Abstract

We have conducted grain growth and creep experiments on the forsterite (Fo)-enstatite (En) system at 1 atmosphere and temperatures of 1260-1360 °C, and with variable volumetric fractions of the two minerals (Fo(100) to Fo3En(97)).

Firtsly, grain growth of these samples are focused. The grain size ratios of forsterite and enstatite in simply annealed (reference) and deformed samples follow the Zener relationship of dI/dII = β/fIIz.For samples where fEn < 0.5, I (the first phase) is forsterite, II (the second phase) is enstatite, β = 0.67, and z = 0.52; for samples where fEn > 0.5, I is enstatite, II is forsterite,β = 0. 73, and z = 0.53; fII is the volume fraction of the second phase, and d is the grain size. Grain growth in the reference samples conforms to the relationship ds4-do4≒k・t (where ds is the grain size under static conditions, d0 is the initial grain size, k is the grain growth coefficient, and t is time), and in the deformed samples the relationship is dΣ≒ ds exp (0.42 ε) (where dε is the grain size under dynamic conditions and ε is the strain). The observed growth coefficient for the first phase (kI) becomes smaller with increasing fII, and this is consistent with the theoretical predictions of kI = (β/fIIz)4・〓where fII is the volume fraction of the second phase, γ is the interfacial energy between the first and second phases, c is the concentration of rate-controlling elements, w is the width of the grain boundary, DiGB is the diffusion coefficient of rate-controlling elements for grain boundary diffusion, V is the molar volume of the solute, w is a slowly varying quantity with different fractions of the second phase, G is a geometric constant, R is the gas constant, and T is the absolute temperature. Overall, our results are consistent with previously proposed grain growth models for static conditions that use mineral physical parameters such as diffusivity (DiGB) and interfacial energy (γ). We discuss grain size variations within the mantle, with lithologies from dunite to pyroxenite, and we go on to present a method that predicts the grain sizes of different mantle lithologies, provided the mineral parameters of diffusivity and interfacial energy are known.

Secondary, mechanical characteristics of these samples are focused. At constant temperatures and strain rates, the flow stresses of the samples decrease with increasing fEn for samples with 0 < fEn < 0.5, and increase with increasing fEn for samples with 0.5 < fEn < 1. This behavior is explained primarily by grain size changes as a function of fEn under grain size sensitive creep. The pre-exponential term, stress, grain size exponent, and activation energy are flow parameters we have determined for a wide range of enstatite fractions (fEn). Samples with a low fEn (≦0.03) exhibit creep characteristics that correspond to dislocation accommodated grain boundary sliding creep (i.e., a stress exponent of 3), whereas diffusion accommodated grain boundary sliding creep is typical of high fEn samples (i.e., the stress and grain size exponents are ~1 and ~2, respectively). The change of creep mechanism is attributed to changes in grain size with respect to fEn. Overall, in the Fo-En system, the change of flow strength as a function of fEn under grain size sensitive creep can be explained primarily by the change of grain size, and secondarily by changes in the volume fractions of phases that have different flow strengths (enstatite is ca. 10 times stronger than forsterite under our conditions of deformation). Viscosities of all samples can be reproduced in a viscosity model that takes into account, (1) the grain sizes that have been estimated by the grain growth laws, and (2) the flow laws for monomineralic systems of forsterite and enstatite. Furthermore, we demonstrate that our model can be extended to make predictions of viscosity in other mineral assemblages.

Finally geological applications especially on fine-grained mylonite are presented. Microstructures of peridotitic ultramylonite from Oman ophiolite are compared with that of our experimental products. Around 0.5 mm thick layers with different amount of pyroxenes (orthopyroxene and clinopyroxene) in each layer are developed well in the rocks forming their foliation. Average grain sizes and grain size ratio of olivine and pyroxenes from each layer are compared with respect to the fraction of pyroxenes (fpx) in the layers. Grain size of the pyroxenes are almost constant among different fpx layers, whereas olivine grain size decreases significantly with increasing fpx, both of which were the characteristic features found in forsterite + enstatite aggregates after the grain growth experiments. Further, Zener relation (log dol/dpx versus log fpx) in the mylonite is remarkably comparable to that found in the experiments. These observations indicate the operation of effective pinning of pyroxene grains on grain growth of olivine grains during the deformation of the rocks. Olivine grains in high fpx (fpx > 0.04) layers do not exhibit lattice preferred orientation (LPO) whereas the grains in low fpx (fpx < 0.04) layers exhibit LPO indicating the deformation proceeded via diffusion and dislocation accommodated creep in the former and the latter layers, respectively. The observation of finer olivine grain size as predicted from the experiments and theory in the low fpx layers is explained well by operation of dislocation accommodated creep where the size is determined by stress. Based on our flow laws obtained for diffusion creep of forsterite + enstatite in our series of work, pyroxene-rich layers are 1 to 3 orders of magnitude weaker relative to essentially pyroxene-free layers. We simulated grain growth in forsterite + enstatite aggregates with fpx < 0.5 under temperature and stress conditions of 700 C and 120 MPa, which are estimated from the chemical compositions of pyroxenes and olivine grain size in the low fpx layers, respectively, based on our grain growth and flow laws. The results show that the observed grains grew from < 1 μm grains with spending ~104 years to reach presently observed grain size.

本論文は、地球の上部マントル物質の8割以上を占めるかんらん石と斜方輝石からなる2相系の流動特性を明らかにすることをめざし、フォルステライト(Fo)とエンスタタイト(En)で構成される超細粒合成物質を用いて2相の堆積比率変化が粒径と粘性に与える影響を徹底的かつ綿密な粒成長実験と変形実験によって解明した論文であり、5章からなる。

第一章はイントロダクションで,マントルの変形機構を概観し、粒径に依存する拡散律速変形機構の粒成長と多相系の効果の理解がこれまで不十分であることを述べ、それを実験的に解明するという本研究の目的を述べている。

第二章では、2相の体積分率X(En)=En/(Fo+En)を0~1、温度を1260~1360°C、応力を0~100MPaまで変えた粒成長実験結果がまとめられ、主要第1相がFoの場合とEnの場合両方について、第2相の割合が増加するにつれて第1相の粒径が減少することを明らかにしている。さらに、この粒径のX(En)依存性がゼナー則に従っていることを明らかにし、微細構造の特徴とあわせて、主要第1相の成長を粒界や3重点に位置する第2相がピン留めし阻害する機構がFo-En系の粒成長過程でも起きていることを明らかにし、粒成長モデルに基づいて具体的な成長支配パラメータを推定している。

第三章では、変形実験の結果を第二章で得た粒成長則を適用し、粘性率の粒径、温度、X(En)依存性を明らかにしている。その結果、主要第1相分率が減少すると、第1相の粘性率が減少することを明らかにし、粘性率を支配しているのは粒径とそれぞれの相の硬さであると推測している。さらに、歪み速度、粒径、温度の関係に基づいて流動則パラメータを決定し、変形が拡散律速型の粒界すべりクリープによっていることを明らかにしている。

第四章では、第二、第三章の結果を天然の岩石に適用している。オマーンオフィオライトのマントル層に見られミクロンサイズの細粒鉱物からなるウルトラマイロナイトを対象とし、粒径、相分率、結晶方位定向性を測定し、結晶方位定向性が顕著で転位クリープで変形したとわかる試料を除外した上で、かんらん石と輝石の粒径比がゼナー則で説明できることを明らかにしている。さらに、ミクロンサイズの初期粒径を仮定し、マントル剪断帯の粒径、粘性率、歪みの時間発展をモデル化し、本論文で得た結果の有用性を示している。

第五章では結論の章であり、本研究で明らかになった結果をまとめている。

本研究は2相よりなる系の粒成長と変形挙動の2相の比率への依存性をこれまでにない質量ともに優れた実験によって明らかにし、マントル物質の流動の理解を大きくすすめる成果をあげた点で、地球科学的意義は大きい。よって本審査委員会は、全員一致で本論文が本学の博士(理学)の学位を授与するに値するものと認定した。

なお本研究の一部は、平賀岳彦氏、David L. Kohlstedt氏、Mark E. Zimmerman氏、道林克禎氏との共同研究であるが、論文提出者が主体となって行ったもので、その寄与が十分であると判断する。