榴辉岩部分熔融过程中的同位素分馏

本文对榴辉岩部分熔融过程中不同同位素体系是否存在分馏这一当前研究热点进行了综述。榴辉岩作为研究洋陆俯冲、超高压变质以及壳幔相互作用的主要岩石类型,其部分熔融与地壳增生、板片折返过程以及俯冲隧道中元素的迁移分配等具有紧密的联系。作为典型的高压-超高压变质岩石,榴辉岩可通过俯冲带将壳源信息携带至地幔深部,影响地幔的化学组成,并可在大洋玄武岩中得以表现。近些年,随着仪器分析技术的发展,实验研究和理论计算均表明榴辉岩部分熔融过程中稳定同位素可以产生显著的分馏。作为常见的放射性成因子体同位素和传统稳定同位素Sr-Nd-Hf-O被广泛应用于源区示踪、岩浆混合以及结晶分异等过程。但目前有研究指出,在非平衡熔融过程中,熔体和源区的Sr-Nd-Hf-O同位素可发生解耦,导致二者的同位素组成不均一。另外,通过研究榴辉岩及其熔融产物的金属稳定同位素特征,发现榴辉岩部分熔融过程中,由于石榴石效应,会造成Ca、Mg、Fe、Li等金属稳定同位素的分馏。因此,当利用稳定同位素示踪榴辉岩熔体的源区时,需要考虑其分馏的影响。

     

Mineral/melt partitioning of trace elements during hydrous peridotite partial melting

This experimental study examines the mineral/melt partitioning of incompatible trace elements among high-Ca clinopyroxene, garnet, and hydrous silicate melt at upper mantle pressure and temperature conditions. Experiments were performed at pressures of 1.2 and 1.6 GPa and temperatures of 1,185 to 1,370 °C. Experimentally produced silicate melts contain up to 6.3 wt% dissolved H ₂O, and are saturated with an upper mantle peridotite mineral assemblage of olivine+orthopyroxene+clinopyroxene+spinel or garnet. Clinopyroxene/melt and garnet/melt partition coefficients were measured for Li, B, K, Sr, Y, Zr, Nb, and select rare earth elements by secondary ion mass spectrometry. A comparison of our experimental results for trivalent cations (REEs and Y) with the results from calculations carried out using the Wood-Blundy partitioning model indicates that H ₂O dissolved in the silicate melt has a discernible effect on trace element partitioning. Experiments carried out at 1.2 GPa, 1,315 °C and 1.6 GPa, 1,370 °C produced clinopyroxene containing 15.0 and 13.9 wt% CaO, respectively, coexisting with silicate melts containing ~1–2 wt% H ₂O. Partition coefficients measured in these experiments are consistent with the Wood-Blundy model. However, partition coefficients determined in an experiment carried out at 1.2 GPa and 1,185 °C, which produced clinopyroxene containing 19.3 wt% CaO coexisting with a high-H ₂O (6.26±0.10 wt%) silicate melt, are significantly smaller than predicted by the Wood-Blundy model. Accounting for the depolymerized structure of the H ₂O-rich melt eliminates the mismatch between experimental result and model prediction. Therefore, the increased Ca ²⁺ content of clinopyroxene at low-temperature, hydrous conditions does not enhance compatibility to the extent indicated by results from anhydrous experiments, and models used to predict mineral/melt partition coefficients during hydrous peridotite partial melting in the sub-arc mantle must take into account the effects of H ₂O on the structure of silicate melts.     

Equivalence of equations describing trace element distribution during equilibrium partial melting

The mathematical equivalence of several equations used to calculate the evolution of trace element abundances during equilibrium partial melting is demonstrated.     

Kinetic effects on trace element partitioning

Kinetic effects on trace element partitioning have been measured for anorthite, forsterite, and diopside grown from synthetic compositions doped with REE. A seeding technique allowed determination of crystal growth rates and partitioning information was obtained from electron microprobe analyses. Compositional deviations from equilibrium values were sought in the crystals and as gradients in the quenched liquids adjacent to the crystals. The principal result is that large deviations in trace element distribution coefficients from equilibrium values do not occur because of a compensating effect. Rapid growth depletes the melt adjacent to the crystal in the elements of which the crystal is composed, leading to different values for apparent distribution coefficients. However, as the boundary layer melt becomes depleted in the components of the crystal, growth slows and the size of the compositional perturbations decreases. Crystals grown at very high rates (e.g., > 0.2 μm/sec for diopside) tended to be too small for accurate microprobe analyses, but are probably not compositionally extreme since the melts adjacent to the crystals did not acquire sizable compositional gradients. At moderately high growth rates (e.g., 0.02 μm/sec), crystals form in the presence of boundary layer compositions perturbed by as much as 10% from bulk melt values and, in diopside, attain concentrations for excluded trace elements about 70% higher than equilibrium values for crystals plus bulk melt. At the slower growth rates typical of igneous systems, kinetic effects on trace element partitioning are probably negligible.     

Patterns of trace element partitioning

When a phenocryst and its lava matrix have been analysed for both major and trace elements, plots of partition coefficient (mineral/matrix) against ionic radius give a family of sub-parallel curves, one univalent, one divalent, etc. Onuma et al. (1968) demonstrated this with two analyses and concluded, from the shape of these curves, that trace element partition between phenocryst and groundmass is determined primarily by crystal structure of the phenocryst. In this study, over 50 such analyses, taken from the literature, have been plotted on modified Onuma diagrams, in order to analyse the constant and variable factors in the curves. It is demonstrated that these curves can provide important additional information in trace element studies by revealing, for example, the site or sites in a mineral which a given element is occupying, the valency state of the element and even, in some cases, the proportion of different valency states present.     

Trace element partitioning during high-P partial melting and melt-rock interaction; an example from northern Fiordland, New Zealand

Pods of granulite facies dioritic gneiss in the Pembroke Valley, Milford Sound, New Zealand, preserve peritectic garnet surrounded by trondhjemitic leucosome and vein networks, that are evidence of high‐P partial melting. Garnet‐bearing trondhjemitic veins extend into host gabbroic gneiss, where they are spatially linked with the recrystallization of comparatively low‐P two‐pyroxene‐hornblende granulite to fine‐grained high‐P garnet granulite assemblages in garnet reaction zones. New data acquired using a Laser Ablation Inductively Coupled Plasma Mass Spectrometer (LA‐ICPMS) for minerals in various textural settings indicate differences in the partitioning of trace elements in the transition of the two rock types to garnet granulite, mostly due to the presence or absence of clinozoisite. Garnet in the garnet reaction zone (gabbroic gneiss) has a distinct trace element pattern, inherited from reactant gabbroic gneiss hornblende. Peritectic garnet in the dioritic gneiss and garnet in trondhjemitic veins from the Pembroke Granulite have trace element patterns inherited from the melt‐producing reaction in the dioritic gneiss. The distinct trace element patterns of garnet link the trondhjemitic veins geochemically to sites of partial melting in the dioritic gneiss.     

An experimental study of the grain-scale processes of peridotite melting: implications for major and trace element distribution during equilibrium and disequilibrium melting

The grain-scale processes of peridotite melting were examined at 1,340°C and 1.5 GPa using reaction couples formed by juxtaposing pre-synthesized clinopyroxenite against pre-synthesized orthopyroxenite or harzburgite in graphite and platinum-lined molybdenum capsules. Reaction between the clinopyroxene and orthopyroxene-rich aggregates produces a melt-enriched, orthopyroxene-free, olivine + clinopyroxene reactive boundary layer. Major and trace element abundance in clinopyroxene vary systematically across the reactive boundary layer with compositional trends similar to the published clinopyroxene core-to-rim compositional variations in the bulk lherzolite partial melting studies conducted at similar P–T conditions. The growth of the reactive boundary layer takes place at the expense of the orthopyroxenite or harzburgite and is consistent with grain-scale processes that involve dissolution, precipitation, reprecipitation, and diffusive exchange between the interstitial melt and surrounding crystals. An important consequence of dissolution–reprecipitation during crystal-melt interaction is the dramatic decrease in diffusive reequilibration time between coexisting minerals and melt. This effect is especially important for high charged, slow diffusing cations during peridotite melting and melt-rock reaction. Apparent clinopyroxene-melt partition coefficients for REE, Sr, Y, Ti, and Zr, measured from reprecipitated clinopyroxene and coexisting melt in the reactive boundary layer, approach their equilibrium values reported in the literature. Disequilibrium melting models based on volume diffusion in solid limited mechanism are likely to significantly underestimate the rates at which major and trace elements in residual minerals reequilibrate with their surrounding melt. Electronic supplementary material  The online version of this article (doi:) contains supplementary material, which is available to authorized users.     

Experimental geochemistry of Pu and Sm and the thermodynamics of trace element partitioning

The partitioning of Pu and Sm between diopside/liquid and whitlockite/liquid has been investigated experimentally to evaluate the geochemical coherence of Pu and the light REEs. ${}^{\mathrm{Pu}}\text{D}{}^{\mathrm{Sm}}\text{D}$ is 2̃ for both diopsidic pyroxene and whitlockite. This small amount of fractionation would be decreased further if Pu were compared to Ce or Nd. Our experimental results thus validate the suggestion that Pu behaves as a LREE during igneous processes in reducing environments.Our data and the data of Ray et al. (1983) indicate that temperature rather than melt composition is the most important control on elemental partitioning. This is true even though we demonstrate that additions of only 1–2 wt.% of P₂O₅ to the diopside-anorthite-albite system change ^PuD_cpx by a factor of two. Our data suggest that P₂O₅ in aluminosilicate melts serves as a complexing agent for the actinides and lanthanides.     

Fluid-mobile trace element constraints on the role of slab melting and implications for Archaean crustal growth models

We use published and new trace element data to identify element ratios which discriminate between arc magmas from the supra-subduction zone mantle wedge and those formed by direct melting of subducted crust (i.e. adakites). The clearest distinction is obtained with those element ratios which are strongly fractionated during refertilisation of the depleted mantle wedge, ultimately reflecting slab dehydration. Hence, adakites have significantly lower Pb/Nd and B/Be but higher Nb/Ta than typical arc magmas and continental crust as a whole. Although Li and Be are also overenriched in continental crust, behaviour of Li/Yb and Be/Nd is more complex and these ratios do not provide unique signatures of slab melting. Archaean tonalite-trondhjemite-granodiorites (TTGs) strongly resemble ordinary mantle wedge-derived arc magmas in terms of fluid-mobile trace element content, implying that they did not form by slab melting but that they originated from mantle which was hydrated and enriched in elements lost from slabs during prograde dehydration. We suggest that Archaean TTGs formed by extensive fractional crystallisation from a mafic precursor. It is widely claimed that the time between the creation and subduction of oceanic lithosphere was significantly shorter in the Archaean (i.e. 20 Ma) than it is today. This difference was seen as an attractive explanation for the presumed preponderance of adakitic magmas during the first half of Earth's history. However, when we consider the effects of a higher potential mantle temperature on the thickness of oceanic crust, it follows that the mean age of oceanic lithosphere has remained virtually constant. Formation of adakites has therefore always depended on local plate geometry and not on potential mantle temperature.     

Aluminum-dependent trace element partitioning in clinopyroxene

As technical advances have dramatically increased our ability to analyze trace elements, the need for more reliable data on the compositional dependence of trace element partitioning between minerals and melt has become increasingly important. The late-Cretaceous Carmacks Group of south central Yukon comprises a succession of primitive high-Mg ankaramitic lavas characterized by shoshonitic chemical affinities and containing large complexly zoned clinopyroxene phenocrysts. The compositional zonation of the clinopyroxene phenocrysts is characterized by relatively Fe-rich (Mg^# = Mg/(Mg + Fe) = 0.85), but mottled, cores surrounded by mantles of cyclically-zoned clinopyroxene whose Mg^# varies repeatedly between 0.9 and 0.80. These cyclically zoned clinopyroxene mantles appear to record the repeated influx and mixing of batches of primitive with more evolved magma in a deep sub-crustal (∼1.2 GPa) magma chamber(s). Laser ablation ICP-MS was used to analyze the trace element variation in these zoned clinopyroxenes. The results indicate more than a threefold variation in the absolute concentrations of Th, Zr, rare earth elements (REE), and Y within individual clinopyroxene phenocrysts, with no apparent change in the degree of REE or high field strength element (HFSE) fractionation. The variation in absolute abundances of trace elements correlates closely with the major element composition of the clinopyroxene, with the most enriched clinopyroxene having the lowest Mg^# and highest Al contents. The problem is that the amount of crystal fractionation required to explain the major element variation (∼20%) in these clinopyroxene phenocrysts cannot explain the increase in the abundance of the incompatible trace elements, which would require more than 70% crystal fractionation, if constant partition coefficients are assumed. The anomalous increase in incompatible trace elements appears to reflect an increase in their partition coefficients with increasing Al^IV in the clinopyroxene; with an increase in Al₂O₃ from 1.5 to 4.0 wt.% during ∼20% crystal fractionation over a temperature decrease of ∼100°C being associated with more that a threefold increase in the partition coefficients of Th, Zr, REE, and Y. The magnitude of these increases may indicate that the substitution of these trace elements into clinopyroxene is better modeled in some natural systems by a local charge balance model, rather than the distributed charge model that better replicates the results of annealed experiments. These findings indicate that the effect of Al on the partition coefficients of incompatible trace elements in clinopyroxene may be under appreciated in natural magmatic systems and that the application of experimentally determined clinopyroxene partition coefficients to natural systems must be done with caution.     

The influence of carbon on trace element partitioning behavior

Carbon has been proposed as a potential light element in planetary cores, included in models of planetary core formation, and found in meteoritic samples and minerals. To better understand the effect of C on the partitioning behavior of elements, solid/liquid partition coefficients (D = (solid metal)/(liquid metal)) were determined for 17 elements (As, Au, Co, Cr, Cu, Ga, Ge, Ir, Ni, Os, Pd, Pt, Re, Ru, Sb, Sn, and W) over a range of C contents in the Fe-Ni-C system at 1 atm. The partition coefficients for the majority of the elements increased as the C content of the liquid increased, an effect analogous to that of S for many of the elements. In contrast, three of the elements, Cr, Re, and W, were found to have anthracophile (C-loving) preferences, partitioning more strongly into the metallic liquid as the C content increased, resulting in decreases to their partition coefficients. For half of the elements examined, the prediction that partitioning in the Fe-Ni-S and Fe-Ni-C systems could be parameterized using a single set of variables was not supported. The effects of S and C on elemental partitioning behavior can be quite different; consequently, the presence of different non-metals can result in different fractionation patterns, and that uniqueness offers the opportunity to gain insight into the evolution of planetary bodies.     

Manganese partitioning during hydrous melting of peridotite

Manganese contents and the iron/manganese ratio of igneous rocks have been used as a method of probing the heterogeneity in the Earth’s mantle during melting of peridotite and pyroxenite lithologies. Most previous work has assumed that changes in these parameters require differences in either source lithology or composition based on experiments indicating that manganese is slightly incompatible during melting and that the iron/manganese ratio is fixed by the presence of olivine. However, the presence of volatiles in the mantle drives melting at lower temperatures and with different compositions than in volatile-free systems, and thus the partitioning of Fe and Mn may in fact vary. We have produced silicate liquids in equilibrium with a peridotite assemblage under hydrous conditions at 3 GPa that show that Mn can also be unexpectedly compatible in garnet at 1375 °C and that Mn partitioning between solids and liquids can be strongly affected by temperature and liquid composition. The compatibility of Mn in garnet provides a mechanism for large variations of Mn contents and the Fe/Mn ratio in silicate melts that solely involves melting of mantle peridotite with only small compositional changes. Correlations between Mn variations and other indices indicative of melting in the presence of garnet may provide a means of more completely understanding the role of garnet at high pressures in peridotite melting.     

Quartz diorites derived by partial melting of eclogite or amphibolite at mantle depths

Rare earth elements, Rb, Sr, Ba and K have been determined in tonalite, trondhjemite, dacite, tholeiite and graywacke from the 2700 m.y. old Early Precambrian greenstone-granite terrane of northeastern Minnesota-northwestern Ontario, and also in trondhjemite from the 3550 m.y. old Morton Gneiss, southwestern Minnesota; and the Mesozoic Craggy Peak Pluton, Klamath Mountains, California.The Early Precambrian tholeiites have trace element compositions similar to modern oceanic tholeiites, while the quartz dioritic rocks, regardless of age, have total rare earth contents lower than that of tholeiitic basalts, with near chondritic heavy rare earth contents. Rb, Sr, Ba and K contents of the quartz diorites are about five times that of oceanic tholeiites, with similar alkali and alkaline earth ratios. The Early Precambrian graywacke has a rare earth content intermediate between greenstone and quartz diorite, reflecting its provenance.It is proposed that the analyzed quartz dioritic rocks, whether plutonic tonalite, dacite porphyry, gneissic or plutonic trondhjemite, or trondhjemite dikes had similar modes of origin, and were derived by partial melting of amphibolite or eclogite of basaltic or gabbroic composition at depths greater than thirty kilometers, leaving a residue consisting predominantly of garnet and clinopyroxene.     

The stability and major element partitioning of ilmenite and armalcolite during lunar cumulate mantle overturn

Ilmenite has played an important role in the petrogenesis of lunar high-Ti picritic magmas, and armalcolite is another high-Ti oxide that was first discovered on the moon. In this study, we examined the thermodynamic stability of ilmenite and armalcolite in the context of lunar cumulate mantle overturn. Two starting compositions were explored, an ilmenite-bearing dunite (olivine + ilmenite) and an ilmenite-bearing harzburgite (olivine + orthopyroxene + ilmenite). Experiments were conducted using a 19.05 mm piston-cylinder apparatus at temperatures of 1235-1475 °C and pressures of 1-2 GPa. In runs with the ilmenite-bearing dunite mixture, ilmenite is stable in the subsolidus assemblage at least up to 1450 °C and 2 GPa. In runs with the ilmenite-bearing harzburgite starting mixture, ilmenite is stable at pressures greater than 1.4 GPa, and armalcolite is stable at lower pressures. Solidi for both starting compositions were determined, and the phase boundary between ilmenite- and armalcolite-bearing harzburgite was shown to have little dependence on temperature. During lunar cumulate overturn, sinking ilmenite formed near the end of lunar magma ocean solidification transforms into armalcolite when in contact with harzburgite cumulates at depths of less than 280 km in the lunar mantle. Inefficient overturn could leave isolated, inhomogeneously distributed pockets of armalcolite-bearing harzburgite in the upper lunar mantle, underlain by an ilmenite-bearing lower lunar mantle. These high-Ti oxide-bearing harzburgitic pockets can serve as potential sources for the generation of high-Ti magmas through partial melting or through assimilation of high-Ti minerals during transport of low-Ti picritic magmas in the lunar mantle.FeO-MgO exchange between olivine and either ilmenite or armalcolite was also examined in this study. We found the FeO-MgO distribution coefficient to be effectively independent of temperature for the pressures, temperatures, and compositions explored, with an average value of 0.179 ± 0.008 for olivine/ilmenite and 0.319 ± 0.021 for olivine/armalcolite. Given the bulk composition of an overturned lunar cumulate mantle, our measured FeO-MgO distribution coefficients can be used to estimate the Mg# of coexisting minerals in armalcolite- or ilmenite-bearing harzburgite and dunite in the overturned lunar mantle. Finally, the transformation from ilmenite-bearing harzburgite to armalcolite-bearing harzburgite results in a density increase of up to 2%. Large armalcolite-bearing cumulate bodies in the upper lunar mantle may be detectable in future lunar geophysical experiments.     

Age and geochemical constraints for partial melting of granulites in Estonia

Summary The rocks of the crystalline basement of the East European Craton in southern Estonia show effects of partial melting under granulite facies conditions. Zircons extracted from partial melting products (tonalite from the Tapa Zone – 1824 ± 26, tonalite from the South Estonian Zone – 1788 ± 16 Ma and charnockite from the Tapa Zone – 1761 ± 11 Ma) yield U–Pb crystallisation ages that span over approximately 80 Ma, suggesting a prolonged high-grade metamorphism or several separate events. U–Pb zircon age of one sample of charnockite is concordant with the Nd model age of partial melting of its host mafic granulite facies gneiss (intercept at 1.76 Ga). Linear geochemical trends and similar initial Nd isotopic compositions of mafic granulites and charnockites suggest their possible genetic relationship. From our new and previously published data it follows that the peak granulite metamorphic conditions and formation of tonalites and charnockites (850 °C and 6 kbar) in the Estonian basement occurred at 1788–1778 Ma. Then, the rocks cooled down, passing through the garnet closure temperature of approximately 650–700 °C at 1728 ± 24 Ma. The age of metamorphism of the Estonian granulites is lower than the metamorphic ages known from southern Finland, but it is similar to the age of metamorphism reported from the Belarus-Baltic Granulite Belt in Latvia.     

Experimental and major element constraints on the evolution of lavas from Lihir Island,Papua New Guinea

The major element chemistry of SiO₂-undersaturated arc lavas from Lihir Island, Papua New Guinea, and 1 atmosphere experiments on an alkali basalt from this island show complex polybaric fractionation affected this suite of lavas. Low Ni and MgO are typical of these arc lavas and result from olivine fractionation, probably at high pressure. Fractionation at low pressure (<5 kb) produces two evolutionary trends. Separation of clinopyroxene, plagioclase and minor olivine from the primitive lavas results in increasing normative nepheline contents and major element trends similar to those of the experiments. In contrast, addition of magnetite and amphibole to the fractionating assemblage in the evolved lavas results in decreasing normative nepheline and major element trends which are markedly different from those of the experiments. The composition of experimental glasses and 1 atmosphere liquid lines of descent, derived from anhydrous melting experiments run at the fayalite-magnetite-quartz (FMQ) buffer and at higher oxygen fugacities, are displaced from the lavas on oxide-oxide plots. HighfO₂ produces high Fe³⁺/Fe²⁺ and the early crystallization of abundant magnetite, and high H₂O contents are responsible for crystallization of amphibole. Crystal fractionation of these phases and the high Fe³⁺/Fe²⁺ are responsible for the displacement of the lavas and experimental glasses in mineral projection schemes from the 1 atmosphere olivine-clinopyroxene-plagioclase saturation boundary of Sack et al. (1987).     

西藏松多榴辉岩的矿物原位微量元素地球化学研究

对松多榴辉岩中单矿物进行的LA-ICP-MS原位微区微量元素分析研究结果表明,石榴石主要富集中、重稀土元素和Y,同时具有高丰度的Sc、V、Cr和Co等元素;绿辉石中的微量元素以中稀土元素、Sr、Sc、V、Cr、Co、Ni和Ti为主,含有一定量的Zr、Hf等。石榴石、绿辉石、角闪石和绿帘石中均显示轻稀土元素亏损的特点,表明在退变质过程中没有发生明显的富轻稀土元素的外来流体交代作用,因而其微量元素矿物地球化学的某些特点不同于苏鲁地区的榴辉岩。石榴石变斑晶中某些元素(如Ti、Zr)的分带性暗示了榴辉岩在紧随峰期变质之后的折返过程中发生了降压增温过程。榴辉岩主要变质矿物中微量元素的分配显然受到矿物主量元素的分配所控制,如MgO在石榴石和绿辉石之间的分配对Ni、Co、Ti分配的控制以及CaO的分配对Sr、Y、REE分配的控制等。退变质过程中矿物的形成或分解以及物理化学条件的改变都可以引起矿物间微量元素的重新分配。由绿辉石退变质而形成的角闪石,较之原先的绿辉石,其微量元素配分曲线总体特征会发生变化,但元素总体丰度相近,某些元素特点相似,又反映了绿辉石和角闪石之间的成生联系。金红石是Ti、Nb、Ta、Zr、Hf的主要赋存矿物,而与之共生的绿帘石所表现出来的高场强元素的亏损特征表明了金红石的存在所带来的影响。     

Trace element partitioning between type B CAI melts and melilite and spinel: Implications for trace element distribution during CAI formation

Calcium- and aluminum-rich inclusions (CAIs), occurring in chondritic meteorites and considered the oldest materials in the solar system, can provide critical information about the environment and time scale of creation of planetary materials. However, interpretation of the trace element and isotope compositions of CAIs, particularly the light elements Li, Be, and B, is hampered by the lack of constraint on melilite-melt and spinel-melt partition coefficients. We determined melilite-melt and spinel-melt partition coefficients for 21 elements by performing controlled cooling rate (2 °C/h) experiments at 1 atmosphere pressure in sealed platinum capsules using a synthetic type B CAI melt. Trace element concentrations were measured by secondary ion mass spectrometry (SIMS) and/or laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS). Melilites vary only slightly in composition, ranging from Åk_31-43. Results for the partitioning of trace elements between melilite and melt in three experiments and between spinel and melt in two experiments show that partition coefficients are independent of trace element concentration, are in good agreement for different analytical techniques (SIMS and LA-ICP-MS), and are in agreement with previous measurements in the literature. Partition coefficients between intermediate composition melilites and CAI melt are the following: Li, 0.5; Be, 1.0; B, 0.22; Rb, 0.012; Sr, 0.68; Zr, 0.004; Nb, 0.003; Cs, 0.002; Ba, 0.018; La, 0.056; Nd, 0.065; Sm, 0.073; Eu, 0.67; Er, 0.037; Yb, 0.018; Hf, 0.001; Ta, 0.003; Pb, 0.15; U, 0.001; Th, 0.002. Site size energetics analysis is used to assess isovalent partitioning into the different cation sites. The Young’s modulus deduced from +2 cations partitioning into the melilite X site agrees well with the bulk modulus of melilite based on X-ray diffraction methods. The changes in light element partitioning as melilite composition varies are predicted and used in several models of fractional crystallization to evaluate if the observed Li, Be, and B systematics in Allende CAI 3529-41 are consistent with crystallization from a melt. Models of crystallization agree reasonably well with observed light element variations in areas previously interpreted to be unperturbed by secondary processes [Chaussidon, M., Robert, F., McKeegan, K.D., 2006. Li and B isotopic variations in an Allende CAI: Evidence for the in situ decay of short-lived ¹⁰Be and for the possible presence of the short-lived nuclide ⁷Be in the early solar system. Geochim. Cosmochim. Acta70, 224-245], indicating that the trends of light elements could reflect fractional crystallization of a melt. In contrast, areas interpreted to have been affected by alteration processes are not consistent with crystallization models.