地球の形成と初期進化： 生命誕生の場ができるまで

京都大学 宇宙物理学教室 佐々木貴教

　本日の発表内容

✤ 太陽系形成論の簡単なレビュー!

✤ “ハビタブルゾーン” と “ハビタブルプラネット”!

✤ いかにして “地球” をつくるか？

−地球の初期進化についての新しいシナリオの提案−

太陽系形成論の簡単なレビュー

太陽系形成標準理論（京都モデル）

�������

�����

�������

�������

������

©Newton Press

巨大氷惑星形成

原始太陽系円盤の組成一般に円盤質量の99%はガス（水素・ヘリウム） 残りの1%がダスト（固体成分）

・現在の太陽系の惑星の固体成分（約10-4M太陽） 　　→ すりつぶして円盤状にならす ・固体成分の約100倍の質量のガス成分を加える

最小質量円盤モデル（京都モデル）

原始太陽系円盤の初期質量は約10-2M太陽 重力と遠心力の釣り合いから半径は約100AU

���������

������������� !10-4g/cm3)

#"$��

��

[g/cm3]

roll

imp23~40.2~

EE

NdNd

ff ��� �

�

Eimp = - p dV�V0

V

Suyama et al. submitted to ApJ

���������� 微惑星の形成

ダストの合体成長 　→ 微惑星形成

微惑星の円盤が形成

微惑星の合体成長数kmサイズの 微惑星が形成

互いに衝突・合体 を繰り返し成長

↓

暴走的成長 　大きい粒子ほど成長が速い !秩序的成長 　全ての粒子が同じ速度で成長

20 KOKUBO AND IDA

FIG. 3. Snapshots of a planetesimal system on the a–e plane. The circlesrepresent planetesimals and their radii are proportional to the radii of planetesi-mals. The system initially consists of 3000 equal-mass (1023 g) planetesimals.The numbers of planetesimals are 2215 (t = 50,000 years), 1787 (t = 100,000years), and 1322 (t = 200,000 years). In the t = 200,000 years panel, the filledcircle represents a protoplanet (runaway body) and lines from the center of theprotoplanet to both sides have the length of 5rH.

the mass ratio keeps increasing when α <−2. If α >−2, meanmass should be similar to themaximummass, so that the increasein the mass ratio would stall soon. Note that our definition ofrunaway growth does not necessarily mean that the growth timedecreases with the mass of a body, but that the mass ratio of anytwo bodies increases with time as shown below.The evolution of the distributions of the RMS eccentricity and

the RMS inclination is plotted in Fig. 6. Let us focus on the massrange 1023 ≤m≤ 1024 g. The values for the mass range larger

FIG. 4. Time evolution of the maximum mass (solid curve) and the meanmass (dashed curve) of the system.

than this range are not statistically valid since eachmass bin oftenhas only a few bodies. First, the distribution tends to relax to adecreasing function of mass through dynamical friction among(energy equipartition of) bodies (t = 50,000, 100,000 years).Second, the distributions tend to flatten (t = 200,000 years). Thisis because as a runaway body grows, the system ismainly heatedby the runaway body (Ida and Makino 1993). In this case, theeccentricity and inclination of planetesimals are scaled by the

FIG. 5. The cumulative number of bodies is plotted against mass att = 50,000 years (dotted curve), 100,000 years (dashed curve), and 200,000years (solid curve). A runaway body at t = 200,000 years is shown by a dot.

暴走的成長の様子

平均値

最大の天体

微惑星の暴走的成長 　→ 原始惑星が誕生する

20 KOKUBO AND IDA

FIG. 3. Snapshots of a planetesimal system on the a–e plane. The circlesrepresent planetesimals and their radii are proportional to the radii of planetesi-mals. The system initially consists of 3000 equal-mass (1023 g) planetesimals.The numbers of planetesimals are 2215 (t = 50,000 years), 1787 (t = 100,000years), and 1322 (t = 200,000 years). In the t = 200,000 years panel, the filledcircle represents a protoplanet (runaway body) and lines from the center of theprotoplanet to both sides have the length of 5rH.

the mass ratio keeps increasing when α <−2. If α >−2, meanmass should be similar to themaximummass, so that the increasein the mass ratio would stall soon. Note that our definition ofrunaway growth does not necessarily mean that the growth timedecreases with the mass of a body, but that the mass ratio of anytwo bodies increases with time as shown below.The evolution of the distributions of the RMS eccentricity and

the RMS inclination is plotted in Fig. 6. Let us focus on the massrange 1023 ≤m≤ 1024 g. The values for the mass range larger

FIG. 4. Time evolution of the maximum mass (solid curve) and the meanmass (dashed curve) of the system.

than this range are not statistically valid since eachmass bin oftenhas only a few bodies. First, the distribution tends to relax to adecreasing function of mass through dynamical friction among(energy equipartition of) bodies (t = 50,000, 100,000 years).Second, the distributions tend to flatten (t = 200,000 years). Thisis because as a runaway body grows, the system ismainly heatedby the runaway body (Ida and Makino 1993). In this case, theeccentricity and inclination of planetesimals are scaled by the

FIG. 5. The cumulative number of bodies is plotted against mass att = 50,000 years (dotted curve), 100,000 years (dashed curve), and 200,000years (solid curve). A runaway body at t = 200,000 years is shown by a dot.

軌道長半径 [AU]

軌道離心率

質量 [1023 g]

時間 [年]

FORMATION OF PROTOPLANETS FROM PLANETESIMALS 23

FIG. 7. Snapshots of a planetesimal system on the a–e plane. The cir-cles represent planetesimals and their radii are proportional to the radii ofplanetesimals. The system initially consists of 4000 planetesimals whose to-tal mass is 1.3× 1027 g. The initial mass distribution is given by the power-law mass distribution with the power index α = −2.5 with the mass range2× 1023 ≤m≤ 4× 1024 g. The numbers of planetesimals are 2712 (t = 100,000years), 2200 (t = 200,000 years), 1784 (t = 300,000 years), 1488 (t = 400,000years), and 1257 (t = 500,000 years). The filled circles represent protoplanetswith mass larger than 2× 1025 g and lines from the center of the protoplanet toboth sides have the length of 5rH.

there is no supply of small bodies. Figure 9 shows the snapshotof another run at t = 500,000 years. Two large protoplanets withmass 1–3× 1026 g with an orbital separation of about 10rH areformed. The two protoplanets contain 38% of the system mass.

FIG. 8. The number of bodies in linear mass bins is plotted for t = 100,000,200,000, 300,000, 400,000, and 500,000 years.

In Fig. 10, we plot the maximum mass and the mean mass ofthe system against time. The time evolution of the RMS eccen-tricity and inclination of the system is plotted in Fig. 11. Thevalues are scaled by the reduced Hill radius of the maximumbody. The reduced Hill radius is given by h= rH/a. After about20,000 years, the scaled RMS eccentricity and inclination are al-most constant, ⟨e2⟩1/2 ≃ 6hmax and ⟨i2⟩1/2 ≃ 3.5hmax. This resultagrees with the estimation of Ida and Makino (1993). They esti-mated the equilibrium eccentricity and inclination of a planetes-imal system perturbed dominantly by a protoplanet under gasdrag. They found that when the mass of a protoplanet becomes

寡占的成長の様子軌道離心率

各場所で微惑星が暴走的成長 　→ 等サイズの原始惑星が並ぶ

寡占的成長とよぶ

＝

各軌道での原始惑星質量 [kg] 形成時間 [yr]

地球軌道 1×10 7×10

木星軌道 3×10 4×10

天王星軌道 8×10 2×10軌道長半径 [AU]

原始惑星から惑星へ������)-/

�%���5�• M ≃ 0.1M⊕

<∼ M�%�• �' =�%���*�

����������,(

�����5�• M ≃ 10M⊕ ≪ M���• ! =�����*�

��������1&

�$����5�• M ≃ 15M⊕ ≃ M�$��• ! =�$����*�

�$����������2��.3 (��� )

����-/

(�%-/

)

0 �

#

�

�

�"

4+

原始惑星の質量 [地球質量]

軌道長半径 [AU]

地球型惑星 　原始惑星同士の合体 !

巨大ガス惑星 　原始惑星のガス捕獲 !

巨大氷惑星 　原始惑星そのまま

snow line

ジャイアントインパクト

軌道長半径 [AU]

軌道離心率

planets is hnM i ’ 2:0 ! 0:6, which means that the typical result-ing system consists of two Earth-sized planets and a smallerplanet. In thismodel, we obtain hnai ’ 1:8 ! 0:7. In other words,one or two planets tend to form outside the initial distribution ofprotoplanets. In most runs, these planets are smaller scatteredplanets. Thus we obtain a high efficiency of h fai ¼ 0:79 ! 0:15.The accretion timescale is hTacci ¼ 1:05 ! 0:58ð Þ ; 108 yr. Theseresults are consistent with Agnor et al. (1999), whose initial con-ditions are the same as the standard model except for !1 ¼ 8.

The left and right panels of Figure 3 show the final planets onthe a-M andM–e, i planes for 20 runs. The largest planets tend to

cluster around a ¼ 0:8 AU, while the second-largest avoid thesame semimajor axis as the largest, shown as the gap around a ¼ha1i. Most of these are more massive thanM%/2. The mass of thelargest planet is hM1i ’ 1:27 ! 0:25M%, and its orbital elementsare ha1i ’ 0:75 ! 0:20 AU, he1i ’ 0:11 ! 0:07, and hi1i ’0:06 ! 0:04. On the other hand, the second-largest planet hashM2i’ 0:66 ! 0:23M%, ha2i ’ 1:12 ! 0:53AU, he2i ’ 0:12 !0:05, and hi2i ’ 0:10 ! 0:08. The dispersion of a2 is large, sincein some runs, the second-largest planet forms inside the largestone, while in others it forms outside the largest. In this model, wefind a1 > a2 in seven runs.

Fig. 2.—Snapshots of the system on the a-e (left) and a-i (right) planes at t ¼ 0, 106, 107, 108, and 2 ; 108 yr for the same run as in Fig. 1. The sizes of the circlesare proportional to the physical sizes of the planets.

Fig. 3.—All planets on the a-M (left) and M–e, i (right) planes formed in the 20 runs of the standard model (model 1). The symbols indicate the planets first(circles), second (hexagons), third (squares), and fourth (triangles) highest in mass. The filled symbols are the final planets, and the open circles are the initialprotoplanets in the left panel. The filled and open symbols mean e and i in the right panel, respectively. [See the electronic edition of the Journal for a color versionof this figure.]

KOKUBO, KOMINAMI, & IDA1134 Vol. 642

長い時間をかけて原始惑星同士の軌道が乱れる 　→ 互いに衝突・合体してより大きな天体に成長

巨大天体衝突による月形成

原始地球に火星サイズの原始惑星が衝突飛び散った破片が地球の 周囲に円盤を形成円盤中で月が誕生！

巨大ガス惑星の形成

原始惑星に円盤ガスが暴走的に流入 → ガス惑星へ

巨大氷惑星の形成

円盤散逸後に原始惑星が形成 → ガスを纏えず氷惑星へ

“ハビタブルゾーン” と

“ハビタブルプラネット”

「ハビタブルゾーン」 惑星表面に液体の水が存在できる領域

Why H2O?195ハビタブルプラネットの起源と進化 第１回／阿部

水素であり，次はヘリウムである．その次に酸素，炭素，ネオン，窒素と来る．固体の惑星の主成分であるマグネシウム，シリコン，鉄が次に続くが，酸素の存

在量はこれらよりも 1桁以上多い．元素合成の過程を考えても水素と酸素が多い元素であることには違いがないであろう．反応性がないヘリウムを除けば，水は最も多い二つの元素の組み合わせでできている．そのことから考えても水という物質が非常に普遍性のある物質であるということがわかる．なお，他の恒星系では酸素よりも炭素の方が多い，というようなこともあるかもしれない．この場合には，炭素がどのような形態をとるかによっては，水は作りにくくなってしまうかもしれない．　次に水の性質に注目してみよう．図1.2に 1 気圧におけるいろいろな化合物の融点と沸点を示した．これは大雑把に液体の状態をとる温度範囲を表すことになる．ここには化合物の分子量も示してある．一般に分子量が大きい物質ほど融点・沸点とも高くなる傾向がある．その中にあって水は分子量が小さいにもかかわらず融点と沸点が高いことが分かるであろう．水並みに融点・沸点が高い物質はどれもかなり複雑な物質である．言い換えればそのような化合物は作りにくい．こう見ていくと水は単純な物質，すなわち存在量が多

図1.1： 太陽系の元素存在度．太陽組成ガス（Solar）と炭素質コンドライト（CI）に含まれる元素の存在度を，珪素の存在度を106に規格化して示した．

図1.2： 1 気圧におけるさまざまな化合物の融点および沸点．線分で示された部分が，その化合物が液体の状態をとる温度範囲を表している．また，その化合物の分子量を白丸で示している．水は分子量が小さい化合物の中で，ひと際融点と沸点が高い．

195ハビタブルプラネットの起源と進化 第１回／阿部

水素であり，次はヘリウムである．その次に酸素，炭素，ネオン，窒素と来る．固体の惑星の主成分であるマグネシウム，シリコン，鉄が次に続くが，酸素の存

在量はこれらよりも 1桁以上多い．元素合成の過程を考えても水素と酸素が多い元素であることには違いがないであろう．反応性がないヘリウムを除けば，水は最も多い二つの元素の組み合わせでできている．そのことから考えても水という物質が非常に普遍性のある物質であるということがわかる．なお，他の恒星系では酸素よりも炭素の方が多い，というようなこともあるかもしれない．この場合には，炭素がどのような形態をとるかによっては，水は作りにくくなってしまうかもしれない．　次に水の性質に注目してみよう．図1.2に 1 気圧におけるいろいろな化合物の融点と沸点を示した．これは大雑把に液体の状態をとる温度範囲を表すことになる．ここには化合物の分子量も示してある．一般に分子量が大きい物質ほど融点・沸点とも高くなる傾向がある．その中にあって水は分子量が小さいにもかかわらず融点と沸点が高いことが分かるであろう．水並みに融点・沸点が高い物質はどれもかなり複雑な物質である．言い換えればそのような化合物は作りにくい．こう見ていくと水は単純な物質，すなわち存在量が多

図1.1： 太陽系の元素存在度．太陽組成ガス（Solar）と炭素質コンドライト（CI）に含まれる元素の存在度を，珪素の存在度を106に規格化して示した．

図1.2： 1 気圧におけるさまざまな化合物の融点および沸点．線分で示された部分が，その化合物が液体の状態をとる温度範囲を表している．また，その化合物の分子量を白丸で示している．水は分子量が小さい化合物の中で，ひと際融点と沸点が高い．存在度の大きい単純な分子の中で

圧倒的に高い融点・沸点を持つ

・Habitable zone の内側境界@present S.S.

　暴走温室条件：0.97A [Kopparapu et al. 2013]

・Habitable zone の外側境界@present S.S.

　CO2 凝縮条件：1.70AU [Kopparapu et al.. 2013]45367298�

� � &��

���� &��

;0.9AU� ;1.5AU�$%���

;110%����'��� ;50%�

����/ ��&��

CO2/ ���

��!).��/�+�-,1!��*0/#(/��:�

from Kasting et al. (1993) "���� 436�

H.Z. for Ocean Planets

・Continuously habitable zone: 0.99AU-1.1AU

次々に発見される太陽系外の惑星たち

宇宙は地球であふれてる！

(c) NASA

「第二の地球」を発見！？Quintana et al., Science (2014)

いかにして “地球” をつくるか

All the Water on Planet Earth

Habitable Trinity

Dohm & Maruyama (2014)

その他の地球特有の事項on the parameters a, b and c in Eq. (20) as listed in Table 6.In Fig. 6b–e the same relationships are shown as in Fig. 6a,except that log D uncertainty envelopes are included foreach element. In each figure, results for Ru, Rh, Re andPt respectively are plotted together with extrapolated parti-tion coefficients for Ir (the most siderophile element) and Pd(the least siderophile element). Including uncertainties, thepressure range where the Pd mantle concentration couldhave been produced by metal–silicate equilibration is 30–45 GPa. Total uncertainties are largest for Ir, because itsconcentration in the silicate melt was below the detectionlimit in several samples and fewer data were available forregression (Table 6). But even at the lower limit of uncer-tainty, the Ir mantle concentration could only have beenachieved at pressures of P53 GPa, based on the (unlikely)assumption of oxidising redox conditions throughout accre-tion. Similarly, the Pt and Rh partition coefficients couldonly have the required D core

mantle values at pressures of 50–80 GPa and 50–100 GPa respectively if the oxygen fugacityremained at a constant value of IW -2. The latter value isinconsistent with the observed depletions of moderately sid-erophile elements (especially V and Cr) which require thatcore formation initially occurred under highly-reducingconditions, such as IW -4 (Wade and Wood, 2005; Rubieet al., 2011). In the case of Ru and Re, at IW -2 even con-ditions of 75 and 85 GPa respectively could not have de-creased their partition coefficients sufficiently (Figs. 6b, d).At lower oxygen fugacities, the trends for all elements inFig. 6 would shift to even higher pressures such that at60 GPa the partition coefficients of Ru, Rh, Re, Ir and Ptstill would have been P103. The resulting mantle concen-trations (normalized to CI chondrite) of Re, Ir, Ru, Pt,Rh and Pd based on partition coefficients calculated forconditions of the peridotite liquidus at 5, 20, 40 and60 GPa and IW -3 respectively, are plotted in Fig. 7. Thisshows clearly that, except for Pd, the HSEs considered inthis study would have been strongly depleted by core for-mation even at 60 GPa. Thus the observed mantle concen-trations could not have been produced by metal–silicatesegregation. Additionally, any such conditions would haveleft a strongly fractionated HSE abundance pattern. This isfurther supported by the recent results of Brenan andMcDonough (2009) on the partitioning behaviour of Irand two HSEs, Au and Os, not considered in the presentwork. At 2 GPa, 2173–2588 K, the partition coefficients ofIr and Os differ by 1 order of magnitude and both are about4–5 orders of magnitude higher than that of Au, whichwould lead to very different degrees of depletion for theseHSEs.

We therefore conclude that the major process thatcaused the observed near-chondritic HSE abundances ofthe Earth’s mantle cannot be high-pressure, high-tempera-ture metal–silicate equilibration during core formation.The most likley alternative explanation is the late accretiontheory, first suggested by Kimura et al. (1974) and Chou(1978), and later established as the ‘late veneer’ stage inthe core formation model of O’Neill (1991). According tothis hypothesis, the HSEs were almost entirely removedfrom a planetary mantle as long as metal–silicate segrega-tion was occurring, but were replenished by the late

accretion of small amounts (e.g. for the Earth 0.09–0.5%of its mass) of some type of primitive meteoritic materialafter core formation had ceased. The late veneer is addi-tionally discussed as a potential source of planetary water,carbon and other highly volatile elements. Although manydetails of the late veneer are poorly understood, it providesa plausible explanation for the enigmatic presence and rel-ative proportions of the HSEs in the silicate portions of theEarth, Moon and Mars (see Walker (2009) for a recentreview).

Models have also been proposed that combine lateaccretion and high-pressure equilibration. Based on parti-tioning data for Pd, Au (Danielson et al., 2005) and Pt(Cottrell and Walker, 2006), Righter et al. (2008) proposedthat mantle concentrations of these HSEs could have beenestablished at high pressure and temperatures in a magmaocean, while the extremely siderophile elements Ir, Os, Ruand Rh would have been depleted from the mantle duringthis equilibration period. Afterwards, a late veneer wouldhave delivered material with about chondritic proportionsof HSEs that overprinted the previously set signature byintroducing Ir, Os, Ru and Rh and slightly increasing theconcentrations of Pd and possibly Au. This would providean explanation for the somewhat suprachondritic concen-trations of Pd and the enstatite chondrite like Au/Ir (e.g.Fischer-Godde et al., 2011). Our data support a partialretention of Pd during high pressure equilibration and itssubsequent supplement by material of the late veneer suchthat the suprachondritic Pd/Ir mantle ratio could be ex-plained. However, at all P–T conditions our data designateRu as the second most siderophile element among the

Fig. 7. Highly siderophile element concentrations, normalised toCI chondrite (Fischer-Godde et al., 2010), predicted for theprimitive upper mantle (PUM, data from Becker et al., 2006; Rhfrom Fischer-Godde et al., 2011) as a result of metal–silicateparititoning at various P–T conditions in a magma ocean. Partitioncoefficients were calculated at the pressures indicated and at the therespective peridotite liquidus temperatures (Herzberg and Zhang,1996; Zerr et al., 1998) and at an oxygen fugacity of IW -3, usingEqs. (19) and (20) together with the regression coefficients fromTable 6. From these partition coefficients the magma oceanconcentrations were calculated employing a mass balance equationand assuming a chondritic bulk composition for the Proto-Earthand a constant molar proportion of 16.6% for the core (all dataemployed are based on molar proportions).

610 U. Mann et al. / Geochimica et Cosmochimica Acta 84 (2012) 593–613

地球マントル

高!圧!実!験

マントル中に強親鉄性元素が過剰に存在

硫黄同位体異常 → 2.5Ga まで低O2, CO2濃度

felsic melts, as indicated by multiple geochemical signatures6,8,9.Oxygen isotopes, for example, demonstrate that the vast majority ofterrestrial Hadean zircons fall outside the zircon mantle equilibriumfield (d18OVSMOW 5 5.3 6 0.3 % (1s))5,20, which is a strong argumentfor the input of supracrustal material into the source melts. Moreover,most zircon crystallization temperatures are consistent with formationin water-saturated or near-water-saturated melts7,8.

Conclusions about the oxidation state of the Hadean mantle basedon zircons of crustal affinity are admittedly indirect. However, a subsetof the Hadean zircon population does appear to be mantle-derived.(We note that direct partial melts of the mantle cannot crystallizezircon, but the fractionation products (residuals) of such melts doeventually saturate in zircon21.) Five of the Hadean crystals haved18O values that fall within the mantle equilibrium field, which hasbeen shown to be constant (60.2%) for the past 4.4 Gyr (ref. 20). Ifthese d18O values are primary, then these zircons crystallized fromuncontaminated, mantle-derived melts that did not interact with thehydrosphere. The five zircons in the d18O mantle equilibrium field—including two crystals with U–Pb ages approaching 4,400 Myr—give acalculated oxygen fugacity of FMQ 1 1.4 (62) (Fig. 3). As a com-parison, zircons crystallized from residual liquids of present-daymantle-derived mid-ocean ridge basalts22–24 yield calculated fO2 valuesof FMQ 1 0.4 (62.6), which agrees with estimates of the oxidationstate of the upper mantle10. This also implies that the redox conditionsin residual, zircon-saturated liquids of present-day mid-ocean ridgesystems did not undergo significant changes in fO2 (relative to thebuffer curve) along the liquid line of descent from basalt. This is alsoconsistent with observations from lunar zircons. Finally, we note thatMartian basalts (for example, shergottite meteorite Dar al Gani 476)exhibiting little or no interaction with a crustal component are onlyabout 1 log unit higher than the estimated fO2 of the upper mantle25.Furthermore, calculated oxygen fugacities from the same meteorite areall within a log unit of the average (FMQ 2 2.5), even though phasescover a crystallization range of .300 uC.

As in all studies in which mantle fO2 is estimated from mantle-derived glasses or minerals3,9, it is important to emphasize that thecalculated fO2 may not directly reflect that of the mantle source region.We note, however, that mantle-derived Hadean zircons would have hadto crystallize from residual melts that underwent ,4 log units of changein fO2 to be reconcilable with mantle source regions in equilibriumwith the IW buffer. At present and in the Archaean, relative fO2

changes along magmatic liquid lines of descent are broadly of theorder of ,1 log unit4,10,26,27. If the evolution of Hadean melts wassimilarly constrained, then our results imply that the mantle reachedits present-day oxidation state ,4,350 Myr ago (Fig. 3). This inter-pretation is consistent with core formation models calling for physicaland chemical isolation from the (upper) mantle ,30 Myr after planetformation28.

If our deductions regarding the oxidation state of Hadean magmasare correct, then the speciation of gases emanating from the Earth at thistime would have been dominated by CO2, SO2, H2O and N2 (ref. 26).An atmosphere of this composition is known to yield a lower abund-ance of sugars and especially amino acids and nucleotides1. If a highlyreduced atmosphere is required for the origin of life, then it may haveoccurred exceptionally early on our planet. However, pre-4,400-Myroutgassing of H2 coupled with slow escape29 may have resulted in anatmosphere out of equilibrium with Earth’s interior. Alternatively, a‘late veneer’ may have served as a source of pre-biotic molecules. Thesealternative scenarios notwithstanding, our results offer the first glimpseinto the oxidation state of terrestrial magmas before 3,900 Myr ago, andthe data presented here open up the possibility of exploring the oxida-tion state of zircon-bearing rocks throughout the geologic record. Thiscalibration will be especially valuable in the early Archaean or Hadeaneons, for which zircon is often the only source of primary chemicalinformation.

METHODS SUMMARYStarting materials were contained in platinum (Pt) or silver (Ag) capsules, and fO2

was buffered by placing ruthenium and ruthenium dioxide (Ru1RuO2; RRO) indirect physical contact with the experimental charge or by use of a separate bufferchamber separated by a Pt or silver-palladium (Ag-Pd) membrane (NNO orMMO). Quenched glasses were analysed by electron microprobe, with at leastone measurement made at the top and bottom of the capsule to assess melthomogeneity. Synthesized zircons were small (,3–20mm) and in some casesthe glass was dissolved with hydrofluoric acid (HF) and the zircons recovered inorder to obtain more accurate measurements.

For the inset plot of Fig. 1, Prx 5 2/3(La)x 1 1/3(Nd)x for ref. 14, where xrepresents a normalized value, either to CHUR or the melt. Ce anomalies fornatural zircons6,19,23–25,30 are calculated in a similar fashion to equation (1) in thetext: (Ce/Ce*)CHUR 5 CeCHUR/(LaCHUR 3 PrCHUR)1/2, where LaCHUR and so onrepresents normalization to the chondritic uniform reservoir rather than the meltconcentration. Ce anomalies are plotted for Hadean zircons with U–Pb ages thatare $94% concordant, and with REE1Ti measurements that reflect primaryigneous chemistry5,6,17. Zircon crystallization temperatures in Fig. 2 are calculatedassuming unity Ti-activity7; sub-unity Ti activity (for example, ,0.5) results insystematic shifts of all data to higher values of fO2, though shifts are within error ofthe standard deviation of each population. Lunar zircons are from Apollo 14polymict breccias 14304, 14305 and 1432119 and are normalized to an averagefO2 of IW at 700 uC, for direct comparison with the other data in the probabilitydistribution plot of Fig. 2. For the inset of Fig. 2b, average fO2 values calculatedfrom zircon Ce anomalies are plotted against independent estimates from lunarbasaltic igneous rocks (IW 2 0.6)18, terrestrial mid-ocean ridge basalts10 (FMQ),and average oxygen fugacities returned from Bishop Tuff titanomagnetite-ilmenitepairs (NNO)27.

Full Methods and any associated references are available in the online version ofthe paper at www.nature.com/nature.

Received 22 March; accepted 6 October 2011.

1. Kasting, J. F. Earth’s early atmosphere. Science 259, 920–926 (1993).2. Canil, D. Vanadium partitioning and the oxidation state of Archaean komatiite

magmas. Nature 389, 842–845 (1997).3. Delano, J.W.Redoxhistory of the Earth’s interior since ,3900Ma: implications for

prebiotic molecules. Orig. Life Evol. Biosph. 31, 311–341 (2001).4. Burgisser, A. & Scaillet, B. Redox evolution of a degassing magma rising to the

surface. Nature 445, 194–197 (2007).5. Cavosie, A. J., Valley, J. W., Wilde, S. A., &. Edinburgh Ion Microprobe Facility.

Magmatic d18O in 4400–3900 Ma detrital zircons: a record of the alteration andrecycling of crust in the early Archean.Earth Planet. Sci. Lett. 235, 663–681 (2005).

6. Cavosie, A. J., Valley, J. W., Wilde, S. A., &. Edinburgh Ion Microprobe Facility.Correlated microanalysis of zircon: trace element, d18O, and U-Th-Pb isotopicconstraints on the igneous origin of complex 3900 Ma detrital grains. Geochim.Cosmochim. Acta 69, 637–648 (2006).

6 Crust-derived

IW

Archaean and present-day upper mantle

Zircon in equil. with mantle (δ18O)

4

2

0

–2

3,800 3,900 4,000 4,100

ΔFM

Q

207Pb/206Pb* age (Myr)

4,200 4,300 4,400

–4

–6

Figure 3 | Oxygen fugacities of Hadean melts plotted against zirconcrystallization age. Errors for individual points are based on the standarddeviation of the entire data set (n 5 20; 1s 5 6 2.3 log units). Zircons withmantle signatures are within error of the estimated oxidation state of the present-day and Archaean upper mantle3,10,26. On average, oxygen fugacities are lower foryounger zircons, though this trend is not robust given the present data set.

LETTER RESEARCH

1 D E C E M B E R 2 0 1 1 | V O L 4 8 0 | N A T U R E | 8 1

Macmillan Publishers Limited. All rights reserved©2011

マントルは酸化的@4.35Ga

地球の軌道がほぼ真円

巨大天体衝突破片（GIF）Giant Impact Fragments

地球型惑星形成理論 �������

106 $10

7 ��

��������

�

107 $10

8 ��

微惑星

原始惑星

地球型惑星（スーパーアース・地球たち）

巨大天体衝突

���/ ��,B#+@�%@��@��C������,B��&@ C!�/�/ �!,B��@&@#+C�

���/��/ ��,B#+@�%@��@��C�01�-2.���/ ��,B�@#+@��@&C�

��/��*��

$)'�3"��

��4=?9���

input&

8A6!��

;4<>7:5A:�

��/��(�3"��

GFD5A:�

E�5A:�

start&

input&

��� ����

�10� �����#���������

�10���� ����#e�0.01���#

GIFによる円軌道化

D;@3A:G�

�� ���>?F=/ �!'264&87!#$ �!�01��!���9�)�5!#$+/0<:��*.��*,�0�7!#$D;@3A:!

�BE@C�0!��!���'��0�!

<:���%��!���9!�5��'��"-.(,!

原始海洋 ＋ GIF → 水素大気発生

H2O+ Fe → FeO + H2

H2大気の形成

マントルへ

岩石 金属鉄

原始海洋

金属鉄 Fe と原始海洋が反応して水素大気を生成

原始海洋が大規模に失われる!大量の水素大気をまとった原始地球の誕生

↓

H2大気

CO2ガス

2H2 + CO2 → H2O + CH2O太陽EUV

GIF降下後の大気進化

原始海洋と GIF が反応した後の水素の分配：f = H2O/(H2+H2O)!

　→ GIF 降下後の原始地球における海洋質量が決まる!!H2 のハイドロダイナミックエスケープ：energy-limited escape!

（太陽 EUV のフラックスで散逸率が決まる [Zahnle et al. 1988]）

海洋質量と水素大気保持期間

00.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

10 20 30 40

Initial ocean volume [x present]

A

B

H2 retention = 1 byr

ocean volume =1

0.5

2

2 byr

3 byr

f = H

2O /

(H2+

H2O

)

B

A

00

100

200

300

400

500

1.0

H2

initial ocean x15initial ocean x10initial ocean x5

f = 0.1

H2O

0.5 1.5 2.0 2.5

Time [billion years after moon-forming impact]

Genda et al., Figure X

Pres

sure

[bar

]

地球は還元的大気を長期間保持

これまで：酸化的な原始地球マントル + 大気とマントルは平衡状態!　　　　　　→ 2.5Ga まで大気を低O2, CO2濃度にしておくことは困難

本シナリオ：大量の H2 大気が系を支配（大気とマントルは非平衡）!　　　　　　　→ 長期間 H2 大気持続, 低O2, CO2濃度を維持

硫黄同位体異常 → 2.5Ga まで低O2, CO2濃度

felsic melts, as indicated by multiple geochemical signatures6,8,9.Oxygen isotopes, for example, demonstrate that the vast majority ofterrestrial Hadean zircons fall outside the zircon mantle equilibriumfield (d18OVSMOW 5 5.3 6 0.3 % (1s))5,20, which is a strong argumentfor the input of supracrustal material into the source melts. Moreover,most zircon crystallization temperatures are consistent with formationin water-saturated or near-water-saturated melts7,8.

Conclusions about the oxidation state of the Hadean mantle basedon zircons of crustal affinity are admittedly indirect. However, a subsetof the Hadean zircon population does appear to be mantle-derived.(We note that direct partial melts of the mantle cannot crystallizezircon, but the fractionation products (residuals) of such melts doeventually saturate in zircon21.) Five of the Hadean crystals haved18O values that fall within the mantle equilibrium field, which hasbeen shown to be constant (60.2%) for the past 4.4 Gyr (ref. 20). Ifthese d18O values are primary, then these zircons crystallized fromuncontaminated, mantle-derived melts that did not interact with thehydrosphere. The five zircons in the d18O mantle equilibrium field—including two crystals with U–Pb ages approaching 4,400 Myr—give acalculated oxygen fugacity of FMQ 1 1.4 (62) (Fig. 3). As a com-parison, zircons crystallized from residual liquids of present-daymantle-derived mid-ocean ridge basalts22–24 yield calculated fO2 valuesof FMQ 1 0.4 (62.6), which agrees with estimates of the oxidationstate of the upper mantle10. This also implies that the redox conditionsin residual, zircon-saturated liquids of present-day mid-ocean ridgesystems did not undergo significant changes in fO2 (relative to thebuffer curve) along the liquid line of descent from basalt. This is alsoconsistent with observations from lunar zircons. Finally, we note thatMartian basalts (for example, shergottite meteorite Dar al Gani 476)exhibiting little or no interaction with a crustal component are onlyabout 1 log unit higher than the estimated fO2 of the upper mantle25.Furthermore, calculated oxygen fugacities from the same meteorite areall within a log unit of the average (FMQ 2 2.5), even though phasescover a crystallization range of .300 uC.

As in all studies in which mantle fO2 is estimated from mantle-derived glasses or minerals3,9, it is important to emphasize that thecalculated fO2 may not directly reflect that of the mantle source region.We note, however, that mantle-derived Hadean zircons would have hadto crystallize from residual melts that underwent ,4 log units of changein fO2 to be reconcilable with mantle source regions in equilibriumwith the IW buffer. At present and in the Archaean, relative fO2

changes along magmatic liquid lines of descent are broadly of theorder of ,1 log unit4,10,26,27. If the evolution of Hadean melts wassimilarly constrained, then our results imply that the mantle reachedits present-day oxidation state ,4,350 Myr ago (Fig. 3). This inter-pretation is consistent with core formation models calling for physicaland chemical isolation from the (upper) mantle ,30 Myr after planetformation28.

If our deductions regarding the oxidation state of Hadean magmasare correct, then the speciation of gases emanating from the Earth at thistime would have been dominated by CO2, SO2, H2O and N2 (ref. 26).An atmosphere of this composition is known to yield a lower abund-ance of sugars and especially amino acids and nucleotides1. If a highlyreduced atmosphere is required for the origin of life, then it may haveoccurred exceptionally early on our planet. However, pre-4,400-Myroutgassing of H2 coupled with slow escape29 may have resulted in anatmosphere out of equilibrium with Earth’s interior. Alternatively, a‘late veneer’ may have served as a source of pre-biotic molecules. Thesealternative scenarios notwithstanding, our results offer the first glimpseinto the oxidation state of terrestrial magmas before 3,900 Myr ago, andthe data presented here open up the possibility of exploring the oxida-tion state of zircon-bearing rocks throughout the geologic record. Thiscalibration will be especially valuable in the early Archaean or Hadeaneons, for which zircon is often the only source of primary chemicalinformation.

METHODS SUMMARYStarting materials were contained in platinum (Pt) or silver (Ag) capsules, and fO2

was buffered by placing ruthenium and ruthenium dioxide (Ru1RuO2; RRO) indirect physical contact with the experimental charge or by use of a separate bufferchamber separated by a Pt or silver-palladium (Ag-Pd) membrane (NNO orMMO). Quenched glasses were analysed by electron microprobe, with at leastone measurement made at the top and bottom of the capsule to assess melthomogeneity. Synthesized zircons were small (,3–20mm) and in some casesthe glass was dissolved with hydrofluoric acid (HF) and the zircons recovered inorder to obtain more accurate measurements.

For the inset plot of Fig. 1, Prx 5 2/3(La)x 1 1/3(Nd)x for ref. 14, where xrepresents a normalized value, either to CHUR or the melt. Ce anomalies fornatural zircons6,19,23–25,30 are calculated in a similar fashion to equation (1) in thetext: (Ce/Ce*)CHUR 5 CeCHUR/(LaCHUR 3 PrCHUR)1/2, where LaCHUR and so onrepresents normalization to the chondritic uniform reservoir rather than the meltconcentration. Ce anomalies are plotted for Hadean zircons with U–Pb ages thatare $94% concordant, and with REE1Ti measurements that reflect primaryigneous chemistry5,6,17. Zircon crystallization temperatures in Fig. 2 are calculatedassuming unity Ti-activity7; sub-unity Ti activity (for example, ,0.5) results insystematic shifts of all data to higher values of fO2, though shifts are within error ofthe standard deviation of each population. Lunar zircons are from Apollo 14polymict breccias 14304, 14305 and 1432119 and are normalized to an averagefO2 of IW at 700 uC, for direct comparison with the other data in the probabilitydistribution plot of Fig. 2. For the inset of Fig. 2b, average fO2 values calculatedfrom zircon Ce anomalies are plotted against independent estimates from lunarbasaltic igneous rocks (IW 2 0.6)18, terrestrial mid-ocean ridge basalts10 (FMQ),and average oxygen fugacities returned from Bishop Tuff titanomagnetite-ilmenitepairs (NNO)27.

Full Methods and any associated references are available in the online version ofthe paper at www.nature.com/nature.

Received 22 March; accepted 6 October 2011.

1. Kasting, J. F. Earth’s early atmosphere. Science 259, 920–926 (1993).2. Canil, D. Vanadium partitioning and the oxidation state of Archaean komatiite

magmas. Nature 389, 842–845 (1997).3. Delano, J.W.Redox history of the Earth’s interior since ,3900Ma: implications for

prebiotic molecules. Orig. Life Evol. Biosph. 31, 311–341 (2001).4. Burgisser, A. & Scaillet, B. Redox evolution of a degassing magma rising to the

surface. Nature 445, 194–197 (2007).5. Cavosie, A. J., Valley, J. W., Wilde, S. A., &. Edinburgh Ion Microprobe Facility.

Magmatic d18O in 4400–3900 Ma detrital zircons: a record of the alteration andrecycling of crust in the early Archean.Earth Planet. Sci. Lett. 235, 663–681 (2005).

6. Cavosie, A. J., Valley, J. W., Wilde, S. A., &. Edinburgh Ion Microprobe Facility.Correlated microanalysis of zircon: trace element, d18O, and U-Th-Pb isotopicconstraints on the igneous origin of complex 3900 Ma detrital grains. Geochim.Cosmochim. Acta 69, 637–648 (2006).

6 Crust-derived

IW

Archaean and present-day upper mantle

Zircon in equil. with mantle (δ18O)

4

2

0

–2

3,800 3,900 4,000 4,100

ΔFM

Q

207Pb/206Pb* age (Myr)

4,200 4,300 4,400

–4

–6

Figure 3 | Oxygen fugacities of Hadean melts plotted against zirconcrystallization age. Errors for individual points are based on the standarddeviation of the entire data set (n 5 20; 1s 5 6 2.3 log units). Zircons withmantle signatures are within error of the estimated oxidation state of the present-day and Archaean upper mantle3,10,26. On average, oxygen fugacities are lower foryounger zircons, though this trend is not robust given the present data set.

LETTER RESEARCH

1 D E C E M B E R 2 0 1 1 | V O L 4 8 0 | N A T U R E | 8 1

Macmillan Publishers Limited. All rights reserved©2011

マントルはFMQ@4.35Ga[Farquhar et al., 2000] [Trail et al., 2011]

いかにして “地球” をつくるか

・大枠は太陽系形成論に則って説明される!!・巨大天体衝突過程で大量の破片（GIF）が形成!!・GIF による地球の円軌道化 & レイトベニア!!・原始海洋 + GIF → 大量の水素大気の発生!!・初期海洋質量が適度に減少（→ 1海洋）!!・酸化的な脱ガスの下で還元的な大気を長期間保持