Chondrules are major constituent of chondrites, occupying up to 70 vol% of chondrites, which are igneous objects cooled rapidly in the early solar system. In spite of numerous analytical and theoretical investigations, the formation mechanism and the origin of compositional diversities are still not known. Because there are systematic differences in the size, abundance of types, and isotopic compositions of constituent minerals in chondrules among different chemical group of chondrites, it is expected that chondrules in different chemical group of chondrites would represent difference in physicochemical environment for their origin.

Recent development of SIMS （secondary ion mass spectrometry） and its application to small chondrules have revealed that chondrules were formed in a fairly long time interval in the early solar system （2 to 3 m.y.）, which makes understanding of their origin more difficult. Although the number of age measurements has been increasing, the information is still limited and more extensive and systematic study is required for better understanding of the origin. In particular, systematic investigation of formation age, chemical compositions, and mineralogical characteristics among different chemical groups of chondrites is crucial in connection with the origin of chemical group of chondrites. Also important is evaluation of the 26A1-26Mg system, which has not been fully done in previous works and the possibility of inclusion of isotopically disturbed ages is contained in literature data. In the present study, I have carried out petrological and 26A1-26Mg chronological study on chondrules in Yamato-81020（Y-81020）, the most primitive CO3.0 chondrite, and compared the results with those in ordinary chondrites. Then, their similarity and difference are used to constrain the differences in physicochemical environments of chondrule formation and evolution of the early solar nebula.

Magnesium isotopic composition of plagioclase is measured on fourteen FeO-poor Type I, two FeO-rich Type II, and one aluminum-rich （A1-rich） chondrule in Y-81020 with SIMS. Excesses in 26Mg with sufficient precision（0.5-2‰） are found in all chondrules studied. Chemical zoning of Na and Mg in anorthite shows that the zoning was produced during crystallization from liquid but not during thermal metamorphism in a parent body. Thus, the chronological data in individual plagioclase grains keeps initial information with regard to the formation age without secondary isotopic disturbance. The 26A1-26Mg ages are obtained to be 1.3-2.4 Myr for Type I, 2.0 Myr forType II, and 2.4 Myr for A1-rich chondrules, assuming homogeneous distributions of 26A1 and the initial 26A1/27A1 ratio of a canonical value （5×10-5）.

The formation age of chondrules in Y-81020 is in good coincidence with that of Semarkona, Bishunpur, and Krymka, LL 3.0 to 3.1 chondrites, in literature, which indicates that Type I chondrules in the CO chondrite were formed contemporaneously with ferromagnesian chondrules in LL chondrites （Figure 1）.The concurrent formation of chondrules of CO and LL chondrites suggest that the cosmochemical differences between CO and LL chondrites might be caused by spatial distinction of their chondrule formation in the protoplanetary disk.

Systematic examination of relationship among texture, mineral assemblage, mineral composition, bulk chemical composition, and 26Mg isotopic compositions shows that there seem to be no correlation among those properties except for the formation age difference between Type I and Type II chondrules. It is worth noting that Type II chondrules tend to have later forming age, and the relative abundance of Type I and Type II chondrules differs between CO and LL chondrites. The bulk chemical compositions of Type I and Type II chondrules obtained by averaging several hundreds of point analyses of EPMA show similarities in almost all elements and the major difference is in the state of iron, either metallic Fe or FeO in silicates, that is, the redox state in another word.

The similar formation age of chondrules in LL and CO chondrites, later generation of Type II chondrules compared to Type I chondrules in Y-81020, and larger abundance of Type I chondrules in CO chondrites than LL chondrites are modeled by taking the stability of silicate, organic materials and ice, temperature evolution of the nebula, and plausible physical process of chondrule formation into consideration. Most plausible way to change the redox state of the solar nebula would be the change of relative abundance of carbon and oxygen, of which main carrier in the solar nebula is organic materials and ice, respectively. Dust grain in the protosolar molecular cloud are thought to have a layered structure with silicate core, organic mantle, and icy crust （Greengberg, 1998）, which were heated at the high temperature stage of the early evolution of the nebula to evaporate partially or totally. The extent of evaporation should be a function of radial distance from the star （the sun）; silicate is most refractory, organic materials next, and icy component is most volatile. The protoplanetary disk was heated in the early stage of evolution, which is a function of radial distance from the protosun, and the disk is divided into three regions: the ice and organic materials were evaporated in the inner region, only icy component evaporated in outer region, and all the components remained in the outermost region. Although the heating mechanism for chondrule formation is one of the most severely debated theme in planetary science which has not yet been in consensus, the present study give some constrains; the heating process took place repeatedly in fairly wide region of the disk and lasted for almost 2 Myr. If shock wave should have attacked the surface layer of the dust enriched midplane of the disk to evaporated the condensed materials, the redox state of gas differed in the three regions because of the difference of the materials existed: the inner most region was oxidizing due to evaporation of the silicate component alone, reducing in the outer region due to evaporation of silicate and organic materials, and intermediate in the outermost region due to evaporation of all the three components. The oxidizing inner region should be responsible for formation of Type II chondrules, and the reducing outer region for Type I chondrules, and the outermost region for Type II chondrules. The boundaries between the regions moved inward with time because the nebula continuously cooled. After 2 to 3 Myr, chondrule forming heating ceased, the relative abundance of Type I and II chondrules varied with distance from the star: chondrules in CO chondrites were formed in the outer or outermost region, where Type I is abundant, whereas those in LL in the inner region, whereType II is abundant （Figure 2）. The present scenario can explain the relationships between chondrule forming ages of Type I and Type II in the CO chondrite, and the difference in the redox states between LL and CO chondrites. Considering the complementarity between chondrule and matrix （e.g. Kong and Palme, 1999） from volatility-controlled fractionated solar nebula （e.g. Wasson and Chou, 1974）, enstatite chondrites were located in the intermediate distance from the sun between ordinary and carbonaceous chondrite regions. The formation process of chondrules and chondrites proposed in this dissertation is successful in explaining all their major properties （e.g. the redox state of chondrule and chondrite, the relative abundance of chondrule types, chondrule formation time） as a function of the distance from the proto-sun and the protoplanetary disk evolution.

Figure l. Comparison of relative 26AI ages ofchondrule formation after CAIs from SIMS study. The datat from [1] Kita et al. （2000）, [2] Mostefaoui et al. （2002）, [3] Hutcheon and Huchison （1989）,[4] McKeegan et af. （2000）,[5] Kita et al. （2005）,[6] Srinivasan et al. （2000）, and [7] Kunihiro et al. （2004）.

Figure 2. Schematic diagrams of protoplanetary disk evolutions relating to LL and CO chondrule formations. See in text for details.

本論文は、ＣＯコンドライト（炭素質隕石）の中で最も変成度の低い隕石のひとつであるＹ―81020に含まれるコンドリュール（隕石中に数十パーセント含まれる0.1-1mmサイズの岩石質の球粒）に対し、イオンマイクロプローブを用いた精密な26Al-26Mg年代測定法を適用し、ＣＯコンドライト中のコンドリュールの詳細な年代学的研究をおこなったものである。さらにその結果をもとに、原始太陽系円盤の進化と、その中での種々のタイプのコンドリュール、種々の隕石グループの形成プロセスの解明をめざした意欲的な論文である。

本論文の最大の成果は、ＣＯコンドライトのコンドリュールに対して非常に精密な26Al-26Mg年代測定をおこなったこと、さらに得られた年代が、隕石母天体における変成の影響を免れたコンドリュールの結晶化年代そのものであることを丹念に検証したことである。コンドリュールの26Al-26Mg年代測定データは、普通コンドライト（ＬＬコンドライト）に関しては木多らのパイオニア的研究があったが、鉄成分の少ないタイプIコンドリュールに関してはデータが少なく、とりわけＣＯコンドライトに関してはデータが大きく欠落していた。本論文では、分析方法そのものにも精度向上の工夫と努力が見られ、コンドリュール形成年代を数十万年（多くは20-30万年）以内の誤差で決定することに成功した。また、分析点の鉱物（主としてアノーサイト）の元素プロファイルを詳細に調べることにより、得られた年代が鉱物の結晶化年代、すなわちコンドリュールの形成年代であることを丹念に実証した。これだけ詳細かつ丹念なコンドリュールの年代学的研究はかつてなく、画期的なものである。

本論文によって、次のような事実が明らかにされた。（1）ＣＯコンドライトのコンドリュールのうち、鉄成分の少ないタイプIコンドリュールの形成年代は、ＣＡＩ（太陽系最古の固体物質で、難揮発性包有物ともよばれる）形成後、百万年〜二百五十万年の範囲にあり、ＬＬコンドライトのコンドリュール形成と同時期である。（2）鉄成分の多いタイプIIコンドリュールの形成年代は、タイプＩコンドリュールと同時期かやや遅い（ＣＡＩ形成後、二百万年〜三百万年）。以上の事実は、ＣＯコンドライトとＬＬコンドライトのコンドリュールの物理的、化学的、あるいは同位体的なバリエーションが、コンドリュールの形成年代差を反映したものではなく、原始太陽系円盤内における形成場所の違いを反映したものであることを強く示唆する。これは非常に重要な結論であり、この結論をもとに本論文は新たな原始太陽系円盤の進化モデルの構築へとむかう。

本論文は、コンドリュールのタイプI、IIの違いがコンドリュール形成場所である原始太陽系円盤内での酸化還元状態を反映していること、その酸化還元状態が、シリケイト微粒子のまわりに付着した有機物や氷の蒸発・凝縮によってコントロールされることに着目し、異なるタイプのコンドリュールの成因および各コンドライトグループの形成過程について、新しいモデルを提唱した。このモデルの重要なポイントは、温度が上昇すると、最初に氷が、次に有機物が蒸発するが、一度蒸発した有機物は化学反応により低分子のガスに分解され、再凝縮できないという点である。本論文は、原始太陽系円盤が太陽に近いシリケイトに富む（有機物や氷がない）酸化的領域、その外側の有機物（炭素）に富む還元的領域、一番遠方の有機物および氷を含む酸化的領域にわかれることを指摘し、原始太陽形星雲の温度低下とともにそれぞれの領域の境界が時間的にどう変化するかを考察し、各場所での酸化還元状態の変化と、それによる異なるタイプのコンドリュールの形成、異なる隕石グループの形成をモデル化した。これは従来のモデルとは全く異なる、非常にオリジナリティーが高いものである。モデルそのものは、まだまだ詳細を検討する余地があるが、本論文が単なるコンドリュールの年代学的研究にとどまらず、隕石学・惑星科学の大きな課題であった、原始太陽系円盤の進化とコンドライト形成過程の解明に向けて、大きな示唆に富む方向性を示したことは、オリジナリティーのある研究として非常に高く評価できる。

以上のように、本論文は、ＣＯコンドライトのコンドリュールの詳細な年代学的成果とともに、その結果に基づく原始太陽系円盤の進化に関する新たなモデルの提唱についても、オリジナリティーに富むすぐれた研究であると評価できる。よって、博士（理学）の学位を授与できると認められる。