastructures of group Another ). n h atrcntttn h u n the and propellants. sun nuclear the as constituting planets us- matter jets, the fusion realized nuclear ing years be of could action thousand projects the these through a of arrive All hence. might then descen- our dants neighborhood whose in Centauri eaae h w otn seso civiliza- that of space islets floating of tion. ocean two vast the separated the of systems across attendant worlds projecting their for with voyage made stars therefore several inter-galactic were an Plans . upon . . it course natural direct its at- from and star an a was detach accident to first tempt the of occasion The speeds, higher instance to for the it accelerating perhaps of is possibility fascinating are Most changes many imaginable. itself, sun the Considering ( 1957 aec ahat2000 Cathcart & Badescu abig A018 USA 02138, MA Cambridge eg rmtehs tr typically star, host the from nergy entgaate ob valid. be to guaranteed not re ytm eepoetemaximum the explore We system. aua srpyia phenomena astrophysical natural , g 6)epiil riuae hssce- this articulated explicitly 260) pg. , rieta pesu to up speeds that ermine oercsreso hypervelocity of surveys rometric dselregnsi h Milky the in engines stellar ed 2 srqie eti assumptions certain requires is 1000 Wright hao 1987 Shkadov mscdrce toward directed km/sec Stapledon ( 2020 and ) , ( 2006 1937 , 1988 igm&Loeb & Lingam ∼ ; hpe XI), Chapter , opr2018 Hooper .However, ). 0 . 1 c α - ; 2

Caplan 2019; Svoronos 2020) and methods for detecting In this work, we explore stellar engines from a generic them during the course of exoplanetary transits were physical standpoint in Sec. 2, and describe how they discussed in Forgan (2013). may attain terminal speeds that are sub-relativistic. We In the classification scheme of Badescu & Cathcart compare these speeds against stars ejected by natural (2000), three major classes of stellar engines were identi- astrophysical phenomena in Sec. 2, and argue that the fied. Class A stellar engines utilize the impulse from stel- latter cannot reach such large values in the Milky Way. lar radiation to generate a thrust force. The quintessen- By making use of this proposition, we examine current tial example of a Class A engine is the so-called Shkadov astrometric surveys to set tentative constraints on the thruster described in Shkadov (1987, 1988). Class B stel- abundance of putative ETIs that develop high-speed lar engines, in contrast, harness the radiation emitted stellar engines in Sec. 3. We end with a summary of by the host star and convert it into mechanical power. our results and prospects for future work in Sec. 4. Class C stellar engines combine elements of both Class A and Class B stellar engines, and thereby generate both 2. MAXIMUM ATTAINABLE STELLAR SPEEDS thrust force and mechanical power. Class D stellar en- We will begin with a synopsis of the maximal speeds gines, mentioned briefly in Badescu & Cathcart (2006, realizable by ejected stars and stellar engines. pg. 121), extract mass from the star by means of “mass 2.1. Maximum speeds of ejected stars in astrophysical lifting” (Criswell 1985) and expel the material to gen- systems erate a rocket effect; such stellar engine were previously termed “stellar rockets” by Fogg (1989). The maximum speed (vmax) achievable by stars The designs suggested for the putative stellar engines ejected after tidal disruption of a stellar binary system vary from one class to another. The Shkadov thruster by Sagittarius A* was computed in the classic analy- (Class A engine) is composed of a collection of reflective sis by Hills (1988), and it was concluded that vmax −2 ≈ statites (Forward 1991), which effectively functions as 1.3 10 c; see also Yu & Tremaine (2003) and Brown × a large mirror in the shape of a spherical arc; the de- (2015). Subsequent numerical simulations by Sari et al. sign for a Class C engine is also similar in this respect (2010, Section 8) determined that the maximum speed (Badescu & Cathcart 2006, Figure 1). Class D engines of the least massive object in a triple system that un- operate on the rocket effect, and can therefore be en- derwent ejection is given by visioned as stellar rockets, but the engineering specifics 1/6 2GM2 M1 differ across proposals (Caplan 2019; Svoronos 2020). vmax 1.3 , (1) ≈ R2 + R3 M2 + M3 We will briefly summarize one such design later in Sec. r   2.2. where M1 denotes the mass of the supermassive black Clearly, stellar engines represent a massive engineer- hole (SMBH), while Mj and Rj are the masses and radii ing feat, and this raises the question of why they would of the stellar binary system that is subjected to tidal be constructed. Before doing do, it should be noted breakup (j = 2, 3), where M3 < M2. We adopt M1 6 ≈ that the planets and moons orbiting the host star would 4 10 M⊙ for Sagittarius A* (Boehle et al. 2016), and need to be accelerated as well, but this feat is relatively substituting× this value into the above equation yields easier in comparison to the star. One possible reason 1/2 −1/2 suggested in Caplan (2019) is that stellar engines may −2 M2 R2 + R3 vmax 3.4 10 c enable ETIs to preemptively escape the vicinity of catas- ≈ × M⊙ R⊙     trophic phenomena (e.g., supernovas) and avoid the ad- −1/6 1 M2 + M3 verse consequences. This issue may be rendered more , (2) × M⊙ prominent near the Galactic center, where higher stellar   densities are associated with higher rates of catastrophic and this expression is considerably simplified when M3 events. Another option is that the ETIs might wish to is at least a few times smaller than M2, and using the undertake intergalactic travel, which is perhaps easier to 0.8 scaling R2 M2 for main-sequence stars (Tout et al. undertake by moving the planetary system as a whole, 1996). In this∝ optimal situation, we end up with in contrast to the alternative concept of building “world −0.07 ships” (Hein et al. 2012). We will not speculate on this −2 M2 vmax 3.4 10 c , (3) topic further, as it partly overlaps with the poorly un- ≈ × M⊙ derstood fields of xenosociology and xenopsychology.   which implies that vmax is nearly independent of M2. Hitherto, we have assumed that M2 is a main-sequence 1 For instance, theoretical models indicate that the habitability of star, but the other extreme is to consider a stellar-mass the Milky Way might be affected over kpc scales due to the cu- black hole instead. The resultant maximum speed was mulative impact of tidal disruption events and a potential quasar phase at some point (Balbi & Tombesi 2017; Forbes & Loeb calculated in Guillochon & Loeb (2015, Equation 5): 2018; Lingam & Loeb 2019b; Lingam et al. 2019; Pacetti et al. 1/6 −0.12 2020). −2 M2 M3 vmax 6.7 10 c , (4) ≈ × 10 M⊙ M⊙     3 where M2 is the mass of a typical stellar-mass black hole condition is fulfilled for M⋆ & M⊙. On the other hand, and M3 is the mass of the ejected star. if this criterion is violated, the stellar engine will not Thus, it is apparent from the preceding formu- be able to achieve v vmax in reality. Since the thrust ≈ lae that achieving vmax > 0.1 c is very unlikely generated by Class A and Class C engines is comparable for gravitational triple interactions in our Galaxy. (Badescu & Cathcart 2006, Section 3), the same upper In fact, even ejected stellar speeds larger than bound also applies to the latter. 0.01 c are rarely attained in numerical simula- Let us consider a generalized stellar engine wherein ∼tions, as seen from the probability distributions of a fraction ε of the total energy radiated by the star is ejected velocities in Bromley et al. (2006); Kenyon et al. harnessed to propel it, as suggested in Hooper (2018). (2008); Ginsburg et al. (2012); Rossi et al. (2014); This would resemble a Class C stellar engine to a certain Generozov & Madigan (2020). Faster speeds of & 0.1 c degree, because there is both energy extraction from the are realizable in theory but require one of the stars in star’s radiation and the generation of thrust force. By the stellar binary system to be replaced by a SMBH applying the conservation of energy, we obtain (Hooper instead (Guillochon & Loeb 2015; Loeb & Guillochon 2018, Equation 2.3): 2016; Darbha et al. 2019); however, this specific scenario 1/2 1/2 −1/2 is manifestly not applicable to the Milky Way. −2 ∆t L M⋆ Aside from the Hills mechanism and its variants men- v 1.2 10 c √ε , ≈ × 1Gyr L⊙ M⊙       tioned earlier, it is necessary to gauge whether other (6) avenues can eject stars at similar speeds. One of the and we can calculate the maximum velocity by invoking most well-known processes entails the disruption of the the relation ∆t t⋆ from earlier, which yields binary when one of the objects undergoes a core-collapse ≈ −2 supernova and leads to the ejection of the other objects vmax 3.8 10 c √ε, (7) (Blaauw 1961; Boersma 1961). Numerical simulations ≈ × indicate, however, that the maximal speeds of ejection and we reiterate that this speed is realizable in reality are . 0.01 c (Tauris 2015; Evans et al. 2020; Neunteufel only when M⋆ & M⊙. As opposed to energy conversion, 2020). Another possibility is the “dynamical ejection if the momentum of radiation is harnessed, the scaling scenario” whereby dynamical ejection from stellar clus- −3 and magnitude of v are akin to (5). ters is facilitated, typically at speeds of order . 10 c We can, however, conceive of more sophisticated sys- (Poveda et al. 1967; Leonard 1991; Oh & Kroupa 2016). tems. Let us suppose, for instance, that instead of utiliz- Lastly, one could replace the SMBH with an intermedi- ing energy, the putative ETI extracts mass at a constant ate mass black hole or a series of stellar-mass black holes, rate (denoted by M˙ ⋆) via “mass lifting” (Criswell 1985). but the resulting speeds of ejected stars are . 0.01 c This scheme does not exhibit a clear one-to-one mapping (Gualandris & Portegies Zwart 2007; O’Leary & Loeb with any of the stellar engines described in Sec. 1 be- 2008; Fragione & Gualandris 2019). In all these pro- cause the focus is not on the energy or momentum of the cesses, the speeds attained are smaller than vmax for the emitted radiation, but rather on mass extraction and its Hills mechanism; see (3) and (4). subsequent conversion into thrust imparted to the star. The mass thus acquired is presumed to be converted 2.2. Stellar engines: potential speeds into energy at an efficiency µ via the mass-energy rela- The Shkadov thruster, which is an example of a Class tionship. In this event, provided that M˙ ⋆∆t M⋆ to A stellar engine, will reach the velocity v after an interval ensure the mass is roughly constant, the speed≪ achieved ∆t as follows (Badescu & Cathcart 2000, Equation 34): over the interval ∆t is estimated to be

−1 5 ∆t L M⋆ 2 ˙ v 3.4 10− c , (5) 2µc M⋆∆t ≈ × 1Gyr L⊙ M⊙ v = . (8)       s M⋆ where L and M are the luminosity and mass of the ⋆ ⋆ In place of working with two variables in the above ex- star. In order to compare the speeds of stellar engines pression, we can rewrite it as follows: we suppose that against those presented in Sec. 2.1, it is necessary to ETI modulates its mass extraction and energy conver- assess their maximal values. sion such that M˙ ∆t = ζM with ζ 1, i.e., the ETI It is reasonable to assume that the maximum speed ⋆ ⋆ ensures that the star’s mass is not significantly≪ depleted is attained when ∆t t , where t denotes the main- 2 ⋆ ⋆ after the interval ∆t. A crude upper bound on ζ is 0.1 sequence lifetime of the≈ star. From the scaling t ⋆ ∝ (larger values would result in non-negligible stellar mass M⋆/L⋆ (Hansen et al. 2004, Equation 1.90), we find that vmax would become independent of stellar properties; −4 after simplification, we obtain vmax 3.4 10 c. Note, 2 ≈ × The depletion of stellar mass, among other things, may aid in however, that this estimate for vmax applies only when controlling the stellar luminosity, and thereby mitigating the shift t⋆ interstellar medium (ISM) is under- undoubtedly complicated) and falls outside the scope of way, it may be feasible to make use of other propulsion this paper. What is clear, however, is that the mass ex- systems to generate additional thrust and couple them traction can continue as long as access to the black hole to the stellar engine. The interstellar ramjet, which is maintained, which implies that the black hole must scoops up interstellar material and converts the accrued also be accelerated. Even setting aside this option, we matter into fuel, represents one such possibility (Bussard remark that ETIs with the technological wherewithal to 1960; Long 2011). We will not delve into the technical build stellar engines might find methods to raise µ by an details of this putative coupling scheme, because ascer- order of magnitude or so compared to the proton-proton taining the engineering technologies adopted by hypo- chain. In case µ 0.1 can be effectuated through a suit- ∼ thetical advanced ETIs is indubitably constrained by able avenue, the above formula leads to vmax 0.1 c. our current level of knowledge and vision. The above analysis, however, ignored the fact∼ that a Lastly, a comment on the effects of the ISM on stellar continual mass loss ought to result in a rocket effect. In engines is in order. The mechanical components com- theory, we can also conceive of thrusters propelled by prising the stellar engine will be subject to erosion and jets in stellar and compact object systems, correspond- damage due to interactions with gas, dust and cosmic ing to the Class D stellar engines mentioned in Sec. 1. rays in the ISM. However, predicting the extent of dam- The final velocity is straightforward to calculate when age is difficult because it depends on the relative ve- the rocket equation holds true (Tsiolkovsky 1903): locity of the stellar engine, the distance travelled by it, the thickness and properties of the materials used to Mi fabricate the components, and many more. There have v = vex ln , (11) Mf been several theoretical studies of the passage of sub-   where Mi and Mf are the initial and final masses of 3 Note, however, that this mass ratio (which is equivalent to ζ ∼ the star, whereas vex is the velocity at which the pro- 0.9) is at odds with the choice of ζ considered previously. We have pellant is expelled. If one considers vex that is a few considered a deliberately high value to highlight the significance times higher than the stellar escape velocity, or equiva- of the rocket effect when it comes to Class D stellar engines. lently the stellar wind velocity, and choose a mass ratio 5 relativistic spacecraft through the ISM, and it might be radial) that are comparable to, but somewhat lower, feasible for them to remain functional over kpc distances than this object (Shen et al. 2018; Bromley et al. 2018; (Hoang et al. 2017; Hoang & Loeb 2017; Hoang 2017; Du et al. 2019; Caffau et al. 2020). Hippke et al. 2018; Lingam & Loeb 2020), although sig- Thus, it would seem a safe bet to contend that all nificant uncertainties still remain. searches conducted to date have failed to yield unam- To summarize, there appear to be plausible designs biguous evidence of sub-relativistic stellar engines. This for stellar engines that seem capable of achieving vmax datum enables us to derive the following constraint: 0.01-0.1 c in principle. At the minimum, it may be ar-∼ Nsurv fT fSE < 1, (12) gued that no compelling a priori grounds exist for dis- · · missing the prospects for sub-relativistic stellar engines where Nsurv represents the number of stars collectively based on core physical principles. encompassed by all astrometric surveys, fT denotes the 3. CONSTRAINTS ON STELLAR ENGINES fraction of all stars that host ETIs with human-level technology, and fSE embodies the fraction of all human- In the preceding Section, we presented arguments as level ETIs that subsequently achieve the capability and to why even the fastest stars ejected from astrophysi- 4 intent to deploy sub-relativistic stellar engines. The rea- cal systems are unlikely to have speeds of & 10 km/s son we select human-level technology as the benchmark (0.03 c), whereas stellar engines could attain such ve- is not due to the notion that we are “special”, but be- locities. Hence, if one were to detect stars moving at cause it constitutes a signpost that is familiar to us. & 0.1 c, it would be strongly indicative of ETI activ- Although (12) may appear simple, the estimation of ity and thus constitute a technosignature. The one false the sample size Nsurv is fraught with several difficulties positive that ought to be taken into consideration is that and ambiguities, some of which are detailed below. a star moving at & 0.1 c may have an extragalactic ori- gin because of ejection during binary SMBH interactions 1. Given that Gaia DR-2 furnished astrometric data 9 7 (Guillochon & Loeb 2015), as explained in Sec. 2.1. A for 10 stars and radial velocity data for 10 ∼ 7 ∼ 9 combination of precise astrometry and chemical tagging stars, it is tempting to specify Nsurv 10 10 . ∼ − should, however, aid in distinguishing between extra- galactic and Galactic hypervelocity stars. This proce- 2. However, the stellar engines we discussed in Sec. 3 reach their peak speeds of 0.01-0.1 c only when dure was utilized to pinpoint the origin of HVS 3 from ∼ the Large Magellanic Cloud (Erkal et al. 2019). the stars have masses close to that of the Sun or higher (M⋆ & M⊙). Among the total population The Gaia mission was designed to accurately pin down 9 the positions and radial velocities of 109 stars in the of 10 objects spanned by the Gaia mission, ∼ only∼ 108 of them are solar-type main-sequence Milky Way. Gaia DR-1 and DR-2 have already yielded ∼4 a wealth of data on this front (Gaia Collaboration et al. stars. Hence, the above numbers for Nsurv must 2016, 2018). As a rigorous assessment of the origin be reduced by roughly an order of magnitude when of hypervelocity stars requires knowledge of their to- we restrict ourselves to solar-type stars. tal velocity (Brown 2015), it is necessary to measure 3. The calcium triplet (CaT) is used by the Gaia mis- both their radial and tangential components. Gaia sion to estimate the radial velocities by measuring DR-2 has provided radial velocity information about 7 the Doppler shift. The spectrograph has a band- million stars (Katz et al. 2019). The number of stars pass of 847-874 nm (Cropper et al. 2018), thereby with data available regarding their proper motion (i.e., providing a “buffer” of 8 nm because the longest tangential velocity) is 1.3 109 (Lindegren et al. ∼ ∼ × CaT spectral line is manifested at 866.2 nm. Stel- 2018). A number of studies have already combed lar engines moving at radial velocities of 0.01- through this sample to unearth evidence for hyperve- 0.1 c would yield Doppler shifts of 8.5-85∼ nm locity stars (Bromley et al. 2018; Hattori et al. 2018; at the CaT wavelength(s). Thus, the∼ Gaia spec- Marchetti et al. 2019; Boubert et al. 2019; Caffau et al. trograph is liable to “reject” such stellar engines 2020; Li et al. 2020). when the shift falls outside its bandpass. In prin- To the best of our knowledge, the fastest hyperveloc- ciple, however, high tangential velocities can be ity star from any survey is S5-HVS1 from the Southern inferred from measurements of the proper motions Stellar Stream Spectroscopic Survey with a velocity of and parallaxes (Shen et al. 2018; Scholz 2018). 6 10−3 c (Koposov et al. 2020), while HVS 22 from the∼ Multiple× Mirror Telescope survey exhibits a similar ve- 4. Last, but not least, (12) is expected to yield an locity of 5 10−3 c (Kreuzer et al. 2020). Among the accurate picture only when the stellar engines are Gaia DR-2∼ catalog,× one of the fastest objects is the can- uniformly distributed in both the spatial and tem- didate Gaia DR-2 6097052289696317952, which appears poral realms. However, it will take . 1 Myr for the to have a tangential velocity of 5.4 10−3 c (Scholz ∼ × 2018). Other analyses of the Gaia DR-2 data have seem- 4 ingly identified stars with velocities (tangential and/or https://www.cosmos.esa.int/web/gaia/faqs 6

stellar engines to exit the 1 kpc region spanned these putative megastructures are operating at veloci- by Gaia DR2 once their∼ maximal velocity is at- ties that fall within the bounds of known hypervelocity tained, and they will eventually cross the entire stars. In this setting, distinguishing between them and Milky Way in 10 Myr. Hence, it is not straight- naturally occurring hypervelocity stars would be an ex- forward a priori∼ to derive proper constraints on tremely challenging endeavour. We will not explore this the unknown parameter(s) unless new stellar en- scenario further because it calls for additional assump- gines are being continually “injected” into the do- tions about the preferred trajectories of stellar engines, main encompassed by surveys like Gaia. and this requires an understanding of the motives of pu- tative ETIs, which is wholly unknown. Bearing these caveats in mind, let us adopt a rough 6 Before moving ahead, we note that stellar engines are estimate of Nsurv 10 with the express purpose of tak- −9 −2 ∼ potentially capable of accelerating at 10 m s ing the argument further; we emphasize that this value (Caplan 2019; Svoronos 2020). In contrast,∼ the cen- is fiducial. By plugging this choice into (12), we obtain tripetal acceleration of the Sun is 10−10 ms−2. Thus, ∼ −1 in principle, detecting anomalously high stellar acceler- −6 Nsurv ations might also be indicative of stellar engine activity, fT fSE < 10 . (13) · 106 although we caution that the ratio of the two accelera-   tions (i.e., an order of magnitude) is not strikingly large. After the Gaia mission is complete, assuming that no stellar engines are found, we may end up with the poten- 4. CONCLUSION −8 tially tightest limit of fT fSE < 10 for speeds & 0.01 c · We examined various mechanisms for the ejection of under optimal circumstances, i.e., when the subtleties stars at high speeds, and concluded that stars ejected in described in points #3 and #4 are set aside. As these the Milky Way are very unlikely to attain speeds over constraints from surveys might prove to be quite strin- & 104 km/s (0.03 c) by any known natural astrophysical gent, the implications are elucidated below. phenomena. Next, we considered some proposed designs for stellar engines, i.e., propulsion systems engineered by • The first possibility is that fT is exceptionally advanced ETIs to accelerate stars, which were conceived small. In this case, the prevalence of human- by Olaf Stapledon and Fritz Zwicky (among others) in level ETIs would be commensurately low. This the mid-20th century. We argued that speeds of 0.01- could arise due to any number of evolution- 0.1 c may be potentially achievable by stellar engines∼ ary bottlenecks ranging from abiogenesis to com- under optimal circumstances. plex multicellularity to technological intelligence Based on the above premises, we examined current (Smith & Szathmary 1995; Lingam & Loeb 2021). surveys for hypervelocity stars including the recent Gaia DR-2 sample. In light of existing studies, we concluded • In the second case, fSE may be minuscule, but that no stars have been conclusively identified that pos- fT may have a moderate magnitude. There are, however, different scenarios at play here. On the sess velocities of & 0.01 c. Taken at face value, stellar one hand, ETIs might have a short technological engines moving at sub-relativistic speeds appear to be lifetime and become extinct before they reach the quite rare but placing stringent constraints on their like- stage where they can build stellar engines. On lihood is rendered difficult because of both instrumental the other hand, it could very well be that weakly limitations and lack of knowledge about their spatio- relativistic stellar engines have hidden engineering temporal distribution. It might be possible that fewer than one in 106 stars is propelled to speeds & 0.01 c by obstacles that render their construction impossi- ∼ ble, or that ETIs possess the capability to build stellar engines, although this statement is not definitive them but opt not to do so for other reasons. in light of the attendant uncertainties. In the future, it would seem worthwhile to pursue • In the third outcome, both fT and fSE are both the search for stars with anomalously high velocities very small. This situation does not warrant sepa- (namely & 0.03 c). This strategy is advantageous for two rate explication, because it represents an amalga- chief reasons. First, it does not necessitate any new re- mation of the above two points. sources, because it can readily piggyback on astrometric surveys like Gaia. Second, if we do stumble upon stars To differentiate between, and indeed shed light on, moving at such anomalously high speeds, their origin the diverse outcomes demarcated above, the practical would be of great interest and significance irrespective importance and necessity of carrying out searches for of whether they have an artificial basis or not. biosignatures and technosignatures on multiple fronts is In this regard, this search exemplifies the phi- self-evident (Frank 2018; Haqq-Misra et al. 2020). losophy underpinning the aptly named “ First There is a fourth option that deserves to be mentioned Law of SETI Investigations” proposed by the at this juncture. In theory, it is possible that stellar en- late Freeman Dyson: “Every search for alien gines might already exist in the Milky Way, but that civilizations should be planned to give interest- 7 ing results even when no aliens are discovered.” ACKNOWLEDGMENTS (NASA Technosignatures Workshop Participants 2018), which itself echoes the earlier sentiments es- We are grateful to the two reviewers for their positive poused by Frank Drake in his neglected early treatise and insightful reports, which helped improve the paper. (Drake 1965, pg. 342): “Thus, any project aimed at This work was supported in part by the Breakthrough the detection of intelligent extraterrestrial life should Prize Foundation, Harvard University’s Faculty of Arts simultaneously conduct more conventional research.”5 and Sciences, and the Institute for Theory and Compu- Thus, in each of these respects alongside a few others, tation (ITC) at Harvard University. this search for technosignatures may score highly on the axes of merit adumbrated in Sheikh (2020).

REFERENCES Badescu, V., & Cathcart, R. B. 2000, J. Br. Interplanet. Cleland, C. E. 2019, Cambridge Astrobiology, Vol. 11, The Soc., 53, 297 Quest for a Universal Theory of Life: Searching for Life —. 2006, Acta Astronaut., 58, 119, As We Don’t Know It (Cambridge: Cambridge University doi: 10.1016/j.actaastro.2005.09.005 Press) Balbi, A., & Tombesi, F. 2017, Sci. Rep., 7, 16626, Criswell, D. R. 1985, Solar System Industrialization: doi: 10.1038/s41598-017-16110-0 Implications for Interstellar Migrations, ed. B. R. Finney Blaauw, A. 1961, Bull. Astron. Inst. Netherlands, 15, 265 & E. M. Jones (Berkeley: University of California Press), Boehle, A., Ghez, A. M., Sch¨odel, R., et al. 2016, 50–87 Astrophys. J., 830, 17, doi: 10.3847/0004-637X/830/1/17 Cropper, M., Katz, D., Sartoretti, P., et al. 2018, Astron. Boersma, J. 1961, Bull. Astron. Inst. Netherlands, 15, 291 Astrophys., 616, A5, doi: 10.1051/0004-6361/201832763 Boubert, D., Strader, J., Aguado, D., et al. 2019, Mon. Not. Darbha, S., Coughlin, E. R., Kasen, D., & Quataert, E. R. Astron. Soc., 486, 2618, doi: 10.1093/mnras/stz253 2019, Mon. Not. R. Astron. Soc., 482, 2132, Bradbury, R. J., Cirkovic, M. M., & Dvorsky, G. 2011, J. doi: 10.1093/mnras/sty2786 Br. Interplanet. Soc., 64, 156 Drake, F. D. 1965, The Radio Search for Intelligent Extraterrestrial Life, ed. G. Mamikunian & M. H. Briggs Bromley, B. C., Kenyon, S. J., Brown, W. R., & Geller, (Oxford: Pergamon Press), 323–345 M. J. 2018, Astrophys. J., 868, 25, Du, C., Li, H., Yan, Y., et al. 2019, Astrophys. J. Suppl. doi: 10.3847/1538-4357/aae83e Ser., 244, 4, doi: 10.3847/1538-4365/ab328c Bromley, B. C., Kenyon, S. J., Geller, M. J., et al. 2006, Dyson, F. 1966, The Search for Extraterrestrial Technology, Astrophys. J., 653, 1194, doi: 10.1086/508419 ed. R. Marshak, J. Blaker, & H. Bethe (New York: Brown, W. R. 2015, Annu. Rev. Astron. Astrophys., 53, 15, Interscience Publishers), 641–655 doi: 10.1146/annurev-astro-082214-122230 Dyson, F. J. 1960, Science, 131, 1667, Bussard, R. W. 1960, Astronaut. Acta, 6, 179 doi: 10.1126/science.131.3414.1667 Caffau, E., Monaco, L., Bonifacio, P., et al. 2020, Astron. Erkal, D., Boubert, D., Gualandris, A., Evans, N. W., & Astrophys., 638, A122, Antonini, F. 2019, Mon. Not. R. Astron. Soc., 483, 2007, doi: 10.1051/0004-6361/202038057 doi: 10.1093/mnras/sty2674 Caplan, M. E. 2019, Acta Astronaut., 165, 96, Evans, F. A., Renzo, M., & Rossi, E. M. 2020, Mon. Not. doi: 10.1016/j.actaastro.2019.08.027 R. Astron. Soc., 497, 5344, doi: 10.1093/mnras/staa2334 Cassibry, J., Cortez, R., Stanic, M., et al. 2015, J. Fogg, M. J. 1989, Specul. Sci. Technol., 12, 153 Spacecraft Rockets, 52, 595, doi: 10.2514/1.A32782 Forbes, J. C., & Loeb, A. 2018, Mon. Not. R. Astron. Soc., ´ Cirkovi´c, M. M. 2015, Serbian Astron. J., 191, 1, 479, 171, doi: 10.1093/mnras/sty1433 doi: 10.2298/SAJ1591001C Forgan, D. H. 2013, J. Br. Interplanet. Soc., 66, 144 Cirkovic, M. M. 2018, The Great Silence: Science and Forward, R. L. 1991, J. Spacecraft Rockets, 28, 606, Philosophy of Fermi’s Paradox (Oxford: Oxford doi: 10.2514/3.26287 University Press) Fragione, G., & Gualandris, A. 2019, Mon. Not. R. Astron. Soc., 489, 4543, doi: 10.1093/mnras/stz2451 Frank, A. 2018, Light of the Stars: Alien Worlds and the 5 On a broader note, the significance of seeking “anomalies” sensu lato is garnering deeper appreciation in the adjacent realm of Fate of the Earth (New York: W. W. Norton & Co.) biosignatures as well (Cleland 2019, Chapter 8). Freitas, R. A., J. 1983, J. Br. Interplanet. Soc., 36, 501 8

Gaia Collaboration, Brown, A. G. A., Vallenari, A., et al. Lacki, B. C., Brzycki, B., Croft, S., et al. 2020, Astrophys. 2016, Astron. Astrophys., 595, A2, J. Suppl. Ser., arXiv:2006.11304. doi: 10.1051/0004-6361/201629512 https://arxiv.org/abs/2006.11304 —. 2018, Astron. Astrophys., 616, A1, Leonard, P. J. T. 1991, Astron. J., 101, 562, doi: 10.1051/0004-6361/201833051 doi: 10.1086/115704 Generozov, A., & Madigan, A.-M. 2020, Astrophys. J., 896, Li, J., Jia, S., Gao, Y., et al. 2020, Res. Astron. Astrophys., 137, doi: 10.3847/1538-4357/ab94bc 20, 042, doi: 10.1088/1674-4527/20/3/42 Ginsburg, I., Loeb, A., & Wegner, G. A. 2012, Mon. Not. Lindegren, L., Hern´andez, J., Bombrun, A., et al. 2018, R. Astron. Soc., 423, 948, Astron. Astrophys., 616, A2, doi: 10.1111/j.1365-2966.2012.20930.x doi: 10.1051/0004-6361/201832727 Gray, R. H. 2020, Astron. J., 159, 228, Lingam, M., Ginsburg, I., & Bialy, S. 2019, Astrophys. J., doi: 10.3847/1538-3881/ab792b 877, 62, doi: 10.3847/1538-4357/ab1b2f Gualandris, A., & Portegies Zwart, S. 2007, Mon. Not. R. Lingam, M., & Loeb, A. 2019a, Astrobiology, 19, 28, Astron. Soc., 376, L29, doi: 10.1089/ast.2018.1936 doi: 10.1111/j.1745-3933.2007.00280.x —. 2019b, Rev. Mod. Phys., 91, 021002, Guillochon, J., & Loeb, A. 2015, Astrophys. J., 806, 124, doi: 10.1103/RevModPhys.91.021002 doi: 10.1088/0004-637X/806/1/124 —. 2020, Astrophys. J., 894, 36, Hansen, C. J., Kawaler, S. D., & Trimble, V. 2004, Stellar doi: 10.3847/1538-4357/ab7dc7 interiors : physical principles, structure, and evolution, —. 2021, Life in the Cosmos: From Biosignatures to 2nd edn. (New York: Springer-Verlag) Technosignatures (Cambridge: Harvard University Press). Haqq-Misra, J., Kopparapu, R. K., & Schwieterman, E. https://www.hup.harvard.edu/catalog.php?isbn=9780674987579 2020, Astrobiology, 20, 572, doi: 10.1089/ast.2019.2154 Loeb, A., & Guillochon, J. 2016, Ann. Math. Sci. Appl., 1, Hattori, K., Valluri, M., Bell, E. F., & Roederer, I. U. 2018, 183, doi: 10.4310/AMSA.2016.v1.n1.a5 Astrophys. J., 866, 121, doi: 10.3847/1538-4357/aadee5 Long, K. F. 2011, Deep Space Propulsion: A Roadmap to Hein, A. M., Pak, M., Putz, D., Buhler, C., & Reiss, P. Interstellar Flight (New York: Springer-Verlag), 2012, J. Br. Interplanet. Soc., 65, 119 doi: 10.1007/978-1-4614-0607-5 Hills, J. G. 1988, Nature, 331, 687, doi: 10.1038/331687a0 Marchetti, T., Rossi, E. M., & Brown, A. G. A. 2019, Mon. Hippke, M., Leyland, P., & Learned, J. G. 2018, Acta Not. R. Astron. Soc., 490, 157, Astronaut., 151, 32, doi: 10.1016/j.actaastro.2018.05.037 doi: 10.1093/mnras/sty2592 Hoang, T. 2017, Astrophys. J., 847, 77, NASA Technosignatures Workshop Participants. 2018, doi: 10.3847/1538-4357/aa88a7 arXiv e-prints, arXiv:1812.08681. Hoang, T., Lazarian, A., Burkhart, B., & Loeb, A. 2017, https://arxiv.org/abs/1812.08681 Astrophys. J., 837, 5, doi: 10.3847/1538-4357/aa5da6 Neunteufel, P. 2020, Astron. Astrophys., 641, A52, Hoang, T., & Loeb, A. 2017, Astrophys. J., 848, 31, doi: 10.3847/1538-4357/aa8c73 doi: 10.1051/0004-6361/202037792 Holmlid, L., & Zeiner-Gundersen, S. 2020, Acta Astronaut., Novikov, I. D., & Thorne, K. S. 1973, in Black Holes (Les 175, 32, doi: 10.1016/j.actaastro.2020.05.034 Astres Occlus), 343–450 Hooper, D. 2018, Phys. Dark Universe, 22, 74, Oh, S., & Kroupa, P. 2016, Astron. Astrophys., 590, A107, doi: 10.1016/j.dark.2018.09.005 doi: 10.1051/0004-6361/201628233 Kardashev, N. S. 1964, Sov. Astron., 8, 217 O’Leary, R. M., & Loeb, A. 2008, Mon. Not. R. Astron. Katz, D., Sartoretti, P., Cropper, M., et al. 2019, Astron. Soc., 383, 86, doi: 10.1111/j.1365-2966.2007.12531.x Astrophys., 622, A205, Pacetti, E., Balbi, A., Lingam, M., Tombesi, F., & doi: 10.1051/0004-6361/201833273 Perlman, E. 2020, Mon. Not. R. Astron. Soc., 498, 3153, Kenyon, S. J., Bromley, B. C., Geller, M. J., & Brown, doi: 10.1093/mnras/staa2535 W. R. 2008, Astrophys. J., 680, 312, doi: 10.1086/587738 Poveda, A., Ruiz, J., & Allen, C. 1967, Bol. Obs. Koposov, S. E., Boubert, D., Li, T. S., et al. 2020, Mon. Tonantzintla Tacubaya, 4, 86 Not. R. Astron. Soc., 491, 2465, Rossi, E. M., Kobayashi, S., & Sari, R. 2014, Astrophys. J., doi: 10.1093/mnras/stz3081 795, 125, doi: 10.1088/0004-637X/795/2/125 Kreuzer, S., Irrgang, A., & Heber, U. 2020, Astron. Sari, R., Kobayashi, S., & Rossi, E. M. 2010, Astrophys. J., Astrophys., 637, A53, doi: 10.1051/0004-6361/202037747 708, 605, doi: 10.1088/0004-637X/708/1/605 9

Scholz, R.-D. 2018, Res. Notes AAS, 2, 211, Tarter, J. C., Agrawal, A., Ackermann, R., et al. 2010, in doi: 10.3847/2515-5172/aaef89 Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, Vol. 7819, Instruments, Sheikh, S. Z. 2020, Int. J. Astrobiol., 19, 237, Methods, and Missions for Astrobiology XIII, 781902, doi: 10.1017/S1473550419000284 doi: 10.1117/12.863128 Tauris, T. M. 2015, Mon. Not. R. Astron. Soc., 448, L6, Shen, K. J., Boubert, D., G¨ansicke, B. T., et al. 2018, doi: 10.1093/mnrasl/slu189 Astrophys. J., 865, 15, doi: 10.3847/1538-4357/aad55b Tout, C. A., Pols, O. R., Eggleton, P. P., & Han, Z. 1996, Shkadov, L. M. 1987, in Brighton International Mon. Not. R. Astron. Soc., 281, 257, doi: 10.1093/mnras/281.1.257 Astronautical Federation Congress, Vol. 38th Congress of Tsiolkovsky, K. E. 1903, The Science Review, 5 IAF, paper IAA–87–613 Winterberg, F. 2019, Acta Astronaut., 161, 241, Shkadov, L. M. 1988, Astronomicheskii Vestnik, 22, 333 doi: 10.1016/j.actaastro.2019.05.017 Worden, S. P., Drew, J., Siemion, A., et al. 2017, Acta Shklovskii, I. S., & Sagan, C. 1966, Intelligent life in the Astronaut., 139, 98, doi: 10.1016/j.actaastro.2017.06.008 universe (San Francisco: Holden Day) Wright, J. T. 2020, Serbian Astron. J., 200, 1, Smith, J. M., & Szathmary, E. 1995, The Major Transitions doi: 10.2298/SAJ2000001W Wright, J. T., Kanodia, S., & Lubar, E. 2018, Astron. J., in Evolution (Oxford: Oxford University Press) 156, 260, doi: 10.3847/1538-3881/aae099 Stapledon, O. 1937, Star Maker (London: Methuen) Wright, J. T., Mullan, B., Sigurdsson, S., & Povich, M. S. 2014, Astrophys. J., 792, 26, Svoronos, A. A. 2020, Acta Astronaut., 176, 306, doi: 10.1088/0004-637X/792/1/26 doi: 10.1016/j.actaastro.2020.07.005 Wurden, G. A., Weber, T. E., Turchi, P. J., et al. 2016, J. Tarter, J. 2001, Annu. Rev. Astron. Astrophys., 39, 511, Fusion Energy, 35, 123, doi: 10.1007/s10894-015-0034-1 Yu, Q., & Tremaine, S. 2003, Astrophys. J., 599, 1129, doi: 10.1146/annurev.astro.39.1.511 doi: 10.1086/379546 Tarter, J. C. 2007, Highlights of Astronomy, 14, 14, Zwicky, F. 1957, Morphological astronomy (Berlin: doi: 10.1017/S1743921307009829 Springer-Verlag), doi: 10.1007/978-3-642-87544-1