Recent near-Earth supernovae probed by global deposition of interstellar radioactive (60)Fe.

The rate of supernovae in our local Galactic neighbourhood within a distance of about 100 parsecs from Earth is estimated to be one every 2-4 million years, based on the total rate in the Milky Way (2.0 ± 0.7 per century). Recent massive-star and supernova activity in Earth's vicinity may be traced by radionuclides with half-lives of up to 100 million years, if trapped in interstellar dust grains that penetrate the Solar System. One such radionuclide is (60)Fe (with a half-life of 2.6 million years), which is ejected in supernova explosions and winds from massive stars. Here we report that the (60)Fe signal observed previously in deep-sea crusts is global, extended in time and of interstellar origin from multiple events. We analysed deep-sea archives from all major oceans for (60)Fe deposition via the accretion of interstellar dust particles. Our results reveal (60)Fe interstellar influxes onto Earth at 1.5-3.2 million years ago and at 6.5-8.7 million years ago. The signal measured implies that a few per cent of fresh (60)Fe was captured in dust and deposited on Earth. Our findings indicate multiple supernova and massive-star events during the last ten million years at distances of up to 100 parsecs.

The density and temperature distribution of the interstellar medium (ISM) is highly variable, with typical substructures of ~50–150 pc (superbubbles) having life-times of some 10 Myr. Several SN explosions over the last ~14 Myr shaped the present structure of the local superbubble (LB)[12-14]. The SS, now embedded in the LB, is expected to have faced fronts of SN ejecta and accumulated material from massive stars. To enter the SS, any material from the ISM must be condensed into larger dust grains to avoid being deflected away by the solar wind and interplanetary magnetic field[3,10,11]. ISM dust particles were indeed identified at Earth orbit[15] and may accumulate on Earth in archives such as deep-sea sediments and ferromanganese (FeMn) crusts and nodules which retain time information over millions of years. 60Fe as well as 26Al (t1/2=0.71 Myr) are observed[1,9] in the ISM as a result of many SNe and emission from massive stars. Direct detection of ‘live’ radionuclides[3-5,10,11] on Earth would provide insight into recent and nearby nucleosynthesis in massive stars[14,16,17], dust formation and transport into the SS. Extraterrestrial 60Fe was in fact already observed in FeMn crusts in pioneering studies at TU Munich[10,11], and interpreted as being of SN[10,11,18] or (micro)meteoritic origin[19,20].

In the present work, the 60Fe contents of three different deep-sea archives (four sediment cores, two FeMn-crusts and two FeMn-nodules) recovered from the Indian, Pacific and Atlantic Oceans respectively (Supplementary Figure S1) were determined. All were dated via their 10Be (t1/2=1.39 Myr) content, complemented by 26Al for the sediments[21]. All radionuclides (60Fe, 26Al and 10Be) were counted using accelerator mass spectrometry (AMS) (Supplementary Information). The sediment cores provided a record from 1.7-3.2 Myr BP (before present) with a time resolution of <30 kyr, bracketed by recent and ~5-7 Myr old samples. Pacific ‘Crust-1’ extends from present to 10.9 Myr with ~2.2 Myr time resolution and ‘Crust-2’ from 1.2–7 Myr BP (~100 kyr resolution). Two nodules covered 5.4 Myr BP (~2 Myr resolution).

In the sediment, 288 60Fe-events were registered for the time period 1.71-3.18 Myr (45 individual samples) with a mean isotopic ratio 60Fe/Fe=(1.79±0.10)×10−15, a factor of ~40 above the measurement background of (0.042±0.015)×10−15. None of the recent or old sediment samples show evidence for 60Fe above background (3 60Fe-events). The first two layers in Crust-1 gave 60Fe-signals 4σ and 7σ above background; layers 3 and 5 are close to the measurement background, but layer 4, which spans the period 6.5-8.7 Myr, has a significantly higher ratio (~4σ above background, Table 2). For Crust-2 a clear 60Fe-signal was also found at <3.5 Myr. The nodules support this finding (Table 3, Supplementary Tables S3-S5).

Table 2

60Fe/Fe-ratios from AMS measurements at ANU of layered samples of the two Pacific FeMn crust samples (Crust-1 and Crust-2) and of the two Atlantic FeMn nodules (no. 21 and 24).

Crust-1
layer	depth (mm)	time period (Myr)	60Fe counts detected	60Fe/Fe (10−15 at/at)[a]	60Fe/Fe│d.c. (10−15 at/at)[b]	Fe conc. (10−2 g g−1)[c]	60Fe conc. (106 at g−1)	60Fe incorp. rates (atoms·cm−2yr−1)	60Fe incorporation (106 at·cm−2/layer)
Layer 1	0 – 5	0 – 2.17	23	0.96±0.25	1.19±0.33	12.8±0.1	1.64±0.45	0.72±0.20	1.56±0.43
Layer 2	5 – 10	2.17 – 4.35	74	1.58±0.23	3.60±0.55	11.8±0.1	4.56±0.68	2.00±0.30	4.34±0.65
Layer 3	10 – 15	4.35 – 6.52	2	0.14±0.12	0.36±0.51	11.4±0.1	0.34±0.62	0.19±0.27	0.40±0.59
Layer 4	15 – 20	6.52 – 8.70	26	0.38±0.09	2.49±0.97	13.7±0.2	3.67±1.43	1.61±0.63	3.49±1.36
Layer 5	20 – 25	8.70 – 10.87	3	0.15±0.11	1.20±1.50	11.2±0.2	1.45±1.82	0.64±0.80	1.38±1.73
total	0–4.35 & 6.52-8.70		123		--			--	9.4±0.9
Crust-2
Layer 1-2	0 – 1.0	1.20 – 1.41	5	1.06±0.48	1.50±0.67	11.6±0.1	1.87±0.83	1.67±0.75	0.36±0.16
Layer 3-5	1.0 – 2.5	1.41 – 1.73	37	1.28±0.21	2.01±0.33	11.9±0.1	2.60±0.43	2.33±0.38	0.74±0.12
Layer 6-8	2.5 – 4.0	1.73 – 2.05	32	0.84±0.15	1.39±0.25	11.8±0.1	1.73±0.31	1.55±0.27	0.50±0.09
Layer 9-10	4.0 – 5.8	2.05 – 3.05	20	0.55±0.12	1.09±0.24	11.1±0.1	1.35±0.30	0.46±0.10	0.46±0.10
Layer 11-12	5.8 – 7.7	3.05 – 4.11	2	0.12±0.08	0.29±0.21	10.4±0.1	0.37±0.26	0.13±0.09	0.13±0.09
Layer 13-26	7.7 – 21.0	4.11 –7	1	<0.03	<0.1	9	<0.1	<0.09	<0.06
60Fe:	0 – 7.7	1.20 – 3.05	94		--				2.19±0.22
Nodule 21
Layer 1	0 – 3	0 – 1.8	3	0.16±0.11	<0.23	15±2	<0.4	<0.13	<0.23
Layer 2	3 – 6	1.8 – 3.3	13	0.47±0.16	0.60±0.22	15±2	0.97±0.36	0.40±0.15	0.55±0.21
Layer 3/4	6 – 17	3.3 – 5.4	5	0.18±0.06	<0.10	15±2	<0.16	<0.1	<0.23
60Fe:	3 – 6	1.8–3.3	13	0.47±0.16	0.60±0.22	15±2	0.97±0.36	0.40±0.15	0.55±0.21
Nodule 24
Layer 1	0 – 4	0–1.9	15	0.56±0.18	0.71±0.23	15±2	1.13±0.37	0.34±0.11	0.86±0.41
Layer 2	4 – 8	1.9–3.3	5	0.23±0.12	0.45±0.23	15±2	0.71±0.37	0.27±0.15	0.54±0.34
Layer 3/4	8 – 19	3.3–5.4	1	0.05±0.05	<0.15	15±2	<0.24	<0.1	<0.25
60Fe:	0 – 8	0–3.3	20	0.40±0.10	0.58±0.13	15±2	0.92±0.21	0.31±0.09	1.40±0.50

measured 60Fe/Fe ratios. The uncertainties from 60Fe denote statistical uncertainties only (1σ, using Poisson statistics).

background & decay corrected (d.c.) 60Fe/Fe data. The machine background of 60Fe/Fe=(0.042±0.015)×10−15 was used for the subsequent background correction. Ages are based on the 10Be data and have an uncertainty of ±0.3 Myr (Crust-1) and ±0.5 Myr (Crust-2 and nodules).

stable iron content of the dissolved crust material and of the leachate in case of the nodules, as measured via ICP-MS. The mean dry density of the crust and nodule samples was 1.9·g·cm−3. ‘conc.’ means ‘concentration’. The Fe concentration listed for Crust-2, layers 13-26 is the average; individual data range between 5.5 and 13.1%.

Table 3

Summary of 60Fe-deposition into various archives as obtained in this work and given in the literature[10,11,22] (no correction for incorporation efficiency). Uncertainties are 1σ.

Deep-sea archive	cores	location	time period (Myr)	60Fe detector events	60Fe deposition (106 at·cm−2)[a]
Sediment	4	Indian Ocean	1.71 – 3.18	288	35.4±2.6
FeMn Crust 1	2	Pacific Ocean	0 – 4.35	97	5.9±0.8
FeMn Crust 1			6.52 – 8.70	26	3.5±1.0
FeMn Crust 2			1.2 – 3.1	94	2.2±0.2
FeMn nodules	2	Atlantic Ocean	1.8 – 3.30 – 3.3	1320	0.6±0.21.4±0.5
FeMn Mona Pihoa[10]	1	Pacific ocean	0 – 5.9	21	~9[b]
FeMn 237KD[11]	1	Pacific ocean	1.74 – 2.61[b]	69	1.5±0.4[b]
Lunar material[22]	1	Moon	integral	--[c]	~10

for Crust-1 and Crust-2 an incorporation efficiency of 17 and 7% respectively, has to be taken into account to calculate the 60Fe fluence from the deposition values; similarly 2 and 4% for the nodules.

adjusted for a revised 60Fe and 10Be half-live values[7,8]

not listed in Ref. 22.

In summary, two clear 60Fe signals with a total of 538 60Fe-events were observed. In the sediments, the signal covers the time period 1.7–3.2 Myr. In the crusts, 60Fe is found up to 3.5 and ~4 Myr, with a second influx between 6.5 and 8.7 Myr. The nodules confirm the presence of 60Fe at <3.3 Myr. No 60Fe signal is found in recent (<0.2 Myr) or older (≥5 Myr) sediments and nodules, or in crusts between 4.4–6.5 and 8.7–10.9 Myr.

Between 1.7 and 3.1 Myr, the 60Fe deposition rate into the sediments was ~11–35 60Fe atoms·cm−2yr−1 (300-kyr averages), whereas incorporation rates into crust material were significantly lower at 1–2 atoms·cm−2·yr−1 (Figure 1, all data are decay-corrected). This suggests an incorporation-efficiency into Crust-1 and Crust-2 of 17% and 7%, respectively. The deposition in the 1.5-Myr interval covered by the signal in the sediment is (35±2)×106 atoms·cm−2. For the second 60Fe signal (6.5-8.7 Myr, Crust-1, 17% incorporation) it is (21±6)×106 atoms·cm−2 (Tables 2-3).

Figure 1

Deposition rates for sediment (150 kyr averaged data) and incorporation rates for two crust samples

60Fe concentrations (60Fe/g) for the sediment are given in the inset; they were on average 6.7×104 atoms/g between 1.7 and 3.2 Myr, but 260×104 atoms/g crust and 95×104 atoms/g nodule, reflecting the difference in growth rate and incorporation efficiency (see Supplement). The error bars (1σ) include all uncertainties and scale with decay correction, thus upper limits are becoming larger for older samples. The absolute ages for the sediment are uncertain by 0.1 Myr, but for the 5.5-Myr sediments ~1 Myr. Ages of Crust-1 are 0.3 and of Crust-2 0.5 Myr uncertain.

Although the 1.5 Myr time-spread of 60Fe influx measured in the present work exceeds the ~0.8 Myr previously reported for crust 237KD[11,18], the two time profiles are not inconsistent given the lower counting statistics and signal-to-background in Ref. 11. Furthermore, the marginally positive result for the same time period for an Atlantic sediment[18] is consistent with our data, considering their higher sedimentation rates and stable Fe-contents. 60Fe has also been reported in lunar material, though without time information[22] and recently in Pacific sediments[23].

Clearly, our data are incompatible with a constant 60Fe production or deposition. A terrestrial origin can be ruled out, because there is no suitable target for cosmic-ray induced production and anthropogenic input would be concentrated in the surface layer. Since 60Fe was found in each of the major oceans, it is reasonable to assume a uniform global distribution. A micro-meteoritic or meteoritic origin can be excluded, since the measured cosmic-dust flux is 400 times lower than would be required (Supplementary Information and Figure S6). Similarly a hypothetical break-up of a single object, comparable to the asteroid invoked in relation to the K/T event 65 Myr ago, would have delivered 4,500 times less 60Fe.

We assume that the extraterrestrial 60Fe flux through Earth’s cross-section is homogenously distributed over Earth’s surface. Thus, the measured mean deposition of ~24.5 atoms·cm−2yr−1 (1.7–3.2 Myr signal) corresponds to a 60Fe-flux of 98 atoms·cm−2yr−1 into the inner SS or integrated over 1.5 Myr to an 60Fe-fluence of (1.46±0.15)×108 atoms·cm−2 at Earth orbit; the fluence for the older event is (1.2±0.4)×108 atoms·cm−2. Interstellar grains, filtered by the SS in size to an average of ~0.5 μm, were detected by space missions[15], suggesting that (6±3)% (εdust) in mass of ISM dust reaches the inner SS[6]. These grains follow the flow velocity of the ISM. Assuming the 60Fe-loaded grains follow the same mass-distribution as determined for ISM grains at Earth orbit, we deduce an interstellar 60Fe-concentration in dust of (2.8±1.4)×10−11 60Fe atoms·cm−3 for 1.7-3.2 Myr and integrated over the full period of 11 Myr an average concentration of ~(5-15)×10−12 atoms·cm−3. Observations of 60Fe-decay[1,9] and nucleosynthesis models[2] suggest an average Galaxy concentration of ~6×10−12 60Fe atoms·cm−3 (Supplementary Information), in agreement with the 11-Myr local-data reported here.

60Fe is produced in massive stars[2,24-27] in their late phases, predominantly just before SN-explosions, and then ejected into space. (Super)AGB stars also produce and eject 60Fe through their stellar winds during ~50 kyr, leading to a time profile similar to SNe; however, their contribution to the galactic 60Fe inventory is small[28].

Models suggest a travel time of ~200 kyr with a time spread of ~100–400 kyr[5] for ejecta from a single SN at ~100pc distance. Our measured spread of ~1.5-Myr is inconsistent with the interpretation in terms of ejecta from a single SN (or AGB-star) moving across the SS (Supplementary Figure S6). It suggests multiple SN- and massive-star activities within the last ~10 Myr in Earth’s vicinity and two distinct periods 1.7–3.2 and ~6.5–8.7 Myr BP. The recent time profile would be compatible with movement across the SS of ejecta in a series of SN-fronts in short succession within 1.5 Myr. This would, however, require a high SN-frequency (~2-3 SNe/Myr) since large fluctuations were not observed in the time profile. Alternatively, the ejecta containing the 60Fe-bearing grains could have come to rest in the ambient ISM and diffused into volumes or clouds, that were then traversed by the SS[18].

The SS is currently embedded in a flow of ISM-material with interstellar grains moving parallel to the flow of neutral interstellar gas in local ISM clouds arguing for a common history or driver[29]. Such clouds were suggested as part of an expanding superbubble-shell driven by SNe and winds from massive stars[29,12-14]. Assuming the ejecta originate from a distance 70-100 pc (~limit of the LB) and 60Fe is equally distributed into the outer shell of size 30 pc (distance representing 1.5 Myr travel), i.e. a spherical shell of mean radius 70-100 pc with a thickness of 30 pc, we deduce a total 60Fe mass trapped in ISM dust of (5–11)×10−5 solar masses (M⊙) in the shell volume. This number represents a lower limit as it reflects the fraction of 60Fe condensed into dust without correction for radioactive decay and neglects the granularity of clumpy ejecta. Models predict core-collapse and electron-capture SN-nucleosynthesis yields for 60Fe to be (0.5–14)×10−5 M⊙ for 8–25 M⊙-stars[24-27,2], depending on the progenitor mass with large uncertainties in the nuclear-physics input. (Super)AGB stars produce (0.003–1)×10−5 M⊙ 60Fe[28]. Our observed signals therefore favor SN events. The fraction of 60Fe in dust can be roughly estimated by a comparison of our measured 60Fe deposition with nucleosynthesis yields. Under these assumptions and assuming reasonable distances (20–100pc) ~0.4-9% of 60Fe would be trapped in dust (Supplementary Information, Figures S7 and S8).

Comparing our data with a similar work for ISM-244Pu in sediments and crust samples[6] yields a 244Pu/60Fe atom-ratio of ~3×10−5 or less during periods of elevated 60Fe deposition over the last 10 Myr which agrees with the recently reported low 244Pu SN-yields[6] (Supplementary Information).

Our broad and global 60Fe-influx on Earth demonstrates recent (<10Myr) and wide-spread massive-star ejections in our near galactic neighborhood (<100pc), most likely from SN-explosions. Interestingly, the older event coincides with a strong increase in 3He and temperature change ~8 Myr BP30, while the more recent activity starting ~3 Myr BP occurred at the same time as Earth’s temperature started to decrease during the Plio-Pleistocene transition.

Table 1

Averaged 60Fe/Fe atom ratios from AMS measurements at ANU

52 sediment samples from four sediment cores (Eltanin) from the Indian Ocean were analyzed (individual data are listed in the Supplementary Information) as well as a series of blank samples (commercial iron).

sediment cores	sediment samples	time period (Myr)	60Fe counts detected	60Fe/Fe (10−15 at/at)[a]	60Fe/Fe│d.c. (10−15 at/at)[b]	Fe conc. (10−2 g g−1)[c]	60Fe conc. (104 at g−1)	60Fe-deposition rates (at cm−2yr−1)	60Fe deposition (106 at·cm−2 layer−1)[d]
45-21 / 50-02	5	< 0.2	2	0.06±0.04	0.02±0.02	0.30±0.10	<0.2	<0.2	--
49-53 / 45-21	14	1.71 – 2.0	123	1.67±0.15	2.52±0.23	0.23±0.01	6.0±0.6	22.8±2.3	6.5±0.7
49-53 / 45-21 / 50-02	11	2.0 – 2.3	51	1.51±0.21	2.48±0.35	0.24±0.01	6.7±1.0	24.8±3.6	7.4±1.1
49-53 / 45-21 / 50-02	7	2.3 – 2.6	33	1.96±0.34	3.50±0.61	0.17±0.01	6.5±1.2	27.1±5.0	8.1±1.5
49-53 / 45-16	7	2.6 – 2.9	54	3.40±0.46	6.61±0.90	0.16±0.01	10.3±1.5	34.8±5.2	10.4±1.5
49-53 / 45-16	6	2.9 – 3.18	27	1.18±0.23	2.41±0.47	0.13±0.01	3.4±0.7	11.4±2.4	3.0±0.6
45-16	2	~4–7[e]	1	0.11±0.11	0.20±0.30	0.14±0.01	<0.4	<1	--
commercial iron	99	background	7	0.042±0.015	--

measured 60Fe/Fe ratios. The uncertainties from 60Fe denote statistical uncertainties only (1σ, using Poisson statistics).

background & decay-corrected (d.c.) 60Fe/Fe-data. Surface and old layers are compatible with the measurement background obtained from chemistry blank samples. The age is based on the 26Al and 10Be data and tie-points of magnetic reversals.

‘conc’ means ‘concentration’. Stable iron content of the leached material as measured via ICP-MS (averaged values, see Methods). The average leachable Fe content for the four cores was measured to 0.18% (core Eltanin 49-53), 0.25% (45-16), 0.3% (45-21) and 0.45% (50-02). The mean dry density of the sediments was 1.16·g·cm−3.

for sediments with 100% incorporation efficiency the Fe deposition equals the terrestrial fluence.

uncertain by ~1 Myr.