Summary Critique of Discovery of Spherules in CNEOS 2014-01-08 arxiv.org
14,002 words - PDF document - View PDF document
One Line
This critique questions the validity of an interstellar bolide, pointing out uncertainties in measurements, lack of statistical evidence, and proposing alternative explanations.
Slides
Slide Presentation (11 slides)
Key Points
- The critique questions the validity of the claim that the 2014-01-08 bolide was interstellar.
- It is concluded that if the bolide were interstellar, only a small fraction of it would have survived entry into the Earth's atmosphere.
- The accuracy of determining the location of the spherules is called into question due to uncertainties in reported latitude and longitude.
- The composition and origin of the spherules are disputed, with arguments suggesting a solar system origin and similarities to other cosmic spherules and coal ash.
- The association of the spherules with the assumed path of the bolide is deemed inconclusive due to a lack of statistical evidence.
- The importance of properly conducting and reviewing scientific research is emphasized, with hopes that the critique will prevent flawed work from entering scientific literature.
Summaries
37 word summary
This critique challenges the claim of an interstellar bolide, citing uncertainties in velocity measurements and survival rate. It questions spherules' location, composition, and origin. The association lacks statistical evidence. The critique finds flaws and proposes alternative explanations.
78 word summary
This critique challenges the claim that a bolide was interstellar, citing uncertainties in velocity measurements and the survival rate of interstellar objects entering Earth's atmosphere. It questions the accuracy of determining the spherules' location and disputes claims about their composition and origin. The association between the spherules and the bolide's path lacks statistical evidence. The critique emphasizes the importance of scientific research and peer review, finding flaws in the arguments for an interstellar origin. Alternative explanations are proposed.
163 word summary
This critique questions the validity of a claim made by Loeb et al. that a bolide was interstellar. Uncertainties in velocity measurements make it difficult to determine the bolide's origin, and it is suggested that if it were interstellar, only a small fraction would have survived entry into the Earth's atmosphere. The accuracy of determining the location of the spherules is also called into question due to uncertainties in reported latitude, longitude, and distance between the fireball and a seismometer. The critique disputes claims made by Loeb et al. regarding the composition and origin of the spherules, suggesting a solar system origin based on Fe isotopic ratios. The association of the spherules with the bolide's path is deemed inconclusive due to a lack of statistical evidence. The critique emphasizes the importance of proper scientific research and peer review. Overall, the arguments made by Loeb et al. regarding the interstellar origin of the spherules are found to be flawed, and alternative explanations are proposed.
321 word summary
A critique of a manuscript by Loeb et al. titled "Discovery of Spherules of Likely Extrasolar Composition in the Pacific Ocean Site of the CNEOS 2014-01-08 (IM1) Bolide" is presented. The critique questions the validity of the claim that the bolide was interstellar, pointing out uncertainties in velocity measurements that make it difficult to determine its origin. It is concluded that if the bolide were interstellar, only a small fraction of it would have survived entry into the Earth's atmosphere and most likely it would have completely vaporized.
The critique also calls into question the accuracy of determining the location of the spherules, pointing out uncertainties in reported latitude and longitude as well as the distance between the fireball and a seismometer on Manus Island. These uncertainties make it challenging to define a reasonable search area for the spherules.
Furthermore, the critique disputes the various claims made by Loeb et al. regarding the composition and origin of the spherules. It argues that the Fe isotopic ratios indicate a solar system origin with high probability. The authors also point out that similar patterns have been observed in other cosmic spherules and even anthropogenic coal ash, undermining the assertion that the spherules have unique chemical compositions. The association of the spherules with the assumed path of the bolide is deemed inconclusive due to a lack of statistical evidence.
The critique concludes by emphasizing the importance of properly conducting and reviewing scientific research. It expresses hope that their critique will serve as a resource for potential peer reviewers and prevent flawed work from entering the scientific literature.
In summary, the critique highlights multiple flaws in the arguments made by Loeb et al. regarding the interstellar origin of spherules recovered from the seafloor. The evidence presented does not support their claims, and alternative explanations for the composition and origin of the spherules are proposed. The importance of rigorous scientific research and peer review is emphasized.
676 word summary
A critique of a manuscript by Loeb et al. titled "Discovery of Spherules of Likely Extrasolar Composition in the Pacific Ocean Site of the CNEOS 2014-01-08 (IM1) Bolide" is presented. The manuscript claimed to have discovered interstellar spherules and made various assertions about their composition and origin. However, the critique highlights multiple flaws in the arguments presented by Loeb et al.
The critique begins by questioning the validity of the claim that the 2014-01-08 bolide was interstellar. The authors argue that uncertainties in velocity measurements make it difficult to determine its origin, and there is a possibility that it originated from our Solar System. The evidence for an interstellar origin is deemed unconvincing.
Next, the critique examines whether any of the bolide would have survived entry into the Earth's atmosphere. It is concluded that if the bolide were interstellar, only a small fraction of it would have survived entry, and most likely it would have completely vaporized.
The accuracy of determining the location of the spherules is also called into question. The authors point out uncertainties in the reported latitude and longitude, as well as the distance between the fireball and a seismometer on Manus Island. These uncertainties make it challenging to define a reasonable search area for the spherules.
Furthermore, the critique addresses the various claims made by Loeb et al. regarding the composition and origin of the spherules. It is argued that the Fe isotopic ratios indicate a solar system origin with a high probability. The authors also dispute the assertion that the spherules have unique chemical compositions, pointing out that similar patterns have been observed in other cosmic spherules and even anthropogenic coal ash. The association of the spherules with the assumed path of the bolide is deemed inconclusive due to a lack of statistical evidence.
Finally, the critique concludes by emphasizing the importance of properly conducting and reviewing scientific research. The authors express the hope that their critique will serve as a resource for potential peer reviewers and prevent the flawed work from entering the scientific literature.
In summary, the critique highlights multiple flaws in the arguments made by Loeb et al. regarding the interstellar origin of spherules recovered from the seafloor. The evidence presented does not support their claims, and alternative explanations for the composition and origin of the spherules are proposed. The importance of rigorous scientific research and peer review is emphasized.
A recent study by Loeb et al. claims that the spherules they found in Papua New Guinea are likely of interstellar origin. However, upon closer examination, their arguments do not hold up. First, the authors suggest that the 2014-01-08 bolide was interstellar, but this is highly unlikely as interstellar objects would have completely vaporized upon entry into our atmosphere. Additionally, the number of ablation spherules produced would have been small and spread out, greatly outnumbered by background spherules. Furthermore, there is no evidence to support the claim that the spherules were concentrated in the path of the bolide, as the trajectory of the bolide is poorly known. The triple compound spherule S21, which the authors suggest formed in an airburst, is more likely to have originated in an impact plume tens of thousands of years ago. The composition of the spherules, including the "BeLaU" spherules enriched in Be, La, and U, is not unique and can be found in other cosmic spherules and anthropogenic coal ash. The Be abundance in the spherules is also likely due to terrestrial contamination rather than cosmic-ray spallation. Finally, the Fe isotopes in the spherules are consistent with a terrestrial origin rather than vaporization during entry into our atmosphere. Overall, the arguments put forth by Loeb et al. do not withstand scrutiny and fail to provide convincing evidence for an interstellar origin of the spherules. The authors did not consider alternative hypotheses or properly design their experiment to avoid inconclusive results, demonstrating confirmation bias. It is clear that the spherules are part of the background of cosmic spherules deposited on the seafloor over thousands of years and have undergone chemical reactions with seawater.
Raw indexed text (84,044 chars / 14,002 words / 1,010 lines)
Critique of arXiv submission 2308.15623, “Discovery of Spherules of Likely
Extrasolar Composition in the Pacific Ocean Site of the CNEOS 2014-01-08 (IM1)
Bolide” by A. Loeb et al.
Steve Desch * and Alan Jackson **
*
Arizona State University, School of Earth and Space Exploration, PO Box 876004, Tempe,
Arizona, 85287-6004
**
Towson University, Department of Physics, Astronomy and Geosciences, 8000 York Rd.,
Towson, Maryland, 21252
Abstract
Recently a manuscript by Loeb et al. was uploaded to arXiv 1 (preprint 2308.15623) that
asserted that the CNEOS bolide 2014-01-08 was interstellar; that spherules recovered from the
seafloor near the airburst were associated with this bolide; that they had Fe isotopic ratios
indicating origin as micrometeorites; that they had unusual chemical compositions enriched in
Be, La and U, never seen before in micrometeorite spherules; that these compositions were
formed in the magma ocean stage of a differentiated extrasolar planet; and that the Be
abundance reflected passage through the interstellar medium. Despite not being peer-
reviewed, this uploaded manuscript has been reported by media outlets as “published”, and its
conclusions have been widely distributed as fact. The purpose of this manuscript is to provide
potential peer reviewers and the general public with an appreciation of the multiple fatal flaws
with the manuscript’s arguments. We discuss the published evidence that the 2014-01-08
bolide is not interstellar. We show that there is no statistical spatial correlation of a chemical
signature or even number of recovered spherules with the 2014-01-08 bolide. We demonstrate
that the Fe isotopic ratios decisively indicate an origin in our Solar System, with > 99.995%
probability. We demonstrate that the unusual enrichments in La, U, etc., have in fact been
observed in micrometeorites before and attributed to terrestrial contamination; and that the
Be abundances are similarly consistent with those of ferromanganese nodules, after reacting
with sea water. Far from being exotic particles from an extrasolar planet, the spherules
collected and analyzed by Loeb et al. appear to be just like those found around the world, with
a Solar System origin and compositions modified by tens of thousands of years residence at the
ocean bottom.
1
https://arxiv.org/pdf/2308.15623.pdf
1I. Introduction
On August 29, 2023, Loeb et al. uploaded a manuscript, “Discovery of Spherules of Likely
Extrasolar Composition in the Pacific Ocean Site of the CNEOS 2014-01-08 (IM1) Bolide” to the
arXiv preprint server 2 . In it, the authors reported on their recovery of spherules from seafloor
near Papua New Guinea, and their subsequent analysis. This manuscript claimed that these
spherules were likely of interstellar origin, the first such samples ever recovered. This claim, if
verified, would represent a significant scientific advance, allowing the first laboratory analysis
of macroscopic quantities of material from another solar system, enabling direct tests of
theories about how stars and planets are born in our Galaxy.
Because of the boldness of the claim, it might be expected that Loeb et al. exercised due
diligence to eliminate alternative, more-prosaic explanations for the data, and would report
their results carefully and deliberately. They did not. Loeb et al. hardly considered, and did not
test, the simplest alternative hypothesis: the spherules they found were of a common and well-
understood type found worldwide, from Solar System asteroids. Apparently, the authors did
not seek expert opinion, and posted the “discovery” manuscript before it had been peer-
reviewed and accepted, or even submitted to a science journal. The Galileo Project headed by
Loeb did, however, put out a press release the same day touting the manuscript 3 , picked up by
national media such as USA Today 4 .
While the manuscript focused on the possibility that the spherules are from a natural object, it
also suggested they “may reflect an extraterrestrial technological origin.” The framing of the
USA Today article advances this by repeating claims that the spherules are “unmatched to any
existing alloys.” About the paper, on August 31 Loeb said to Salon 5 : “The ideal scenario is that in
addition to tiny fragments, we would find a piece of an advanced technological device, like the
hundredth version of the iPhone… I would love to press a button on such an object.” Not
coincidentally, Loeb’s book, Interstellar: The Search for Extraterrestrial Life and Our Future in
the Stars, was published the same day the manuscript and press releases were posted.
The net result is that Loeb et al. have affixed in the public consciousness the unsubstantiated—
and probably false—idea that they have recovered interstellar spherules; and they have
irresponsibly led the public to conflate “interstellar” with “extraterrestrial technology.”
Ideally, the peer review process would identify the fatal flaws in their conclusions, finding no
evidence for an interstellar origin, the manuscript would not be published in a peer-reviewed
journal or given legitimacy, and it would be forgotten. However, by posting the manuscript on
the arXiv preprint server before peer review, by issuing premature press releases, and by
2
https://arxiv.org/pdf/2308.15623.pdf
https://projects.iq.harvard.edu/galileo/news/spherule-analysis-finds-evidence-extrasolar-composition
4
https://www.usatoday.com/story/news/nation/2023/08/29/metallic-spheres-interstellar-origin-avi-loeb-
finds/70699783007/
5
https://www.salon.com/2022/08/31/cneos-2014-01-08-loeb-alien-asteroid/
3
2making very public and unsubstantiated claims, Loeb et al. have circumvented the normal peer-
review process. This demands a response to their points in an equally public venue.
This article represents our attempt to make public the many clear and fatal flaws in the
arguments made by Loeb et al. Far from being interstellar objects, their own data demand the
spherules be from our Solar System. Although the authors do not recognize it, the spherules are
of a common type, and evidently have reacted with seawater for tens of thousands of years,
explaining their enrichments in Be, La, and U, among other elements. They are probably not
even associated with the 2014-01-08 bolide, which was probably not interstellar anyway.
We hope this manuscript will serve as a resource for multiple audiences. For the general public,
we hope to communicate in a clear and accessible way that Loeb et al. have not discovered
objects from beyond our Solar System, let alone anything technological. For Loeb and co-
authors, we hope to provide a referee report that identifies the major flaws that should be
rectified, if that is even possible, before their work should be published. A pillar of the scientific
method is self-correction—the refinement or even abandonment of theories that simply don’t
match observations. However, anticipating that Loeb et al. will not abandon their premature
and unwarranted claims, we hope that this manuscript will serve as a resource for potential
peer reviewers, so that this deeply flawed work does not enter the scientific literature.
Below we summarize the claims made by Loeb et al. We provide an overview of the major
flaws, then discuss each one in turn.
II. Summary of the manuscript by Loeb et al., and outline of this article
We begin with a summary of the claims made by Loeb et al., and the logic of their argument.
We then summarize the critiques of each point and outline this paper.
Summary of the claims by Loeb et al.
The manuscript describes the recovery and analysis of spherules from the ocean floor in the
vicinity of where the authors believe the Center for Near-Earth Object Studies (CNEOS) 2014-
01-08 bolide 6 disintegrated and deposited material. A magnetic sled was employed, resulting in
the collection of about 700 spherules, of which 57 were analyzed for bulk and trace chemistry,
including 11 for Fe isotopic compositions. It is argued that a subset of at least 5 of these
spherules, with distinctive enrichments in Be, La, U, etc., which they term “BeLaU” spherules,
are from the 2014 bolide and are of extrasolar origin. This claim of extrasolar material is based
on the following chain of logic.
1) The authors claim that the 2014-01-08 bolide was interstellar in origin, moving too fast
to have been bound in orbit around the Sun when it struck the Earth.
6
https://cneos.jpl.nasa.gov/fireballs/
32) The authors assert that significant quantities of material from this bolide would have
survived atmospheric entry, to be deposited on the seafloor, mostly as millimeter-sized
ablation spherules.
3) The authors claim to know the position where these spherules would be deposited, to
within an area 1 km wide by 10 km long, the bolide’s assumed path.
4) The authors claim that a statistically significant excess of spherules (twice as many) were
collected along the assumed path, compared to two control regions far from the path,
implying that a large fraction (about half) of the spherules collected along the assumed
path are from the 2014 bolide.
5) The authors measured Fe isotopic ratios in 11 spherules, including 3 “BeLaU” spherules,
and find all of them lie along the terrestrial fractionation line within errors (< 1‰), and
thus show fractionations consistent with Fe loss by vaporization during entry.
6) One of these “BeLaU” spherules (“S21”) was found to be a compound spherule, three
spherules stuck together while molten. The authors present a back-of-the-envelope
calculation to suggest this is also consistent with formation in a fireball.
7) The authors measured elemental compositions of 57 spherules and found 5 “BeLaU”
spherules with distinctive enrichments in Be, La, U, and other elements (and depletions
in Mg, Ca, etc.) relative to CI chondrites. The authors assert that this compositional
pattern is not found in any terrestrial samples or other cosmic spherules. By implication
the “BeLaU” spherules would have to be extrasolar.
8) The authors claim the “BeLaU” spherules are associated only with the assumed path of
the bolide and are absent from regions far from the assumed path, and therefore
associated with the 2014 fireball.
9) The authors argue that the “BeLaU” compositions would be (uniquely?) consistent with
formation in the magma ocean of a differentiated (extrasolar?) planet.
10) The authors suggest the high abundance of Be, which is produced by spallation of atoms
in meteoroids by Galactic cosmic rays (GCRs), is indicative of GCR irradiation during
passage through interstellar space.
To claim an interstellar origin for the spherules they collected, all the links in the chain of logic
above would have to withstand scrutiny.
4Summary of the critiques to these claims
In fact, not a single link above withstands scrutiny.
1) The 2014-01-08 bolide probably wasn’t interstellar. Uncertainties in the velocities of
objects in the CNEOS database are not reported, but these can be estimated separately
and they are large. There is a > 0.1% probability the 2014-01-08 bolide is not interstellar,
which sounds small, but in a catalog of about a thousand (Solar System) bolides, the
odds are high that one would be that significantly mismeasured and appear to have
interstellar velocities. 2014-01-08 appears to be that one.
2) If interstellar, practically none of the 2014-01-08 bolide would have survived entry. If it
were traveling at the speeds that were reported (and necessary to be interstellar), then
at least 99.8%, and probably > 99.9999% of it would have vaporized in the atmosphere,
leaving insignificant quantities to be deposited on the seafloor.
3) The location of the bolide’s path is far more uncertain than Loeb et al. realize and also
almost certainly miscalculated. The CNEOS database reports two locations for the 2014-
01-08 fireball, 55 km apart. Even if we pick one location, there are still large inherent
uncertainties in trying to pinpoint the location using sound waves reaching a
seismometer on Manus Island some 90 km away.
4) There is no evidence the density of collected spherules varied across their search region.
The authors did not collect a statistically significant number of spherules from regions
off the assumed path to test this. Even if there were differences, it isn’t clear they could
be attributed to regional concentrations versus differences in collection efficiencies.
5) The Fe isotopic ratios are a smoking gun for the spherules originating in our Solar
System. The fact that all the samples lie within < 1‰ of the terrestrial fractionation line
(including all the “BeLaU” spherules they measured) indicates a > 99.995% probability
that all the spherules (including the “BeLaU” spherules) are from our Solar System. And
while the Fe isotopic ratios may indicate fractionation during atmospheric entry, they
appear more consistent with terrestrial contamination.
6) The triple spherule S21 almost certainly did not form in a fireball event like the 2014
bolide. It requires a denser environment such as an impact plume. The back-of-the-
envelope calculation meant to demonstrate that compound spherules are consistent
with a fireball origin actually is in error by a factor of 10 4 . If it were true, compound
cosmic spherules would be common, which they are not.
7) Despite the authors’ assertions, the distinctive “BeLaU” patterns have been seen before,
in cosmic spherules (of Solar System origin) from the Indian Ocean (Rudraswami et al.
2016) and ancient micrometeorites from Antarctica (van Ginneken et al. 2021). They are
not even dissimilar to (anthropogenic) coal ash (Gallardo 2023).
58) The authors collected too few spherules off the assumed path of the bolide to make any
statistical arguments about the association of “BeLaU” spherules with the path. The fact
that they found no “BeLaU” spherules off the path is meaningless because they
measured only 10 and would expect to find only 1.
9) The “BeLaU” pattern seen in other cosmic spherules has been attributed to terrestrial
contamination (Rudraswami et al. 2016; van Ginneken et al. 2021). Few details were
provided of how this pattern would arise in an extrasolar magma ocean, but this
interpretation has flaws. Contamination by sea water is likely.
10) Finally, the Be abundance seen in these spherules is far too high to be created by
cosmic-ray spallation, but is very similar to that seen in ferromanganese nodules, and
therefore also consistent with reaction with seawater over tens of thousands of years.
Outline
Loeb et al. have presented a hypothesis that many of the spherules they’ve located are
interstellar in origin. For this to be tenable, each of the claims above would have to resist
falsification, and the chain of logic withstand scrutiny. One might expect a few links above to be
weak or hard to verify. Instead, upon minimal investigation, every single link in the chain is
unsubstantiated. In many cases (points 5, 6, and 7 especially), their own data negate their
conclusions.
In the sections below, we investigate each of these points one-by-one. We then close with
conclusions about how to properly search for interstellar micrometeorites.
III. Was the 2014-01-08 bolide interstellar?
The recovery of interstellar materials is, of course, predicated on the 2014-01-08 bolide itself
originating outside the Solar System. The manuscript by Loeb et al. asserts this and repeatedly
assumes that it is, supporting this with the statement: “In 2022 the US Space Command issued
a formal letter to NASA certifying a 99.999% likelihood that the [2014-01-08] object was
interstellar in origin.”
This is not quite true. US Space Command never certified the likelihood was 99.999%. The
March 1, 2022 memo from Lt. Gen. John Shaw of Space Force states 7 : “Dr. [sic] Amir Siraj and
Dr. Abraham Loeb… authored a paper… [that] reported the meteor as originating from an
unbound hyperbolic orbit… with 99.999% confidence,” and continued, “Dr. Mozer confirmed
that the velocity estimate reported to NASA is sufficiently accurate to indicate an interstellar
7
https://lweb.cfa.harvard.edu/~loeb/DoD.pdf
6trajectory.” But US Space Command never stated a confidence level for their trajectory
calculation, and they did not confirm the asserted 99.999% probability.
The number 99.999% does not appear in the refereed papers by Siraj & Loeb (2022a, b). All
citations are to its first mention, in the unrefereed work by Siraj & Loeb (2019, arXiv
1904.07224) 8 . There, the probability was estimated assuming relatively small measurement
uncertainties in each of the velocity components. This underestimates, by orders of magnitude,
the probability the bolide 2014-01-08 originated in our Solar System.
This probability is entirely dependent on the measurement uncertainties in the velocities. The
geocentric velocities reported by the Department of Defense to the Jet Propulsion Laboratory
(JPL) Center for Near-Earth Object Studies (CNEOS) fireballs database for the 2014-01-08 bolide
are 9 : Vx=-3.4 km/s, Vy=-43.5 km/s, and Vz=-10.2 km/s, for a total speed relative to Earth of 44.8
km/s. Considering the time of impact (17 h 05 m 34 s , Jan. 8, 2014) and the rotation of the
geocentric frame, this translates to a heliocentric speed of 61.2 km/s at the time of impact.
Heliocentric speeds > 42.5 km/s, like this one, indicate an object is not gravitationally bound to
the Sun and must be interstellar. It is straightforward to show, though, that if the Vy velocity
were mismeasured and really -19.0 km/s, then the speed relative to Earth would be 22 km/s,
and the heliocentric speed would be 42 km/s, meaning it would have been in an orbit bound to
the Sun. The question of whether the 2014-01-08 bolide was interstellar rests entirely on how
unlikely it is for its Vy velocity to have been mismeasured (or misreported) by 24.5 km/s.
Such a large error sounds too large to be made by US Government sensors, but what exactly are
the errors in the reported velocities? Siraj & Loeb (2019, arXiv 1904.07224) assumed the
measurement errors were only ±10% (i.e., ±4.4 km/s uncertainty in Vy), based on an
unverifiable personal communication (i.e., someone’s guess). In other words, they accept that
the velocities are imprecise, but at a level equivalent to the determination of an interstellar
trajectory for 2014-01-08 being a 5.7σ result (24.5/4.35≈5.7), i.e., only a 0.001% chance of it
being from our Solar System.
Already in 2019, though, Brown et al. (2016) and Devillepoix et al. (2019) had investigated the
accuracy of the velocities reported in the CNEOS database, by checking them against precise
velocities measured for 10 meteors also observed by ground-based camera networks. Only 4 of
the 10 meteors’ orbits could be accurately reproduced from CNEOS data. Two meteors had
radiants (directions of travel) fully 90° off from what was reported in the CNEOS database. Two
had speeds off by significant amounts: Buzzard Coulee was reported at 12.9 km/s but really
traveled at 18.1 km/s (Milley et al. 2010); and a Romanian bolide was reported at 35.7 km/s but
really traveled at 27.8 km/s (Borovicka et al. 2017). Those works suggested, and careful analysis
by Brown & Borovicka (2023) and Hajdukova et al. (2023, submitted to Nature Astronomy)
subsequently confirmed, that the 1σ uncertainties in velocity measurements in the CNEOS
catalog are > 8 km/s for fast meteors like the 2014-01-08 bolide.
8
9
https://arxiv.org/pdf/1904.07224.pdf
https://cneos.jpl.nasa.gov/fireballs/
7Therefore, instead of it being a > 5σ detection, the interstellar nature of 2014-01-08 is more like
a 3σ detection (24.5/8). Instead of a 0.001% chance it is from our Solar System, there is a 0.1%
chance that it is. That is, there is a one-in-1000 chance the Vy velocity has been mismeasured
so badly (-43.5 km/s instead of -19.0 km/s) that this meteor could have been in a Sun-bound
orbit yet appeared as “obviously” interstellar as 2014-01-08. In isolation, the 2014-01-08 bolide
clearly appears interstellar. But the CNEOS catalog contains 966 fireballs. If all are from our
Solar System but have measurement or reporting errors of 8 km/s, then we should expect (with
66% probability) at least one to have been so badly mismeasured it would appear just as clearly
interstellar as 2014-01-08. All the other potentially interstellar candidates in the database are
much less convincing, and 2014-01-08 happens to be that one fluke.
The situation is exactly analogous to the probabilities of rolling five dice and getting a Yahtzee
(all the same number). On any single roll, the odds are 1/1296, or 0.08%. But if you roll the dice
966 times, there is a 52% probability that one of those rolls is a Yahtzee. Let’s say you happened
to roll five sixes on the 610 th roll. Should you be accused of cheating and of turning the dice
over by hand? After all, the odds of that 610 th roll being a Yahtzee were only 0.08%!
It should be noted that a mismeasured velocity, yielding a speed relative to Earth of 22 km/s
from a Sun-bound orbit, would imply a lower and more common material strength ~50 MPa
(see below), and would be more consistent with the light curve of the 2014-01-08 bolide, as an
object at 45 km/s would have been visible by satellites at a higher altitude (Brown & Borovicka
2023). The 2014-01-08 bolide appears more consistent with an origin in our Solar System.
In a catalog with a thousand objects from our Solar System, there was bound to be at least one
that was mismeasured (or misreported) and appeared interstellar at the 99.9% level. Loeb et al.
have cherry-picked that one object and assigned it near-certainty of being interstellar. They
choose to dismiss the analyses of Brown et al. (2016), Devillepoix et al. (2019), and Brown &
Borovicka (2023) and ignore the evidence that the velocity uncertainties in the CNEOS database
are simply larger than the estimate they used. When Loeb was questioned about the reliability
of CNEOS data, he is quoted by the New York Times on July 24, 2023 as saying, “They are
responsible for national security, I think they know what they are doing.” 10 Appealing to
authority instead of examining evidence and checking one’s work is an abrogation of a
scientist’s duties.
In summary, velocities of bolides are just not reported very accurately in the CNEOS
database. It’s entirely possible one component of the velocity of the 2014-01-08 bolide was
mismeasured or misreported. Its interstellar status is a 3σ detection, or one-in-a-thousand
chance of it being from the Solar System but appearing interstellar. But there was bound to
be one object like this among the 966 objects in the CNEOS database, and this is that one.
The evidence for the 2014-01-08 bolide being interstellar is entirely unconvincing.
10
https://www.nytimes.com/2023/07/24/science/avi-loeb-extraterrestrial-life.html
8IV. Would any of the 2014-01-08 bolide (if interstellar) survive entry?
The 2014-01-08 meteoroid was not that big. In the spirit of recent conventions 11 , it was about
half the mass of a giraffe. Its mass was estimated from its measurements brightness and
standard ‘luminous efficiencies’, which yield a kinetic energy ½ M V 2 = 4.6 x 10 11 J. Assuming a
speed V = 44.8 km/s, this implies a mass M ≈ 450 kg. That’s the mass of an iron ball only 48 cm
(19 inches) in diameter.
At the terrific speeds it is assumed to have traveled at—140 times the speed of sound—the
meteoroid would have completely vaporized in the atmosphere. The friction between a
meteoroid and the air first melts its surface, causing millimeter-sized melt droplets (‘ablation
spherules’) to spray off; but further heating can vaporize this surface, and the spherules too. At
first this process is gradual, but as the meteor penetrates deeper into the atmosphere, the
density of air, ρ, encountered by the meteor increases. The air pushing against the meteor
creates a ‘ram pressure’ P RAM = ρ V 2 that eventually exceeds the strength of the material, and
the meteoroid is crushed. It disintegrates into myriad fragments, each of them continuing at
the same speeds and melting and vaporizing, but now with greater surface area. So much
kinetic energy is suddenly converted into melting and vaporization, that the air heats and
expands, creating a loud boom. The 2014-01-08 bolide suffered three such disruption events
(‘airbursts’) in the span of 0.22 seconds. After disintegrating, the material continued traveling at
45 km/s, almost certainly vaporizing before it slowed down.
As an aside, the ram pressure at the point where the 2014-01-08 bolide disrupted was
calculated by Siraj & Loeb (2022b) to be P RAM = 194 MPa. They then claimed this exceeded the
material strength of even iron meteorites, which they said was 50 MPa, citing Petrovic (2001). It
is clear Loeb et al. did not actually read this paper, however; even the abstract of Petrovic
(2001) says “the average iron meteorite compressive strength is 430 MPa.” Again, it is likely
that the 2014-01-08 velocity was mismeasured, its speed was actually 22 km/s instead of 44
km/s, and that it broke up at a ram pressure ~50 MPa. That would be consistent with the
strengths of more-common rocky bodies, which typically don’t survive pressures greater than
tens of MPa (e.g., Pohl & Britt 2020). But if the 2014-01-08 bolide were interstellar, and
traveling at 44.8 km/s, its breakup is much more consistent with an iron meteorite.
How much of a bolide survives entry vs. vaporizes in the atmosphere depends on the material it
is made of and the velocity with which it enters the atmosphere. This is determined by
integrating the equations of frictional slowing and heating, and heat transfer within the
meteoroid. The standard result is that the fraction of mass that does not vaporize is
M F / M I = exp [ -σ (V I2 – V F2 ) / 2 ],
11
(1)
https://physicsworld.com/a/astronomers-launch-new-asteroid-classification-system-based-on-animal-sizes/
9where M F and V F ≈ 0 are the final mass and velocity, M I and V I the initial mass and velocity, and
σ is the ‘ablation parameter’, a parameter combining information about the heat capacity and
thermal conductivity, etc., of the material (e.g., Popova et al. 2019). The ablation parameter has
been determined for meteors of various materials, and is typically σ > 0.01 km 2 s -2 for rocky
material (Revelle & Ceplecha 2001), and up to 0.07 km 2 s -2 for iron (Revelle & Ceplecha, 1994).
When the ablation parameter σ is large, and especially when the initial velocity V I is large, the
surviving fraction becomes exponentially small. If the 2014-01-08 bolide were interstellar and
traveling at V I ≈ 45 km/s—a speed significantly higher than most meteors experience—then
vaporization would have been fierce. If it was a rocky body, only a fraction ≈ 4 x 10 -5 , or 20
grams, could have survived entry. If it were an iron body, it would have completely vaporized.
In contrast, Tillinghast-Raby, Loeb & Siraj 12 (2022, arXiv 2212.00839) concluded that a
significant fraction (8-21%) of the mass of the bolide could survive entry. The discrepancy could
have been due to their implicitly assuming a small ablation parameter σ = Λ / (2 Q Γ).
Tillinghast-Raby et al. assumed a heat transfer coefficient Λ = 0.04, a heat of ablation Q = 6.549
x 10 10 erg/g, and a drag coefficient Γ = 0.5, yielding σ = 0.0061 km 2 s -2 . This is smaller than
observed for natural stony meteorites and much lower than for iron meteorites, so less
material would be predicted to vaporize. However, we note that equations 2-6 of Tillinghast-
Raby et al. reproduce the same equations that, when integrated, have the solution in Equation
1 above. Thus, they should have predicted that 0.2% of the mass, or 1000 g of spherules, would
survive entry. We believe the authors mistakenly decreased the entry velocities of a large
percentage of the fragments in their calculations.
In summary, if the 2014-01-08 bolide were interstellar, mere grams of it would have
survived entry (if rocky), and more likely all of it vaporized (if iron). Loeb et al. make a
number of mathematical mistakes and misreadings of the literature in order to claim (in
unrefereed work) that tens of kilograms might survive.
V. Do we even know where spherules would have fallen?
Very few spherules would have been created, but in principle these could be recovered (along
with a bunch of background spherules), if one knew where to look. As we’ll show, there are far
too many uncertainties to define a reasonable search area.
Loeb et al. used the longitude and latitude (1.3°S, 147.6°E) and altitude (18.7 km) reported into
the CNEOS database to fix the location of the fireball. In the map in their Figure 3, reproduced
in Figure 1 below, this places the fireball within a red box 0.1° (11 km) on a side. They then tried
to fix the location within that box by knowing the distance r to a seismometer on Manus Island,
about 90 km to the south. They claim this distance is r = 83.9 ± 0.7 km. This places the fireball
within an orange band on their map. This is where they concentrated their search.
12
https://arxiv.org/pdf/2212.00839.pdf
10Figure 1. Reproduction of Figure 3 of Loeb et al., showing the location where they believe the
fireball occurred, according to one set of coordinates on the CNEOS website (red box) and their
(flawed) estimate of the distance to the Manus Island seismometer. The actual location is much
more uncertain, as is the location along the path where the spherules were produced.
There are several obstacles to this approach. One, it’s not clear how accurate the reported
latitude and longitude are (i.e., where to place the red box). Two, the distance to the fireball
(i.e., where to place the orange band) is possibly very different and also so uncertain to be
meaningless (i.e., the orange band is as wide as the red box). It’s also unclear which of the three
fireball disruption events is being recorded by the satellites and the seismometer, or even if
they are recording the same events (where along the orange band spherules were produced).
Location of the fireball / where to place the red box?
The latitude and longitude on the main list of fireballs on the CNEOS website lists a location
(1.3°S, 147.6°E). Uncertainties are not reported, but we presume the position is measured and
reported accurately, implying uncertainties in latitude and longitude of ±5.5 km, i.e., a red box
11 km on a side centered on those coordinates. However, the location of the flash also is
11reported elsewhere on the CNEOS website 13 to have taken place at a latitude of 1.2°S, and a
longitude of 147.1°E, as shown in Figure 2. This implies a red box 11 km on a side and centered
on those coordinates, 11 km to the north and 55 km to the east of where Loeb et al. searched.
Uncertainties in altitude are not reported by CNEOS, but comparison with ground-based ob-
servations shows the average discrepancy is ±3 km (Brown et al. 2016; Devillepoix et al. 2019).
Figure 2. This text accompanied the light curve data on the CNEOS website. The location
reported here contradicts the information reported elsewhere on the site.
Path of the fireball?
Presuming the velocity components Vx, Vy, and Vz reported in the CNEOS database are correct,
the direction of travel of the 2014-01-08 bolide can be reconstructed. Zuluaga (2019) reported
an azimuth of 285.6° (i.e., moving from the northwest, to 16° south of due east, as is evident
from the geometry and negative Vz value) and an elevation of 26.8°. We note that in Figure 4 of
Siraj & Loeb (2023; arXiv2303.07357), they mistakenly show the direction of travel of the
meteor being exactly in the opposite direction.
The direction of travel is the most uncertain quantity for fireballs because the velocities are so
uncertain (Devillepoix et al. 2019). For example, if we assume Vy was mismeasured and was -19
km/s (as above), the azimuth would be off by 12°. The travel directions of several meteors in
the CNEOS database are off by as much as 90° (Brown et al. 2016; Devillepoix et al. 2019). We
may have no practical idea of the meteor’s path.
As the meteor traveled along its path, it suffered three disintegration events, with the first and
last spaced about 0.22 s apart, as shown by the light curve depicted in Figure 3. If traveling at
45 km/s, the first and third events would have been separated by 9 km on the ground, and 4.5
km in altitude. It is not clear whether the position reported in the CNEOS database refers to the
first disruption event at 0.35 s, or the last and brightest event at 0.57 s. If it’s the first, but most
spherules and the earliest-arriving sound waves are generated by the third, then the
seismometer data is referring to an event 9 km to the east of the location reported in the
CNEOS database (whichever set of coordinates is right).
13
https://cneos.jpl.nasa.gov/fireballs/lc/bolide.2014.008.170534.pdf
12Figure 3. Light curve of the 2014-01-08 bolide reported by CNEOS.
Distance between the Manus Island seismometer and “the” fireball?
Siraj & Loeb (2023; arXiv2023.07357) 14 attempted to fix the distance to the fireball (more
precisely, to one of the three disintegration events) using the timing of the first sound wave
from the fireball to reach a seismometer on Manus Island in Papua New Guinea (located at
2.0432°S, 147.3662°E). This seismometer is roughly 86 ± 6 km away, if the fireball were at 1.3°S
and 147.6°E; or 97 ± 6 km away, if the fireball were at 1.2°S and 147.1°E. If the sound speed
were 347 m/s, the signal would arrive about 248 ± 17 s or 280 ± 17 s later. This gives a ~1
minute window in which sound waves might arrive at the seismometer. If the seismometer
recorded a signal in this window, it might be attributed to the fireball, and the time of arrival of
sound waves used to refine the distance between the seismometer and “the” fireball.
As an aside, Loeb et al. state these locations are outside the 22 km limit for the “territorial
waters” of Papua New Guinea; but they clearly are within its Exclusive Economic Zone, as set by
the United Nations Convention on the Law of the Sea (UNCLOS). Papua New Guinea officials
have accused the authors of removing materials without permits, as described in The Post on
July 10, 2023 15 .
Refining the location of the fireball necessitates confidently identifying a shaking of the
seismometer with the arrival of sound waves from the fireball. This cannot be done with fully
certainty. Figure 4 shows the signal from the Manus Island seismometer, downloaded from the
International Federation of Digital Seismograph Networks website 16 . The event attributed to
the fireball is labeled. Evidently, events comparable in seismic energy to the purported fireball
event occur on average every 5-10 minutes or so, including a similar event that occurred only 4
14
https://arxiv.org/pdf/2303.07357.pdf
https://www.thepost.co.nz/a/world-news/350033283/scientists-snatched-our-alien-beads-say-authorities-in-
papua-new-guinea
16
https://www.fdsn.org/networks/detail/AU/
15
13minutes earlier. Within the ~1 minute window in which Siraj & Loeb sought a signal from a
sound wave, there is a decent chance (10-20%) that a background event would occur.
Figure 4. Data from the Manus Island seismometer during the 24-hour period of Jan 8, 2014.
Siraj & Loeb identified the labeled event with the fireball and essentially assumed that the
distance to the fireball is an average sound speed V A (about 347 m/s near the ground)
multiplied by the time after the flash for the sound waves to reach Manus Island, Δt ≈ 270 s.
They concluded that the ground-distance between the seismometer and the fireball (at least,
the nearest or loudest one) was r = 83.9 ± 0.7 km. Based on that presumed uncertainty, the
orange band in Figure 3 of Loeb et al. is made only 1.4 km thick, much smaller than the red box.
However, this uncertainty, only ±0.8%, is remarkably small considering all the complications.
They also concluded the altitude of the fireball was 16.9 ± 0.9 km. It is curious that they accept
this discrepancy between this value and the 18.7 km reported in the CNEOS database.
Siraj & Loeb provide hardly any details about how the arrival times of sound waves at the
seismometer were calculated, so it is difficult to judge its correctness. One issue is that,
because the sound speed varies with temperature and therefore height, as V A = [ 347.2 – 1.861
(z / km) ] m/s, sound waves are refracted away from the ground or ducted through the
stratosphere. Notably, Siraj & Loeb did not recognize that this formula does not apply in the
14stratosphere (z > 10 km), where the sound speed is uniform with height; this makes a difference
of several seconds in arrival times, and at least 1 km in distance.
Because we could not check the details of their model, we reproduced their calculation as well
as they described it. We find that the first sound waves would arrive at the seismometer 270.0 s
after the fireball if they traveled a distance r = 88.5 km along the ground after originating at an
altitude of 18.7 km; or a distance r = 89.2 km if they originated at 16.9 km. These values of r
differ from those of Siraj & Loeb; we would shift the position of the orange band 5 km to the
north of where Loeb et al. searched. Either we and Siraj & Loeb are making different
assumptions, or one or both of us has miscalculated. Either way, reasonable differences about
how the calculation is conducted probably lead to differences in r of about 5 km.
This calculation also shows there is a trade-off between ground-distance and altitude; for every
1 km variation in the fireball’s altitude, there is a 0.4 km difference in ground-distance.
Altitudes of fireballs in the CNEOS database are uncertain at the ±3 km level, and the first and
last disruption events differed in altitude by 4.5 km. This translates into additional uncertainties
in r of ±1 km, and possibly a shift in 2 km in where to place the path on the ground.
The biggest source of uncertainty, though, concerns the arrival time of sound waves after the
flash. Siraj & Loeb assumed Δt = 270.0 ± 0.5 s, a very small uncertainty. Reported times of
fireballs in the CNEOS database are accurate to “within a few seconds” according to Devillepoix
et al. (2019). In addition, inspection of Figure 5, which shows the intensity of sound waves at
the seismometer, suggests the uncertainty is at least ±5 seconds. This translates into additional
uncertainties ±2 km of where to place the path.
Figure 5. Figure 3
from Siraj & Loeb
(2023). The blue
curves denote the
intensity of sound
waves at the Manus
Island seismometer.
The orange curve is an
unexplained model.
Did the first signals
arrive at 260 s, 270 s,
or 280 s?
15An additional issue is, what would sound waves provide a distance to? As depicted in Figure 3,
there were three disruption events that could produce explosions and sound waves. The path
of the bolide had an azimuth of 16° south of east (assuming the velocities are correct), while
Manus Island lay roughly 17° west of south, so that the bolide traveled almost perpendicular to
the line of sight to the seismometer. In this geometry, the first sound waves would arrive at the
seismometer from the third and final disruption event, because it is at the lowest altitude and
also the loudest. However, if the velocities are incorrect and the path substantially to the
northeast instead, then sound waves from the first disruption event would arrive first.
There is an additional factor to consider: solid material generated atmospheric entry tends to
be dropped along a long path called a strewn field. Strewn fields can be very long (> 10 km),
and that certainly would be the case here, if each of the three disintegration events dropped
spherules along the path. Strewn fields are also very wide: > 2 km for iron meteorites, and > 4
km for stony meteorites (e.g., Moilanen et al. 2021). In addition, spherules would take varying
times (≈ 1-30 minutes) to sink to the seafloor, during which time the Equatorial UnderCurrent
would sweep them westward several km (Desch et al. 2023).
In summary, the distance is very uncertain, and the thickness of the orange band where
spherules could be found should be about 4 times wider than Loeb et al. drew, almost as large
as the red box. In addition, the orange band may need to be shifted 5 km to the north. If the
flash reported in the CNEOS database refers to the first disruption event, then the meteor
would have traveled roughly 9 km to the east before generating most of the spherules. All in all,
the uncertainties are basically comparable to the size of the red box.
And of course, there is still the issue that red box may be 55 km to the west, depending on
which coordinates from the CNEOS website one uses. The red box would be at a distance of
97±6 km, and my estimate for the distance of the orange band, 89±3 km, would allow some
overlap (r ≈ 92 km) at that location, so that is also a viable solution to all the available data.
Summary
The uncertainties in when sound waves would arrive at the Manus Island seismometer
make the distance r to the fireball almost as uncertain as the error in the reported position,
±5.5 km. The trajectory of the fireball, being perpendicular to the line-of-sight to Manus
Island, makes it difficult to fix where along the path the bolide disintegrated; the search
area could be shifted 9 km to the east from the fireball. Of course, two locations are
reported in the CNEOS database, 55 km apart, and both are consistent within uncertainties
with the seismometer data. It’s just incredibly naïve and overconfident to claim to know
where spherules would have fallen with sufficient accuracy to mount a search.
16VI. Did the concentration of spherules vary across the expedition site?
Loeb et al. claimed that the density of spherules (number per km 2 ) was much higher in the
regions they associated with the path of the 2014-01-08 bolide than in regions definitely off the
path. They argue that a significant fraction of the spherules they recovered along the assumed
path therefore must have come from the bolide. They specifically claim: “Given that the
highest-yield (yellow) regions in the heat map account for roughly twice the background yield
owing to the excess spherules added by IM1, we expect to find a significant numbers [sic] of
IM1’s spherules relative to background spherules in the highest-yield regions.” However, Loeb
et al. collected far too few spherules to make this claim. It also would be wildly unexpected for
even a small fraction of spherules collected to be from the 2014-01-08 bolide.
Density of background spherules
The ocean floor is littered with a background concentration of cosmic spherules and
micrometeorites, as well as volcanic ash and anthropogenic coal ash. Cosmic spherules have
been collected from the seafloor for decades (e.g., Millard & Finkelman 1970; Blanchard et al.
1980; Brownlee 1985; Rudraswami et al. 2011; Prasad et al. 2015; etc.) Millard & Finkelman
(1970) collected 1200 spherules from 750 kg of Pacific red clay. Assuming both they and Loeb et
al. collected material from the same depths (~10 cm), that equates to roughly 4 x 10 6 spherules
per km 2 . The flux of micrometeorites is uniform across the globe, so any seafloor site with
similarly slow deposition rates ~1 mm/kyr (including this one) should have similar and uniform
background concentrations.
As an aside, Loeb et al. collected from 0.26 km 2 area, meaning they must have been ~10 6
spherules in their search area. The fact that they gathered only ~700 suggests a surprisingly low
collection efficiency, perhaps because their collection was skewed to the very largest particles.
Spatial density of spherules from the 2014-01-08 bolide
The additional contributions from the 2014-01-08 bolide would have been small in comparison
to the background. Assuming a typical spherule mass ~1 mg, at most ~1 x 10 6 spherules would
have been created, if 0.2% of the bolide survived vaporization. These would have been spread
out over a path at least > 10 km long and > 4 km wide (the typical dimensions of strewn fields),
not even accounting for the additional spreading due to ocean currents. This yields a maximum
density < 2 x 10 4 spherules per km 2 . Even assuming the most generous ablation parameter, so
that 0.2% of the mass survives vaporization, spherules from the bolide would make up < 1% of
the background. But for a realistic ablation parameter, far, far less of the bolide would survive.
There is every expectation that > 99% (probably 100%) of the spherules Loeb et al. collected
were simply part of the background of micrometeorites deposited on the seafloor over the last
tens of thousands of years.
17Are there more spherules in the presumed path of the bolide?
Loeb et al. have argued that the concentration varied spatially, being about twice as high in the
regions they associate with the path of the 2014-01-08 bolide than in the two “control” regions
corresponding to tracks 17 and 22 in their Figure 3. The path of the bolide, we’ve argued, is
more uncertain that Loeb et al. realize; but it is fair to say that Tracks 17 and 22 probably do not
include spherules from the 2014 bolide, and reflect the background concentration. Are the
concentrations in tracks 17 and 22 lower than in the other regions?
Loeb et al. report that 622 spherules were collected in runs 1-24: 105 in runs 17 and 22, and
517 in the other runs. Unfortunately, Loeb et al. did not provide any useful information about
the collection area per run, including for how long they dragged the seafloor, or where along
each run the spherules were collected. Instead, one must estimate the collected spherules per
area of seafloor from the map in their Figure 3. Roughly 11 pixels of area (each pixel in Figure 3
has area 3.43 km 2 ) can be associated with runs 17 and 22, meaning there are 2.8 collected
spherules per km 2 . About 56 pixels of are associated with the other runs, yielding or 2.7
collected spherules per km 2 . Even in regions sampled by runs 13, 14, and 15, supposedly of high
spherule concentration (as shown by the heatmap density depicted in their Figure 5), there are
183 spherules in 20 pixels of area, or 2.7 spherules per km 2 . Within the uncertainties (±10%
from counting statistics alone), these are the same. It’s not clear how the heatmap was
calculated, what data went into it, or what it’s measuring, but there is no evidence of spatial
variations in spherule concentration in or out of the assumed path of the 2014-01-08 bolide.
Even if there had been variations, it’s not clear they could be attributed to the bolide rather
than variations in sampling depths, collection efficiencies, deposition times, etc.
In summary, Loeb et al. did not hit some sort of motherlode of spherules from the 2014-
01-08 bolide. Even under the most generous assumptions about survival of material, it
would have barely made a dent in the abundances of spherules. Almost all, if not literally
all, of the cosmic spherules collected by Loeb et al. are from the background population of
micrometeorites deposited on the seafloor over the last tens of thousands of years.
VII. What do the Fe isotopes mean?
Loeb et al. measured the iron isotopic ratios 57 Fe/ 54 Fe and 56 Fe/ 54 Fe in 9 of the spherules (and in
S21 twice). They found them to array along a line of slope ≈3/2 (1.46) in a “three-isotope” plot.
They interpret spherules with high 57 Fe/ 54 Fe and 56 Fe/ 54 Fe to have lost Fe by vaporization
during atmospheric entry, and conclude they formed in a fireball. They miss that the slope itself
is not diagnostic of vaporization per se and that the magnitudes of the deviations from
terrestrial iron isotopic ratios suggest loss or gain of Fe by chemical means other than
vaporization. Most importantly, the line connecting the spherules passes directly through the
terrestrial values, indicating that almost all the iron in these spherules is from our Solar System.
18Figure 6 reproduces Figure 12a of Loeb et al. and shows the fractional deviations of Fe isotopic
ratios from a terrestrial standard. Here, δ 56/54 Fe = 1000 x [ ( 56 Fe/ 54 Fe) sample – ( 56 Fe/ 54 Fe) standard ] /
( 56 Fe/ 54 Fe) standard , and δ 57/54 Fe = 1000 x [ ( 57 Fe/ 54 Fe) sample – ( 57 Fe/ 54 Fe) standard ] / ( 57 Fe/ 54 Fe) standard .
Samples that match the terrestrial standard exactly have δ 56/54 Fe = 0 and δ 57/54 Fe = 0. These
quantities can be used to diagnose chemical processes. If a sample loses Fe by any chemical
reaction, including vaporization, will tend to lose its lightest isotopes fastest, as chemical
reaction rates vary as 1/(m Fe ) 1/2 . As such, loss of Fe will make the 56 Fe/ 54 Fe ratio increase, but it
will make the 57 Fe/ 54 Fe ratio increase even more. In a plot of δ 57/54 Fe vs. δ 56/54 Fe, samples that
started with a terrestrial value (at δ 56/54 Fe = 0 and δ 57/54 Fe = 0) will move along a line of slope
≈1.48, called the ‘terrestrial fractionation line’ or ‘TFL’, to the upper right. Samples that gained
Fe through a chemical process will move along the line to the lower left. The deviations typically
are in parts per thousand, ‘permil’, symbolized ‰.
Figure 6. Deviations of iron isotopic ratios from terrestrial ratios in nine spherules, reproduced
from Figure 12a of Loeb et al. The pink line adjoining them is the terrestrial fractionation line
(TFL) on which samples should array if they have a Solar System origin and experienced
chemically-driven mass-dependent isotopic fractionation by gain or loss of Fe, for example by
vaporization of Fe. All nine spherules are exactly consistent with a Solar System origin.
Figure 6 shows the δ 56/54 Fe and δ 57 / 54 Fe values for nine spherules, including two of the five
“BeLaU” spherules. All are found to lie on the TFL (the pink line), on which all samples on Earth
19lie. Indeed, within the blue box lie most samples from Earth, Mars, Vesta, and other meteorites;
all samples from our Solar System lie on the TFL. This raises two questions. First, why do these
samples, purportedly from another solar system, have Fe isotopes that are such an exact match
to Solar System materials? Second, why is it that, even compared to Solar System materials,
they have such a small range of δ 56/54 Fe values between -1 and +3?
Why are the spherules’ δ 56/54 Fe and δ 57/54 Fe so close to Solar System values?
Figure 6 is quite damning against the interstellar hypothesis. If the “BeLaU” spherules truly
were samples of another stellar system, they would not have Fe isotopic ratios so close to
terrestrial. While all Solar System materials lie very close (<< 1‰) to the TFL, one would expect
materials in another system to have different starting values (hundreds of ‰ away from the
TFL), reflecting their time and place of formation in the Galaxy.
Because there is no reason to expect a meter-sized iron meteoroid to be destroyed over billions
of years, an interstellar meteoroid could originate in system from almost any time over the
Galaxy’s 12 Gyr history, from any star orbiting at about the same distance (within ~1 kiloparsec)
as the Sun around the Galaxy. Galactic Chemical Evolution (GCE) models show that materials
forming at various times and places would sample metallicities (heavy element-to-hydrogen
ratios) Z and combinations of δ 57/54 Fe and δ 56/54 Fe very different from the Sun. Stars in the
Milky Way over the last 7 Gyr sample metallicities Z ≈ 0.5 - 1.3 × Z Sun (Carrillo et al. 2023).
According to Kobayashi et al. (2011), such stars should sample δ 56/54 Fe ≈ -100 to +1500, and
δ 57/54 Fe ≈ 0.9 δ 56/54 Fe ≈ -100 to +1300. Stars with any Fe isotopic ratios across these ranges
could contribute to the background of interstellar meteoroids.
Assuming any star formed over the last 7 Gyr could contribute equally to interstellar
meteoroids, there would be a < 5% chance its δ 56/54 Fe value would lie within 50‰ of the Sun’s
value (a generous maximum variation due to vaporization or aqueous alteration). More
importantly, there would be a < 0.1% chance, whatever its δ 56/54 Fe value, that its δ 57/54 Fe value
would lie within < 1‰ of the TFL. On the face of it, then, < 0.005% of all systems formed over
the last 7 Gyr would produce materials whose Fe isotopes so closely matched Solar System
values. A more careful analysis will show the exact fraction to be different from 0.005%, but it
will surely be similarly small.
It thus would appear there is a > 99.995% chance that the spherules measured here, including
the “BeLaU” spherules S21 and S10, reflect a Solar System Fe isotopic ratio. There is only a <
0.005% chance another solar system would match ours so exactly.
Variation of spherules’ δ 56/54 Fe
Loeb et al. argued that spherules in Figure 6 array along the TFL because of loss of Fe during
vaporization. Rayleigh distillation (loss of Fe by vaporization) has the effect of increasing
δ 56/54 Fe and moving samples up along the TFL. Cosmic spherules of Solar System origin mostly
derive from materials at δ 56/54 Fe ≈0 and δ 56/54 Fe ≈0 (placing them on the TFL); after
20vaporization, S-type (rocky) spherules tend to have δ 56/54 Fe ≈ +1.4 to +3.2, and I-type (iron)
spherules tend to have δ 56/54 Fe ≈ +20 to +36.
Isotopic shifts due only to vaporization can be ruled out. Notably, the “BeLaU” spherule S10 and
the I-type spherules S5 and S19 have δ 56/54 Fe < 0. All Solar System sources (including Mars,
angrites, Vesta, ureilites, etc.) happen to lie on the TFL, but because none have δ 56/54 Fe < 0, the
three spherules S10, S5 and S19 could not derive from Solar System materials (if one only
considers vaporization). These three spherules would have to derive from the same extrasolar
body, presumably 2014-01-08. It is exceedingly improbable that this source would like on the
TFL so close to terrestrial values (see above), and it’s also very unlikely that all these samples
would lie on the same line if only half were from the bolide (as Loeb et al. suggest); but also the
magnitude of the variations in δ 56/54 Fe are too small to be consistent with vaporization alone.
The δ 56/54 Fe values of the S-type, I-type, and “BeLaU” spherules all lie between -1 and +3, a
spread of only 4‰. In contrast, the spread among cosmic spherules is much larger: 16‰ among
the I-types alone, and about 35‰ between the S-type and I-type spherules. These variations
have to do with how Fe vaporizes from these samples, and even interstellar materials should
see similar ranges of δ 56/54 Fe values. This strongly suggests that all of these spherules later
exchanged about 90% of their Fe atoms with a terrestrial reservoir near δ 56/54 ≈ 0, after falling
into the ocean. Presumably they reacted with seawater.
Reaction with seawater also seems consistent with the correlation of δ 56/54 Fe values with Fe/Ni
ratios, suggesting they lost or even gained Fe, after atmospheric entry. In terrestrial samples,
samples with δ 56/54 Fe < 0 tend to result from gain of Fe by aqueous alteration, as in terrestrial
magnetite with negative δ 56/54 Fe. Gain and loss of Fe is suggested among the I-type spherules
analyzed here. S11, with positive δ 56/54 Fe, has lower Fe abundance (3.25x CI) and a lower Fe/Ni
ratio (0.62x CI). That’s consistent with some mild loss of Fe during atmospheric entry. Spherules
S5 and S19, with negative δ 56/54 Fe, have higher Fe abundances (3.75x CI) and have very high
Fe/Ni ratios (145x and 1880x CI). Such extreme fractionation of Fe and Ni is not expected during
vaporization. All of this strongly suggests gain or loss of Fe by aqueous alteration instead.
In summary, the Fe isotopic ratios of the spherules definitively identify them as from our Solar
System, and not an extrasolar source. They could have experienced isotopic fractionation
during vaporization, moving them up the TFL, but this cannot be the only effect. Their Fe
isotopic ratios appear most consistent with massive exchange of Fe with a terrestrial water,
presumably via aqueous alteration while sitting on the seafloor for tens of thousands of years.
VIII. What does it mean that S21 is a triple compound spherule?
Some of the spherules, including one of the “BeLaU” spherules (S21), are found to be
compound, the fusion of three smaller spherules while still partially molten. Loeb et al. argue
“the existence of a triple-merger like S21 can be explained as a product of IM1’s airburst.” and
they provide a back-of-the-envelope calculation to support this claim. However, their
21calculation is flawed, and if it were true, then compound spherules like S21 would be common,
which they are not. The existence of compound spherules like S21 is actually proof that the
“BeLaU” spherules did not form in an airburst like the 2014-01-08 bolide.
It is possible, in theory, for some ablation spherules coming off the bolide to collide while in
mid-air before cooling off. Loeb et al. estimate the probability of two spherules colliding to be
τ ~ n σ R ~ 0.3, assuming the number density of spherules is n ~ 0.2 cm -3 , the cross-sectional
area is σ ~ 0.03 cm 2 , and the typical distance traveled through a cloud of spherules is R ~ 50 cm.
The probability of two collisions, producing a triple spherule, would then be τ 2 ~ 0.1, suggesting
spherules would be common in the 2014-01-08 bolide and all others.
In fact, compound spherules are very rare among cosmic spherules, and it is notable when they
are found, such as in the impact event in Antarctica 430 kyr ago (van Ginneken et al. 2021).
Their presence in that event appears to require formation in a very large impact plume, not a
typical fireball. This implies S21 was not formed in the 2014-01-08 fireball, either.
Indeed, the back-of-the-envelope calculation of Loeb et al. seems fundamentally wrong. First,
the spherules have an average diameter 600 μm, mutual cross sections σ ~ 0.01 cm 2 , and
masses about 1 mg. The entire mass of a 500 kg meteoroid could be converted into 5 x 10 5 such
spherules. But a number density as high as n ~ 0.2 cm -3 would require they be distributed in a
volume no larger than 2.5 m 3 , a sphere with radius ~1 m. While this is consistent with the
assumption by Loeb et al. that spherules travel ~50 cm through a cloud of other spherules, it is
certainly wrong.
Spherules would have been generated during the 0.22 s the 2014-01-08 bolide was ablating,
during which they would have escaped into a volume with cross-sectional area 2000 cm 2 (the
meteoroid’s cross-sectional area) times the length of the channel, 10 km (how far the
meteoroid traveled in 0.22 s). The channel has volume 2000 m 3 , and the number density n of
spherules therefore is three orders of magnitude smaller than Loeb et al. assumed. A value of R
~ 1 m is appropriate, but only because spherules have at most t ~ 1 s to collide with another
spherule before they cool and solidify, and the proper value of R is V REL t, where V REL ~ 1 m/s is
the maximum relative velocities between spherules that allows them to collide and stick
instead of shattering. The net result is that the probability of collision is at most τ ~ 3 x 10 -6 ,
even if none of the material vaporized. The probability of a triple spherule would be a
vanishingly small < 10 -11 .
In summary, a proper calculation shows why compound spherules are so rarely produced
in regular fireballs, and practically rules out a fireball origin for the “BeLaU” spherule S21. It
is likely to have an origin in a relatively rare impact plume, like the one in Antarctica, which
occur on ~10 5 yr timescales (von Ginneken et al. 2021). Very likely, S21—and by extension,
other “BeLaU” spherules—have sat on the ocean floor for tens of thousands of years.
22IX. Are the “BeLaU” compositions new and unique?
Loeb et al. analyzed the chemical compositions of 57 spherules and found that many did not
match their expectations, including a subset of 5 spherules enriched in Be, La and U that they
call “BeLaU” spherules. Loeb et al. assert “these have never been described in the cosmic
spherule literature.” Their argument that these have an extrasolar origin hinges entirely on
there not being any examples of similar compositions among cosmic spherules from asteroidal
micrometeorites in our Solar System. It is not clear how exhaustive their search of the entire
cosmic spherule literature was; it appears to have been limited to the single paper by Folco and
Cordier (2015). It was shockingly easy to find examples in the literature of (Solar System)
spherules with surprisingly similar compositions, indicating that the “BeLaU” spherules almost
certainly have a Solar System origin as well.
Anthropogenic coal ash is a very likely candidate, as recently pointed out by Gallardo (2023).
Even half a century ago, 5% of all magnetic spherules on the mid-ocean seafloor were coal ash,
and 60% of magnetic spherules near land (Parkin et al. 1970); these fractions are surely higher
now. It has been recognized for just as long that diagnosing between coal ash and cosmic
spherules is difficult and cannot be done morphologically, and requires a comprehensive
chemical analysis (Doyle et al. 1976). It is stunningly negligent that Loeb et al. did not present a
plan for distinguishing between cosmic spherules and coal ash, or even consider this possibility.
Loeb et al. analyzed the chemical compositions of 57 spherules. They found that 18 matched
the “S-type” (silicate-rich) spherules and 18 more matched the “I-type” (iron-rich) spherules
described by “Folco et al. (2015)” (they meant Folco and Cordier 2015). Another 21 did not
match types specifically described by “Folco et al. (2015)”, and—it bears repeating—at that
point the authors concluded “these have never been described in the cosmic spherule
literature.” They then proceeded to assert that the chemical compositions of these spherules,
being depleted in Mg, “are clearly derived from material that has gone through planetary
differentiation.” Ignoring the ability of sea water to leach Mg out of minerals, Loeb et al. give
these a new name (“D-type”). Some are depleted in La (relative to CI chondrites) and others
(including the spherule S21) are enriched in Be, La, and U; the latter they call “BeLaU”.
What is this unusual chemical pattern? Figure 7 shows Figure 10a of Loeb et al., with the mass
fractions of various elements in the five “BeLaU” spherules, normalized to their abundance in CI
chondrites. CI chondrites are the most primitive types of meteorites, matching the composition
of the Sun. Other chondrites (meteorites from unmelted asteroids) roughly match CI chondrites
in composition. Any number of processes can alter compositions of meteoritic materials,
including vaporization, melting, aqueous alteration, etc. The elements in Figure 7 are arrayed
from left to right in order of increasing volatility (ability to be vaporized).
It is common in the meteoritics literature to report molar ratios of elements relative to a
chemically similar element, normalized to the same ratio in CI chondrites, e.g., (U/Lu) sample /
(U/Lu) CI . (Uranium and lutetium are similarly non-volatile, rock-loving elements.) In contrast, it
is common in the cosmic spherules literature to report mass fractions of elements relative to
23the mass fractions in CI chondrites, i.e., X(U) sample / X(U) CI . The difference matters because mass
fractions are easier to change than elemental ratios. For example, 40% of the mass of CI
chondrites are Mg, Si, and Ca, and the O atoms that go with them (Lodders 2003). These
elements are also easily leached out of minerals by water. Losing most of the Mg, SI and Ca
from a CI chondrite would raise the mass fractions of U, Lu, etc., by a factor of 1.6, even if they
didn’t participate in any chemistry. The (U/Lu) ratio, in contrast, would be unchanged.
Figure 7. A reproduction of Figure 10a of Loeb et al., showing the mass fractions of various
elements in the five “BeLaU” spherules, normalized to the mass fractions of the same elements
in CI chondrites.
With this in mind, we focus first on the elemental abundance pattern, the ratios of elements
with respect to each other. Most rare earth elements (REEs) like Lu, Er, etc., are present at 80×
CI in spherule S10, and 300× CI in spherule S21, but in both, the relative enrichments are: U/REE
≈ 7.5× CI; La/REE, Ba/REE, Th/REE all ≈ 4× CI; plus slight enrichments in Ce, Hf, and Pb. The
pattern also includes severe depletions in Mg, with MgO mass fractions ≈ 0.5wt%.
Delving into the causes of these patterns later, the immediate question is whether similar
abundance patterns exist in other cosmic spherules. It turns out that a very similar pattern,
reproduced in Figure 8, is found in spherules recovered from the Indian Ocean by Rudraswami
et al. (2016). For example, in the relict-bearing spherules, most REEs are present at levels ≈ 2×
CI. The most enriched elements are (in order): Th (Th/REE ≈ 10× CI), U (U/REE ≈ 6× CI), Ba
(Ba/REE ≈ 6× CI), La (La/REE ≈ 4× CI), then Ce, Hf, and Pb. The MgO mass fraction is ≈35wt%.
24Figure 8. Mass fractions of elements in cosmic spherules recovered from the Indian Ocean,
normalized to mass fractions in CI chondrites, from Figure 2 of Rudraswami et al. (2016).
These patterns match remarkably well. They have the same maximum enrichments in the same
four elements (U, La, Ba, Th). The enrichments are similar, on the order of 4-10× CI. The
elements enriched to lesser degrees are the same: Ce, Hf, Pb.
As for the absolute mass fractions, the Indian Ocean spherules have 35wt% MgO, compared to
0.03wt% in the “BeLaU” spherules. The Papua New Guinea spherules appear to have been
depleted in Mg and other major rock-forming elements, by a factor of 100 relative to the Indian
Ocean spherules. It follows that the mass fractions of REEs, La, U, Th, Ba, etc., would be
elevated by a factor of 100 times.
Notably, micrometeorites from the impact plume in Antarctica 430 kyr ago also are enriched in
La, Th, and U, attributed to terrestrial contamination (van Ginneken et al. 2021).
In summary, the claim that these abundance patterns “have never been described in the
cosmic spherule literature” is laughable. Loeb et al. failed to do due diligence in their
literature review. This is a chemical pattern seen in spherules in very different parts of the
world and are not unique to the spherules from the 2014-01-08 bolide (if they could even be
associated with it).
25X. Were the “BeLaU” spherules concentrated in the presumed path of the bolide?
Loeb et al. claim that the spherules with the distinctive “BeLaU” compositions had a higher
concentration in the assumed path of the bolide than out of it. If this were true, this would
strongly suggest an association of that composition with an origin in the bolide. However, there
is no evidence for variations in the “BeLaU” fraction among collected spherules, and the data
are consistent with the “BeLaU” composition being common among background spherules, and
uniformly distributed.
Loeb et al. collected 622 spherules. Of these, 57 were analyzed: 10 from off the path (runs 17
and 22) and 47 from the path (all other runs). Among these analyzed spherules, 5 were found
to have the BeLaU pattern: all 5 were (potentially) on the path, and none were found off the
path. It’s interesting that they did not find “BeLaU” spherules where they didn’t expect to find
any, but how hard did they look, really?
The frequency of “BeLaU” spherules among analyzed spherules appears to be 5/57, or 8.8%. If
this is characteristic of background spherules, we’d expect the same frequency of “BeLaU”
compositions in spherules off the path. However, only 10 such spherules were even analyzed,
so we’d expect to find only 0.9, i.e., less than 1 on average. According to Poisson statistics, the
probability of finding k=0 “BeLaU” spherules among t=10 analyzed spherules when the
frequency is r=0.088, is (rt) k exp(-rt) / k! = 42%. That is, there could be just as many
“BeLaU”spherules in the control regions, but there’s a 42% chance Loeb et al. wouldn’t find
any, because they hardly looked in those regions.
In fact, even if the frequency of “BeLaU” spherules off the path were 17.5%—double the
frequency on the path—there would still be a 17% chance of finding zero “BeLaU” spherules
among the 10 analyzed! The typical threshold for significance of a null result is 5%, so the
authors cannot even rule out an increased abundance of “BeLaU” spherules off the path.
Given the frequency found in spherules on the path, a finding of zero “BeLaU” spherules off the
path would not be significant unless at least 34 spherules were analyzed, not just a paltry 10.
Only then do the different scenarios predict meaningfully different numbers of “BeLaU”
spherules (0 if none are found off the path; 3 if the distribution is uniform everywhere).
In summary, to do a proper control, Loeb et al. should have collected and analyzed many,
many more spherules from runs 17 and 22. The data are completely consistent with the
“BeLaU” spherules being part of the background of spherules deposited on the seafloor
over the last tens of thousands of years, with nothing to do with the 2014-01-08 bolide.
26XI. What is the cause of the enrichments of La, U, etc., in the “BeLaU” spherules?
Loeb et al. speculate considerably about how the chemical compositions of the “BeLaU”
spherules could arise. They suggest vaporization played a major role in setting depletions, that
siderophile (iron-loving) elements are depleted, and that the most likely explanation is origin
from an (extrasolar) planet that had differentiated (melted and separated rock from an iron
core). This interpretation has a number of problems. What’s striking is that the authors never
once considered the chemical effects on the spherules from sitting in sea water.
There is no doubt that vaporization can deplete a spherule in easily-evaporated (‘volatile’)
elements. This probably explains why elements more volatile than Zn generally are relatively
depleted. It’s also clear that volatility is not the only control: Sc is much more depleted than
similarly non-volatile (‘refractory’) Hf and Y; refractory Ca is much more depleted than similarly
refractory Ta and Nb; V is not nearly as depleted as similarly volatile Ni and Cr.
They conclude that “This “BeLaU” abundance pattern found in IM1’s spherules could have
possibly originated from a highly differentiated planetary magma ocean,” and “the BeLaU
samples possibly reflect an extremely evolved composition from a planetary magma ocean, but
not the bodies in the Solar System.” It’s not clear which part of the planet they mean. The
magma ocean would be rich in Mg and highly depleted in Fe, which seems inconsistent with the
fact that S21 is < 2wt% Mg and > 60wt% Fe. The rock that crystallizes last from a cooling magma
ocean is basalt, but that doesn’t seem to be what they mean, either. They compare to Earth’s
continental crust, which is formed by planetary processes quite distinct from magma oceans.
This is the best match, but even so, the “BeLaU” spherules are depleted in Mg and Ca by an
order of magnitude relative to continental crust.
Rather than arising in an unusual magma ocean on an extrasolar planet, it is likely that the
abundance pattern observed in these spherules is due to the same process that gave rise to the
abundance patterns in the Indian Ocean spherules. The sediments in which the Indian Ocean
spherules were found were also measured by Prasad et al. (2015) for their elemental
compositions and found to be enriched in these same elements (La, U, Th, etc.). Rudraswami et
al. (2016) interpreted their spherules to have chemically interacted with these sediments
(mediated by seawater) over the ~50 kyr residence time at the seafloor. Likewise, van Ginneken
et al. (2021) attributed enrichments in La, Th, and U in their micrometeorites to terrestrial
contamination. Given the low sedimentation rates in the site where the authors recovered their
samples, it is to be expected that they had a similarly long residence time and chemically
interacted with their surrounding sediments. A much greater loss of Mg and Ca is the only
discernible difference between these samples and the Indian Ocean spherules, and is easily
explained by longer residence in the ocean.
27In summary, the “BeLaU” abundance patterns in the Papua New Guinea spherules probably
have the same origin as the Indian Ocean spherules: water-driven chemical reactions with
the sea water and the seafloor sediments the spherules are in. The “BeLaU” spherules like
S21 probably sat on the seafloor for tens of thousands of years.
XII. Does the Be in the spherules indicate GCR irradiation in interstellar space?
Loeb et al. attach significance to the abundance of Be in their 5 “BeLaU” spherules. The mass
fraction of Be in them averages about 300 times the concentration in CI chondrites, or ~8 ppm.
They claim this abundance is not only consistent with, but an indicator of creation of Be by
Galactic cosmic rays (GCRs) striking nuclei in the sample over billions of years while traveling
through interstellar space: “… the enhanced Be abundance might represent a flag of cosmic-ray
spallation on IM1’s surface along a [sic] extended interstellar journey… this constitutes a fourth
indicator of an interstellar origin to IM1…”
In principle, billions of years of exposure to GCRs would convert a fraction of the oxygen nuclei
into Li, Be, and B nuclei; but the fraction is very small. After millions of years of GCR exposure,
meteorites acquire 9 Be and 10 Be abundances on the order of 10 -6 ppm (equivalent to tens of
disintegrations per minute of radioactive 10 Be). This is much, much smaller than the ~0.03 ppm
abundances of Be meteorites typically have due to normal chemical processes. This is why only
radioactive 10 Be, not overall Be itself, is the flag of GCR exposure After all, gems of emerald and
ruby and aquamarine on Earth are 5% beryllium, but this is not an indicator of an interstellar
origin. After billions of years, the GCR-produced Be abundance might be as large as ~10 -3 ppm,
as corroborated by a simple calculation [assuming a flux of > 30 MeV GCRs F ~ 10 cm -2 s -1 , a
cross section for Be production of σ ~ 6 mbarns, and an exposure time t ~ 1 Gyr, a fraction (F σ
t) ~ 2 x 10 -9 of the oxygen nuclei would be converted into 9 Be and 10 Be nuclei, yielding a mass
fraction of Be ~ 0.0006 ppm]. But even exposure to GCRs for billions of years would not
produce even one part in 10 4 of the Be measured in the “BeLaU” cosmic spherules.
The enhanced (above CI chondrite levels) abundances of Be in the spherules clearly is due to
chemistry. There is ample opportunity for to add Be to the cosmic spherules via aqueous
alteration while sitting on the seafloor for tens of thousands of years. Deep ocean water
worldwide contains Be at an abundance of 20-30 pM (picomolar), supplied by river particulates
(von Blanckenburg et al. 1996). Be atoms precipitate into ferromanganese nodules after
spending < 1 kyr in the oceans, yielding Be concentrations in the nodules of 2-3 ppm (Merrill et
al. 1960; von Blanckenburg et al. 1996). Cosmic spherules exposed to sea water would have
comparable Be concentrations, consistent with the ~8 ppm measured in the “BeLaU” spherules.
In summary, even billions of years of GCR irradiation wouldn’t produce even a tiny fraction
of the Be seen in the “BeLaU” spherules. It must be due to chemistry like that experienced
by nodules on the seafloor, and indeed the Be concentration in the spherules is comparable
to the Be concentration in ferromanganese nodules.
28XIII. Conclusions
Loeb et al. claim the spherules they found are “likely” to be interstellar in origin. They advanced
arguments to this effect, but upon slight scrutiny, all of their arguments fall apart. There is no
evidence for interstellar materials. The 2014-01-08 probably wasn’t interstellar. If it were, it
would have completely vaporized. Even if ablation spherules were produced, these would have
been few and spread out, and vastly outnumbered by background spherules. There is no
evidence that spherules overall, or “BeLaU” spherules, were concentrated anywhere, let alone
the path of the bolide, which is very poorly known. The “BeLaU” triple spherule S21 most likely
formed in an impact plume tens of thousands of years ago, and not in a fireball like the 2014-
01-08 bolide. The “BeLaU” pattern is seen in other cosmic spherules in the Indian Ocean and
Antarctica and is attributed to terrestrial contamination. The Be abundance also is attributable
to terrestrial contamination, not cosmic-ray spallation. Finally, the Fe isotopes are more
consistent with terrestrial contamination than vaporization during entry, and are a smoking gun
for a Solar System origin for all their spherules. Not a single one of their arguments holds water.
The reason their arguments fall apart so easily is because they did not follow the scientific
method. They made associations between their data and their favored hypothesis (interstellar
origin), but at no point did they consider any competing alternative hypothesis and ask whether
the data are better explained by that model. It’s a textbook example of confirmation bias.
The simplest competing hypothesis is this: the spherules they collected are part of the
copious background of cosmic spherules deposited on the seafloor over the world.
During the tens of thousands of years spherules spend on the seafloor before being buried by
sediments, they may chemically react with the sea water and the sediments. This alternative
hypothesis could have been tested by sampling the sediments along with the spherules, or by
collecting a large and statistically meaningful number of spherules from regions far from the
bolide’s path. Because they didn’t consider alternative hypotheses, Loeb et al. did not properly
design their experiment to avoid inconclusive results.
On January 31, 2021, Loeb is quoted by The Guardian 17 as saying: “If someone comes to me and
says, ‘For these scientific reasons, I have a scenario that makes much more sense than yours,’
then I’d rip that paper up and accept it,” he says. “But most of the people who attacked, they
hadn’t even looked at my paper, or read the issues, or referred to the items we discussed…”
We have read and considered the manuscript they uploaded to arXiv, and we have a scenario
that makes much more sense than theirs. Will they now metaphorically rip up their paper?
17
https://www.theguardian.com/science/2021/jan/31/professor-avi-loeb-it-would-be-arrogant-
to-think-were-alone-in-the-universe-
29References
Blanchard, M. B., Brownlee, D. E., Bunch, T. E., Hodge, P. W., and Kyte, F. T. (1980) “Meteoroid
ablation spheres from deep-sea sediments” Earth & Planet. Sci. Lett. 46, 178-190.
Borovicka, J., Spurny, P., Grigore, V. I., and Svoren, J. (2017) “The January 7, 2015 superbolide
over Romania and structural diversity of meter-sized asteroids”, Planet. & Space Sci. 143, 147-
158.
Brown, P. G., Wiegert, P., Clark, D. and Tagliaferri, E. (2016) “Orbital and physical characteristics
of meter-scale impactors from airburst observations”, Icarus 266, 96-111.
Brown, P. G. and Borovicka, J. (2023) “On the Proposed Interstellar Origin of the USG 20140108
Fireball”, Ap.J. 953, 167-179.
Brownlee, D. E. (1985) “Cosmic dust – Collection and research”, Ann. Rev. Earth & Planet. Sci.
13, 147-173.
Carrillo, A., Ness, M. K., Hawkins, K., Sanderson, R. E., Wang, K., Wetzel, A., and Bellardini, M. A.
(2023) “The Relationship between Age, Metallicity, and Abundances for Disk Stars in a
Simulated Milky Way”, Ap. J. 942, 35-56.
Desch, S. J., Jackson, A. P. and Hartnett, H. E. (2023), “The Challenges of Recovering Interstellar
Meteorites”, Asteroids, Comets and Meteors Conference 14, #2244 (abstract).
Devillepoix, H. A. R., Bland, P. A., Sansom, E. K., Towner, M. C., Cupak, M., Howie, R. M., Hartig,
B. A. D., Jansen-Sturgeon, T. & Cox, M. A. (2019) “Observations of metre-scale impactors by the
Desert Fireball Network”, Mon. Not. Roy. Astron. Soc. 483, 5166-5178.
Doyle, L. J., Hopkins, T. L. and Betzer, P. R. (1976) “Black magnetic spherule fallout in the
eastern Gulf of Mexico”, Science 194, 1157-1159.
Folco, L., and Cordier, C. (2015) “Micrometeorites” in Planetary Mineralogy, M. R. Lee and H.
Leroux, eds., EMU Notes in Mineralogy 15, 253-297.
Gallardo, P. A. (2023) “Anthropogenic coal ash as a contaminant in a micro-meteoritic
underwater search”, Res. Notes. Amer. Astron. Soc. 7, 220.
Kobayashi, C., Karakas, A. I., and Umeda, H. (2011) “The evolution of isotope ratios in the Milky
Way Galaxy”, Mon. Not. Roy. Astron. Soc. 414, 3231-3250.
Lodders, K. (2003) “Solar System Abundances and Condensation Temperatures of the
Elements”, Ap.J. 591, 1220-1247.
30Merrill, J. R., Lyden, E. F. X., Honda, M., and Arnold, J. R. (1960) “The sedimentary geochemistry
of the beryllium isotopes”, Geochim. Cosmochim. Acta 18, 108-129.
Millard, H. T., Jr., and Finkelman, R. B. (1970) “Chemical and mineralogical compositions of
cosmic and terrestrial spherules from a marine sediment”, J. Geophys. Res. 75, 2125-2134.
Milley, E. P. (2010) “Physical Properties of Fireball-Producing Earth-Impacting Meteoroids and
Orbit Determination through Shadow Calibration of the Buzzard Coulee Meteorite Fall”, Ph.D.
Thesis (University of Calgary).
Moilanen, J., Gritsevich, M. and Lyytinen, E. (2021) “Determination of strewn fields for
meteorite falls”, Mon. Not. Roy. Astron. Soc. 503, 3337-3350.
Petrovic, J. J. (2001) “Mechanical properties of meteorites and their constituents”, J. Mat. Sci.
36, 1579-1583.
Pohl, L. and Britt, D. T. (2020) “Strengths of meteorites—An overview and analysis of available
data”, Meteoritics & Plan. Sci. 55, 962-987.
Popova, O., Borovicka, J., and Campbell-Brown, M. D. (2019) “Modeling the Entry of
Meteoroids”, in Meteoroids: Sources of Meteors on Earth and Beyond, G. O. Ryabova, D. J.
Asher, and M. D. Campbell-Brown, eds. (Cambridge Univ. Press), pp. 9-36.
Prasad, M. Shyam, Rudraswami, N. G., de Araujo, A., Babu, E. V. S. S. K., and Vijaya Kumar, T.
(2015) “Ordinary chondritic micrometeorites from the Indian Ocean”, Meteoritics & Planet. Sci.
50, 1013-1031.
Revelle, D. O. and Ceplecha, Z. (1994) “Analysis of identified iron meteoroids: possible relation
with M-type Earth-crossing asteroids?”, Astron. & Astrophys. 292, 330-336.
Revelle, D. O. and Ceplecha, Z. (2001) “Bolide physical theory with application ot PN and EN
fireballs”, in Proceedings of the Meteoroids 2001 Conference (ESA SP-2495), 507-512.
Rudraswami, N. G., Parashar, K., and Shaym Prasad, M. (2011) “Micrometer- and nanometer-
sized platinum group nuggets in micrometeorites from deep-sea sediments of the Indian
Ocean”, Meteoritics. & Planet. Sci. 46, 470-491.
Rudraswami, N. G., Shaym Prasad, M., Babu, E. V. S. S. K., and Vijaya Kumar, T. (2016) “Major
and trace element geochemistry of S-type cosmic spherules”, Meteoritics & Planet. Sci. 51, 718-
742.
Siraj, A. & Loeb, A. (2022a) “A Meteor of Apparent Interstellar Origin in the CNEOS Fireball
Catalog”, Ap.J. 939, 53-57.
31Siraj, A. & Loeb, A. (2022b) “Interstellar Meteors are Outliers in Material Strength”, Ap.J.L. 941,
L28-31.
Siraj, A. & Loeb, A. (2023) “Localizing the First Interstellar Meteor with Seismometer Data”,
arXiv:2303.07357 (unpublished).
van Ginneken, M., Goderis, S., Artemieva, N., Debaille, V., Decree, S., Harvey, R. P., Huwig, K. A.,
Hecht, L., Yang, S., Kaufmann, F. E. D., Soens, B., Humayun, M., van Maldeghem, Genge, M. J.,
and Claeys, P. (2021) “A large meteoritic event over Antarctica ca. 430 ka ago inferred from
chondritic spherules from the Sor Rondane Mountains”, Sci. Adv. 7, eabc1008.
von Blanckenburg, F., O’Nions, R. K., Belshaw, N. S., Gibb, A., and Hein, J. R. (1996), “Global
distribution of beryllium isotopes in deep ocean water as derived from Fe-Mn crusts”, Earth &
Planet. Sci. Lett. 141, 213-226.
Zuluaga, J. I. (2019) “Speed Thresholds for Hyperbolic Meteors: The Case of the 2014 January 08
CNEOS Meteor”, Res. Notes Amer. Astron. Soc. 3, 68.
32
