Earth and Planetary Science Letters

Calcium isotopes track volatile components in mantle sources of alkaline rocks

and associated carbonatites
--Manuscript Draft--

Manuscript Number: EPSL-D-23-00599R1

Article Type: Letters

Keywords: Ca isotopes, volatiles, alkaline rocks, carbonate, K-richterite, carbonatites

Abstract: The volatile components CO2 and H2O induce mantle melting and thus exert major
controls on mantle heterogeneity. Primitive intraplate alkaline magmatic rocks are the
closest analogues for incipient mantle melts and provide the most direct method to
assess such mantle heterogeneity. Given the considerable Ca isotope differences
among carbonate, clinopyroxene, garnet, and orthopyroxene in the mantle (up to 1‰
for δ 44/40Ca), δ44/40Ca of alkaline rocks is a promising tracer of lithological
heterogeneity. We present stable Ca isotope data for ca. 1.4 Ga lamproites, 590-555
Ma ultramafic lamprophyres and carbonatites, and 142 Ma nephelinites from Aillik Bay
in Labrador, eastern Canada. These primitive alkaline rock suites are the products of
three stages of magmatism that accompanied lithospheric thinning and rifting of the
North Atlantic craton. The three discrete magmatic events formed by melting of
different lithologies in a metasomatized lithospheric mantle column at various depths:
(1) MARID-like components (mica-amphibole-rutile-ilmenite-diopside) in the source of
the lamproites; (2) phlogopite-carbonate veins were an additional source component
for ultramafic lamprophyres during the second event; and (3) wehrlites at shallower
depths were an important source component for nephelinites during the final event.
The Mesoproterozoic lamproites show lower δ44/40Ca values (0.58 to 0.66‰) than
MORBs (0.84 ± 0.03‰, 2se). This cannot be explained by fractional crystallization or
melting of the clinopyroxene-dominated source but can be attributed to a source
enriched in the alkali amphibole K-richterite, which has characteristically low δ44/40Ca.
The δ44/40Ca values of the ultramafic lamprophyre suite during the second rifting
stage are remarkably uniform, with overlapping ranges for primary carbonated silicate
melts (aillikite: 0.67 to 0.75‰), conjugate carbonatitic liquids (0.71 to 0.82‰) and
silicate-dominated damtjernite liquid (primary damtjernite: 0.68 to 0.72‰). This
suggests negligible Ca isotope fractionation during liquid immiscibility of carbonate-
bearing magmas. Combined with previously reported δ44/40Ca values for carbonatites
and kimberlites, our data suggest that carbonated silicate melts in Earth’s mantle have
δ44/40Ca compositions resolvably lower than those for MORB (0.74 ± 0.02‰ versus
0.84 ± 0.03‰, 2se). The δ44/40Ca values of the Cretaceous nephelinites (0.72 to
0.78‰) are homogenous and similar to those of the 590-555 Ma ultramafic
lamprophyres, suggesting that the wehrlitic source component for the nephelinites
formed by mantle metasomatism during interaction with rising aillikite magmas during
the second rifting stage. Our results highlight that both K-richterite and carbonate
components in mantle sources can result in the systematically low δ44/40Ca values of
alkaline magmas, which may explain previously reported low δ44/40Ca values of
continental alkaline rocks and some OIBs. Our study indicates that Ca isotopes are a
robust tracer of lithological variation caused by volatiles in the Earth’s upper mantle.

Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation
Revision Notes

Response to referees´ comments, manuscript EPSL-D-23-00599

Chunfei Chen, Stephen F. Foley, Sebastian Tappe, Huange Ren, Lanping Feng, Yongsheng Liu

" Calcium isotopes trace volatile-rich components in the mantle sources of alkaline rocks "

Referees’ comments in black

Replies in blue (All line numbers mentioned below are referred to the file “Manuscript with
Track Changes in Red”, NOT the file “Manuscript”.)

Reviewer #1: General Takeaways:

Chen and colleagues make a great attempt to address a significant gap in our knowledge regarding
the Ca isotope systematics of igneous rocks: how do the Ca isotope compositions of silica-
undersaturated continental alkaline rocks compare to other igneous rocks? They report Ca isotope
compositions for a respectable suite of lamproites, aililikites, damtjernites, nephelinites, and
associated carbonatites from the North Atlantic Craton, ranging in age from 1.4 Ga to 142 Ma.
Relative to MORBs, they report isotopically light Ca isotope compositions, ranging from 0.58 to
0.90 permille. Considering theoretically-predicted fractionation factors and assumptions about the
magnitude of Ca isotope fractionation between cpx and melt, they predict the expected Ca isotope
composition of melts from average carbonated peridotites (~0.74 permille). They also infer the Ca
isotope composition of volatile-rich cratonic sources, which they expect to be significantly lighter
than the BSE value of 0.94 permille, largely due to the presence of K-richterite in these mantle
assemblages. On the basis of their modeling efforts, they suggest there to be negligible fractionation
during carbonatite-silicate melt liquid immiscibility in the mantle. They finally suggest that K-
richterite may play an important role in producing the isotopically light Ca isotope compositions
observed in other cratonic igneous rocks, and perhaps even OIBs.

The manuscript is well-written (apart from several small typos) and reasonably well-organized;
however, there are several sections that are unclear and somewhat confusing (see more in "Line-
specific comments"). The data and interpretations are very interesting; however, I have several
misgivings with the quality of the data and the way that it is interpreted. Moreover, there are several
claims about the data that are completely invalid or contradictory to the associated figures, which I
find very concerning.

First off, many of their data points fall off the expected mass-dependent fractionation line (in a plot
of δ43/42Ca versus δ44/42Ca - Fig. S3), which makes me question the quality of their data. Secondly,
there are several modeling assumptions that I do not agree with, and the documentation and
explanation of some of their modeling efforts can be quite poor, to the point where I found it unclear
what actually went into one of their calculations. Moreover, one claim (that there is no apparent
trend between δ44Ca and CaO) is completely invalid, which undermines their conclusion that the
Ca isotope signatures of their lamproite and nephelinite samples reflect primary magmatic processes.
This, combined with my misgivings about their data quality, makes it quite difficult to accept their
interpretations and the associated implications. Lastly, I found the extension of their implications to
explain the Ca isotope systematics of OIBs to be unnecessary and invalid, especially since their only
support was found to be misleading and incorrect.
Reply: Thank you very much for your constructive comments, which greatly improved the
manuscript. Firstly, to double-check data quality, we have remeasured Ca isotopes of four samples
that fall slightly off the fractionation line in the original manuscript. The remeasured δ44/42Ca values
are consistent with previous values, but the remeasured δ43/42Ca values are higher than the previous
values. Now all samples fall on the fractionation line (Fig. S3 in the revised manuscript). Secondly,
we have investigated the fractional crystallization history of these alkaline rocks through
petrography, major elements, and modelling by MELTS. Results show the Mesoproterozoic
lamproites mainly experienced fractional crystallization of olivine and the Mesozoic nephelinites
experienced fractional crystallization of both olivine and clinopyroxene. Therefore, the effect of
fractional crystallization on Ca isotopes is discussed in the revised manuscript. Lastly, we have
deleted the discussion on the Ca isotope systematics of OIBs.

Overall, I think that this study has a lot of promise, but the potential issues with the quality of the
data must be addressed. Additionally, I am not convinced that their samples, especially the
lamproites and nephelinites, reflect primary Ca isotope signatures. And lastly, and potentially most
concerning, several of their misleading (or false) claims need to be addressed and corrected before
I can recommend this manuscript for publication in EPSL. At this time, I recommend major
revisions, because I do think that these data could eventually be very important for the field.

Line-specific comments:
L 25: The author mentions the acronym "MARID" in the abstract — before reading the full text, I
had never come across this acronym before, so I would encourage the authors to spell out the
acronym in the abstract.
Reply: Done (Lines 24-25). Note: All line numbers are from the file “Manuscript with Track
Changes in Red”, NOT the file “Manuscript”.

L 50: Very nice introduction! It sets the stage well and motivates the problem. I think that the authors
could make it even better by briefly motivating why we care about compositional/lithological
heterogeneity in the mantle (e.g., relating it to the chemical evolution of the Earth as a whole) before
getting too deep into the weeds.
Reply: We have added an introduction to the importance of the compositional/lithological
heterogeneity in the mantle (Lines 50-51).

L 164-166 and Fig. S3: Upon examination of Fig. S3, it is quite clear that many samples actually
fall off of the exponential fractionation line. This is especially clear for damtjernite and nephelinite
samples. This is concerning, as this could point to issues with the quality of the data, potentially
resulting from analytical artifacts or interferences. I think it is critical that the authors address why
many of their data fall off the expected fractionation line.
Reply: The natural abundance of 43Ca is much lower than that of 44Ca (0.135 wt.% versus 2.086
wt.%), resulting in a much lower signal intensity for 43Ca than for 44Ca during isotope measurement
by MC-ICP-MS. Therefore, the uncertainty of δ43/42Ca value is much larger than that of δ44/42Ca.
The four samples falling slightly off the exponential fractionation line in the original manuscript
were attributed to this large uncertainty of δ43/42Ca. Our data quality is robust because these δ44/42Ca
values (and δ44/40Ca) are accurate regardless of the large uncertainty of δ43/42Ca. 43Ca is not used in
the Figures (2, 5-8) nor in the interpretations.

To double-check data quality, we have remeasured Ca isotopes of the four samples that fall slightly
off the fractionation line in the original manuscript. The remeasured δ44/42Ca values are consistent
with previous values, but the remeasured δ43/42Ca values are higher than the previous values. Now
all samples fall on the fractionation line (Fig. S3 in the revised manuscript). This confirms that our
data quality is robust regardless of the large uncertainty of δ43/42Ca. We have updated Fig. S3 and
Table 1.

Figure caption: Comparison of the remeasured Ca isotopes in the revised manuscript with Ca
isotopes in the original manuscript.

Fig. S3 δ44/42Ca versus δ43/42Ca in all measured samples and reference materials.

L 173: The reported 2 SE values for delta44/40Ca should only be reported to two significant digits,
to remain consistent with the rest of the table.
Reply: Done (See Table 1).

L 174-175: The fact that some of their samples do indeed fall off the expected fractionation line
could point to radiogenic ingrowth of 40Ca from 40K decay. Fu et al. (2022) recently showed the
importance of correcting for 40Ca ingrowth in ~300 Ma rocks, so I could imagine that in these rocks,
many of which are much older, radiogenic decay could have a measurable effect. I think it is critical
that the authors address this point in significantly more detail (e.g., calculate what the expected
deviation (Epsilon 40Ca) would be for their rocks, given their age and K/Ca ratios). Fu et al. (2022)
is a great reference for this.
Reply: Due to the large interference of 40Ar+ on 40Ca from the argon gas in MC-ICP-MS, the
44Ca/40Ca ratio cannot be determined in this study. All δ44/40Ca values were obtained by conversion

of measured δ44/42Ca applying the mass-dependent fractionation law. Thus, quoted δ44/40Ca
variations reflect mass-dependent fractionation without the effect of 40Ca ingrowth from the
radioactive decay of 40K.
We can calculate δ44/40Ca values with the addition of 40Ca ingrowth via the decay of 40K to 40Ca
(δ44/40CaRadiogenic in Table 1; calculation details see Fu et al. (2022)): δ44/40CaRadiogenic values of our
samples show no difference with δ44/40Ca values of the Late Neoproterozoic ultramafic
lamprophyres and the Cretaceous nephelinites, but are 0.10-0.17 lower than δ44/40Ca values for the
Mesoproterozoic olivine lamproites. Nevertheless, we only use δ44/40Ca values of our samples in
this study to discuss the mass-dependent fractionation without the effect of the radiogenic excess of
40Ca. We have clarified this point in the revised manuscript (Lines 175-177 and Lines 180-183).

Fig. 2: I believe the large variability and isotopically heavy values shown in the OIB section of this
figure are misleading. The most recent OIB measurements (e.g., Eriksen & Jacobsen, 2022) are
generally isotopically lighter than MORBs (~0.85 permille) and all significantly lighter than BSE
(0.94 permille). I would encourage the authors to carefully select the data they report here, instead
of just reporting everything that exists in the literature.
Reply: The data we plot is compiled OIB data from publications since 2015 (See Fig. 2).

L 192: I do not see a plot of delta44Ca versus H2O, so I would suggest that the authors do not
mention H2O here.
Reply: We have deleted it.

L 205-209: ~8 wt.% MgO is not particularly high for primitive magmas. It seems likely that these
rocks have indeed experienced fractional crystallization. I would encourage the authors to more
quantitatively show why we can believe they are more or less primary before plotting them in major
element diagrams to determine their source lithology (Fig. 3a). I am not necessarily doubting their
interpretation, I just think it makes it easier on the reader if the author does this work for them.
Reply: We agree that the fractional crystallization history of these alkaline rocks needs to be assessed
before decoding their mantle source lithologies using the major element and Ca isotope
geochemistry. Previous papers by co-authors of this manuscript have investigated these rocks in
detail (Tappe et al., 2006, 2007). The phenocrysts in the Mesoproterozoic lamproites are composed
of olivine with a small amount of clinopyroxene, phlogopite, and apatite, suggesting that they
mainly experienced the fractional crystallization of olivine during the early stages (Tappe et al.,
2007). The uniform CaO/Al2O3 ratios regardless of various MgO contents in the lamproites also
indicate fractional crystallization of olivine without clinopyroxene (Fig. 3a). This is supported by
modeling of the fractional crystallization of an experimental lamproite melt from Foley et al. (2022)
using MELTS (Gualda and Ghiorso, 2015) which show that lamproite melts mainly experience
olivine fractionation of until their MgO content < 3.5 wt.% (Fig.3a). The high MgO contents of the
olivine lamproites (mostly >8.1 wt.%) suggest that they represent primitive magmas (Tappe et al.,
2007). The Mesozoic nephelinites contain olivine and clinopyroxene phenocrysts. Positive
correlation between CaO/Al2O3 ratio and MgO content in the nephelinites indicates that they
underwent fractional crystallization of both olivine and clinopyroxene, consistent with fractional
crystallization modeling of an experimental nephelinite melt from Dasgupta et al. (2007) using
MELTS (Gualda and Ghiorso, 2015) (Fig. 3a). We have added a new paragraph (Lines 209-224)
and figure (new Fig. 3) to discuss the fractional crystallization in the revised manuscript.

L 231: I suggest that the authors indicate in which section of the Supplementary Material the reader
can find these details.
Reply: We have numbered the contents in the supplementary information. It is clear for the reader
to find the details.

L 231: Regarding their thermobarometric calculation, in the Supplementary Material, the authors
say, "Thus, we compiled olivine compositions in mantle xenoliths of calcite-bearing peridotite and
amphibole wehrlites from other craton regions according to the following criteria: (1) dolomite
peridotite xenoliths are excluded; (2) only amphibole wehrlites were corrected without amphibole-
free wehrlites and phlogopite wehrlites."
I think it is good practice to report any data compilation used in a calculation as a supplementary
table, and I encourage the authors to do so. I know some of this information is presented in Fig. S6,
but I think an accompanying table would definitely help. I also think their assumptions regarding
the Mg# of the source should be more clearly discussed in the main text, rather than buried in the
supplements.
Reply: We have added the supplementary table about the compiled Mg# of olivines.

L 246-247: How much is a "small amount" of fractional crystallization?

Reply: It is difficult to quantify the degree of fractional crystallization because we cannot closely
constrain the starting point, i.e.the chemical compositions of their primary magmas. However, we
have investigated the fractional crystallization history of the lamproites and nephelinites combined
with the modeling calculation using MELTS (Lines 209-224 and Fig. 3). This knowledge is enough
for us to constrain Ca isotope fractionation caused by the fractional crystallization (Lines 268-272).

L 247-249: The fractionation factor for cpx is close to one, but significantly different from one
(Zhang et al 2018, Antonelli et al 2019, Zhu et al. 2021a, Antonelli et al 2021). I think it is important
that the authors rethink this assumption during modeling, as cpx should be slightly heavier than the
melt. It is heavy enough that with significant cpx fractionation, there should be an observed effect
in the Ca isotope composition of the melt. This is an important point to quantitatively consider,
especially since cpx fractionation should push the melt (slowly) toward lighter Ca isotope
compositions.
Reply: Thanks for this comment. Fractional crystallization has been constrained and thus its effect
on Ca isotopes of the lamproites and nephelinites have been investigated in the revised manuscript
(Lines 268-272 and 403-405). The lamproites experienced fractional crystallization of olivine
without clinopyroxene. Due to the extremely low CaO content of olivine phenocrysts (CaO=0.007-
0.012 wt.%) (Tappe et al., 2007), fractional crystallization of olivine would not change CaO/Al2O3
ratios or δ44/40Ca values of magmas (Fig. 3). The nephelinites show a positive correlation between
MgO content and CaO/Al2O3 ratios, suggesting a history of fractional crystallization that includes
both olivine and clinopyroxene. However, δ44/40Ca values are uniform regardless of various
CaO/Al2O3 ratios, indicating that fractional crystallization of clinopyroxene has negligible effect on
Ca isotopes. This also indicates that the fractionation factor of Ca isotopes between silicate melt and
clinopyroxene is indeed close to one.

L 274-276: I am a bit confused here. I am not exactly sure what was done during this modeling
exercise (the one depicted in Fig. 4c). Is it a mixing calculation to approximate metasomatism?
There are a lot of details here that are missing, making it very difficult to understand, and therefore
be convinced by, their modeling efforts. I suggest that the authors dramatically overhaul this
paragraph so that the reader can understand what went into this modeling calculation.
Reply: We have clarified the modelling of the metasomatic process in detail (Lines 296-301). K-
richterite-bearing MARID was formed by reaction between the refractory cratonic lithosphere
mantle (CaO content of refractory harzburgite < 0.5 wt.%, Pearson et al. (2003)) and hydrous silicate
melt. Considering the much lower CaO content of refractory harzburgite than that of hydrous silicate
melt (the average CaO contents of worldwide continental basalts are 7.3-11.1 wt.%, Farmer (2014)),
the Ca isotopes of K-richterite-bearing MARID are not influenced by the refractory cratonic
lithosphere mantle, but are controlled by the metasomatic melts. We modelled the Ca isotope
compositions of K-richterite-bearing MARID material formed by mantle metasomatism at the base
of the cratonic lithosphere mantle for different geothermal gradients (Fig. 5c) assuming that the
infiltrating hydrous silicate melts have a MORB-like source δ44/40Ca value of 0.84‰.

L 275: The value of 0.82 permille used for carbonated peridotite must be motivated here. I know it
is discussed later on in the manuscript, but on my first read, I was lost as to where this number came
from. Moreover, I think the compilation of carbonated peridotite values quoted in the text (L 314)
should be reported as a supplementary table as well.
Reply: We assumed that the metasomatic hydrous silicate melts have MORB-like source δ44/40Ca
value of 0.84‰ (Lines 304-305). We have also added a supplementary table for compiled Ca
isotopes of carbonated peridotite (Table S3 in the revised manuscript)

Fig. 4c: I believe the x-axis label should be "Temperature" here instead of "Melting degree."
Reply: We have revised it (Fig. 5c).

L 293-296: This is an interesting conclusion, very cool!

Reply: Thanks.

L 304-305: I think this is a reasonable conclusion, but is the direction of fractionation consistent
with what we would expect from the crystallizing assemblage? This could be good to note, at least
qualitatively.
Reply: Following liquid immiscibility, these evolved damtjernites experienced fractional
crystallization of olivine, clinopyroxene and apatite, as indicated by their olivine phenocrysts and
clinopyroxene-apatite needles (Tappe et al., 2006). Clinopyroxene and apatite have heavier Ca
isotopes than dolomite (may also carbonate-bearing damtjernitic magmas), thus fractional
crystallization of clinopyroxene and apatite may result in lighter Ca isotopes in damtjernitic magmas,
and can explain slightly low δ44/40Ca values in the four evolved damtjernites. This is clarified in the
revised text at lines 333-339.

L 307-309: I agree that this suggests there is negligible fractionation during immiscibility reactions,
but the authors note in the previous few sentences how fractional crystallization could have a minor
effect. Therefore, liquid immiscibility may not have much of an effect, but fractional crystallization
can. I would suggest that the authors reword this part to make this clear.
Reply: The answer to this is the same as to the previous questions. Fig. 6 also makes this point
clearer.

L 325-326: This sentence is a bit confusing to me. It seems like they are only going to qualitatively
estimate a fractionation factor between solid and melt for carbonated peridotite, but then later they
go on to actually calculate this number. This paragraph seems a bit redundant to me, especially since
they report the calculated solid-melt fractionation in the next paragraph.
Reply: We have deleted this sentence and have merged two paragraphs to one.

L 355-357: This is an important conclusion, especially considering the controversy around the Ca
isotope composition of carbonatites. Great work!
Reply: Thanks.

L 364-365: I don't think this sentence is necessary, especially since Sun et al. (2021) called into
question the quality of the data in Amsellem et al. (2020), based off potential analytical interferences.
I think the carbonatite data in this manuscript further support Sun et al. (2021)'s conclusions,
showing that average carbonatites are probably not as light as ~0.3 permille.
Reply: We have deleted it.

L 370-372: Once again, I would encourage the authors to revisit this assumption. The Ca isotope
fractionation between cpx and melt may be small, but it can be significant.
Reply: Positive correlation between CaO/Al2O3 ratio and MgO contents in the Mesozoic
nephelinites indicates that they underwent fractional crystallization of clinopyroxene besides olivine,
agreeing well with fractional crystallization models for an experimental nephelinite melt (Dasgupta
et al., 2007) using MELTS (Gualda and Ghiorso, 2015) (Fig. 3a). However, δ44/40Ca values are
uniform despite variation in CaO/Al2O3 ratios, indicating that negligible Ca isotope fractionation
during fractional crystallization of clinopyroxene occurred. This also implies that the fractionation
factor for Ca isotopes between silicate melt and clinopyroxene is close to one. We have clarified
this point in Lines 400-405.

L 372-374: I am very concerned with how misleading this sentence is. The authors motivate their
interpretation that the delta44Ca values of the nephelinites/lamproites represent primitive Ca isotope
compositions by reporting that there is no correlation between delta44Ca and CaO. Upon closer
examination of Fig. S4b, there is in fact an exceptional correlation between delta44Ca and CaO for
lamproites and nephelinites, with an R2 value of 0.93. This completely undermines the rest of their
interpretation in section 5.4, and this also makes me question their other interpretations and
assumptions throughout the rest of this manuscript.
Because of this, I think that cpx fractionation may be having a very real effect on the Ca isotope
compositions of these rocks. I think that it's important that the authors also plot delta44Ca versus
other chemical indices for cpx fractionation, like Sc, as a test for this. This is very concerning, and
it is critical that the authors address this point before considering these delta44Ca values as primary
signatures.
Reply: CaO/Al2O3 is a sensitive indicator of fractional crystallization. We have discussed the
fractional crystallization history in detail in the revised manuscript. The nephelinites underwent
fractional crystallization of clinopyroxene as well as olivine, but the lamproites did not. The uniform
δ44/40Ca values despite various CaO/Al2O3 ratios in the nephelinite indicate negligible Ca isotope
fractionation during fractional crystallization of clinopyroxene (See Lines 403-405 in the revised
manuscript).

L 400-406: I am now questioning how robust these conclusions are, after noting that these Ca isotope
signatures may not reflect primary signatures, especially for lamproites and nephelinites.
Reply: We have addressed this issue. See our answers to previous questions.

L 423-428: A more appropriate benchmark for what? This is unclear to me. Are the authors talking
about the melts or the sources?
Reply: a benchmark for primary carbonate-rich mantle melts. We have clarified this at Line 458 in
the revised manuscript.

L 429-434: This is a potentially important conclusion! As stated above, however, I am still not
convinced that a significant amount of these data don't reflect secondary processes like fractional
crystallization, which could push initial d44Ca values toward lighter compositions.
Reply: We have addressed this issue. See our answers to previous questions.

L 441-445: I think it's good to try to tie the implications of a given study into the bigger picture, but
I have a number of issues with how it's done in this paragraph. First off, while OIBs are often silica
undersaturated and alkaline in nature, directly comparing them to continental alkaline rocks is
problematic, as OIB petrogenesis occurs beneath much younger and thinner oceanic lithosphere
(<250 million years old and <90 km thick), compared to the much thicker cratonic lithosphere that
likely sourced the highly silica undersaturated rocks considered in this study. Secondly, once again
in this manuscript, there are references to trends that are wholly incorrect. The OIB "trends" in plots
of d44Ca versus SiO2 (Fig. 7a) and CaO/Al2O3 (Fig. 7b) are not statistically significant trends at
all, with R2 values of 0.14 and 0.15, respectively. I suggest that the authors reconsider this paragraph,
as there seems to be little evidence to support their claims that OIB Ca isotope systematics can be
explained by the same mechanisms that they propose to explain continental alkaline rocks.
Reply: We have deleted this paragraph, leaving this issue open for further work (Lines 467-471 in
the revised manuscript).

Fig. 8c: I find the yellow text difficult to discern against the blue background.
Reply: We have modified the color of these texts to make them clearer (see Fig.9 in the revised
manuscript).
Reviewer #2: The authors measured and reported the Ca isotope composition of several samples
from the North Atlantic craton at Aillik Bay. The data is useful to understand the Ca isotope
behaviors in the mantle. However, thoroughly modifications are need for this manuscript to be
accepted published in this journal. Firstly, the introduction part is weak to produce the main word
of this word. Secondly, there are many typos through the manuscript should be corrected, I just list
few in the following. Finally, the depth of the discussion part is insufficient. There are many general
descriptions and direct deduction which is lack of enough evidence. Thus, this manuscript is needed
to be modified thoroughly.
Reply: Thanks for your comments. We have strengthened the relationship between volatiles and Ca
isotopes in the introduction. Secondly, we have thoroughly checked for typos in the manuscript.
Lastly, we have added discussion on the petrogenesis of the nephelinites based on major and trace
elements and Ca isotopes in the revised manuscript.

1. The introduction is weak to introduce the correlation between Ca isotope fractionation and
volatile components (eg., H2O, CO2). There need more details to refill the introduction part.
Reply: We have strengthened the relationship between volatiles and Ca isotopes in the introduction
(Lines 73-74 and 79-81).

2. line 189-191, please give the evidence why these samples are very fresh with only little or no
signs of hydrothermal overprinting.
Reply: The detailed petrographic observation in previous studies (Tappe et al., 2006; Tappe et al.,
2007) indicates no secondary alteration in these samples. We have emphasized this point in the
revised manuscript (Lines 199-200). Additionally, we preclude the effect of alteration on the Ca
isotopes through their geochemical characteristics (Lines 200-207).

3. Line 202, 213, 220, 256, 268, the words "Mesoproterozoic" might have typos.
4. Line 205, 237, 247, 248, 254, the word "crystallization" also has typos
5. Line 215, the word "metasomatized" has a typo
6. Line 216, the word "Neoproterozoic" has a typo
Reply: We have double-checked these words: “Mesoproterozoic”, “crystallization”,
“metasomatized”, and “Neoproterozoic” are spelled correctly.

7. Line 268-273, from Fig 4a, at near 1300 °C, the Ca isotope fractionation between K-richterite
and CPX is about -0.05 ‰, however, the Mesoproterozoic lamproites show a variation δ44/40Ca
values from 0.58 to 0.66‰, while the MORB is 0.85±0.09‰. The difference between lamproites
and MORB is larger than 0.2 ‰, which is could not be caused by K-richterite in the MARID source
with the low δ44/40Ca signature. Have you measured the component of K-richterite in your samples?
Reply: ∆44/40Ca between K-richterite and Cpx are from -0.15 to -0.20‰ at 950-1100 °C and -0.11‰
at 1300 °C, not -0.05‰. Such ∆44/40Ca values may produce the about 0.2‰ difference of δ44/40Ca
between lamproites and MORB as we modelling in Fig. 5c.

8. The x axis table in Fig 4c should be "temperature".

Reply: We have revised it in Fig. 5c.
9. Line 291-296, the aillikites and mela-aillikites samples in Fig 5 are few (total 7 samples), which
is difficult to deduce that they are from the separation of a carbonate-rich component from more
primitive aillikite magma. More samples and evidence are needed to prove this hypothesis, thus, the
following deduction about the separation of carbonate from silicate components is also weak.
Reply: The evolution processes of the aillikites, mela-aillikites, carbonatites, damtjernites are
discussed in detail in previous publication (Tappe et al., 2006) where they have more samples. The
evolution trends in Fig. 6 are based on Tappe et al. (2006).

10. Line 329-331, Chen et al. (2020b) and Wang et al. (2017a) give the inter-mineral Ca isotope
fractionation using different methods, the former is deduction from natural samples and the latter is
first principles calculation. I suggest choose one method to reduce the system error as much as
possible.
Reply: Ca isotope fractionation factor between Grt and Cpx from Chen et al. (2020b) considered
the effect of composition, pressure, and temperature, however, Chen et al. (2020b) did not report
the Ca isotope fractionation factor between carbonate and Cpx. Wang et al. (2017a) reported the Ca
isotope fractionation factor between carbonate and Cpx using first principles, but without
considering the effect of pressure and complex compositions. Therefore, we choose appropriate Ca
isotope fractionation factor from these publications.

11. Line 332, I am confused that the Fig 3 has only two graphs, there is no Fig 3d.
Reply: The figure was referred to wrongly – it is Fig. 5d in the revised manuscript.

12. Line 383-387, the discussion about nephelinites samples is lack of foundations, try to explain it
more clearly.
Reply: We have strengthened the discussion on the nephelinites based on major and trace elements
and Ca isotopes in the revised manuscript (Lines 414-416 and 416-420). We have also emphasized
that the detailed petrogenesis of the nephelinites are given in Tappe et al. (2007).

13. The part 5.5 is just the combination of 5.1-5.4, I can not find more new discussions, which is
similar with part 6 Summary and Conclusions and is redundant.
Reply: Section 5.5 is the implication of our finding to Ca isotope compositions of metasomatized
mantle at the base of the cratonic lithosphere mantle and the origin of the origin of alkaline
magmatism. We think these are important for Ca isotope geochemistry. Furthermore, we have
condensed the conclusion to avoid repetition.

References:
Dasgupta, R., Hirschmann, M.M. and Smith, N.D. (2007) Partial Melting Experiments of Peridotite +
CO2 at 3 GPa and Genesis of Alkalic Ocean Island Basalts. Journal of Petrology 48, 2093-2124.
Farmer, G.L. (2014) 4.3 - Continental Basaltic Rocks, in: Holland, H.D., Turekian, K.K. (Eds.), Treatise
on Geochemistry (Second Edition). Elsevier, Oxford, pp. 75-110.
Foley, S.F., Ezad, I.S., van der Laan, S.R. and Pertermann, M. (2022) Melting of hydrous pyroxenites
with alkali amphiboles in the continental mantle: 1. Melting relations and major element compositions
of melts. Geoscience Frontiers 13, 101380.
Fu, H., Jacobsen, S.B., Larsen, B.T. and Eriksen, Z.T. (2022) Ca-isotopes as a robust tracer of magmatic
differentiation. Earth and Planetary Science Letters 594, 117743.
Gualda, G.A.R. and Ghiorso, M.S. (2015) MELTS_Excel: A Microsoft Excel-based MELTS interface
for research and teaching of magma properties and evolution. Geochemistry, Geophysics, Geosystems
16, 315-324.
Pearson, D.G., Canil, D. and Shirey, S.B. (2003) 2.05 - Mantle Samples Included in Volcanic Rocks:
Xenoliths and Diamonds, in: Holland, H.D., Turekian, K.K. (Eds.), Treatise on Geochemistry. Pergamon,
Oxford, pp. 171-275.
Tappe, S., Foley, S.F., Jenner, G.A., Heaman, L.M., Kjarsgaard, B.A., Romer, R.L., Stracke, A., Joyce,
N. and Hoefs, J. (2006) Genesis of Ultramafic Lamprophyres and Carbonatites at Aillik Bay, Labrador:
a Consequence of Incipient Lithospheric Thinning beneath the North Atlantic Craton. Journal of
Petrology 47, 1261-1315.
Tappe, S., Foley, S.F., Stracke, A., Romer, R.L., Kjarsgaard, B.A., Heaman, L.M. and Joyce, N. (2007)
Craton reactivation on the Labrador Sea margins: 40Ar/39Ar age and Sr–Nd–Hf–Pb isotope constraints
from alkaline and carbonatite intrusives. Earth and Planetary Science Letters 256, 433-454.
Manuscript with Track Changes in Red Click here to view linked References

1 Calcium isotopes track volatile components in the mantle

2 sources of alkaline rocks and associated carbonatites

3 Chunfei Chena,b*, Stephen F. Foleya, Sebastian Tappec, Huange Renb, Lanping Fengb,
4 Yongsheng Liub*
a
5 School of Natural Sciences, Macquarie University, North Ryde, New South Wales 2109, Australia
b
6 State Key Laboratory of Geological Processes and Mineral Resources, School of Earth Sciences,

7 China University of Geosciences, Wuhan 430074, China

c
8 Department of Geosciences, UiT - The Arctic University of Norway, N-9037 Tromsø, Norway

10 *Corresponding authors

11 Chunfei Chen (chfchen2016@hotmail.com) and Yongsheng Liu (yshliu@hotmail.com)

12

13 Abstract

14 The volatile components CO2 and H2O induce mantle melting and thus exert major controls on

15 mantle heterogeneity. Primitive intraplate alkaline magmatic rocks are the closest analogues for

16 incipient mantle melts and provide the most direct method to assess such mantle heterogeneity.

17 Given the considerable Ca isotope differences among carbonate, clinopyroxene, garnet, and

18 orthopyroxene in the mantle (up to 1‰ for δ44/40Ca), δ44/40Ca of alkaline rocks is a promising

19 tracer of lithological heterogeneity. We present stable Ca isotope data for ca. 1.4 Ga lamproites,

20 590-555 Ma ultramafic lamprophyres and carbonatites, and 142 Ma nephelinites from Aillik Bay

21 in Labrador, eastern Canada. These primitive alkaline rock suites are the products of three stages

22 of magmatism that accompanied lithospheric thinning and rifting of the North Atlantic craton. The

23 three discrete magmatic events formed by melting of different lithologies in a metasomatized

24 lithospheric mantle column at various depths: (1) MARID-like components (mica-amphibole-

25 rutile-ilmenite-diopside) in the source of the lamproites; (2) phlogopite-carbonate veins were an

26 additional source component for ultramafic lamprophyres during the second event; and (3)

27 wehrlites at shallower depths were an important source component for nephelinites during the

28 final event.

Page 1/28
29 The Mesoproterozoic lamproites show lower δ44/40Ca values (0.58 to 0.66‰) than MORBs

30 (0.84 ± 0.03‰, 2se). This cannot be explained by fractional crystallization or melting of the

31 clinopyroxene-dominated source but can be attributed to a source enriched in the alkali amphibole

32 K-richterite, which has characteristically low δ44/40Ca. The δ44/40Ca values of the ultramafic

33 lamprophyre suite during the second rifting stage are remarkably uniform, with overlapping ranges

34 for primary carbonated silicate melts (aillikite: 0.67 to 0.75‰), conjugate carbonatitic liquids

35 (0.71 to 0.82‰) and silicate-dominated damtjernite liquid (primary damtjernite: 0.68 to 0.72‰).

36 This suggests negligible Ca isotope fractionation during liquid immiscibility of carbonate-bearing

37 magmas. Combined with previously reported δ44/40Ca values for carbonatites and kimberlites, our

38 data suggest that carbonated silicate melts in Earth’s mantle have δ44/40Ca compositions resolvably

39 lower than those for MORB (0.74 ± 0.02‰ versus 0.84 ± 0.03‰, 2se). The δ44/40Ca values of the

40 Cretaceous nephelinites (0.72 to 0.78‰) are homogenous and similar to those of the 590-555 Ma

41 ultramafic lamprophyres, suggesting that the wehrlitic source component for the nephelinites

42 formed by mantle metasomatism during interaction with rising aillikite magmas during the second

43 rifting stage. Our results highlight that both K-richterite and carbonate components in mantle

44 sources can result in the systematically low δ44/40Ca values of alkaline magmas, which may

45 explain previously reported low δ44/40Ca values of continental alkaline rocks and some OIBs. Our

46 study indicates that Ca isotopes are a robust tracer of lithological variation caused by volatiles in

47 the Earth’s upper mantle.

48

49 1. Introduction

50 Lithological and geochemical heterogeneity in Earth’s mantle archives mantle evolution,

51 including mantle melting and recycling of crustal materials caused by plate tectonics. Besides

52 pressure and temperature, CO2 and H2O contents are the two most important controls on mantle

53 melting because they cause a significant reduction in the solidus of mantle rocks and a large

54 expansion in the compositional range of melts (Dasgupta, 2018; Foley et al., 2009; Stagno and

55 Frost, 2010), especially at depths greater than 140 km, where melting occurs only when promoted

56 by volatiles (Dasgupta and Hirschmann, 2006; Foley and Pintér, 2018). The volatile-bearing melts

57 produced solidify elsewhere to cause mineralogical and geochemical variations in the Earth’s

Page 2/28
58 mantle (Aiuppa et al., 2021; Gaillard et al., 2008). It is increasingly recognized that volatile-rich

59 dykes/veins or domains (carbonate, phlogopite, amphibole, and/or fluid) co-exist in the mantle

60 together with the dominant lithology peridotite (Dawson and Smith, 1977; Foley, 1992; Grégoire

61 et al., 2002; Smart et al., 2019; Tappe et al., 2008). Geochemically, volatile-bearing mantle is

62 enriched in fusible components (e.g., K, Na, Ca). Alkaline igneous rocks are generally formed by

63 volatile-triggered mantle melting and hold key information about the role that volatile evolution

64 plays in governing the spatial and temporal variation in mantle rocks, as revealed by major and

65 trace elements, radiogenic isotopes (e.g., Sr, Nd, Hf and Pb isotopes), and olivine geochemistry

66 (Foley et al., 2013; Tappe et al., 2006; Tappe et al., 2007).

67 Stable calcium isotopes (δ44/40Ca = ((44Ca/40Ca)SAMPLE/(44Ca/40Ca)SRM915a - 1) × 1000) can be

68 used to trace large-scale geological processes, opening a new window in our understanding of

69 alkaline magma origins, including magma processes and the lithology of mantle sources

70 (Antonelli et al., 2023; Eriksen and Jacobsen, 2022; Wang et al., 2019). Considerable equilibrium

71 Ca isotope fractionation (up to 1‰) has been observed between the mantle minerals carbonate,

72 clinopyroxene (Cpx), K-richterite, garnet (Grt), and orthopyroxene (Opx) (Antonelli et al., 2019;

73 Chen et al., 2019; Chen et al., 2020b; Xiao et al., 2022): carbonate and hydrous K-richterite have

74 lower δ44/40Ca values than Cpx. Such inter-mineral Ca isotope differences may result in: (1) large

75 Ca isotope fractionation between melts and residual solids during partial melting (Antonelli et al.,

76 2021; Eriksen and Jacobsen, 2022; Wang et al., 2019); and (2) significant Ca isotope differences

77 between different mantle lithologies, as known for example from the high Ca isotope values for

78 dunite and orthopyroxenite (1.11-1.81‰ and 1.13‰ δ44/40Ca, respectively) (Chen et al., 2019) and

79 low δ44/40Ca values for carbonated peridotites (0.56-0.95‰, Ionov et al., 2019; Zhu et al., 2021).

80 Therefore, Ca isotopes are potentially a powerful tool for tracing lithological variations caused by

81 volatiles in the mantle sources of alkaline magmas.

82 The calcium isotope fractionation factor between Cpx and silicate melt is close to 1 (Chen et

83 al., 2019; Zhang et al., 2018), so that δ44/40Ca values of mid-ocean ridge basalts (MORBs: 0.84 ±

84 0.03‰, 2se) are similar to those of Cpx in spinel peridotites. Recently, Eriksen and Jacobsen

85 (2022) emphasized that melting of a garnet-rich source may result in the light Ca isotope

86 compositions shown by some oceanic island basalts (OIBs) due to the heavier Ca isotope

87 compositions of Grt compared with Cpx (Chen et al., 2020b; Li et al., 2022; Wang et al., 2019).

Page 3/28
 88 However, possible Ca isotope fractionation in alkaline magmas caused by volatile-bearing

89 minerals or melts in their mantle sources is largely unexplored. Calcium is abundant in several

90 volatile-bearing minerals (e.g., carbonate and richteritic amphiboles) and mantle-derived melts

91 (e.g., carbonatites), and Ca isotope differences between Cpx and volatile-bearing minerals are

92 significant (Wang et al., 2017a; Xiao et al., 2022). Knowledge of Ca isotope fractionation related

93 to the action of volatiles in the mantle is therefore important for understanding the large observed

94 Ca isotope variations and thus the origin of volatile-rich igneous rocks such as silica-

95 undersaturated alkaline rocks, kimberlites, and carbonatites (Amsellem et al., 2020; Antonelli et

96 al., 2023; Banerjee et al., 2021; Sun et al., 2021).

97 The North Atlantic craton (NAC) was split by continental rifting, ultimately marked by the

98 opening of the Labrador Sea and Davis Strait with short periods of basaltic oceanic crust

99 formation during the early Cenozoic (Fig. 1). Before this, a sequence of ca. 1.4 Ga lamproites,

100 590-555 Ma ultramafic lamprophyres, and ca. 142 Ma nephelinites at Aillik Bay on the western

101 Labrador Sea margin record a long history of craton breakup during which lithospheric thinning

102 occurred in three main stages (Tappe et al., 2006; Tappe et al., 2007). The passive continental

103 margin of Labrador was unaffected by subduction (Keen et al., 2012), and the rifting processes at

104 each of the three stages resulted largely from intraplate continental stretching. The trace elements

105 and radiogenic isotope compositions of these rift-related alkaline rocks reveal the important role of

106 volatile-rich metasomatized lithospheric mantle sources without notable contributions from

107 subducted crustal materials (Tappe et al., 2006; Tappe et al., 2007). The protracted history of this

108 alkaline rock occurrence makes it an ideal target to explore Ca isotope fractionation induced by

109 volatiles in the upper mantle. We have undertaken a detailed Ca isotope study on the diverse

110 magmatic products of the three rifting stages and our results show how the lithological variations

111 caused by volatiles in Earth’s upper mantle may fractionate the Ca isotopes of alkaline and

112 carbonate-rich magmatic rocks.

113 2. Geological setting and samples

114 The NAC was split during the Mesozoic-Cenozoic into two continental blocks – the Nain

115 Province of Labrador and the Archaean terranes of West Greenland, separated by the Labrador Sea

116 (Fig. 1). The Archaean blocks on either side of the Labrador Sea preserve tonalitic crust as old as

Page 4/28
117 3.8 Ga and thick cratonic mantle roots persist to the present day (Windley and Garde, 2009) (Fig.

118 1). The Archaean cores were augmented by the accretion of micro-continents assembled during the

119 Paleoproterozoic (1.9-1.7 Ga) marking the final phase of cratonization through a sequence of

120 subduction and collision events (Wardle and Hall, 2002). The NAC has experienced numerous

121 mantle-derived alkaline igneous events since the Proterozoic (Larsen and Rex, 1992), which

122 eventually accompanied rifting that proceeded to the production of oceanic crust in the Labrador

123 Sea for a short time interval during the early Cenozoic. The uplifted basement blocks of the NAC

124 record one of the longest craton splitting histories known: the first stage corresponds to

125 impregnation of the mantle lithosphere with K-rich silicate melts >2000 Myr ago, which was

126 reactivated and melted at ca. 1370 Ma to produce olivine lamproite magmas. The second stage is

127 represented by widespread emplacement of ultramafic lamprophyres and carbonatites in the late

128 Neoproterozoic (Fig. 1), and the third episode led to successful rifting and the production of new

129 oceanic crust, preceded by the eruption of kimberlites, ultramafic lamprophyres and carbonatites

130 in western Greenland between ca. 200-150 Ma (Tappe et al., 2017), and eventually nephelinite and

131 melilitite magmas along the rift flanks at ca. 150-100 Ma (Tappe et al., 2006; Tappe et al., 2007).

132 The samples selected for Ca isotopic analysis are from the Aillik Bay locality (Fig. 1), which

133 is the only known area where all three rifting stages are recorded by the occurrence of alkaline

134 magmatic rocks. The samples include four 1.37 Ga old lamproites, 20 ultramafic lamprophyres

135 and carbonatites emplaced between 590 and 555 Ma, and five nephelinites and melilitites dated at

136 ca. 142 Ma. The ultramafic lamprophyres are primitive SiO2-poor potassic igneous rocks

137 comprising aillikites (for which Aillik Bay is the type locality), mela-aillikites, and damtjernites

138 following the classification of Tappe et al. (2005), in which aillikites are the archetypal carbonate-

139 rich member of the ultramafic lamprophyre group (Malpas et al., 1986; Rock, 1986). Detailed

140 descriptions and age determinations of the rocks analyzed in this study, along with their major and

141 trace element contents and Sr-Nd-Hf-Pb isotopic compositions, can be found in Tappe et al. (2006)

142 and Tappe et al. (2007).

143 3. Analytical methods

144 Chemical purification and measurement of Ca isotope ratios were performed at the State Key

145 Laboratory of Geological Processes and Mineral Resources, China University of Geosciences,

Page 5/28
146 Wuhan, China, with methods following Feng et al. (2018). Briefly, Ca was separated from the

147 sample matrix on micro-columns filled with Ca-selective DGA resin. High recovery (> 99%),

148 efficient separation of Ca, and a low total procedural blank of <10 ng were achieved. Following

149 purification, calcium isotopic compositions were analyzed on a Neptune Plus (Thermo Scientific,

150 Bremen, Germany) MC-ICP-MS collecting 42Ca+, 43Ca+, and 44Ca+ isotopes using Faraday cups at

151 L3, L2, and central positions, respectively. The signal at m/z 43.5 was also collected at L1 to

152 monitor the doubly charged interference of Sr on Ca isotopes and was found to be negligible for

153 purified samples. Isotope measurements were performed using standard-sample bracketing to

154 correct for instrumental drift and results are defined using δ-notation: δn/42Ca =

155 [(nCa/42Ca)sample/(nCa/42Ca)SRM915a-1]×1000 where n=44 or 43. All measured Ca isotope values

156 (δ44/42Ca) were converted to δ44/40Ca using a factor of 2.048 calculated by the mass-dependent

157 fractionation law (Heuser et al., 2016). The long-term external 2sd reproducibility of 0.08‰ for

158 δ44/40Ca in this study is assessed based on replicate measurements of NIST SRM 915a and Alfa Ca

159 over the course of half a year (Fig. S1 in Supplementary Information: Section 1).

160 Three reference materials (COQ-1, BHVO-2, and BCR-2) and three duplicate samples were

161 processed as unknowns to assess accuracy and reproducibility. We obtained δ44/40Ca of 0.74 ±

162 0.06‰ (2sd, n=4), 0.77 ± 0.07‰ (2sd, n=6), and 0.83 ± 0.11‰ (2sd, n=3) for COQ-1, BHVO-2,

163 and BCR-2, respectively (Fig. S2 and Table 1), consistent with previous studies within the

164 analytical uncertainty (Amini et al., 2009; Amsellem et al., 2017; Feng et al., 2017; He et al., 2017;

165 Schiller et al., 2012; Valdes et al., 2014). Three duplicate samples show good reproducibility

166 within analytical uncertainty (Table 1). Furthermore, Antonelli et al. (2023) showed the generally

167 good agreement on analytical results for the same kimberlite powders between our laboratory and

168 the laboratory at Chengdu University of Technology using the double-spike thermal ionization

169 mass-spectrometry. Measured δ44/42Ca and δ43/42Ca of the alkaline rocks and reference materials

170 plot on a line of theoretical kinetic fractionation with a slope of 0.506 within uncertainty (Heuser

171 et al., 2016) (Fig. S3), reflecting mass-independent fractionation without analytical artifacts from

172 mass spectrometric interferences.

173 4. Results

174 Calcium isotopic compositions of the alkaline rocks and carbonatites from Aillik Bay and

Page 6/28
175 reference materials are reported in Table 1. Due to the large interference of 40Ar+ from the argon
40 44
176 carrier gas on Ca in analytes within the ICP-MS instrument, the Ca/40Ca ratio could not be

177 determined directly in this study. All δ44/40Ca values were obtained by conversion of measured

178 δ44/42Ca applying the mass-dependent fractionation law. Thus, δ44/40Ca variations in our samples
40
179 reflect mass-dependent fractionation without the effect of Ca ingrowth from the radioactive

180 decay of 40K. We also calculated δ44/40Ca values accounting for 40Ca addition from the radioactive
40
181 decay of K (δ44/40CaRadiogenic in Table 1; for calculation see Fu et al. (2022)); however, in this

182 study we only use the uncorrected δ44/40Ca values for our samples to discuss the mass-dependent

183 fractionation without the effect of radiogenic excess 40Ca.

184 The Mesoproterozoic olivine lamproites show a small variation of δ44/40Ca values from 0.58

185 to 0.66‰, lower than MORB values (0.84‰, Chen et al., 2020a; Eriksen and Jacobsen, 2022; Zhu

186 et al., 2018) (Fig. 2). The δ44/40Ca values of the Late Neoproterozoic ultramafic lamprophyres

187 range from 0.56 to 0.75‰ (Fig. 2). Among them, aillikites and mela-aillikites have homogeneous

188 Ca isotopic compositions of 0.67‰ to 0.75‰, also systematically lower than MORB values but

189 higher than the olivine lamproites. The δ44/40Ca values of the Neoproterozoic carbonatites are also

190 homogenous (0.71 to 0.82‰, except for one carbonatite at 0.90‰), consistent with those of

191 aillikites and mela-aillikites. The δ44/40Ca values of the damtjernites show a larger range from 0.56

192 to 0.72‰, with five relatively primitive damtjernites at 0.68-0.72‰ and four more evolved

193 damtjernites at 0.56-0.62‰ (Fig. 2). The δ44/40Ca values of the Cretaceous nephelinites vary from

194 0.72 to 0.78‰ (Fig. 2), similar to the aillikites and mela-aillikites.

195 5. Discussion

196 The calcium isotopic compositions of magmatic rocks may be significantly modified by

197 precipitation of Ca-rich secondary minerals such as hydrothermal carbonates. The lamproite,

198 ultramafic lamprophyre, carbonatite, and nephelinite samples studied here are very fresh with only

199 little or no signs of hydrothermal overprinting: no secondary alteration was observed during

200 petrographic analysis (Tappe et al., 2006; Tappe et al., 2007). For magmatic silicate rocks, the lack

201 of correlation between δ44/40Ca values and CO2 contents in the lamproites and nephelinites

202 indicates a negligible effect of alteration on their Ca isotopic compositions (Fig. S4a). The

203 magmatic carbonate-bearing aillikites and damtjernites have overlapping mantle-like δ13C and

Page 7/28
204 δ18O values (Tappe et al., 2006), in contrast to strongly fractionated carbonatites (Fig. S4b).

205 Importantly, the lamproite, aillikite, and nephelinite suites each have homogenous Ca isotope

206 compositions, which suggests that their δ44/40Ca values are unaffected by hydrothermal processes

207 (Fig. 2).

208 5.1. Mantle source lithologies and melting conditions at Aillik Bay

209 The fractional crystallization history of the alkaline rocks needs to be assessed before

210 decoding their mantle source lithologies using Ca isotope geochemistry alongside standard

211 petrological tools. The Mesoproterozoic lamproites contain olivine phenocrysts with a small

212 amount of clinopyroxene, phlogopite and apatite, suggesting that they mainly experienced

213 fractional crystallization of olivine (Tappe et al., 2007). Uniform CaO/Al2O3 ratios with variation

214 in MgO contents in the lamproites also indicates fractional crystallization of olivine without

215 clinopyroxene (Fig. 3a). This is supported by fractional crystallization modelling of an

216 experimental lamproite melt from Foley et al. (2022) using MELTS (Gualda and Ghiorso, 2015),

217 which shows that fractional crystallization of lamproite melt is dominated by olivine until <3.5

218 wt.% MgO is reached (Fig. 3a; for modelling parameters see Table S1). The high MgO contents of

219 the olivine lamproites (mostly >8.1 wt.%) indicate that they represent relatively primitive magmas

220 (Tappe et al., 2007). The Mesozoic nephelinites contain olivine and clinopyroxene phenocrysts.

221 The positive correlation between the CaO/Al2O3 ratio and MgO content for the nephelinites

222 indicates that they underwent fractional crystallization of both olivine and clinopyroxene,

223 consistent with fractional crystallization modelling of an experimental nephelinite melt from

224 Dasgupta et al. (2007) using MELTS (Gualda and Ghiorso, 2015) (Fig. 3a and Table S1).

225 The major element compositions of mantle-derived magmas can be used to constrain source

226 lithologies. Low CaO contents and high FC3MS values (FeO/CaO-3*MgO/SiO2) are typical

227 features of basaltic melts derived from pyroxenites (Lambart et al., 2013; Yang et al., 2016). The

228 Mesoproterozoic lamproites are SiO2-saturated (44.6-47.0 wt.%) and have high FC3MS values but

229 low CaO contents (6.3-7.8 wt.%), different from any experimental melts of peridotite but similar

230 to melts of pyroxenite or MARID metasomes (mica-amphibole-rutile-ilmenite-diopside) (Fig. 4).

231 Although fractional crystallization of clinopyroxene could produce high FC3MS values and low

232 CaO contents, these lamproites did not experience clinopyroxene fractionation. The high K2O

Page 8/28
233 contents (5.1-7.7 wt.%) of the lamproites require a source containing K-rich minerals (e.g.,

234 phlogopite or potassic richterite). Their EM1-like Sr-Nd-Hf-Pb isotope compositions also

235 fingerprint long-term enriched cratonic mantle components, which can be explained by K-rich

236 ultramafic veins in peridotite (Tappe et al., 2007). Furthermore, the Mesoproterozoic lamproites

237 resemble the experimental melts of MARID assemblages (Foley et al., 2022) and are different

238 from experimental melts of phlogopite clinopyroxenite (Funk and Luth, 2013; Lloyd et al., 1985),

239 supporting a MARID-style metasomatized mantle source (Fig. 4b).

240 Among the Neoproterozoic ultramafic lamprophyres, primitive aillikites have low SiO2 and

241 high CaO and MgO contents, similar to experimental melts of carbonated peridotite at ≥3 GPa

242 (Fig. 4a). Their olivine phenocryst geochemistry (Veter et al., 2017) and bulk chemical and

243 isotopic compositions (Tappe et al., 2006; Tappe et al., 2007) are consistent with melting of a

244 carbonated mantle source. The Mesozoic nephelinite suite at Aillik Bay comprises dykes and sills

245 of nephelinite and basanite, plus rare melilitite (Tappe et al., 2007). The high MgO (6.9–11.3

246 wt.%), Ni (76–198 ppm), and Cr contents (123–485 ppm) of this alkaline rock suite indicate

247 derivation from primitive mantle-sourced magmas, and their high CaO contents (10.3-14.6 wt.%)

248 and relatively low FC3MS values are also consistent with experimental melts of peridotite (Fig.

249 4a). Their low SiO2 (34.7–44.3 wt.%) contents resemble experimental melts of peridotite in the

250 presence of CO2 at 2-3 GPa (Green and Falloon, 1998), indicating the involvement of wehrlite in

251 their mantle source (Tappe et al., 2007).

252 Temperatures and pressures of melt formation are estimated using the method of Sun and

253 Dasgupta (2020), which was calibrated for CO2-rich, silica-undersaturated magmatic systems.

254 This thermobarometer treats the Mg-number of olivine in the mantle source and the H2O content

255 of primary magmas as key parameters (Supplementary Information: Section 2). The results in Fig.

256 S5 and Table 1 indicate melting pressures of approximately 4-6 GPa for both lamproites and

257 aillikites, and 2-3 GPa for the nephelinites, broadly consistent with the petrogenetic model for the

258 evolution of alkaline magmatism at Aillik Bay (Tappe et al., 2007). The temperature estimates are

259 approximately 1370-1500 °C for the lamproites and aillikites, and 1250-1360 °C for the

260 nephelinites. To examine the reliability of these temperature models, we calculated the

261 crystallization temperature of nephelinite sample ST103 (~1200 °C) using Al-in-olivine

262 thermometry (Coogan et al., 2014). Given that the nephelinite magma crystallization temperature

Page 9/28
263 is ~100 °C lower than the calculated melting temperature of the mantle source, we place

264 confidence in these results. The melting temperatures are higher than the warmest known cratonic

265 geotherms (Lee et al., 2011) and correspond broadly to the thermal regime of the underlying

266 asthenosphere, indicating that melting was initiated by asthenospheric flow.

267 5.2. Origin of lamproites from K-richterite-bearing metasomes with low 44/40Ca

268 Due to the extremely low CaO contents of the olivine phenocrysts (<0.012 wt.% CaO, Tappe

269 et al., 2007), fractional crystallization of olivine changes neither CaO/Al2O3 nor δ44/40Ca of the

270 magma (Fig. 3). Furthermore, the lamproites show homogenous Ca isotopic compositions

271 regardless of variable CaO and MgO contents, demonstrating the limited effect of fractional

272 crystallization on the δ44/40Ca compositions. Apatite has slightly lower 44/40Ca than Cpx (Xiao et

273 al., 2022), so that fractionation of minor apatite would result in slightly heavier Ca isotope

274 compositions of the residual melt, contrasting with the light Ca isotopes of the Mesoproterozoic

275 lamproites. Therefore, the δ44/40Ca values of the lamproites must reflect the Ca isotopic

276 compositions of their parental magmas.

277 The Ca isotope compositions of the Mesoproterozoic lamproites are lighter than those of

278 MORBs (Chen et al., 2020a; Eriksen and Jacobsen, 2022; Zhu et al., 2018). This can be attributed

279 to either isotope fractionation during the melting of MARID or inheritance of Ca isotope

280 compositions from the MARID component that already possessed low δ44/40Ca. K-richterite and

281 clinopyroxene are the two major Ca-phases in MARID assemblages. K-richterite is fully

282 consumed in melting reactions within 50 °C of the MARID solidus (Foley et al., 2022), and it has

283 much lower δ44/40Ca values than Cpx (Xiao et al., 2022). Given that the fractionation factor for Ca

284 isotopes between clinopyroxene and silicate melt is close to 1, K-richterite would have lighter Ca

285 isotopes than its equilibrium silicate melts (Fig. 5a). The δ44/40Ca values of melts produced from

286 MARID assemblages are higher than the residue and starting material if K-richterite is not

287 completely consumed, and would approach the source values if K-richterite is completely

288 consumed (Fig. 5b, Supplementary information: Section 3.1, and Table S2). Calcium isotope

289 fractionation induced by partial melting cannot explain the light Ca isotopes in the

290 Mesoproterozoic lamproites. We therefore suggest that the K-richterite in the MARID source

291 carried the low δ44/40Ca signature. The Ca isotope compositions of mantle rocks are also dependent

Page 10/28
292 on their mineralogy; for example, dunite and orthopyroxenite have heavy Ca isotopes due to the

293 modal dominance of olivine and Opx, respectively (Chen et al., 2019). Similarly, K-richterite-

294 bearing MARID assemblages would have much lower Ca isotope compositions than peridotites

295 due to the low δ44/40Ca nature of K-richterite (Xiao et al., 2022).

296 K-richterite-bearing MARID was formed by reaction of infiltrating hydrous silicate melts

297 with the refractory cratonic mantle lithosphere. Considering the much lower CaO content of

298 refractory harzburgite (<0.5 wt.%, Pearson et al. (2003)) than that of hydrous silicate melt (for

299 example, the average CaO contents of worldwide continental basalts are 7.3-11.1 wt.%, Farmer

300 (2014)), the Ca isotopes in K-richterite-bearing MARID are not affected by the refractory cratonic

301 mantle, but are controlled by metasomatic melts. We modelled the Ca isotope compositions of K-

302 richterite-bearing MARID material formed by mantle metasomatism at the base of the cratonic

303 lithosphere for different geothermal gradients (Fig. 5c). These calculations assume that the

304 hydrous silicate melts that caused the metasomatism had a MORB-like δ44/40Ca value of 0.84‰

305 (Supplementary Information: Section 3.2). Results show that the K-richterite-bearing MARID has

306 a theoretical δ44/40Ca value of 0.65‰ at 950-1100 °C. This model temperature range provides a

307 realistic analogue to natural MARID assemblage formation by melt/fluid metasomatism in the

308 relatively cold Kaapvaal cratonic mantle (Fitzpayne et al., 2018a; Fitzpayne et al., 2018b).

309 Therefore, the light Ca isotope compositions of the Mesoproterozoic lamproites from the NAC

310 indicate a mantle source containing K-richterite-bearing MARID-type metasomes.

311

312 5.3. Calcium isotope systematics of ultramafic lamprophyres and associated carbonatites

313 5.3.1. Negligible calcium isotope fractionation during differentiation of carbonated silicate magma

314 The aillikites, mela-aillikites, damtjernites, and carbonatites studied here have identical Sr-

315 Nd-Hf-Pb isotopic compositions, indicating that they are related to a common parental magma

316 formed by melting of a carbonate-rich mantle domain (Tappe et al., 2006). The low SiO2 (22.5-

317 26.8 wt.%) and high MgO (17.3-20.9 wt.%), Ni (483-719 ppm), and Cr contents (577-734 ppm) of

318 the aillikites testify to their near-primary nature (Tappe et al., 2006; Tappe et al., 2008). In contrast,

319 the mela-aillikites show higher SiO2 (31.6-35.9 wt.%) contents but low CaO (8.2-11 wt.%) and

320 CO2 (2.0-5.4 wt.%) contents, falling on a trend that may correspond to the separation of a

Page 11/28
321 carbonate-rich component from more primitive aillikite magma, with the mela-aillikites

322 representing the silicate-enriched portion (Fig. 6) (Tappe et al., 2006). The mela-aillikites show the

323 same range of δ44/40Ca values (0.67 to 0.72 ‰) as primitive aillikites (0.71-0.75‰), indicating that

324 separation of carbonate from silicate components, which may include decarbonation and CO2-

325 degassing, causes negligible fractionation of Ca isotopes.

326 Although the damtjernites do not lie on the same trend of carbonate liquid separation from

327 the more primitive aillikites, they are generally SiO2-richer (similar to the mela-aillikites) and

328 show large variations in their SiO2 (30.2-38 wt.%), CaO (10.3-21.2 wt.%), and CO2 (0.2-8.1 wt.%)

329 contents (Fig. 6 and Table 1). Damtjernites have been interpreted to result from liquid

330 immiscibility from an alkali-rich proto-aillikite magma at relatively shallow depths (Tappe et al.,

331 2006). Four evolved damtjernite samples with low Mg# values and high Al2O3 contents have

332 slightly lower δ44/40Ca values (0.56 to 0.62 ‰), whereas the more primitive damtjernites show

333 higher δ44/40Ca values (0.67 to 0.72 ‰). Following liquid immiscibility, these evolved damtjernites

334 experienced fractional crystallization of olivine, clinopyroxene and apatite, as indicated by olivine

335 phenocrysts and clinopyroxene-apatite prisms (Tappe et al., 2006). Clinopyroxene and apatite may

336 have heavier Ca isotopes than the carbonate-bearing damtjernitic magmas, just as clinopyroxene

337 and apatite have heavier Ca isotopes than dolomite (Fig. 5a). Therefore, fractional crystallization

338 of clinopyroxene and apatite may induce minor Ca isotope fractionation, explaining the slightly

339 lower δ44/40Ca values of the four evolved damtjernite samples. The δ44/40Ca values (0.67 to 0.72 ‰)

340 of the primitive damtjernites and carbonatites (except for ST203 with a high δ44/40Ca value of

341 0.90‰) are similar to those of aillikites and mela-aillikites. Together, these observations suggest

342 negligible Ca isotope fractionation during differentiation of carbonated silicate magmas including

343 liquid immiscibility and CO2-degassing.

344 5.3.2. Calcium isotope fractionation during low-degree partial melting of carbonated peridotite

345 The δ44/40Ca values of the most primitive aillikite samples represent those of primary

346 carbonated silicate melts (0.72 ± 0.05 ‰, 2sd) produced by incipient melting of carbonated

347 peridotites. We compiled bulk Ca isotope compositions of carbonated peridotite xenoliths (Ionov

348 et al., 2019; Zhu et al., 2021), which show an average δ44/40Ca value of 0.82 ± 0.05 ‰, 2sd (Fig.

349 7a and Table S3). The significantly different δ44/40Ca values for carbonated silicate melts and

Page 12/28
350 carbonated peridotites could potentially reflect Ca isotope fractionation during low-degree partial

351 melting of carbonated mantle rocks. Carbonate, clinopyroxene, and garnet are the main Ca-phases

352 in carbonated mantle peridotite domains at the pressure and temperature conditions relevant to the

353 lowermost cratonic lithosphere and underlying convecting upper mantle. First-principles

354 calculations can be used to predict Ca isotope fractionation between clinopyroxene and various

355 types of carbonate minerals, including calcite, dolomite, and magnesite (Wang et al., 2017a; Wang

356 et al., 2017b). Results show that Δ44/40CaCarbonate-Cpx is -0.04 to -0.10 when the Ca/(Mg+Ca) of the

357 carbonate is higher than 1/6 at 1050 ℃ (Fig. 5a). Furthermore, garnet and Opx that are residual

358 after low-degree melting of carbonated peridotite have heavier Ca isotope compositions than Cpx

359 and carbonate. We quantitatively modelled the behaviour of Ca isotopes during partial melting of

360 magnesite lherzolite at 6.6 GPa using an incremental batch melting model (Chen et al., 2019;

361 Williams and Bizimis, 2014). The inter-mineral fractionation factors for carbonate, garnet, and

362 pyroxenes are from Chen et al. (2020b) and Wang et al. (2017a), and melting parameters are from

363 Dasgupta et al. (2007). Details of the modelling are presented in the Supplementary Information:

364 Section 3.1 and Table S2, and the results are illustrated in Fig. 5d. Our results show that isotope

365 fractionation of about 0.08‰ δ44/40Ca occurs between carbonated melt and its peridotitic residue,

366 which explains the measured differences (about 0.1‰) between carbonated peridotites and the

367 carbonated silicate melts at Aillik Bay (Fig. 7a).

368 5.3.3. Calcium isotope compositions of carbonate-rich mantle melts

369 Previously reported δ44/40Ca values for global carbonatites show a large variation (Figs. 2 and

370 7). The origin of this large Ca isotope variation is unclear, and competing ideas invoke Rayleigh

371 fractionation during late-stage precipitation of carbonates from fluid-rich carbonatite magmas

372 (Sun et al., 2021) and the involvement of recycled sedimentary carbonate in the mantle source

373 (Amsellem et al., 2020; Banerjee et al., 2021). The calcium isotope compositions of primitive

374 carbonate-rich magmas from the mantle can serve as a benchmark to evaluate directions and

375 degrees of isotope variation caused by later magmatic-hydrothermal processes or originally by

376 mantle source contamination with recycled sedimentary carbonate components. Previous studies

377 used the Bulk Silicate Earth (BSE) composition as the benchmark for δ44/40Ca (Amsellem et al.,

378 2020; Banerjee et al., 2021). However, experiments show that melts of carbonated peridotite

Page 13/28
379 at >140 km depth have a carbonated silicate structure and composition (Dasgupta and Hirschmann,

380 2006; Pintér et al., 2021), with aillikites being the best natural analogues. Therefore, the Ca

381 isotope compositions of the type aillikites and associated pristine carbonatites can be used to

382 constrain the calcium isotope compositions of primitive carbonate-rich magmas in the mantle.

383 The average δ44/40Ca value of the carbonate-rich igneous rocks at Aillik Bay (aillikites and

384 carbonatites) is 0.72 ± 0.05‰ (2sd, n=11). This value is consistent with the δ44/40Ca of 0.72 ±

385 0.06‰ (2sd, n=18) for pristine global carbonatites with mantle-like C-O isotope compositions

386 (Sun et al., 2021). It is also similar to the average δ44/40Ca composition of magmatic kimberlites at

387 0.77 ± 0.11‰ (2sd, n=19) (Antonelli et al., 2023). Therefore, we propose that the average value of

388 0.74‰ (2sd = 0.10‰; 2se = 0.02‰; n = 48) represents the Ca isotope composition of carbonate-

389 rich magmas in the mantle.

390 Carbonatite sample ST203 from Aillik Bay has a δ13C composition of -2.8‰ and this

391 deviation from the mantle range (-4 to -6‰ δ13C) was explained by high-temperature Rayleigh

392 fractionation (Tappe et al., 2006). The high δ44/40Ca value of 0.90‰ for ST203 suggests that

393 complex carbonatite magma evolution can fractionate the Ca and C isotope compositions, as

394 suggested by Sun et al. (2021). This may explain the origin of high δ44/40Ca values in some

395 carbonatites from occurrences worldwide (much higher than 0.72 ‰, Fig. 7b-c), which must await

396 future detailed investigations.

397

398 5.4. Ca isotope constraints on the metasomatic source component in the Mesozoic

399 nephelinites

400 The Mesozoic nephelinites experienced fractional crystallization of olivine and

401 clinopyroxene (see Section 5.1). The fractionation factor for Ca isotopes between clinopyroxene

402 and silicate melt is close to 1, so that fractional crystallization of clinopyroxene does not

403 significantly affect Ca isotopes (Chen et al., 2020a; Zhang et al., 2018). The lack of correlation

404 between δ44/40Ca values and CaO/Al2O3 ratios also indicates that fractional crystallization of

405 clinopyroxene did not change the Ca isotopic compositions of the nephelinites (Fig. 3b).

406 Therefore, the δ44/40Ca values of the nephelinites reflect the Ca isotopic compositions of their

407 parental magmas.

Page 14/28
408 The δ44/40Ca values of the nephelinites are similar to those of the Late Neoproterozoic

409 ultramafic lamprophyres and carbonatites, slightly lower than average MORB (Figs. 2 and 8). This

410 could be attributed to either the direct effect of melting of wehrlite or to the wehrlite source

411 already possessing low δ44/40Ca. Partial melting of shallow spinel-facies wehrlite (melting pressure

412 = 2-3 GPa) cannot explain the systematically low δ44/40Ca values of the nephelinites relative to

413 MORBs because the Ca budget is dominated by Cpx during partial melting (Chen et al., 2019;

414 Zhang et al., 2018). The wehrlite-rich source of the nephelinites formed by metasomatism by

415 carbonate-rich melts, as indicated by the low SiO2 contents and high CaO/Al2O3 ratios of the

416 nephelinites (Tappe et al., 2007). Trace element patterns of the nephelinites resemble those of

417 aillikites, and their Sr-Nd-Hf-Pb isotope compositions indicate that their source formed by

418 metasomatic reactions between cratonic lherzolite and aillikitic melts during the Neoproterozoic

419 (Tappe et al., 2007). This is consistent with similar δ44/40Ca values for the nephelinites and the

420 Neoproterozoic aillikites. Therefore, we suggest that the wehrlites in their mantle source already

421 had δ44/40Ca values of about 0.75‰ as a result of metasomatism by carbonated silicate melts

422 derived from Neoproterozoic aillikite magmatism.

423

424 5.5. A calcium isotope perspective on the composition of metasomatized cratonic mantle and

425 the origin of alkaline magmatism

426 The lithosphere-asthenosphere boundary (LAB) beneath cratons represents a transition zone

427 that undergoes episodic rejuvenation caused by the impregnation with asthenosphere-derived

428 volatile-rich melts (Foley, 2008; O'Reilly and Griffin, 2010). The cratonic LAB is recognized as a

429 key source of ultramafic alkaline and carbonatitic magmas, and as a zone that modulates the

430 chemical compositions of deeper-derived magmas during their ascent (Foley et al., 2019; Giuliani

431 et al., 2020; Tappe et al., 2023). Our results reveal that the Ca isotope compositions of

432 metasomatized lithosphere at the base of the NAC are different from those typical for refractory

433 cratonic mantle and the asthenosphere (Chen et al., 2019; Kang et al., 2017) (Fig. 9). For example,

434 δ44/40Ca values of about 0.60‰ characterize the MARID-type metasomes tapped during

435 Mesoproterozoic lamproite magmatism, and δ44/40Ca values of about 0.72-0.82‰ characterize the

436 carbonated peridotite components that gave rise to Neoproterozoic ultramafic lamprophyre and

Page 15/28
437 carbonatite magmatism. Undoubtedly, the multiply metasomatically overprinted LAB beneath

438 cratons has an important effect on the Ca isotope compositions of deep-sourced magmas (Tappe et

439 al., 2021).

440 The δ44/40Ca values of some continental SiO2-saturated basalts range from 0.65 to 0.80‰ (Liu

441 et al., 2017), notably lower than the MORB average (Fig. 8d). This has been attributed to

442 recycling of sedimentary carbonate into their mantle sources (Liu et al., 2017). However, the SiO2-

443 saturated nature of these basalts at low CaO/Al2O3 is in conflict with magma origins from a

444 carbonated mantle source (Fig. 8). Our results for the Mesoproterozoic lamproites from the NAC

445 indicate a key role for K-richterite in determining the Ca isotope compositions of alkaline

446 magmas. We suggest that low-δ44/40Ca K-richterite in the cold metasomatized lithospheric mantle

447 should be considered as a possible source of the low δ44/40Ca signature of certain SiO2-saturated

448 basalts. Potassic richterite is a readily fusible component enriched in Ca, Na and K, and will be an

449 early contributor to melts generated within metasomatised lithosphere (Foley et al., 2022), but also

450 to asthenosphere-derived melts upon their entry into metasomatised lithosphere, particularly where

451 the lithosphere is very thick (Gao et al., 2023). The high Na2O+K2O contents of continental SiO2-

452 saturated basalts indicate that K-richterite may have been involved in producing the low δ44/40Ca

453 values (Fig. 8d).

454 Carbonated silicate melts are stable and prevalent at >140 km depths (Dasgupta and

455 Hirschmann, 2006) because deep mantle melting occurs only where promoted by carbon and

456 water (Dasgupta and Hirschmann, 2006; Foley and Pintér, 2018). Our δ44/40Ca value of 0.74 ±

457 0.02 ‰ (2se) for ultramafic lamprophyres and carbonatites from a rifted craton setting serves as a

458 benchmark for primary carbonate-rich melts in the mantle to evaluate the directions and degrees of

459 isotope variations observed in carbonate-rich magmas, and it provides a more appropriate choice

460 than the δ44/40Ca value for BSE (Chen et al., 2019; Kang et al., 2017), as used in previous studies

461 (Banerjee et al., 2021; Qi et al., 2022; Zhao et al., 2022). Furthermore, the low δ44/40Ca values of

462 the Mesozoic nephelinites from Aillik Bay suggest a role for carbonated silicate melts in the origin

463 of SiO2-undersaturated alkaline rocks. We suggest that the low δ44/40Ca values for previously

464 reported SiO2-undersaturated alkaline basalts (0.65 to 0.77 ‰) and some carbonatites may not

465 trace recycled sedimentary carbonate in their mantle source as previously proposed, but rather that

466 such Ca isotope signatures are a hallmark geochemical feature of asthenosphere-derived

Page 16/28
467 carbonated melts and their metasomatic products (Fig. 8). Furthermore, SiO2-undersaturated

468 alkaline OIBs also show lower δ44/40Ca values than the MORB average (Eriksen and Jacobsen,

469 2022; He et al., 2017; Jacobson et al., 2015; Zhang et al., 2018) (Figs. 2 and 8). The influence of

470 carbonated silicate melts on OIB mantle sources leading potentially to low δ44/40Ca values should

471 be explored in the near future.

472

473 6. Conclusions

474 To explore the effects of lithological mantle heterogeneity on the Ca isotope systematics of

475 magmas, we present new Ca isotope data for lamproites, ultramafic lamprophyres, carbonatites,

476 and nephelinites from the North Atlantic craton at Aillik Bay on the margin of the Labrador Sea

477 rift. The Mesoproterozoic lamproites show lower δ44/40Ca values (0.58 to 0.66‰) than MORBs,

478 suggesting an origin by melting of a MARID-type source containing K-richterite, a mineral that

479 can cause the low-δ44/40Ca signature. The δ44/40Ca values of the Neoproterozoic ultramafic

480 lamprophyres are relatively uniform (0.67 to 0.75‰) and similar to those of the carbonatites (0.71

481 to 0.82‰), discounting significant Ca isotope fractionation during liquid immiscibility of a

482 parental carbonated silicate melt. The average δ44/40Ca value of the carbonate-bearing magmas is

483 slightly, but consistently, lower than the MORB average. The δ44/40Ca values of the Mesozoic

484 nephelinites (0.72 to 0.78‰) are similar to those of the Neoproterozoic ultramafic lamprophyres,

485 indicating a wehrlitic metasomatic source that was formed by interaction between cratonic mantle

486 peridotite and rising carbonated silicate magmas during the Neoproterozoic rifting stage. The new

487 data indicate that both K-richterite and carbonate components in upper mantle environments can

488 cause systematically low δ44/40Ca values in alkaline melts produced from metasomatized sources,

489 which may explain the frequently observed low-δ44/40Ca signature of some intraplate continental

490 and oceanic basalts. Our study highlights Ca isotopes as a robust tracer of lithological variations in

491 the mantle sources of alkaline magmas, promising to constrain the sources and mobility of volatile

492 components within the crust-mantle system.

493

494 Acknowledgments

Page 17/28
495 This research is supported by Key R&D Program of China (2019YFA0708400), NSFC

496 (41530211) and SKL-GPMR (MSFGPMR01). SFF and CC are funded by ARC grant

497 FL180100134. Mapping and sampling by ST and SFF in Labrador during 2003-2004 were

498 financially supported by the German Research Foundation (DFG). We appreciate the constructive

499 comments from two anonymous reviewers that significantly improved the manuscript and Editor

500 Rosemary Hickey-Vargas for the efficient editing of the paper. We thank Dr. Kai Wang for help

501 with the map in Figure 1.

502 CRediT authorship contribution statement

503 Chunfei Chen: Conceptualization, Investigation, Methodology, Formal analysis, Writing –

504 original draft. Stephen F. Foley: Conceptualization, Funding acquisition, Writing – review &

505 editing. Sebastian Tappe: Investigation, Writing – review & editing. Huange Ren: Methodology,

506 Resources. Lanping Feng: Methodology, Resources. Yongsheng Liu: Conceptualization, Funding

507 acquisition.

508 Competing interests

509 The authors declare no competing interests.

510 Data availability

511 The data that support the findings of this study are available in the paper or in the supplementary

512 files.

513

514 References

515 Aiuppa, A., Casetta, F., Coltorti, M., Stagno, V., Tamburello, G., 2021. Carbon concentration increases
516 with depth of melting in Earth’s upper mantle. Nature Geoscience 14, 697-703.

517 Amini, M., Eisenhauer, A., Böhm, F., Holmden, C., Kreissig, K., Hauff, F., Jochum, K.P., 2009.
518 Calcium Isotopes (δ44/40Ca) in MPI-DING Reference Glasses, USGS Rock Powders and Various Rocks:
519 Evidence for Ca Isotope Fractionation in Terrestrial Silicates. Geostandards and Geoanalytical
520 Research 33, 231-247.

521 Amsellem, E., Moynier, F., Bertrand, H., Bouyon, A., Mata, J., Tappe, S., Day, J.M., 2020. Calcium
522 isotopic evidence for the mantle sources of carbonatites. Science advances 6, eaba3269.

523 Amsellem, E., Moynier, F., Pringle, E.A., Bouvier, A., Chen, H., Day, J.M.D., 2017. Testing the
524 chondrule-rich accretion model for planetary embryos using calcium isotopes. Earth and Planetary

Page 18/28
525 Science Letters 469, 75-83.

526 Antonelli, M.A., Giuliani, A., Wang, Z., Wang, M., Zhou, L., Feng, L., Li, M., Zhang, Z., Liu, F.,
527 Drysdale, R.N., 2023. Subducted carbonates not required: Deep mantle melting explains stable Ca
528 isotopes in kimberlite magmas. Geochimica et Cosmochimica Acta 348, 410-427.

529 Antonelli, M.A., Kendrick, J., Yakymchuk, C., Guitreau, M., Mittal, T., Moynier, F., 2021. Calcium
530 isotope evidence for early Archaean carbonates and subduction of oceanic crust. Nature
531 Communications 12, 2534.

532 Antonelli, M.A., Schiller, M., Schauble, E.A., Mittal, T., DePaolo, D.J., Chacko, T., Grew, E.S., Tripoli,
533 B., 2019. Kinetic and equilibrium Ca isotope effects in high-T rocks and minerals. Earth and Planetary
534 Science Letters 517, 71-82.

535 Banerjee, A., Chakrabarti, R., Simonetti, A., 2021. Temporal evolution of δ 44/40Ca and 87
Sr/86Sr of
536 carbonatites: Implications for crustal recycling through time. Geochimica et Cosmochimica Acta 307,
537 168-191.

538 Chen, C., Ciazela, J., Li, W., Dai, W., Wang, Z., Foley, S.F., Li, M., Hu, Z., Liu, Y., 2020a. Calcium
539 isotopic compositions of oceanic crust at various spreading rates. Geochimica et Cosmochimica Acta
540 278, 272-288.

541 Chen, C., Dai, W., Wang, Z., Liu, Y., Li, M., Becker, H., Foley, S.F., 2019. Calcium isotope
542 fractionation during magmatic processes in the upper mantle. Geochimica et Cosmochimica Acta 249,
543 121-137.

544 Chen, C., Huang, J.-X., Foley, S.F., Wang, Z., Moynier, F., Liu, Y., Dai, W., Li, M., 2020b.
545 Compositional and pressure controls on calcium and magnesium isotope fractionation in magmatic
546 systems. Geochimica et Cosmochimica Acta 290, 257-270.

547 Coogan, L.A., Saunders, A.D., Wilson, R.N., 2014. Aluminum-in-olivine thermometry of primitive
548 basalts: Evidence of an anomalously hot mantle source for large igneous provinces. Chemical Geology
549 368, 1-10.

550 Dasgupta, R., 2018. Volatile-bearing partial melts beneath oceans and continents–Where, how much,
551 and of what compositions? American Journal of Science 318, 141-165.

552 Dasgupta, R., Hirschmann, M.M., 2006. Melting in the Earth's deep upper mantle caused by carbon
553 dioxide. Nature 440, 659-662.

554 Dasgupta, R., Hirschmann, M.M., 2007. Effect of variable carbonate concentration on the solidus of
555 mantle peridotite. American Mineralogist 92, 370-379.

556 Dasgupta, R., Hirschmann, M.M., Smith, N.D., 2007. Partial Melting Experiments of Peridotite + CO 2
557 at 3 GPa and Genesis of Alkalic Ocean Island Basalts. Journal of Petrology 48, 2093-2124.

558 Dawson, J.B., Smith, J.V., 1977. The MARID (mica-amphibole-rutile-ilmenite-diopside) suite of

Page 19/28
559 xenoliths in kimberlite. Geochimica et Cosmochimica Acta 41, 309-323.

560 Eriksen, Z.T., Jacobsen, S.B., 2022. Calcium isotope constraints on OIB and MORB petrogenesis: The
561 importance of melt mixing. Earth and Planetary Science Letters 593, 117665.

562 Farmer, G.L., 2014. 4.3 - Continental Basaltic Rocks, in: Holland, H.D., Turekian, K.K. (Eds.), Treatise
563 on Geochemistry (Second Edition). Elsevier, Oxford, pp. 75-110.

564 Feng, L.-P., Zhou, L., Yang, L., DePaolo, D.J., Tong, S.-Y., Liu, Y.-S., Owens, T.L., Gao, S., 2017.
565 Calcium Isotopic Compositions of Sixteen USGS Reference Materials. Geostandards and
566 Geoanalytical Research 41, 93-106.

567 Feng, L., Zhou, L., Yang, L., Zhang, W., Wang, Q., Tong, S., Hu, Z., 2018. A rapid and simple single-
568 stage method for Ca separation from geological and biological samples for isotopic analysis by MC-
569 ICP-MS. Journal of Analytical Atomic Spectrometry 33, 413-421.

570 Fitzpayne, A., Giuliani, A., Hergt, J., Phillips, D., Janney, P., 2018a. New geochemical constraints on
571 the origins of MARID and PIC rocks: Implications for mantle metasomatism and mantle-derived
572 potassic magmatism. Lithos 318-319, 478-493.

573 Fitzpayne, A., Giuliani, A., Phillips, D., Hergt, J., Woodhead, J.D., Farquhar, J., Fiorentini, M.L.,
574 Drysdale, R.N., Wu, N., 2018b. Kimberlite-related metasomatism recorded in MARID and PIC mantle
575 xenoliths. Mineralogy and Petrology 112, 71-84.

576 Foley, S., 1992. Vein-plus-wall-rock melting mechanisms in the lithosphere and the origin of potassic
577 alkaline magmas. Lithos 28, 435-453.

578 Foley, S.F., 2008. Rejuvenation and erosion of the cratonic lithosphere. Nature Geoscience 1, 503-510.

579 Foley, S.F., Ezad, I.S., van der Laan, S.R., Pertermann, M., 2022. Melting of hydrous pyroxenites with
580 alkali amphiboles in the continental mantle: 1. Melting relations and major element compositions of
581 melts. Geoscience Frontiers 13, 101380.

582 Foley, S.F., Pintér, Z., 2018. Chapter 1 - Primary Melt Compositions in the Earth's Mantle, in: Kono, Y.,
583 Sanloup, C. (Eds.), Magmas Under Pressure. Elsevier, pp. 3-42.

584 Foley, S.F., Prelevic, D., Rehfeldt, T., Jacob, D.E., 2013. Minor and trace elements in olivines as probes
585 into early igneous and mantle melting processes. Earth and Planetary Science Letters 363, 181-191.

586 Foley, S.F., Yaxley, G.M., Kjarsgaard, B.A., 2019. Kimberlites from Source to Surface: Insights from
587 Experiments. Elements 15, 393-398.

588 Foley, S.F., Yaxley, G.M., Rosenthal, A., Buhre, S., Kiseeva, E.S., Rapp, R.P., Jacob, D.E., 2009. The
589 composition of near-solidus melts of peridotite in the presence of CO2 and H2O between 40 and 60 kbar.
590 Lithos 112, Supplement 1, 274-283.

591 Fu, H., Jacobsen, S.B., Larsen, B.T., Eriksen, Z.T., 2022. Ca-isotopes as a robust tracer of magmatic
592 differentiation. Earth and Planetary Science Letters 594, 117743.

Page 20/28
593 Funk, S.P., Luth, R.W., 2013. Melting phase relations of a mica–clinopyroxenite from the Milk River
594 area, southern Alberta, Canada. Contributions to Mineralogy and Petrology 166, 393-409.

595 Gaillard, F., Malki, M., Iacono-Marziano, G., Pichavant, M., Scaillet, B.J.S., 2008. Carbonatite melts
596 and electrical conductivity in the asthenosphere. Science 322, 1363-1365.

597 Gao, M., Xu, H., Foley, S.F., Zhang, J., Wang, Y., 2023. Ultrahigh-pressure mantle metasomatism in
598 continental collision zones recorded by post-collisional mafic rocks. GSA Bulletin, In press.

599 Giuliani, A., Pearson, D.G., Soltys, A., Dalton, H., Phillips, D., Foley, S.F., Lim, E., Goemann, K.,
600 Griffin, W.L., Mitchell, R.H., 2020. Kimberlite genesis from a common carbonate-rich primary melt
601 modified by lithospheric mantle assimilation. Science Advances 6, eaaz0424.

602 Green, D.H., Falloon, T.J., 1998. Pyrolite: A Ringwood Concept and Its Current Expression, in:
603 Jackson, I. (Ed.), The Earth's mantle: composition, structure, evolution, p. 311.

604 Grégoire, M., Bell, D., Le Roex, A., 2002. Trace element geochemistry of phlogopite-rich mafic mantle
605 xenoliths: their classification and their relationship to phlogopite-bearing peridotites and kimberlites
606 revisited. Contributions to Mineralogy and Petrology 142, 603-625.

607 Gualda, G.A.R., Ghiorso, M.S., 2015. MELTS_Excel: A Microsoft Excel-based MELTS interface for
608 research and teaching of magma properties and evolution. Geochemistry, Geophysics, Geosystems 16,
609 315-324.

610 He, Y., Wang, Y., Zhu, C., Huang, S., Li, S., 2017. Mass-Independent and Mass-Dependent Ca Isotopic
611 Compositions of Thirteen Geological Reference Materials Measured by Thermal Ionisation Mass
612 Spectrometry. Geostandards and Geoanalytical Research 41, 283-302.

613 Heuser, A., Schmitt, A.-D., Gussone, N., Wombacher, F., 2016. Analytical methods: Calcium stable
614 isotope geochemistry., in: Gussone, N., Schmitt, A.-D., Heuser, A., Wombacher, F., Dietzel, M., Tipper,
615 E., Schiller, M., Bohm, F. (Eds.), Calcium stable isotope geochemistry. Springer.

616 Ionov, D.A., Qi, Y.-H., Kang, J.-T., Golovin, A.V., Oleinikov, O.B., Zheng, W., Anbar, A.D., Zhang, Z.-
617 F., Huang, F., 2019. Calcium isotopic signatures of carbonatite and silicate metasomatism, melt
618 percolation and crustal recycling in the lithospheric mantle. Geochimica et Cosmochimica Acta 248, 1-
619 13.

620 Jacobson, A.D., Grace Andrews, M., Lehn, G.O., Holmden, C., 2015. Silicate versus carbonate
621 weathering in Iceland: New insights from Ca isotopes. Earth and Planetary Science Letters 416, 132-
622 142.

623 Kang, J.-T., Ionov, D.A., Liu, F., Zhang, C.-L., Golovin, A.V., Qin, L.-P., Zhang, Z.-F., Huang, F., 2017.
624 Calcium isotopic fractionation in mantle peridotites by melting and metasomatism and Ca isotope
625 composition of the Bulk Silicate Earth. Earth and Planetary Science Letters 474, 128-137.

626 Keen, C.E., Dickie, K., Dehler, S.A., 2012. The volcanic margins of the northern Labrador Sea:
627 Insights to the rifting process. Tectonics 31, TC1011.

Page 21/28
628 Lambart, S., Laporte, D., Schiano, P., 2013. Markers of the pyroxenite contribution in the major-
629 element compositions of oceanic basalts: Review of the experimental constraints. Lithos 160-161, 14-
630 36.

631 Larsen, L.M., Rex, D.C., 1992. A review of the 2500 Ma span of alkaline-ultramafic, potassic and
632 carbonatitic magmatism in West Greenland. Lithos 28, 367-402.

633 Lee, C.-T.A., Luffi, P., Chin, E.J., 2011. Building and Destroying Continental Mantle. Annual Review
634 of Earth and Planetary Sciences 39, 59-90.

635 Li, Y., Wu, Z., Huang, S., Wang, W., 2022. Pressure and concentration effects on intermineral calcium
636 isotope fractionation involving garnet. Chemical Geology 591, 120722.

637 Liu, F., Li, X., Wang, G., Liu, Y., Zhu, H., Kang, J., Huang, F., Sun, W., Xia, X., Zhang, Z., 2017.
638 Marine Carbonate Component in the Mantle Beneath the Southeastern Tibetan Plateau: Evidence From
639 Magnesium and Calcium Isotopes. Journal of Geophysical Research: Solid Earth 122, 9729-9744.

640 Lloyd, F.E., Arima, M., Edgar, A.D., 1985. Partial melting of a phlogopite-clinopyroxenite nodule from
641 south-west Uganda: an experimental study bearing on the origin of highly potassic continental rift
642 volcanics. Contributions to Mineralogy and Petrology 91, 321-329.

643 Malpas, J., Foley, S.F., F., K.A., 1986. Alkaline mafic and ultramafic lamprophyres from the Aillik Bay
644 area, Labrador. Canadian Journal of Earth Sciences 23, 1902-1918.

645 O'Reilly, S.Y., Griffin, W.L., 2010. The continental lithosphere–asthenosphere boundary: Can we
646 sample it? Lithos 120, 1-13.

647 Pearson, D.G., Canil, D., Shirey, S.B., 2003. 2.05 - Mantle Samples Included in Volcanic Rocks:
648 Xenoliths and Diamonds, in: Holland, H.D., Turekian, K.K. (Eds.), Treatise on Geochemistry.
649 Pergamon, Oxford, pp. 171-275.

650 Pintér, Z., Foley, S.F., Yaxley, G.M., Rosenthal, A., Rapp, R.P., Lanati, A.W., Rushmer, T., 2021.
651 Experimental investigation of the composition of incipient melts in upper mantle peridotites in the
652 presence of CO2 and H2O. Lithos 396-397, 106224.

653 Qi, Y., Wang, Q., Wei, G.-J., Li, J., Wyman, D.A., 2022. Magnesium and Calcium Isotopic
654 Geochemistry of Silica-Undersaturated Alkaline Basalts: Applications for Tracing Recycled Carbon.
655 Geochemistry, Geophysics, Geosystems 23, e2022GC010463.

656 Rock, N.M.S., 1986. The Nature and Origin of Ultramafic Lamprophyres: Alnöites and Allied Rocks.
657 Journal of Petrology 27, 155-196.

658 Schiller, M., Paton, C., Bizzarro, M., 2012. Calcium isotope measurement by combined HR-MC-
659 ICPMS and TIMS. Journal of Analytical Atomic Spectrometry 27, 38-49.

660 Smart, K.A., Tappe, S., Ishikawa, A., Pfänder, J.A., Stracke, A., 2019. K-rich hydrous mantle
661 lithosphere beneath the Ontong Java Plateau: Significance for the genesis of oceanic basalts and

Page 22/28
662 Archean continents. Geochimica et Cosmochimica Acta 248, 311-342.

663 Stagno, V., Frost, D.J., 2010. Carbon speciation in the asthenosphere: Experimental measurements of
664 the redox conditions at which carbonate-bearing melts coexist with graphite or diamond in peridotite
665 assemblages. Earth and Planetary Science Letters 300, 72-84.

666 Sun, C., Dasgupta, R., 2020. Thermobarometry of CO2-rich, silica-undersaturated melts constrains
667 cratonic lithosphere thinning through time in areas of kimberlitic magmatism. Earth and Planetary
668 Science Letters 550, 116549.

669 Sun, J., Zhu, X., Belshaw, N., Chen, W., Doroshkevich, A., Luo, W., Song, W., Chen, B., Cheng, Z., Li,
670 Z., 2021. Ca isotope systematics of carbonatites: Insights into carbonatite source and evolution.
671 Geochemical Perspectives Letters 17, 11-15.

672 Tappe, S., Foley, S.F., Jenner, G.A., Heaman, L.M., Kjarsgaard, B.A., Romer, R.L., Stracke, A., Joyce,
673 N., Hoefs, J., 2006. Genesis of Ultramafic Lamprophyres and Carbonatites at Aillik Bay, Labrador: a
674 Consequence of Incipient Lithospheric Thinning beneath the North Atlantic Craton. Journal of
675 Petrology 47, 1261-1315.

676 Tappe, S., Foley, S.F., Jenner, G.A., Kjarsgaard, B.A., 2005. Integrating Ultramafic Lamprophyres into
677 the IUGS Classification of Igneous Rocks: Rationale and Implications. Journal of Petrology 46, 1893-
678 1900.

679 Tappe, S., Foley, S.F., Kjarsgaard, B.A., Romer, R.L., Heaman, L.M., Stracke, A., Jenner, G.A., 2008.
680 Between carbonatite and lamproite—diamondiferous Torngat ultramafic lamprophyres formed by
681 carbonate-fluxed melting of cratonic MARID-type metasomes. Geochimica et Cosmochimica Acta 72,
682 3258-3286.

683 Tappe, S., Foley, S.F., Stracke, A., Romer, R.L., Kjarsgaard, B.A., Heaman, L.M., Joyce, N., 2007.
684 Craton reactivation on the Labrador Sea margins: 40Ar/39Ar age and Sr–Nd–Hf–Pb isotope constraints
685 from alkaline and carbonatite intrusives. Earth and Planetary Science Letters 256, 433-454.

686 Tappe, S., Massuyeau, M., Smart, K.A., Woodland, A.B., Gussone, N., Milne, S., Stracke, A., 2021.
687 Sheared Peridotite and Megacryst Formation Beneath the Kaapvaal Craton: a Snapshot of
688 Tectonomagmatic Processes across the Lithosphere–Asthenosphere Transition. Journal of Petrology 62,
689 1-39.

690 Tappe, S., Ngwenya, N.S., Stracke, A., Romer, R.L., Glodny, J., Schmitt, A.K., 2023. Plume–
691 lithosphere interactions and LIP-triggered climate crises constrained by the origin of Karoo lamproites.
692 Geochimica et Cosmochimica Acta 350, 87-105.

693 Tappe, S., Romer, R.L., Stracke, A., Steenfelt, A., Smart, K.A., Muehlenbachs, K., Torsvik, T.H., 2017.
694 Sources and mobility of carbonate melts beneath cratons, with implications for deep carbon cycling,
695 metasomatism and rift initiation. Earth and Planetary Science Letters 466, 152-167.

696 Valdes, M.C., Moreira, M., Foriel, J., Moynier, F., 2014. The nature of Earth's building blocks as
697 revealed by calcium isotopes. Earth and Planetary Science Letters 394, 135-145.

Page 23/28
698 Veter, M., Foley, S.F., Mertz-Kraus, R., Groschopf, N., 2017. Trace elements in olivine of ultramafic
699 lamprophyres controlled by phlogopite-rich mineral assemblages in the mantle source. Lithos 292-293,
700 81-95.

701 Wang, W., Qin, T., Zhou, C., Huang, S., Wu, Z., Huang, F., 2017a. Concentration effect on equilibrium
702 fractionation of Mg-Ca isotopes in carbonate minerals: Insights from first-principles calculations.
703 Geochimica et Cosmochimica Acta 208, 185-197.

704 Wang, W., Zhou, C., Qin, T., Kang, J.-T., Huang, S., Wu, Z., Huang, F., 2017b. Effect of Ca content on
705 equilibrium Ca isotope fractionation between orthopyroxene and clinopyroxene. Geochimica et
706 Cosmochimica Acta 219, 44-56.

707 Wang, Y., He, Y., Wu, H., Zhu, C., Huang, S., Huang, J., 2019. Calcium isotope fractionation during
708 crustal melting and magma differentiation: Granitoid and mineral-pair perspectives. Geochimica et
709 Cosmochimica Acta 259, 37-52.

710 Wardle, R.J., Hall, J., 2002. Proterozoic evolution of the northeastern Canadian Shield: Lithoprobe
711 Eastern Canadian Shield Onshore–Offshore Transect (ECSOOT), introduction and summary. Canadian
712 Journal of Earth Sciences 39, 563-567.

713 Williams, H.M., Bizimis, M., 2014. Iron isotope tracing of mantle heterogeneity within the source
714 regions of oceanic basalts. Earth and Planetary Science Letters 404, 396-407.

715 Windley, B.F., Garde, A.A., 2009. Arc-generated blocks with crustal sections in the North Atlantic
716 craton of West Greenland: Crustal growth in the Archean with modern analogues. Earth-Science
717 Reviews 93, 1-30.

718 Xiao, Z., Zhou, C., Kang, J., Wu, Z., Huang, F., 2022. The factors controlling equilibrium inter-mineral
719 Ca isotope fractionation: Insights from first-principles calculations. Geochimica et Cosmochimica Acta
720 333, 373-389.

721 Yang, Z.-F., Li, J., Liang, W.-F., Luo, Z.-H., 2016. On the chemical markers of pyroxenite contributions
722 in continental basalts in Eastern China: Implications for source lithology and the origin of basalts.
723 Earth-Science Reviews 157, 18-31.

724 Yuan, H., French, S., Cupillard, P., Romanowicz, B., 2014. Lithospheric expression of geological units
725 in central and eastern North America from full waveform tomography. Earth and Planetary Science
726 Letters 402, 176-186.

727 Zhang, H., Wang, Y., He, Y., Teng, F.-Z., Jacobsen, S.B., Helz, R.T., Marsh, B.D., Huang, S., 2018. No
728 Measurable Calcium Isotopic Fractionation During Crystallization of Kilauea Iki Lava Lake.
729 Geochemistry, Geophysics, Geosystems 19, 3128-3139.

730 Zhao, K., Dai, L.-Q., Fang, W., Zheng, Y.-F., Zhao, Z.-F., Zheng, F., 2022. Decoupling between Mg and
731 Ca isotopes in alkali basalts: Implications for geochemical differentiation of subduction zone fluids.
732 Chemical Geology 606, 120983.

Page 24/28
733 Zhu, H., Ionov, D.A., Du, L., Zhang, Z., Sun, W., 2021. Ca-Sr isotope and chemical evidence for
734 distinct sources of carbonatite and silicate mantle metasomatism. Geochimica et Cosmochimica Acta
735 312, 158-179.

736 Zhu, H., Liu, F., Li, X., Wang, G., Zhang, Z., Sun, W., 2018. Calcium Isotopic Compositions of Normal
737 Mid-Ocean Ridge Basalts From the Southern Juan de Fuca Ridge. Journal of Geophysical Research:
738 Solid Earth 123, 1303-1313.
739
740
741
742
743

Page 25/28
744

745 Figure captions

746 Figure 1. The locations of alkaline rocks at the flanks of the rift and seismic shear wave velocities

747 beneath the North Atlantic craton and the Labrador Sea at the depth of 150 km. Seismic

748 tomography is from Yuan et al. (2014). The lithosphere beneath Greenland and Labrador shows

749 high velocities similar to the Superior craton, indicating its cratonic nature. The mantle under the

750 Labrador Sea shows low velocity, indicating rifted lithosphere with a higher-velocity ledge in the

751 Davis Strait between Baffin Bay and the Labrador Sea. The locations of the alkaline rocks (UMLs

752 = Ultramafic Lamprophyres and Kim. and carb. = Kimberlites and carbonatites) are from Tappe et

753 al. (2007) and are distributed along the relatively low-velocity zone at the margins of Greenland

754 and Labrador.

755

756 Figure 2. δ44/40Ca values of the alkaline rocks from Aillik Bay compared to average MORBs

757 (Chen et al., 2020a; Eriksen and Jacobsen, 2022; Zhu et al., 2018), kimberlites (Antonelli et al.,

758 2023; Tappe et al., 2021), carbonatites (Banerjee et al., 2021; Sun et al., 2021), continental silica-

759 saturated basalts (Liu et al., 2017) and silica-unsaturated basalts (Qi et al., 2022; Zhao et al., 2022),

760 and oceanic island basalts OIBs in publications since 2015 (Eriksen and Jacobsen, 2022; He et al.,

761 2017; Jacobson et al., 2015; Zhang et al., 2018).

762

763 Figure 3. Bulk rock MgO content versus CaO/Al2O3 ratio (a) and δ44/40Ca value versus

764 CaO/Al2O3 ratio (b) for Mesoproterozoic lamproites and Cretaceous nephelinites from Aillik Bay.

765 Modelled crystallization of an experimental lamproite melt (blue line) from Foley et al. (2022) and

766 an experimental nephelinite melt (magenta line) from Dasgupta et al. (2007) using MELTS

767 (Gualda and Ghiorso, 2015) are shown for comparison. The phases during crystallization

768 modelling (L-liquid, Ol-olivine, Cpx-clinopyroxene) are shown along the lines.

769

770 Figure 4. (a-b) Bulk rock MgO content versus FC3MS (FeO/CaO-3*MgO/SiO2) and bulk rock

771 K2O versus CaO content for Mesoproterozoic lamproites, Neoproterozoic ultramafic

772 lamprophyres, and Cretaceous nephelinites from Aillik Bay. Data for experimental pyroxenite

773 melt and peridotite melt were compiled by Yang et al. (2016). In (a), the dotted line indicates the

Page 26/28
774 upper boundary for peridotite melts (Yang et al., 2016), discriminating melts from pyroxenite and

775 peridotite. Experimental melts of MARID (Foley et al., 2022) and phlogopite clinopyroxenites

776 (Funk and Luth, 2013; Lloyd et al., 1985) are also shown for comparison.

777

778 Figure 5. (a) Calcium isotope β-factors for clinopyroxene (Ca/(Ca+Mg) = 7/16), carbonate

779 minerals (calcite, dolomite, and magnesites with Ca/(Ca+Mg) of 1/16 and 1/36), apatite, and K-

780 richterite from first-principles calculations (Wang et al., 2017a; Wang et al., 2017b; Xiao et al.,

781 2022). (b) Modelled δ44/40Ca of lamproitic melts and residues with increasing degree of partial

782 melting of MARID at 6.6 GPa using the melting reaction from Foley et al. (2022). The initial

783 δ44/40Ca value is assumed to be the same as MORB (Chen et al., 2020a; Eriksen and Jacobsen,

784 2022). (c) Modelled δ44/40Ca of MARID (with 30 wt/% K-richterite and 30 wt/% K-richterite + 5

785 wt.% clinopyroxene, respectively) crystallized from silicate melt/fluid using fractionation factors

786 in (a). The cratonic geotherms are summarized in Lee et al. (2011). (d) Modelled δ44/40Ca of

787 carbonatite melts and residues with an increasing degree of partial melting of magnesite lherzolite

788 at 6.6 GPa using the melting reaction from (Dasgupta and Hirschmann, 2007). The δ44/40Ca in the

789 starting carbonated peridotite is the average value of carbonated peridotite xenoliths from the

790 Siberian craton (Ionov et al., 2019; Zhu et al., 2021). Details of all modelling calculations are in

791 Supplementary information: Section 3.1 and Table S2.

792

793 Figure 6. Plots of SiO2 versus CaO (a) and Al2O3 (b) contents and plots of δ44/40Ca versus SiO2 (c)

794 and Al2O3 contents (d) for the Late Neoproterozoic ultramafic lamprophyres. Magma evolution

795 trends (arrows) from Tappe et al. (2006). Igneous carbonatites are shown for comparison

796 (Banerjee et al., 2021; Sun et al., 2021).

797

798 Figure 7. (a) Comparison of δ44/40Ca values of aillikites and mela-aillikites with carbonated

799 peridotites (Table S3, Ionov et al., 2019; Zhu et al., 2021). (b-c) δ44/40Ca values versus 87Sr/86Sr

800 and δ13C of carbonate-rich igneous rocks including aillikites, mela-aillikites, and carbonatites.

801 Previously reported δ44/40Ca values of global carbonatites are from Sun et al. (2021) and Banerjee

802 et al. (2021). Sun et al. (2021) defined primary carbonatites based on their C-O isotopes (-8‰ <

803 δ13C < -4‰ and 4‰ < δ18O < 10‰). δ44/40Ca values of kimberlites are also shown for comparison

Page 27/28
804 (Antonelli et al., 2023).

805

806 Figure 8. Bulk rock δ44/40Ca versus bulk rock SiO2 (a), CaO/Al2O3 (b), La/Lu ratio (c), and

807 Na2O+K2O content (d) for the alkaline rocks in this study, OIBs (Eriksen and Jacobsen, 2022),

808 continental silica-saturated (CSS basalts) (Liu et al., 2017) and silica-undersaturated (CSU basalts)

809 alkaline basalts (Qi et al., 2022; Zhao et al., 2022), and kimberlites (Antonelli et al., 2023).

810 Average MORBs are shown for comparison (Chen et al., 2020a; Eriksen and Jacobsen, 2022; Zhu

811 et al., 2018).

812

813 Figure 9. Evolution of calcium isotope compositions of the metasomatic mantle lithosphere at the

814 base of the North Atlantic craton over geological history (a-d). Calcium isotope compositions of

815 the refractory cratonic lithosphere mantle (Chen et al., 2019; Kang et al., 2017), upper mantle

816 (Chen et al., 2019; Kang et al., 2017), and average MORB (Chen et al., 2020a; Eriksen and

817 Jacobsen, 2022; Zhu et al., 2018) are shown for comparison. (b-c) Schematic illustration of

818 progressive rifting of the North Atlantic craton by episodic metasomatism and asthenospheric

819 convection (modified after Tappe et al. (2007)). Mesoproterozoic lamproites were formed by the

820 melting of phlogopite pyroxenite and peridotite at the base of the lithosphere in reduced conditions

821 (b). Late Neoproterozoic ultramafic lamprophyres and carbonatites originate from the melting of

822 phlogopite-carbonate veins derived from the solidification of carbonate-rich melts from the

823 asthenosphere (c). The aillikitic magmas from this episode played an important role in the

824 alteration and weakening of the shallow lithosphere mantle. The Cretaceous nephelinites were

825 produced by melting of the lithosphere mantle (wehrlite) modified by aillikitic melt (d).

Page 28/28
Highlights (for review)

Highlights:
1. Ca isotopes do not fractionate during magmatic differentiation of carbonate-rich magmas.
2. Carbonated silicate melts have δ44/40Ca lower than MORB.
3. K-richterite and carbonate in mantle sources produce low δ44/40Ca values of magmas.
4. Ca isotopes are a robust tracer of volatile-induced mantle lithological variation.
Manuscript Click here to view linked References

1
2 Calcium isotopes track volatile components in the mantle
3
4
5 sources of alkaline rocks and associated carbonatites
6
7
8 Chunfei Chena,b*, Stephen F. Foleya, Sebastian Tappec, Huange Renb, Lanping Fengb,
9
10 Yongsheng Liub*
11
12 a
School of Natural Sciences, Macquarie University, North Ryde, New South Wales 2109, Australia
13
14 b
State Key Laboratory of Geological Processes and Mineral Resources, School of Earth Sciences,
15
16 China University of Geosciences, Wuhan 430074, China
17
18 c
Department of Geosciences, UiT - The Arctic University of Norway, N-9037 Tromsø, Norway
19
20
21
22 *Corresponding authors
23
24 Chunfei Chen (chfchen2016@hotmail.com) and Yongsheng Liu (yshliu@hotmail.com)
25
26
27
28 Abstract
29
30
31 The volatile components CO2 and H2O induce mantle melting and thus exert major controls on
32
33 mantle heterogeneity. Primitive intraplate alkaline magmatic rocks are the closest analogues for
34
35 incipient mantle melts and provide the most direct method to assess such mantle heterogeneity.
36
37
Given the considerable Ca isotope differences among carbonate, clinopyroxene, garnet, and
38
39
40 orthopyroxene in the mantle (up to 1‰ for δ44/40Ca), δ44/40Ca of alkaline rocks is a promising
41
42 tracer of lithological heterogeneity. We present stable Ca isotope data for ca. 1.4 Ga lamproites,
43
44 590-555 Ma ultramafic lamprophyres and carbonatites, and 142 Ma nephelinites from Aillik Bay
45
46 in Labrador, eastern Canada. These primitive alkaline rock suites are the products of three stages
47
48 of magmatism that accompanied lithospheric thinning and rifting of the North Atlantic craton. The
49
50 three discrete magmatic events formed by melting of different lithologies in a metasomatized
51
52 lithospheric mantle column at various depths: (1) MARID-like components (mica-amphibole-
53
54 rutile-ilmenite-diopside) in the source of the lamproites; (2) phlogopite-carbonate veins were an
55
56 additional source component for ultramafic lamprophyres during the second event; and (3)
57
58 wehrlites at shallower depths were an important source component for nephelinites during the
59
60 final event.
61
62
63 Page 1/28
64
65
 The Mesoproterozoic lamproites show lower δ44/40Ca values (0.58 to 0.66‰) than MORBs
1
2 (0.84 ± 0.03‰, 2se). This cannot be explained by fractional crystallization or melting of the
3
4
clinopyroxene-dominated source but can be attributed to a source enriched in the alkali amphibole
5
6
7 K-richterite, which has characteristically low δ44/40Ca. The δ44/40Ca values of the ultramafic
8
9 lamprophyre suite during the second rifting stage are remarkably uniform, with overlapping ranges
10
11 for primary carbonated silicate melts (aillikite: 0.67 to 0.75‰), conjugate carbonatitic liquids
12
13 (0.71 to 0.82‰) and silicate-dominated damtjernite liquid (primary damtjernite: 0.68 to 0.72‰).
14
15 This suggests negligible Ca isotope fractionation during liquid immiscibility of carbonate-bearing
16
17 magmas. Combined with previously reported δ44/40Ca values for carbonatites and kimberlites, our
18
19 data suggest that carbonated silicate melts in Earth’s mantle have δ44/40Ca compositions resolvably
20
21 lower than those for MORB (0.74 ± 0.02‰ versus 0.84 ± 0.03‰, 2se). The δ44/40Ca values of the
22
23 Cretaceous nephelinites (0.72 to 0.78‰) are homogenous and similar to those of the 590-555 Ma
24
25 ultramafic lamprophyres, suggesting that the wehrlitic source component for the nephelinites
26
27 formed by mantle metasomatism during interaction with rising aillikite magmas during the second
28
29 rifting stage. Our results highlight that both K-richterite and carbonate components in mantle
30
31 sources can result in the systematically low δ44/40Ca values of alkaline magmas, which may
32
33 explain previously reported low δ44/40Ca values of continental alkaline rocks and some OIBs. Our
34
35
study indicates that Ca isotopes are a robust tracer of lithological variation caused by volatiles in
36
37
38 the Earth’s upper mantle.
39
40
41
42
43 1. Introduction
44
45
Lithological and geochemical heterogeneity in Earth’s mantle archives mantle evolution,
46
47
48 including mantle melting and recycling of crustal materials caused by plate tectonics. Besides
49
50 pressure and temperature, CO2 and H2O contents are the two most important controls on mantle
51
52 melting because they cause a significant reduction in the solidus of mantle rocks and a large
53
54 expansion in the compositional range of melts (Dasgupta, 2018; Foley et al., 2009; Stagno and
55
56 Frost, 2010), especially at depths greater than 140 km, where melting occurs only when promoted
57
58 by volatiles (Dasgupta and Hirschmann, 2006; Foley and Pintér, 2018). The volatile-bearing melts
59
60 produced solidify elsewhere to cause mineralogical and geochemical variations in the Earth’s
61
62
63 Page 2/28
64
65
 mantle (Aiuppa et al., 2021; Gaillard et al., 2008). It is increasingly recognized that volatile-rich
1
2 dykes/veins or domains (carbonate, phlogopite, amphibole, and/or fluid) co-exist in the mantle
3
4
together with the dominant lithology peridotite (Dawson and Smith, 1977; Foley, 1992; Grégoire
5
6
7 et al., 2002; Smart et al., 2019; Tappe et al., 2008). Geochemically, volatile-bearing mantle is
8
9 enriched in fusible components (e.g., K, Na, Ca). Alkaline igneous rocks are generally formed by
10
11 volatile-triggered mantle melting and hold key information about the role that volatile evolution
12
13 plays in governing the spatial and temporal variation in mantle rocks, as revealed by major and
14
15 trace elements, radiogenic isotopes (e.g., Sr, Nd, Hf and Pb isotopes), and olivine geochemistry
16
17 (Foley et al., 2013; Tappe et al., 2006; Tappe et al., 2007).
18
19 Stable calcium isotopes (δ44/40Ca = ((44Ca/40Ca)SAMPLE/(44Ca/40Ca)SRM915a - 1) × 1000) can be
20
21 used to trace large-scale geological processes, opening a new window in our understanding of
22
23 alkaline magma origins, including magma processes and the lithology of mantle sources
24
25 (Antonelli et al., 2023; Eriksen and Jacobsen, 2022; Wang et al., 2019). Considerable equilibrium
26
27 Ca isotope fractionation (up to 1‰) has been observed between the mantle minerals carbonate,
28
29 clinopyroxene (Cpx), K-richterite, garnet (Grt), and orthopyroxene (Opx) (Antonelli et al., 2019;
30
31 Chen et al., 2019; Chen et al., 2020b; Xiao et al., 2022): carbonate and hydrous K-richterite have
32
33 lower δ44/40Ca values than Cpx. Such inter-mineral Ca isotope differences may result in: (1) large
34
35
Ca isotope fractionation between melts and residual solids during partial melting (Antonelli et al.,
36
37
38 2021; Eriksen and Jacobsen, 2022; Wang et al., 2019); and (2) significant Ca isotope differences
39
40 between different mantle lithologies, as known for example from the high Ca isotope values for
41
42 dunite and orthopyroxenite (1.11-1.81‰ and 1.13‰ δ44/40Ca, respectively) (Chen et al., 2019) and
43
44 low δ44/40Ca values for carbonated peridotites (0.56-0.95‰, Ionov et al., 2019; Zhu et al., 2021).
45
46 Therefore, Ca isotopes are potentially a powerful tool for tracing lithological variations caused by
47
48 volatiles in the mantle sources of alkaline magmas.
49
50 The calcium isotope fractionation factor between Cpx and silicate melt is close to 1 (Chen et
51
52 al., 2019; Zhang et al., 2018), so that δ44/40Ca values of mid-ocean ridge basalts (MORBs: 0.84 ±
53
54 0.03‰, 2se) are similar to those of Cpx in spinel peridotites. Recently, Eriksen and Jacobsen
55
56 (2022) emphasized that melting of a garnet-rich source may result in the light Ca isotope
57
58 compositions shown by some oceanic island basalts (OIBs) due to the heavier Ca isotope
59
60 compositions of Grt compared with Cpx (Chen et al., 2020b; Li et al., 2022; Wang et al., 2019).
61
62
63 Page 3/28
64
65
 However, possible Ca isotope fractionation in alkaline magmas caused by volatile-bearing
1
2 minerals or melts in their mantle sources is largely unexplored. Calcium is abundant in several
3
4
volatile-bearing minerals (e.g., carbonate and richteritic amphiboles) and mantle-derived melts
5
6
7 (e.g., carbonatites), and Ca isotope differences between Cpx and volatile-bearing minerals are
8
9 significant (Wang et al., 2017a; Xiao et al., 2022). Knowledge of Ca isotope fractionation related
10
11 to the action of volatiles in the mantle is therefore important for understanding the large observed
12
13 Ca isotope variations and thus the origin of volatile-rich igneous rocks such as silica-
14
15 undersaturated alkaline rocks, kimberlites, and carbonatites (Amsellem et al., 2020; Antonelli et
16
17 al., 2023; Banerjee et al., 2021; Sun et al., 2021).
18
19 The North Atlantic craton (NAC) was split by continental rifting, ultimately marked by the
20
21 opening of the Labrador Sea and Davis Strait with short periods of basaltic oceanic crust
22
23 formation during the early Cenozoic (Fig. 1). Before this, a sequence of ca. 1.4 Ga lamproites,
24
25 590-555 Ma ultramafic lamprophyres, and ca. 142 Ma nephelinites at Aillik Bay on the western
26
27 Labrador Sea margin record a long history of craton breakup during which lithospheric thinning
28
29 occurred in three main stages (Tappe et al., 2006; Tappe et al., 2007). The passive continental
30
31 margin of Labrador was unaffected by subduction (Keen et al., 2012), and the rifting processes at
32
33 each of the three stages resulted largely from intraplate continental stretching. The trace elements
34
35
and radiogenic isotope compositions of these rift-related alkaline rocks reveal the important role of
36
37
38 volatile-rich metasomatized lithospheric mantle sources without notable contributions from
39
40 subducted crustal materials (Tappe et al., 2006; Tappe et al., 2007). The protracted history of this
41
42 alkaline rock occurrence makes it an ideal target to explore Ca isotope fractionation induced by
43
44 volatiles in the upper mantle. We have undertaken a detailed Ca isotope study on the diverse
45
46 magmatic products of the three rifting stages and our results show how the lithological variations
47
48 caused by volatiles in Earth’s upper mantle may fractionate the Ca isotopes of alkaline and
49
50 carbonate-rich magmatic rocks.
51
52
53 2. Geological setting and samples
54
55
56 The NAC was split during the Mesozoic-Cenozoic into two continental blocks – the Nain
57
58 Province of Labrador and the Archaean terranes of West Greenland, separated by the Labrador Sea
59
60 (Fig. 1). The Archaean blocks on either side of the Labrador Sea preserve tonalitic crust as old as
61
62
63 Page 4/28
64
65
 3.8 Ga and thick cratonic mantle roots persist to the present day (Windley and Garde, 2009) (Fig.
1
2 1). The Archaean cores were augmented by the accretion of micro-continents assembled during the
3
4
Paleoproterozoic (1.9-1.7 Ga) marking the final phase of cratonization through a sequence of
5
6
7 subduction and collision events (Wardle and Hall, 2002). The NAC has experienced numerous
8
9 mantle-derived alkaline igneous events since the Proterozoic (Larsen and Rex, 1992), which
10
11 eventually accompanied rifting that proceeded to the production of oceanic crust in the Labrador
12
13 Sea for a short time interval during the early Cenozoic. The uplifted basement blocks of the NAC
14
15 record one of the longest craton splitting histories known: the first stage corresponds to
16
17 impregnation of the mantle lithosphere with K-rich silicate melts >2000 Myr ago, which was
18
19 reactivated and melted at ca. 1370 Ma to produce olivine lamproite magmas. The second stage is
20
21 represented by widespread emplacement of ultramafic lamprophyres and carbonatites in the late
22
23 Neoproterozoic (Fig. 1), and the third episode led to successful rifting and the production of new
24
25 oceanic crust, preceded by the eruption of kimberlites, ultramafic lamprophyres and carbonatites
26
27 in western Greenland between ca. 200-150 Ma (Tappe et al., 2017), and eventually nephelinite and
28
29 melilitite magmas along the rift flanks at ca. 150-100 Ma (Tappe et al., 2006; Tappe et al., 2007).
30
31 The samples selected for Ca isotopic analysis are from the Aillik Bay locality (Fig. 1), which
32
33 is the only known area where all three rifting stages are recorded by the occurrence of alkaline
34
35
magmatic rocks. The samples include four 1.37 Ga old lamproites, 20 ultramafic lamprophyres
36
37
38 and carbonatites emplaced between 590 and 555 Ma, and five nephelinites and melilitites dated at
39
40 ca. 142 Ma. The ultramafic lamprophyres are primitive SiO2-poor potassic igneous rocks
41
42 comprising aillikites (for which Aillik Bay is the type locality), mela-aillikites, and damtjernites
43
44 following the classification of Tappe et al. (2005), in which aillikites are the archetypal carbonate-
45
46 rich member of the ultramafic lamprophyre group (Malpas et al., 1986; Rock, 1986). Detailed
47
48 descriptions and age determinations of the rocks analyzed in this study, along with their major and
49
50 trace element contents and Sr-Nd-Hf-Pb isotopic compositions, can be found in Tappe et al. (2006)
51
52 and Tappe et al. (2007).
53
54
55 3. Analytical methods
56
57
58 Chemical purification and measurement of Ca isotope ratios were performed at the State Key
59
60 Laboratory of Geological Processes and Mineral Resources, China University of Geosciences,
61
62
63 Page 5/28
64
65
 Wuhan, China, with methods following Feng et al. (2018). Briefly, Ca was separated from the
1
2 sample matrix on micro-columns filled with Ca-selective DGA resin. High recovery (> 99%),
3
4
efficient separation of Ca, and a low total procedural blank of <10 ng were achieved. Following
5
6
7 purification, calcium isotopic compositions were analyzed on a Neptune Plus (Thermo Scientific,
8
9 Bremen, Germany) MC-ICP-MS collecting 42Ca+, 43Ca+, and 44Ca+ isotopes using Faraday cups at
10
11 L3, L2, and central positions, respectively. The signal at m/z 43.5 was also collected at L1 to
12
13 monitor the doubly charged interference of Sr on Ca isotopes and was found to be negligible for
14
15 purified samples. Isotope measurements were performed using standard-sample bracketing to
16
17 correct for instrumental drift and results are defined using δ-notation: δn/42Ca =
18
19 [(nCa/42Ca)sample/(nCa/42Ca)SRM915a-1]×1000 where n=44 or 43. All measured Ca isotope values
20
21 (δ44/42Ca) were converted to δ44/40Ca using a factor of 2.048 calculated by the mass-dependent
22
23 fractionation law (Heuser et al., 2016). The long-term external 2sd reproducibility of 0.08‰ for
24
25 δ44/40Ca in this study is assessed based on replicate measurements of NIST SRM 915a and Alfa Ca
26
27 over the course of half a year (Fig. S1 in Supplementary Information: Section 1).
28
29 Three reference materials (COQ-1, BHVO-2, and BCR-2) and three duplicate samples were
30
31 processed as unknowns to assess accuracy and reproducibility. We obtained δ44/40Ca of 0.74 ±
32
33 0.06‰ (2sd, n=4), 0.77 ± 0.07‰ (2sd, n=6), and 0.83 ± 0.11‰ (2sd, n=3) for COQ-1, BHVO-2,
34
35
and BCR-2, respectively (Fig. S2 and Table 1), consistent with previous studies within the
36
37
38 analytical uncertainty (Amini et al., 2009; Amsellem et al., 2017; Feng et al., 2017; He et al., 2017;
39
40 Schiller et al., 2012; Valdes et al., 2014). Three duplicate samples show good reproducibility
41
42 within analytical uncertainty (Table 1). Furthermore, Antonelli et al. (2023) showed the generally
43
44 good agreement on analytical results for the same kimberlite powders between our laboratory and
45
46 the laboratory at Chengdu University of Technology using the double-spike thermal ionization
47
48 mass-spectrometry. Measured δ44/42Ca and δ43/42Ca of the alkaline rocks and reference materials
49
50 plot on a line of theoretical kinetic fractionation with a slope of 0.506 within uncertainty (Heuser
51
52 et al., 2016) (Fig. S3), reflecting mass-independent fractionation without analytical artifacts from
53
54 mass spectrometric interferences.
55
56
57 4. Results
58
59
60 Calcium isotopic compositions of the alkaline rocks and carbonatites from Aillik Bay and
61
62
63 Page 6/28
64
65
 reference materials are reported in Table 1. Due to the large interference of 40Ar+ from the argon
1
2 carrier gas on 40
Ca in analytes within the ICP-MS instrument, the 44
Ca/40Ca ratio could not be
3
4
determined directly in this study. All δ44/40Ca values were obtained by conversion of measured
5
6
7 δ44/42Ca applying the mass-dependent fractionation law. Thus, δ44/40Ca variations in our samples
8 40
9 reflect mass-dependent fractionation without the effect of Ca ingrowth from the radioactive
10
11 decay of 40K. We also calculated δ44/40Ca values accounting for 40Ca addition from the radioactive
12
40
13 decay of K (δ44/40CaRadiogenic in Table 1; for calculation see Fu et al. (2022)); however, in this
14
15 study we only use the uncorrected δ44/40Ca values for our samples to discuss the mass-dependent
16
17 fractionation without the effect of radiogenic excess 40Ca.
18
19 The Mesoproterozoic olivine lamproites show a small variation of δ44/40Ca values from 0.58
20
21 to 0.66‰, lower than MORB values (0.84‰, Chen et al., 2020a; Eriksen and Jacobsen, 2022; Zhu
22
23 et al., 2018) (Fig. 2). The δ44/40Ca values of the Late Neoproterozoic ultramafic lamprophyres
24
25 range from 0.56 to 0.75‰ (Fig. 2). Among them, aillikites and mela-aillikites have homogeneous
26
27 Ca isotopic compositions of 0.67‰ to 0.75‰, also systematically lower than MORB values but
28
29 higher than the olivine lamproites. The δ44/40Ca values of the Neoproterozoic carbonatites are also
30
31 homogenous (0.71 to 0.82‰, except for one carbonatite at 0.90‰), consistent with those of
32
33 aillikites and mela-aillikites. The δ44/40Ca values of the damtjernites show a larger range from 0.56
34
35
to 0.72‰, with five relatively primitive damtjernites at 0.68-0.72‰ and four more evolved
36
37
38 damtjernites at 0.56-0.62‰ (Fig. 2). The δ44/40Ca values of the Cretaceous nephelinites vary from
39
40 0.72 to 0.78‰ (Fig. 2), similar to the aillikites and mela-aillikites.
41
42
43 5. Discussion
44
45
The calcium isotopic compositions of magmatic rocks may be significantly modified by
46
47
48 precipitation of Ca-rich secondary minerals such as hydrothermal carbonates. The lamproite,
49
50 ultramafic lamprophyre, carbonatite, and nephelinite samples studied here are very fresh with only
51
52 little or no signs of hydrothermal overprinting: no secondary alteration was observed during
53
54 petrographic analysis (Tappe et al., 2006; Tappe et al., 2007). For magmatic silicate rocks, the lack
55
56 of correlation between δ44/40Ca values and CO2 contents in the lamproites and nephelinites
57
58 indicates a negligible effect of alteration on their Ca isotopic compositions (Fig. S4a). The
59
60 magmatic carbonate-bearing aillikites and damtjernites have overlapping mantle-like δ13C and
61
62
63 Page 7/28
64
65
 δ18O values (Tappe et al., 2006), in contrast to strongly fractionated carbonatites (Fig. S4b).
1
2 Importantly, the lamproite, aillikite, and nephelinite suites each have homogenous Ca isotope
3
4
compositions, which suggests that their δ44/40Ca values are unaffected by hydrothermal processes
5
6
7 (Fig. 2).
8
9 5.1. Mantle source lithologies and melting conditions at Aillik Bay
10
11
12 The fractional crystallization history of the alkaline rocks needs to be assessed before
13
14 decoding their mantle source lithologies using Ca isotope geochemistry alongside standard
15
16 petrological tools. The Mesoproterozoic lamproites contain olivine phenocrysts with a small
17
18 amount of clinopyroxene, phlogopite and apatite, suggesting that they mainly experienced
19
20 fractional crystallization of olivine (Tappe et al., 2007). Uniform CaO/Al2O3 ratios with variation
21
22 in MgO contents in the lamproites also indicates fractional crystallization of olivine without
23
24 clinopyroxene (Fig. 3a). This is supported by fractional crystallization modelling of an
25
26 experimental lamproite melt from Foley et al. (2022) using MELTS (Gualda and Ghiorso, 2015),
27
28 which shows that fractional crystallization of lamproite melt is dominated by olivine until <3.5
29
30 wt.% MgO is reached (Fig. 3a; for modelling parameters see Table S1). The high MgO contents of
31
32 the olivine lamproites (mostly >8.1 wt.%) indicate that they represent relatively primitive magmas
33
34 (Tappe et al., 2007). The Mesozoic nephelinites contain olivine and clinopyroxene phenocrysts.
35
36 The positive correlation between the CaO/Al2O3 ratio and MgO content for the nephelinites
37
38 indicates that they underwent fractional crystallization of both olivine and clinopyroxene,
39
40
consistent with fractional crystallization modelling of an experimental nephelinite melt from
41
42
43 Dasgupta et al. (2007) using MELTS (Gualda and Ghiorso, 2015) (Fig. 3a and Table S1).
44
45 The major element compositions of mantle-derived magmas can be used to constrain source
46
47 lithologies. Low CaO contents and high FC3MS values (FeO/CaO-3*MgO/SiO2) are typical
48
49 features of basaltic melts derived from pyroxenites (Lambart et al., 2013; Yang et al., 2016). The
50
51 Mesoproterozoic lamproites are SiO2-saturated (44.6-47.0 wt.%) and have high FC3MS values but
52
53 low CaO contents (6.3-7.8 wt.%), different from any experimental melts of peridotite but similar
54
55 to melts of pyroxenite or MARID metasomes (mica-amphibole-rutile-ilmenite-diopside) (Fig. 4).
56
57 Although fractional crystallization of clinopyroxene could produce high FC3MS values and low
58
59 CaO contents, these lamproites did not experience clinopyroxene fractionation. The high K2O
60
61
62
63 Page 8/28
64
65
 contents (5.1-7.7 wt.%) of the lamproites require a source containing K-rich minerals (e.g.,
1
2 phlogopite or potassic richterite). Their EM1-like Sr-Nd-Hf-Pb isotope compositions also
3
4
fingerprint long-term enriched cratonic mantle components, which can be explained by K-rich
5
6
7 ultramafic veins in peridotite (Tappe et al., 2007). Furthermore, the Mesoproterozoic lamproites
8
9 resemble the experimental melts of MARID assemblages (Foley et al., 2022) and are different
10
11 from experimental melts of phlogopite clinopyroxenite (Funk and Luth, 2013; Lloyd et al., 1985),
12
13 supporting a MARID-style metasomatized mantle source (Fig. 4b).
14
15 Among the Neoproterozoic ultramafic lamprophyres, primitive aillikites have low SiO2 and
16
17 high CaO and MgO contents, similar to experimental melts of carbonated peridotite at ≥3 GPa
18
19 (Fig. 4a). Their olivine phenocryst geochemistry (Veter et al., 2017) and bulk chemical and
20
21 isotopic compositions (Tappe et al., 2006; Tappe et al., 2007) are consistent with melting of a
22
23 carbonated mantle source. The Mesozoic nephelinite suite at Aillik Bay comprises dykes and sills
24
25 of nephelinite and basanite, plus rare melilitite (Tappe et al., 2007). The high MgO (6.9–11.3
26
27 wt.%), Ni (76–198 ppm), and Cr contents (123–485 ppm) of this alkaline rock suite indicate
28
29 derivation from primitive mantle-sourced magmas, and their high CaO contents (10.3-14.6 wt.%)
30
31 and relatively low FC3MS values are also consistent with experimental melts of peridotite (Fig.
32
33 4a). Their low SiO2 (34.7–44.3 wt.%) contents resemble experimental melts of peridotite in the
34
35
presence of CO2 at 2-3 GPa (Green and Falloon, 1998), indicating the involvement of wehrlite in
36
37
38 their mantle source (Tappe et al., 2007).
39
40 Temperatures and pressures of melt formation are estimated using the method of Sun and
41
42 Dasgupta (2020), which was calibrated for CO2-rich, silica-undersaturated magmatic systems.
43
44 This thermobarometer treats the Mg-number of olivine in the mantle source and the H2O content
45
46 of primary magmas as key parameters (Supplementary Information: Section 2). The results in Fig.
47
48 S5 and Table 1 indicate melting pressures of approximately 4-6 GPa for both lamproites and
49
50 aillikites, and 2-3 GPa for the nephelinites, broadly consistent with the petrogenetic model for the
51
52 evolution of alkaline magmatism at Aillik Bay (Tappe et al., 2007). The temperature estimates are
53
54 approximately 1370-1500 °C for the lamproites and aillikites, and 1250-1360 °C for the
55
56 nephelinites. To examine the reliability of these temperature models, we calculated the
57
58 crystallization temperature of nephelinite sample ST103 (~1200 °C) using Al-in-olivine
59
60 thermometry (Coogan et al., 2014). Given that the nephelinite magma crystallization temperature
61
62
63 Page 9/28
64
65
 is ~100 °C lower than the calculated melting temperature of the mantle source, we place
1
2 confidence in these results. The melting temperatures are higher than the warmest known cratonic
3
4
geotherms (Lee et al., 2011) and correspond broadly to the thermal regime of the underlying
5
6
7 asthenosphere, indicating that melting was initiated by asthenospheric flow.
8
9 5.2. Origin of lamproites from K-richterite-bearing metasomes with low 44/40Ca
10
11
12 Due to the extremely low CaO contents of the olivine phenocrysts (<0.012 wt.% CaO, Tappe
13
14 et al., 2007), fractional crystallization of olivine changes neither CaO/Al2O3 nor δ44/40Ca of the
15
16 magma (Fig. 3). Furthermore, the lamproites show homogenous Ca isotopic compositions
17
18 regardless of variable CaO and MgO contents, demonstrating the limited effect of fractional
19
20 crystallization on the δ44/40Ca compositions. Apatite has slightly lower 44/40Ca than Cpx (Xiao et
21
22 al., 2022), so that fractionation of minor apatite would result in slightly heavier Ca isotope
23
24 compositions of the residual melt, contrasting with the light Ca isotopes of the Mesoproterozoic
25
26 lamproites. Therefore, the δ44/40Ca values of the lamproites must reflect the Ca isotopic
27
28 compositions of their parental magmas.
29
30 The Ca isotope compositions of the Mesoproterozoic lamproites are lighter than those of
31
32 MORBs (Chen et al., 2020a; Eriksen and Jacobsen, 2022; Zhu et al., 2018). This can be attributed
33
34 to either isotope fractionation during the melting of MARID or inheritance of Ca isotope
35
36 compositions from the MARID component that already possessed low δ44/40Ca. K-richterite and
37
38 clinopyroxene are the two major Ca-phases in MARID assemblages. K-richterite is fully
39
40
consumed in melting reactions within 50 °C of the MARID solidus (Foley et al., 2022), and it has
41
42
43 much lower δ44/40Ca values than Cpx (Xiao et al., 2022). Given that the fractionation factor for Ca
44
45 isotopes between clinopyroxene and silicate melt is close to 1, K-richterite would have lighter Ca
46
47 isotopes than its equilibrium silicate melts (Fig. 5a). The δ44/40Ca values of melts produced from
48
49 MARID assemblages are higher than the residue and starting material if K-richterite is not
50
51 completely consumed, and would approach the source values if K-richterite is completely
52
53 consumed (Fig. 5b, Supplementary information: Section 3.1, and Table S2). Calcium isotope
54
55 fractionation induced by partial melting cannot explain the light Ca isotopes in the
56
57 Mesoproterozoic lamproites. We therefore suggest that the K-richterite in the MARID source
58
59 carried the low δ44/40Ca signature. The Ca isotope compositions of mantle rocks are also dependent
60
61
62
63 Page 10/28
64
65
 on their mineralogy; for example, dunite and orthopyroxenite have heavy Ca isotopes due to the
1
2 modal dominance of olivine and Opx, respectively (Chen et al., 2019). Similarly, K-richterite-
3
4
bearing MARID assemblages would have much lower Ca isotope compositions than peridotites
5
6
7 due to the low δ44/40Ca nature of K-richterite (Xiao et al., 2022).
8
9 K-richterite-bearing MARID was formed by reaction of infiltrating hydrous silicate melts
10
11 with the refractory cratonic mantle lithosphere. Considering the much lower CaO content of
12
13 refractory harzburgite (<0.5 wt.%, Pearson et al. (2003)) than that of hydrous silicate melt (for
14
15 example, the average CaO contents of worldwide continental basalts are 7.3-11.1 wt.%, Farmer
16
17 (2014)), the Ca isotopes in K-richterite-bearing MARID are not affected by the refractory cratonic
18
19 mantle, but are controlled by metasomatic melts. We modelled the Ca isotope compositions of K-
20
21 richterite-bearing MARID material formed by mantle metasomatism at the base of the cratonic
22
23 lithosphere for different geothermal gradients (Fig. 5c). These calculations assume that the
24
25 hydrous silicate melts that caused the metasomatism had a MORB-like δ44/40Ca value of 0.84‰
26
27 (Supplementary Information: Section 3.2). Results show that the K-richterite-bearing MARID has
28
29 a theoretical δ44/40Ca value of 0.65‰ at 950-1100 °C. This model temperature range provides a
30
31 realistic analogue to natural MARID assemblage formation by melt/fluid metasomatism in the
32
33 relatively cold Kaapvaal cratonic mantle (Fitzpayne et al., 2018a; Fitzpayne et al., 2018b).
34
35
Therefore, the light Ca isotope compositions of the Mesoproterozoic lamproites from the NAC
36
37
38 indicate a mantle source containing K-richterite-bearing MARID-type metasomes.
39
40
41
42 5.3. Calcium isotope systematics of ultramafic lamprophyres and associated carbonatites
43
44
5.3.1. Negligible calcium isotope fractionation during differentiation of carbonated silicate magma
45
46
The aillikites, mela-aillikites, damtjernites, and carbonatites studied here have identical Sr-
47
48
Nd-Hf-Pb isotopic compositions, indicating that they are related to a common parental magma
49
50
51 formed by melting of a carbonate-rich mantle domain (Tappe et al., 2006). The low SiO2 (22.5-
52
53 26.8 wt.%) and high MgO (17.3-20.9 wt.%), Ni (483-719 ppm), and Cr contents (577-734 ppm) of
54
55 the aillikites testify to their near-primary nature (Tappe et al., 2006; Tappe et al., 2008). In contrast,
56
57 the mela-aillikites show higher SiO2 (31.6-35.9 wt.%) contents but low CaO (8.2-11 wt.%) and
58
59 CO2 (2.0-5.4 wt.%) contents, falling on a trend that may correspond to the separation of a
60
61
62
63 Page 11/28
64
65
 carbonate-rich component from more primitive aillikite magma, with the mela-aillikites
1
2 representing the silicate-enriched portion (Fig. 6) (Tappe et al., 2006). The mela-aillikites show the
3
4
same range of δ44/40Ca values (0.67 to 0.72 ‰) as primitive aillikites (0.71-0.75‰), indicating that
5
6
7 separation of carbonate from silicate components, which may include decarbonation and CO2-
8
9 degassing, causes negligible fractionation of Ca isotopes.
10
11 Although the damtjernites do not lie on the same trend of carbonate liquid separation from
12
13 the more primitive aillikites, they are generally SiO2-richer (similar to the mela-aillikites) and
14
15 show large variations in their SiO2 (30.2-38 wt.%), CaO (10.3-21.2 wt.%), and CO2 (0.2-8.1 wt.%)
16
17 contents (Fig. 6 and Table 1). Damtjernites have been interpreted to result from liquid
18
19 immiscibility from an alkali-rich proto-aillikite magma at relatively shallow depths (Tappe et al.,
20
21 2006). Four evolved damtjernite samples with low Mg# values and high Al2O3 contents have
22
23 slightly lower δ44/40Ca values (0.56 to 0.62 ‰), whereas the more primitive damtjernites show
24
25 higher δ44/40Ca values (0.67 to 0.72 ‰). Following liquid immiscibility, these evolved damtjernites
26
27 experienced fractional crystallization of olivine, clinopyroxene and apatite, as indicated by olivine
28
29 phenocrysts and clinopyroxene-apatite prisms (Tappe et al., 2006). Clinopyroxene and apatite may
30
31 have heavier Ca isotopes than the carbonate-bearing damtjernitic magmas, just as clinopyroxene
32
33 and apatite have heavier Ca isotopes than dolomite (Fig. 5a). Therefore, fractional crystallization
34
35
of clinopyroxene and apatite may induce minor Ca isotope fractionation, explaining the slightly
36
37
38 lower δ44/40Ca values of the four evolved damtjernite samples. The δ44/40Ca values (0.67 to 0.72 ‰)
39
40 of the primitive damtjernites and carbonatites (except for ST203 with a high δ44/40Ca value of
41
42 0.90‰) are similar to those of aillikites and mela-aillikites. Together, these observations suggest
43
44 negligible Ca isotope fractionation during differentiation of carbonated silicate magmas including
45
46 liquid immiscibility and CO2-degassing.
47
48
49 5.3.2. Calcium isotope fractionation during low-degree partial melting of carbonated peridotite
50
51 The δ44/40Ca values of the most primitive aillikite samples represent those of primary
52
53 carbonated silicate melts (0.72 ± 0.05 ‰, 2sd) produced by incipient melting of carbonated
54
55 peridotites. We compiled bulk Ca isotope compositions of carbonated peridotite xenoliths (Ionov
56
57 et al., 2019; Zhu et al., 2021), which show an average δ44/40Ca value of 0.82 ± 0.05 ‰, 2sd (Fig.
58
59 7a and Table S3). The significantly different δ44/40Ca values for carbonated silicate melts and
60
61
62
63 Page 12/28
64
65
 carbonated peridotites could potentially reflect Ca isotope fractionation during low-degree partial
1
2 melting of carbonated mantle rocks. Carbonate, clinopyroxene, and garnet are the main Ca-phases
3
4
in carbonated mantle peridotite domains at the pressure and temperature conditions relevant to the
5
6
7 lowermost cratonic lithosphere and underlying convecting upper mantle. First-principles
8
9 calculations can be used to predict Ca isotope fractionation between clinopyroxene and various
10
11 types of carbonate minerals, including calcite, dolomite, and magnesite (Wang et al., 2017a; Wang
12
13 et al., 2017b). Results show that Δ44/40CaCarbonate-Cpx is -0.04 to -0.10 when the Ca/(Mg+Ca) of the
14
15 carbonate is higher than 1/6 at 1050 ℃ (Fig. 5a). Furthermore, garnet and Opx that are residual
16
17 after low-degree melting of carbonated peridotite have heavier Ca isotope compositions than Cpx
18
19 and carbonate. We quantitatively modelled the behaviour of Ca isotopes during partial melting of
20
21 magnesite lherzolite at 6.6 GPa using an incremental batch melting model (Chen et al., 2019;
22
23 Williams and Bizimis, 2014). The inter-mineral fractionation factors for carbonate, garnet, and
24
25 pyroxenes are from Chen et al. (2020b) and Wang et al. (2017a), and melting parameters are from
26
27 Dasgupta et al. (2007). Details of the modelling are presented in the Supplementary Information:
28
29 Section 3.1 and Table S2, and the results are illustrated in Fig. 5d. Our results show that isotope
30
31 fractionation of about 0.08‰ δ44/40Ca occurs between carbonated melt and its peridotitic residue,
32
33 which explains the measured differences (about 0.1‰) between carbonated peridotites and the
34
35
carbonated silicate melts at Aillik Bay (Fig. 7a).
36
37
38 5.3.3. Calcium isotope compositions of carbonate-rich mantle melts
39
40 Previously reported δ44/40Ca values for global carbonatites show a large variation (Figs. 2 and
41
42 7). The origin of this large Ca isotope variation is unclear, and competing ideas invoke Rayleigh
43
44 fractionation during late-stage precipitation of carbonates from fluid-rich carbonatite magmas
45
46
(Sun et al., 2021) and the involvement of recycled sedimentary carbonate in the mantle source
47
48
(Amsellem et al., 2020; Banerjee et al., 2021). The calcium isotope compositions of primitive
49
50
51 carbonate-rich magmas from the mantle can serve as a benchmark to evaluate directions and
52
53 degrees of isotope variation caused by later magmatic-hydrothermal processes or originally by
54
55 mantle source contamination with recycled sedimentary carbonate components. Previous studies
56
57 used the Bulk Silicate Earth (BSE) composition as the benchmark for δ44/40Ca (Amsellem et al.,
58
59 2020; Banerjee et al., 2021). However, experiments show that melts of carbonated peridotite
60
61
62
63 Page 13/28
64
65
 at >140 km depth have a carbonated silicate structure and composition (Dasgupta and Hirschmann,
1
2 2006; Pintér et al., 2021), with aillikites being the best natural analogues. Therefore, the Ca
3
4
isotope compositions of the type aillikites and associated pristine carbonatites can be used to
5
6
7 constrain the calcium isotope compositions of primitive carbonate-rich magmas in the mantle.
8
9 The average δ44/40Ca value of the carbonate-rich igneous rocks at Aillik Bay (aillikites and
10
11 carbonatites) is 0.72 ± 0.05‰ (2sd, n=11). This value is consistent with the δ44/40Ca of 0.72 ±
12
13 0.06‰ (2sd, n=18) for pristine global carbonatites with mantle-like C-O isotope compositions
14
15 (Sun et al., 2021). It is also similar to the average δ44/40Ca composition of magmatic kimberlites at
16
17 0.77 ± 0.11‰ (2sd, n=19) (Antonelli et al., 2023). Therefore, we propose that the average value of
18
19 0.74‰ (2sd = 0.10‰; 2se = 0.02‰; n = 48) represents the Ca isotope composition of carbonate-
20
21 rich magmas in the mantle.
22
23 Carbonatite sample ST203 from Aillik Bay has a δ13C composition of -2.8‰ and this
24
25 deviation from the mantle range (-4 to -6‰ δ13C) was explained by high-temperature Rayleigh
26
27 fractionation (Tappe et al., 2006). The high δ44/40Ca value of 0.90‰ for ST203 suggests that
28
29 complex carbonatite magma evolution can fractionate the Ca and C isotope compositions, as
30
31 suggested by Sun et al. (2021). This may explain the origin of high δ44/40Ca values in some
32
33 carbonatites from occurrences worldwide (much higher than 0.72 ‰, Fig. 7b-c), which must await
34
35
future detailed investigations.
36
37
38
39
40 5.4. Ca isotope constraints on the metasomatic source component in the Mesozoic
41
42 nephelinites
43
44
45 The Mesozoic nephelinites experienced fractional crystallization of olivine and
46
47 clinopyroxene (see Section 5.1). The fractionation factor for Ca isotopes between clinopyroxene
48
49 and silicate melt is close to 1, so that fractional crystallization of clinopyroxene does not
50
51 significantly affect Ca isotopes (Chen et al., 2020a; Zhang et al., 2018). The lack of correlation
52
53 between δ44/40Ca values and CaO/Al2O3 ratios also indicates that fractional crystallization of
54
55 clinopyroxene did not change the Ca isotopic compositions of the nephelinites (Fig. 3b).
56
57 Therefore, the δ44/40Ca values of the nephelinites reflect the Ca isotopic compositions of their
58
59 parental magmas.
60
61
62
63 Page 14/28
64
65
 The δ44/40Ca values of the nephelinites are similar to those of the Late Neoproterozoic
1
2 ultramafic lamprophyres and carbonatites, slightly lower than average MORB (Figs. 2 and 8). This
3
4
could be attributed to either the direct effect of melting of wehrlite or to the wehrlite source
5
6
7 already possessing low δ44/40Ca. Partial melting of shallow spinel-facies wehrlite (melting pressure
8
9 = 2-3 GPa) cannot explain the systematically low δ44/40Ca values of the nephelinites relative to
10
11 MORBs because the Ca budget is dominated by Cpx during partial melting (Chen et al., 2019;
12
13 Zhang et al., 2018). The wehrlite-rich source of the nephelinites formed by metasomatism by
14
15 carbonate-rich melts, as indicated by the low SiO2 contents and high CaO/Al2O3 ratios of the
16
17 nephelinites (Tappe et al., 2007). Trace element patterns of the nephelinites resemble those of
18
19 aillikites, and their Sr-Nd-Hf-Pb isotope compositions indicate that their source formed by
20
21 metasomatic reactions between cratonic lherzolite and aillikitic melts during the Neoproterozoic.
22
23 This is consistent with similar δ44/40Ca values for the nephelinites and the Neoproterozoic aillikites
24
25 (Tappe et al., 2007). Therefore, we suggest that the wehrlites in their mantle source already had
26
27 δ44/40Ca values of about 0.75‰ as a result of metasomatism by carbonated silicate melts derived
28
29 from Neoproterozoic aillikite magmatism.
30
31
32
33
5.5. A calcium isotope perspective on the composition of metasomatized cratonic mantle and
34
35
the origin of alkaline magmatism
36
37
38 The lithosphere-asthenosphere boundary (LAB) beneath cratons represents a transition zone
39
40
that undergoes episodic rejuvenation caused by the impregnation with asthenosphere-derived
41
42
43 volatile-rich melts (Foley, 2008; O'Reilly and Griffin, 2010). The cratonic LAB is recognized as a
44
45 key source of ultramafic alkaline and carbonatitic magmas, and as a zone that modulates the
46
47 chemical compositions of deeper-derived magmas during their ascent (Foley et al., 2019; Giuliani
48
49 et al., 2020; Tappe et al., 2023). Our results reveal that the Ca isotope compositions of
50
51 metasomatized lithosphere at the base of the NAC are different from those typical for refractory
52
53 cratonic mantle and the asthenosphere (Chen et al., 2019; Kang et al., 2017) (Fig. 9). For example,
54
55 δ44/40Ca values of about 0.60‰ characterize the MARID-type metasomes tapped during
56
57 Mesoproterozoic lamproite magmatism, and δ44/40Ca values of about 0.72-0.82‰ characterize the
58
59 carbonated peridotite components that gave rise to Neoproterozoic ultramafic lamprophyre and
60
61
62
63 Page 15/28
64
65
 carbonatite magmatism. Undoubtedly, the multiply metasomatically overprinted LAB beneath
1
2 cratons has an important effect on the Ca isotope compositions of deep-sourced magmas (Tappe et
3
4
al., 2021).
5
6
7 The δ44/40Ca values of some continental SiO2-saturated basalts range from 0.65 to 0.80‰ (Liu
8
9 et al., 2017), notably lower than the MORB average (Fig. 8d). This has been attributed to
10
11 recycling of sedimentary carbonate into their mantle sources (Liu et al., 2017). However, the SiO2-
12
13 saturated nature of these basalts at low CaO/Al2O3 is in conflict with magma origins from a
14
15 carbonated mantle source (Fig. 8). Our results for the Mesoproterozoic lamproites from the NAC
16
17 indicate a key role for K-richterite in determining the Ca isotope compositions of alkaline
18
19 magmas. We suggest that low-δ44/40Ca K-richterite in the cold metasomatized lithospheric mantle
20
21 should be considered as a possible source of the low δ44/40Ca signature of certain SiO2-saturated
22
23 basalts. Potassic richterite is a readily fusible component enriched in Ca, Na and K, and will be an
24
25 early contributor to melts generated within metasomatised lithosphere (Foley et al., 2022), but also
26
27 to asthenosphere-derived melts upon their entry into metasomatised lithosphere, particularly where
28
29 the lithosphere is very thick (Gao et al., 2023). The high Na2O+K2O contents of continental SiO2-
30
31 saturated basalts indicate that K-richterite may have been involved in producing the low δ44/40Ca
32
33 values (Fig. 8d).
34
35
Carbonated silicate melts are stable and prevalent at >140 km depths (Dasgupta and
36
37
38 Hirschmann, 2006) because deep mantle melting occurs only where promoted by carbon and
39
40 water (Dasgupta and Hirschmann, 2006; Foley and Pintér, 2018). Our δ44/40Ca value of 0.74 ±
41
42 0.02 ‰ (2se) for ultramafic lamprophyres and carbonatites from a rifted craton setting serves as a
43
44 benchmark for primary carbonate-rich melts in the mantle to evaluate the directions and degrees of
45
46 isotope variations observed in carbonate-rich magmas, and it provides a more appropriate choice
47
48 than the δ44/40Ca value for BSE (Chen et al., 2019; Kang et al., 2017), as used in previous studies
49
50 (Banerjee et al., 2021; Qi et al., 2022; Zhao et al., 2022). Furthermore, the low δ44/40Ca values of
51
52 the Mesozoic nephelinites from Aillik Bay suggest a role for carbonated silicate melts in the origin
53
54 of SiO2-undersaturated alkaline rocks. We suggest that the low δ44/40Ca values for previously
55
56 reported SiO2-undersaturated alkaline basalts (0.65 to 0.77 ‰) and some carbonatites may not
57
58 trace recycled sedimentary carbonate in their mantle source as previously proposed, but rather that
59
60 such Ca isotope signatures are a hallmark geochemical feature of asthenosphere-derived
61
62
63 Page 16/28
64
65
 carbonated melts and their metasomatic products (Fig. 8). Furthermore, SiO2-undersaturated
1
2 alkaline OIBs also show lower δ44/40Ca values than the MORB average (Eriksen and Jacobsen,
3
4
2022; He et al., 2017; Jacobson et al., 2015; Zhang et al., 2018) (Figs. 2 and 8). The influence of
5
6
7 carbonated silicate melts on OIB mantle sources leading potentially to low δ44/40Ca values should
8
9 be explored in the near future.
10
11
12
13
14 6. Conclusions
15
16
17 To explore the effects of lithological mantle heterogeneity on the Ca isotope systematics of
18
19 magmas, we present new Ca isotope data for lamproites, ultramafic lamprophyres, carbonatites,
20
21 and nephelinites from the North Atlantic craton at Aillik Bay on the margin of the Labrador Sea
22
23 rift. The Mesoproterozoic lamproites show lower δ44/40Ca values (0.58 to 0.66‰) than MORBs,
24
25 suggesting an origin by melting of a MARID-type source containing K-richterite, a mineral that
26
27 can cause the low-δ44/40Ca signature. The δ44/40Ca values of the Neoproterozoic ultramafic
28
29 lamprophyres are relatively uniform (0.67 to 0.75‰) and similar to those of the carbonatites (0.71
30
31 to 0.82‰), discounting significant Ca isotope fractionation during liquid immiscibility of a
32
33 parental carbonated silicate melt. The average δ44/40Ca value of the carbonate-bearing magmas is
34
35 slightly, but consistently, lower than the MORB average. The δ44/40Ca values of the Mesozoic
36
37 nephelinites (0.72 to 0.78‰) are similar to those of the Neoproterozoic ultramafic lamprophyres,
38
39 indicating a wehrlitic metasomatic source that was formed by interaction between cratonic mantle
40
41 peridotite and rising carbonated silicate magmas during the Neoproterozoic rifting stage. The new
42
43
data indicate that both K-richterite and carbonate components in upper mantle environments can
44
45
cause systematically low δ44/40Ca values in alkaline melts produced from metasomatized sources,
46
47
48 which may explain the frequently observed low-δ44/40Ca signature of some intraplate continental
49
50 and oceanic basalts. Our study highlights Ca isotopes as a robust tracer of lithological variations in
51
52 the mantle sources of alkaline magmas, promising to constrain the sources and mobility of volatile
53
54 components within the crust-mantle system.
55
56
57
58
59
Acknowledgments
60
61
62
63 Page 17/28
64
65
 This research is supported by Key R&D Program of China (2019YFA0708400), NSFC
1
2 (41530211) and SKL-GPMR (MSFGPMR01). SFF and CC are funded by ARC grant
3
4
FL180100134. Mapping and sampling by ST and SFF in Labrador during 2003-2004 were
5
6
7 financially supported by the German Research Foundation (DFG). We appreciate the constructive
8
9 comments from two anonymous reviewers that significantly improved the manuscript and Editor
10
11 Rosemary Hickey-Vargas for the efficient editing of the paper. We thank Dr. Kai Wang for help
12
13 with the map in Figure 1.
14
15 CRediT authorship contribution statement
16
17 Chunfei Chen: Conceptualization, Investigation, Methodology, Formal analysis, Writing –
18
19
original draft. Stephen F. Foley: Conceptualization, Funding acquisition, Writing – review &
20
21
22 editing. Sebastian Tappe: Investigation, Writing – review & editing. Huange Ren: Methodology,
23
24 Resources. Lanping Feng: Methodology, Resources. Yongsheng Liu: Conceptualization, Funding
25
26 acquisition.
27
28 Competing interests
29
30 The authors declare no competing interests.
31
32
33 Data availability
34
35 The data that support the findings of this study are available in the paper or in the supplementary
36
37 files.
38
39
40
41
42 References
43
44
45 Aiuppa, A., Casetta, F., Coltorti, M., Stagno, V., Tamburello, G., 2021. Carbon concentration increases
46 with depth of melting in Earth’s upper mantle. Nature Geoscience 14, 697-703.
47
48
Amini, M., Eisenhauer, A., Böhm, F., Holmden, C., Kreissig, K., Hauff, F., Jochum, K.P., 2009.
49
50 Calcium Isotopes (δ44/40Ca) in MPI-DING Reference Glasses, USGS Rock Powders and Various Rocks:
51 Evidence for Ca Isotope Fractionation in Terrestrial Silicates. Geostandards and Geoanalytical
52
53 Research 33, 231-247.
54
55 Amsellem, E., Moynier, F., Bertrand, H., Bouyon, A., Mata, J., Tappe, S., Day, J.M., 2020. Calcium
56
isotopic evidence for the mantle sources of carbonatites. Science advances 6, eaba3269.
57
58
59 Amsellem, E., Moynier, F., Pringle, E.A., Bouvier, A., Chen, H., Day, J.M.D., 2017. Testing the
60 chondrule-rich accretion model for planetary embryos using calcium isotopes. Earth and Planetary
61
62
63 Page 18/28
64
65
 Science Letters 469, 75-83.
1
2
Antonelli, M.A., Giuliani, A., Wang, Z., Wang, M., Zhou, L., Feng, L., Li, M., Zhang, Z., Liu, F.,
3
4 Drysdale, R.N., 2023. Subducted carbonates not required: Deep mantle melting explains stable Ca
5 isotopes in kimberlite magmas. Geochimica et Cosmochimica Acta 348, 410-427.
6
7
8 Antonelli, M.A., Kendrick, J., Yakymchuk, C., Guitreau, M., Mittal, T., Moynier, F., 2021. Calcium
9 isotope evidence for early Archaean carbonates and subduction of oceanic crust. Nature
10 Communications 12, 2534.
11
12
13 Antonelli, M.A., Schiller, M., Schauble, E.A., Mittal, T., DePaolo, D.J., Chacko, T., Grew, E.S., Tripoli,
14 B., 2019. Kinetic and equilibrium Ca isotope effects in high-T rocks and minerals. Earth and Planetary
15
16 Science Letters 517, 71-82.
17
18 Banerjee, A., Chakrabarti, R., Simonetti, A., 2021. Temporal evolution of δ 44/40Ca and 87
Sr/86Sr of
19 carbonatites: Implications for crustal recycling through time. Geochimica et Cosmochimica Acta 307,
20
21 168-191.
22
23 Chen, C., Ciazela, J., Li, W., Dai, W., Wang, Z., Foley, S.F., Li, M., Hu, Z., Liu, Y., 2020a. Calcium
24
25 isotopic compositions of oceanic crust at various spreading rates. Geochimica et Cosmochimica Acta
26 278, 272-288.
27
28 Chen, C., Dai, W., Wang, Z., Liu, Y., Li, M., Becker, H., Foley, S.F., 2019. Calcium isotope
29
30 fractionation during magmatic processes in the upper mantle. Geochimica et Cosmochimica Acta 249,
31 121-137.
32
33
34 Chen, C., Huang, J.-X., Foley, S.F., Wang, Z., Moynier, F., Liu, Y., Dai, W., Li, M., 2020b.
35 Compositional and pressure controls on calcium and magnesium isotope fractionation in magmatic
36 systems. Geochimica et Cosmochimica Acta 290, 257-270.
37
38
39 Coogan, L.A., Saunders, A.D., Wilson, R.N., 2014. Aluminum-in-olivine thermometry of primitive
40 basalts: Evidence of an anomalously hot mantle source for large igneous provinces. Chemical Geology
41
368, 1-10.
42
43
44 Dasgupta, R., 2018. Volatile-bearing partial melts beneath oceans and continents–Where, how much,
45 and of what compositions? American Journal of Science 318, 141-165.
46
47
48 Dasgupta, R., Hirschmann, M.M., 2006. Melting in the Earth's deep upper mantle caused by carbon
49 dioxide. Nature 440, 659-662.
50
51
52 Dasgupta, R., Hirschmann, M.M., 2007. Effect of variable carbonate concentration on the solidus of
53 mantle peridotite. American Mineralogist 92, 370-379.
54
55 Dasgupta, R., Hirschmann, M.M., Smith, N.D., 2007. Partial Melting Experiments of Peridotite + CO 2
56
57 at 3 GPa and Genesis of Alkalic Ocean Island Basalts. Journal of Petrology 48, 2093-2124.
58
59 Dawson, J.B., Smith, J.V., 1977. The MARID (mica-amphibole-rutile-ilmenite-diopside) suite of
60
61
62
63 Page 19/28
64
65
 xenoliths in kimberlite. Geochimica et Cosmochimica Acta 41, 309-323.
1
2
Eriksen, Z.T., Jacobsen, S.B., 2022. Calcium isotope constraints on OIB and MORB petrogenesis: The
3
4 importance of melt mixing. Earth and Planetary Science Letters 593, 117665.
5
6 Farmer, G.L., 2014. 4.3 - Continental Basaltic Rocks, in: Holland, H.D., Turekian, K.K. (Eds.), Treatise
7
8 on Geochemistry (Second Edition). Elsevier, Oxford, pp. 75-110.
9
10 Feng, L.-P., Zhou, L., Yang, L., DePaolo, D.J., Tong, S.-Y., Liu, Y.-S., Owens, T.L., Gao, S., 2017.
11
Calcium Isotopic Compositions of Sixteen USGS Reference Materials. Geostandards and
12
13 Geoanalytical Research 41, 93-106.
14
15 Feng, L., Zhou, L., Yang, L., Zhang, W., Wang, Q., Tong, S., Hu, Z., 2018. A rapid and simple single-
16
17 stage method for Ca separation from geological and biological samples for isotopic analysis by MC-
18 ICP-MS. Journal of Analytical Atomic Spectrometry 33, 413-421.
19
20
Fitzpayne, A., Giuliani, A., Hergt, J., Phillips, D., Janney, P., 2018a. New geochemical constraints on
21
22 the origins of MARID and PIC rocks: Implications for mantle metasomatism and mantle-derived
23 potassic magmatism. Lithos 318-319, 478-493.
24
25
26 Fitzpayne, A., Giuliani, A., Phillips, D., Hergt, J., Woodhead, J.D., Farquhar, J., Fiorentini, M.L.,
27 Drysdale, R.N., Wu, N., 2018b. Kimberlite-related metasomatism recorded in MARID and PIC mantle
28 xenoliths. Mineralogy and Petrology 112, 71-84.
29
30
31 Foley, S., 1992. Vein-plus-wall-rock melting mechanisms in the lithosphere and the origin of potassic
32 alkaline magmas. Lithos 28, 435-453.
33
34
35 Foley, S.F., 2008. Rejuvenation and erosion of the cratonic lithosphere. Nature Geoscience 1, 503-510.
36
37 Foley, S.F., Ezad, I.S., van der Laan, S.R., Pertermann, M., 2022. Melting of hydrous pyroxenites with
38
alkali amphiboles in the continental mantle: 1. Melting relations and major element compositions of
39
40 melts. Geoscience Frontiers 13, 101380.
41
42 Foley, S.F., Pintér, Z., 2018. Chapter 1 - Primary Melt Compositions in the Earth's Mantle, in: Kono, Y.,
43
44 Sanloup, C. (Eds.), Magmas Under Pressure. Elsevier, pp. 3-42.
45
46 Foley, S.F., Prelevic, D., Rehfeldt, T., Jacob, D.E., 2013. Minor and trace elements in olivines as probes
47
into early igneous and mantle melting processes. Earth and Planetary Science Letters 363, 181-191.
48
49
50 Foley, S.F., Yaxley, G.M., Kjarsgaard, B.A., 2019. Kimberlites from Source to Surface: Insights from
51 Experiments. Elements 15, 393-398.
52
53
54 Foley, S.F., Yaxley, G.M., Rosenthal, A., Buhre, S., Kiseeva, E.S., Rapp, R.P., Jacob, D.E., 2009. The
55 composition of near-solidus melts of peridotite in the presence of CO2 and H2O between 40 and 60 kbar.
56
Lithos 112, Supplement 1, 274-283.
57
58
59 Fu, H., Jacobsen, S.B., Larsen, B.T., Eriksen, Z.T., 2022. Ca-isotopes as a robust tracer of magmatic
60 differentiation. Earth and Planetary Science Letters 594, 117743.
61
62
63 Page 20/28
64
65
 Funk, S.P., Luth, R.W., 2013. Melting phase relations of a mica–clinopyroxenite from the Milk River
1 area, southern Alberta, Canada. Contributions to Mineralogy and Petrology 166, 393-409.
2
3
4 Gaillard, F., Malki, M., Iacono-Marziano, G., Pichavant, M., Scaillet, B.J.S., 2008. Carbonatite melts
5 and electrical conductivity in the asthenosphere. Science 322, 1363-1365.
6
7
8 Gao, M., Xu, H., Foley, S.F., Zhang, J., Wang, Y., 2023. Ultrahigh-pressure mantle metasomatism in
9 continental collision zones recorded by post-collisional mafic rocks. GSA Bulletin, In press.
10
11
Giuliani, A., Pearson, D.G., Soltys, A., Dalton, H., Phillips, D., Foley, S.F., Lim, E., Goemann, K.,
12
13 Griffin, W.L., Mitchell, R.H., 2020. Kimberlite genesis from a common carbonate-rich primary melt
14 modified by lithospheric mantle assimilation. Science Advances 6, eaaz0424.
15
16
17 Green, D.H., Falloon, T.J., 1998. Pyrolite: A Ringwood Concept and Its Current Expression, in:
18 Jackson, I. (Ed.), The Earth's mantle: composition, structure, evolution, p. 311.
19
20
Grégoire, M., Bell, D., Le Roex, A., 2002. Trace element geochemistry of phlogopite-rich mafic mantle
21
22 xenoliths: their classification and their relationship to phlogopite-bearing peridotites and kimberlites
23 revisited. Contributions to Mineralogy and Petrology 142, 603-625.
24
25
26 Gualda, G.A.R., Ghiorso, M.S., 2015. MELTS_Excel: A Microsoft Excel-based MELTS interface for
27 research and teaching of magma properties and evolution. Geochemistry, Geophysics, Geosystems 16,
28 315-324.
29
30
31 He, Y., Wang, Y., Zhu, C., Huang, S., Li, S., 2017. Mass-Independent and Mass-Dependent Ca Isotopic
32 Compositions of Thirteen Geological Reference Materials Measured by Thermal Ionisation Mass
33
34 Spectrometry. Geostandards and Geoanalytical Research 41, 283-302.
35
36 Heuser, A., Schmitt, A.-D., Gussone, N., Wombacher, F., 2016. Analytical methods: Calcium stable
37 isotope geochemistry., in: Gussone, N., Schmitt, A.-D., Heuser, A., Wombacher, F., Dietzel, M., Tipper,
38
39 E., Schiller, M., Bohm, F. (Eds.), Calcium stable isotope geochemistry. Springer.
40
41 Ionov, D.A., Qi, Y.-H., Kang, J.-T., Golovin, A.V., Oleinikov, O.B., Zheng, W., Anbar, A.D., Zhang, Z.-
42
43 F., Huang, F., 2019. Calcium isotopic signatures of carbonatite and silicate metasomatism, melt
44 percolation and crustal recycling in the lithospheric mantle. Geochimica et Cosmochimica Acta 248, 1-
45 13.
46
47
48 Jacobson, A.D., Grace Andrews, M., Lehn, G.O., Holmden, C., 2015. Silicate versus carbonate
49 weathering in Iceland: New insights from Ca isotopes. Earth and Planetary Science Letters 416, 132-
50
142.
51
52
53 Kang, J.-T., Ionov, D.A., Liu, F., Zhang, C.-L., Golovin, A.V., Qin, L.-P., Zhang, Z.-F., Huang, F., 2017.
54 Calcium isotopic fractionation in mantle peridotites by melting and metasomatism and Ca isotope
55
56 composition of the Bulk Silicate Earth. Earth and Planetary Science Letters 474, 128-137.
57
58 Keen, C.E., Dickie, K., Dehler, S.A., 2012. The volcanic margins of the northern Labrador Sea:
59
Insights to the rifting process. Tectonics 31, TC1011.
60
61
62
63 Page 21/28
64
65
 Lambart, S., Laporte, D., Schiano, P., 2013. Markers of the pyroxenite contribution in the major-
1 element compositions of oceanic basalts: Review of the experimental constraints. Lithos 160-161, 14-
2
3 36.
4
5 Larsen, L.M., Rex, D.C., 1992. A review of the 2500 Ma span of alkaline-ultramafic, potassic and
6
7 carbonatitic magmatism in West Greenland. Lithos 28, 367-402.
8
9 Lee, C.-T.A., Luffi, P., Chin, E.J., 2011. Building and Destroying Continental Mantle. Annual Review
10 of Earth and Planetary Sciences 39, 59-90.
11
12
13 Li, Y., Wu, Z., Huang, S., Wang, W., 2022. Pressure and concentration effects on intermineral calcium
14 isotope fractionation involving garnet. Chemical Geology 591, 120722.
15
16
17 Liu, F., Li, X., Wang, G., Liu, Y., Zhu, H., Kang, J., Huang, F., Sun, W., Xia, X., Zhang, Z., 2017.
18 Marine Carbonate Component in the Mantle Beneath the Southeastern Tibetan Plateau: Evidence From
19 Magnesium and Calcium Isotopes. Journal of Geophysical Research: Solid Earth 122, 9729-9744.
20
21
22 Lloyd, F.E., Arima, M., Edgar, A.D., 1985. Partial melting of a phlogopite-clinopyroxenite nodule from
23 south-west Uganda: an experimental study bearing on the origin of highly potassic continental rift
24
25 volcanics. Contributions to Mineralogy and Petrology 91, 321-329.
26
27 Malpas, J., Foley, S.F., F., K.A., 1986. Alkaline mafic and ultramafic lamprophyres from the Aillik Bay
28 area, Labrador. Canadian Journal of Earth Sciences 23, 1902-1918.
29
30
31 O'Reilly, S.Y., Griffin, W.L., 2010. The continental lithosphere–asthenosphere boundary: Can we
32 sample it? Lithos 120, 1-13.
33
34
35 Pearson, D.G., Canil, D., Shirey, S.B., 2003. 2.05 - Mantle Samples Included in Volcanic Rocks:
36 Xenoliths and Diamonds, in: Holland, H.D., Turekian, K.K. (Eds.), Treatise on Geochemistry.
37 Pergamon, Oxford, pp. 171-275.
38
39
40 Pintér, Z., Foley, S.F., Yaxley, G.M., Rosenthal, A., Rapp, R.P., Lanati, A.W., Rushmer, T., 2021.
41 Experimental investigation of the composition of incipient melts in upper mantle peridotites in the
42
43 presence of CO2 and H2O. Lithos 396-397, 106224.
44
45 Qi, Y., Wang, Q., Wei, G.-J., Li, J., Wyman, D.A., 2022. Magnesium and Calcium Isotopic
46 Geochemistry of Silica-Undersaturated Alkaline Basalts: Applications for Tracing Recycled Carbon.
47
48 Geochemistry, Geophysics, Geosystems 23, e2022GC010463.
49
50 Rock, N.M.S., 1986. The Nature and Origin of Ultramafic Lamprophyres: Alnöites and Allied Rocks.
51
52 Journal of Petrology 27, 155-196.
53
54 Schiller, M., Paton, C., Bizzarro, M., 2012. Calcium isotope measurement by combined HR-MC-
55 ICPMS and TIMS. Journal of Analytical Atomic Spectrometry 27, 38-49.
56
57
58 Smart, K.A., Tappe, S., Ishikawa, A., Pfänder, J.A., Stracke, A., 2019. K-rich hydrous mantle
59 lithosphere beneath the Ontong Java Plateau: Significance for the genesis of oceanic basalts and
60
61
62
63 Page 22/28
64
65
 Archean continents. Geochimica et Cosmochimica Acta 248, 311-342.
1
2
Stagno, V., Frost, D.J., 2010. Carbon speciation in the asthenosphere: Experimental measurements of
3
4 the redox conditions at which carbonate-bearing melts coexist with graphite or diamond in peridotite
5 assemblages. Earth and Planetary Science Letters 300, 72-84.
6
7
8 Sun, C., Dasgupta, R., 2020. Thermobarometry of CO2-rich, silica-undersaturated melts constrains
9 cratonic lithosphere thinning through time in areas of kimberlitic magmatism. Earth and Planetary
10 Science Letters 550, 116549.
11
12
13 Sun, J., Zhu, X., Belshaw, N., Chen, W., Doroshkevich, A., Luo, W., Song, W., Chen, B., Cheng, Z., Li,
14 Z., 2021. Ca isotope systematics of carbonatites: Insights into carbonatite source and evolution.
15
16 Geochemical Perspectives Letters 17, 11-15.
17
18 Tappe, S., Foley, S.F., Jenner, G.A., Heaman, L.M., Kjarsgaard, B.A., Romer, R.L., Stracke, A., Joyce,
19 N., Hoefs, J., 2006. Genesis of Ultramafic Lamprophyres and Carbonatites at Aillik Bay, Labrador: a
20
21 Consequence of Incipient Lithospheric Thinning beneath the North Atlantic Craton. Journal of
22 Petrology 47, 1261-1315.
23
24
25 Tappe, S., Foley, S.F., Jenner, G.A., Kjarsgaard, B.A., 2005. Integrating Ultramafic Lamprophyres into
26 the IUGS Classification of Igneous Rocks: Rationale and Implications. Journal of Petrology 46, 1893-
27 1900.
28
29
30 Tappe, S., Foley, S.F., Kjarsgaard, B.A., Romer, R.L., Heaman, L.M., Stracke, A., Jenner, G.A., 2008.
31 Between carbonatite and lamproite—diamondiferous Torngat ultramafic lamprophyres formed by
32
carbonate-fluxed melting of cratonic MARID-type metasomes. Geochimica et Cosmochimica Acta 72,
33
34 3258-3286.
35
36 Tappe, S., Foley, S.F., Stracke, A., Romer, R.L., Kjarsgaard, B.A., Heaman, L.M., Joyce, N., 2007.
37
38 Craton reactivation on the Labrador Sea margins: 40Ar/39Ar age and Sr–Nd–Hf–Pb isotope constraints
39 from alkaline and carbonatite intrusives. Earth and Planetary Science Letters 256, 433-454.
40
41
Tappe, S., Massuyeau, M., Smart, K.A., Woodland, A.B., Gussone, N., Milne, S., Stracke, A., 2021.
42
43 Sheared Peridotite and Megacryst Formation Beneath the Kaapvaal Craton: a Snapshot of
44 Tectonomagmatic Processes across the Lithosphere–Asthenosphere Transition. Journal of Petrology 62,
45
46 1-39.
47
48 Tappe, S., Ngwenya, N.S., Stracke, A., Romer, R.L., Glodny, J., Schmitt, A.K., 2023. Plume–
49
lithosphere interactions and LIP-triggered climate crises constrained by the origin of Karoo lamproites.
50
51 Geochimica et Cosmochimica Acta 350, 87-105.
52
53 Tappe, S., Romer, R.L., Stracke, A., Steenfelt, A., Smart, K.A., Muehlenbachs, K., Torsvik, T.H., 2017.
54
55 Sources and mobility of carbonate melts beneath cratons, with implications for deep carbon cycling,
56 metasomatism and rift initiation. Earth and Planetary Science Letters 466, 152-167.
57
58
Valdes, M.C., Moreira, M., Foriel, J., Moynier, F., 2014. The nature of Earth's building blocks as
59
60 revealed by calcium isotopes. Earth and Planetary Science Letters 394, 135-145.
61
62
63 Page 23/28
64
65
 Veter, M., Foley, S.F., Mertz-Kraus, R., Groschopf, N., 2017. Trace elements in olivine of ultramafic
1 lamprophyres controlled by phlogopite-rich mineral assemblages in the mantle source. Lithos 292-293,
2
3 81-95.
4
5 Wang, W., Qin, T., Zhou, C., Huang, S., Wu, Z., Huang, F., 2017a. Concentration effect on equilibrium
6
7 fractionation of Mg-Ca isotopes in carbonate minerals: Insights from first-principles calculations.
8 Geochimica et Cosmochimica Acta 208, 185-197.
9
10 Wang, W., Zhou, C., Qin, T., Kang, J.-T., Huang, S., Wu, Z., Huang, F., 2017b. Effect of Ca content on
11
12 equilibrium Ca isotope fractionation between orthopyroxene and clinopyroxene. Geochimica et
13 Cosmochimica Acta 219, 44-56.
14
15
16 Wang, Y., He, Y., Wu, H., Zhu, C., Huang, S., Huang, J., 2019. Calcium isotope fractionation during
17 crustal melting and magma differentiation: Granitoid and mineral-pair perspectives. Geochimica et
18 Cosmochimica Acta 259, 37-52.
19
20
21 Wardle, R.J., Hall, J., 2002. Proterozoic evolution of the northeastern Canadian Shield: Lithoprobe
22 Eastern Canadian Shield Onshore–Offshore Transect (ECSOOT), introduction and summary. Canadian
23
Journal of Earth Sciences 39, 563-567.
24
25
26 Williams, H.M., Bizimis, M., 2014. Iron isotope tracing of mantle heterogeneity within the source
27 regions of oceanic basalts. Earth and Planetary Science Letters 404, 396-407.
28
29
30 Windley, B.F., Garde, A.A., 2009. Arc-generated blocks with crustal sections in the North Atlantic
31 craton of West Greenland: Crustal growth in the Archean with modern analogues. Earth-Science
32
Reviews 93, 1-30.
33
34
35 Xiao, Z., Zhou, C., Kang, J., Wu, Z., Huang, F., 2022. The factors controlling equilibrium inter-mineral
36 Ca isotope fractionation: Insights from first-principles calculations. Geochimica et Cosmochimica Acta
37
38 333, 373-389.
39
40 Yang, Z.-F., Li, J., Liang, W.-F., Luo, Z.-H., 2016. On the chemical markers of pyroxenite contributions
41
in continental basalts in Eastern China: Implications for source lithology and the origin of basalts.
42
43 Earth-Science Reviews 157, 18-31.
44
45 Yuan, H., French, S., Cupillard, P., Romanowicz, B., 2014. Lithospheric expression of geological units
46
47 in central and eastern North America from full waveform tomography. Earth and Planetary Science
48 Letters 402, 176-186.
49
50
Zhang, H., Wang, Y., He, Y., Teng, F.-Z., Jacobsen, S.B., Helz, R.T., Marsh, B.D., Huang, S., 2018. No
51
52 Measurable Calcium Isotopic Fractionation During Crystallization of Kilauea Iki Lava Lake.
53 Geochemistry, Geophysics, Geosystems 19, 3128-3139.
54
55
56 Zhao, K., Dai, L.-Q., Fang, W., Zheng, Y.-F., Zhao, Z.-F., Zheng, F., 2022. Decoupling between Mg and
57 Ca isotopes in alkali basalts: Implications for geochemical differentiation of subduction zone fluids.
58
Chemical Geology 606, 120983.
59
60
61
62
63 Page 24/28
64
65
 Zhu, H., Ionov, D.A., Du, L., Zhang, Z., Sun, W., 2021. Ca-Sr isotope and chemical evidence for
1 distinct sources of carbonatite and silicate mantle metasomatism. Geochimica et Cosmochimica Acta
2
3 312, 158-179.
4
5 Zhu, H., Liu, F., Li, X., Wang, G., Zhang, Z., Sun, W., 2018. Calcium Isotopic Compositions of Normal
6
7 Mid-Ocean Ridge Basalts From the Southern Juan de Fuca Ridge. Journal of Geophysical Research:
8 Solid Earth 123, 1303-1313.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63 Page 25/28
64
65
 1
2 Figure captions
3
4 Figure 1. The locations of alkaline rocks at the flanks of the rift and seismic shear wave velocities
5
6 beneath the North Atlantic craton and the Labrador Sea at the depth of 150 km. Seismic
7
8
tomography is from Yuan et al. (2014). The lithosphere beneath Greenland and Labrador shows
9
10
11
high velocities similar to the Superior craton, indicating its cratonic nature. The mantle under the
12
13 Labrador Sea shows low velocity, indicating rifted lithosphere with a higher-velocity ledge in the
14
15 Davis Strait between Baffin Bay and the Labrador Sea. The locations of the alkaline rocks (UMLs
16
17 = Ultramafic Lamprophyres and Kim. and carb. = Kimberlites and carbonatites) are from Tappe et
18
19 al. (2007) and are distributed along the relatively low-velocity zone at the margins of Greenland
20
21 and Labrador.
22
23
24
25 Figure 2. δ44/40Ca values of the alkaline rocks from Aillik Bay compared to average MORBs
26
27 (Chen et al., 2020a; Eriksen and Jacobsen, 2022; Zhu et al., 2018), kimberlites (Antonelli et al.,
28
29 2023; Tappe et al., 2021), carbonatites (Banerjee et al., 2021; Sun et al., 2021), continental silica-
30
31 saturated basalts (Liu et al., 2017) and silica-unsaturated basalts (Qi et al., 2022; Zhao et al., 2022),
32
33 and oceanic island basalts OIBs in publications since 2015 (Eriksen and Jacobsen, 2022; He et al.,
34
35 2017; Jacobson et al., 2015; Zhang et al., 2018).
36
37
38
39
Figure 3. Bulk rock MgO content versus CaO/Al2O3 ratio (a) and δ44/40Ca value versus
40
41
42
CaO/Al2O3 ratio (b) for Mesoproterozoic lamproites and Cretaceous nephelinites from Aillik Bay.
43
44 Modelled crystallization of an experimental lamproite melt (blue line) from Foley et al. (2022) and
45
46 an experimental nephelinite melt (magenta line) from Dasgupta et al. (2007) using MELTS
47
48 (Gualda and Ghiorso, 2015) are shown for comparison. The phases during crystallization
49
50 modelling (L-liquid, Ol-olivine, Cpx-clinopyroxene) are shown along the lines.
51
52
53
54 Figure 4. (a-b) Bulk rock MgO content versus FC3MS (FeO/CaO-3*MgO/SiO2) and bulk rock
55
56 K2O versus CaO content for Mesoproterozoic lamproites, Neoproterozoic ultramafic
57
58 lamprophyres, and Cretaceous nephelinites from Aillik Bay. Data for experimental pyroxenite
59
60 melt and peridotite melt were compiled by Yang et al. (2016). In (a), the dotted line indicates the
61
62
63 Page 26/28
64
65
 upper boundary for peridotite melts (Yang et al., 2016), discriminating melts from pyroxenite and
1
2 peridotite. Experimental melts of MARID (Foley et al., 2022) and phlogopite clinopyroxenites
3
4
(Funk and Luth, 2013; Lloyd et al., 1985) are also shown for comparison.
5
6
7
8
9 Figure 5. (a) Calcium isotope β-factors for clinopyroxene (Ca/(Ca+Mg) = 7/16), carbonate
10
11 minerals (calcite, dolomite, and magnesites with Ca/(Ca+Mg) of 1/16 and 1/36), apatite, and K-
12
13 richterite from first-principles calculations (Wang et al., 2017a; Wang et al., 2017b; Xiao et al.,
14
15 2022). (b) Modelled δ44/40Ca of lamproitic melts and residues with increasing degree of partial
16
17 melting of MARID at 6.6 GPa using the melting reaction from Foley et al. (2022). The initial
18
19 δ44/40Ca value is assumed to be the same as MORB (Chen et al., 2020a; Eriksen and Jacobsen,
20
21 2022). (c) Modelled δ44/40Ca of MARID (with 30 wt/% K-richterite and 30 wt/% K-richterite + 5
22
23 wt.% clinopyroxene, respectively) crystallized from silicate melt/fluid using fractionation factors
24
25 in (a). The cratonic geotherms are summarized in Lee et al. (2011). (d) Modelled δ44/40Ca of
26
27 carbonatite melts and residues with an increasing degree of partial melting of magnesite lherzolite
28
29 at 6.6 GPa using the melting reaction from (Dasgupta and Hirschmann, 2007). The δ44/40Ca in the
30
31 starting carbonated peridotite is the average value of carbonated peridotite xenoliths from the
32
33 Siberian craton (Ionov et al., 2019; Zhu et al., 2021). Details of all modelling calculations are in
34
35
Supplementary information: Section 3.1 and Table S2.
36
37
38
39
40 Figure 6. Plots of SiO2 versus CaO (a) and Al2O3 (b) contents and plots of δ44/40Ca versus SiO2 (c)
41
42 and Al2O3 contents (d) for the Late Neoproterozoic ultramafic lamprophyres. Magma evolution
43
44 trends (arrows) from Tappe et al. (2006). Igneous carbonatites are shown for comparison
45
46 (Banerjee et al., 2021; Sun et al., 2021).
47
48
49
50 Figure 7. (a) Comparison of δ44/40Ca values of aillikites and mela-aillikites with carbonated
51
52 peridotites (Table S3, Ionov et al., 2019; Zhu et al., 2021). (b-c) δ44/40Ca values versus 87Sr/86Sr

53
54 and δ13C of carbonate-rich igneous rocks including aillikites, mela-aillikites, and carbonatites.
55
56 Previously reported δ44/40Ca values of global carbonatites are from Sun et al. (2021) and Banerjee
57
58 et al. (2021). Sun et al. (2021) defined primary carbonatites based on their C-O isotopes (-8‰ <
59
60 δ13C < -4‰ and 4‰ < δ18O < 10‰). δ44/40Ca values of kimberlites are also shown for comparison
61
62
63 Page 27/28
64
65
 (Antonelli et al., 2023).
1
2
3
4
Figure 8. Bulk rock δ44/40Ca versus bulk rock SiO2 (a), CaO/Al2O3 (b), La/Lu ratio (c), and
5
6
7 Na2O+K2O content (d) for the alkaline rocks in this study, OIBs (Eriksen and Jacobsen, 2022),
8
9 continental silica-saturated (CSS basalts) (Liu et al., 2017) and silica-undersaturated (CSU basalts)
10
11 alkaline basalts (Qi et al., 2022; Zhao et al., 2022), and kimberlites (Antonelli et al., 2023).
12
13 Average MORBs are shown for comparison (Chen et al., 2020a; Eriksen and Jacobsen, 2022; Zhu
14
15 et al., 2018).
16
17
18
19 Figure 9. Evolution of calcium isotope compositions of the metasomatic mantle lithosphere at the
20
21 base of the North Atlantic craton over geological history (a-d). Calcium isotope compositions of
22
23 the refractory cratonic lithosphere mantle (Chen et al., 2019; Kang et al., 2017), upper mantle
24
25 (Chen et al., 2019; Kang et al., 2017), and average MORB (Chen et al., 2020a; Eriksen and
26
27 Jacobsen, 2022; Zhu et al., 2018) are shown for comparison. (b-c) Schematic illustration of
28
29 progressive rifting of the North Atlantic craton by episodic metasomatism and asthenospheric
30
31 convection (modified after Tappe et al. (2007)). Mesoproterozoic lamproites were formed by the
32
33 melting of phlogopite pyroxenite and peridotite at the base of the lithosphere in reduced conditions
34
35
(b). Late Neoproterozoic ultramafic lamprophyres and carbonatites originate from the melting of
36
37
38 phlogopite-carbonate veins derived from the solidification of carbonate-rich melts from the
39
40 asthenosphere (c). The aillikitic magmas from this episode played an important role in the
41
42 alteration and weakening of the shallow lithosphere mantle. The Cretaceous nephelinites were
43
44 produced by melting of the lithosphere mantle (wehrlite) modified by aillikitic melt (d).
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63 Page 28/28
64
65
Figures Click here to access/download;Figure;Figures.pdf

Figure 1

-80° -60° -40° -20°

Lamproites (ca.1400-1200 Ma)

UMLs (ca.610-550 Ma)
Kim. and carb. (ca.160 Ma)
Nephelinites (ca.150-55 Ma)
Baffin Bay
Greenland
70° 70°

Davis Strait

North Atlantic

La
60° ocean 60°

bra
do
Superior Labrador

rS
craton

ea
Aillik

50° 50°

Vs
4.4 4.6 4.8 km/s Depth: 150km
40° 40°
-80° -60° -40° -20°
 Figure 2

δ44/40Ca (‰)
0.3 0.5 0.7 0.9 1.1 1.3

Lamproites
Aillikites MORBs
Mela-Aillikites
Rift-related alkaline rocks
This study
Damtjernites

Carbonatites

Nephelinites

Kimberlites

Primary-Sun et al. 2021

Carbonatites

SiO2-saturated
Continental basalts
SiO2-unsaturated

OIBs

Carbonated silicate melts: 0.74 ± 0.02, 2se, n=48

 Figure 3

1.0
(a) (b)
L+Ol
1.5

0.8

δ44/40Ca (‰)
CaO/Al2O3

1.0
0.6

L+Ol

0.4
0.5 0.6 0.8 1.0 1.2 1.4
2 5 8 11 14
MgO (wt.%) CaO/Al2O3
 Figure 4

1.5 20
(a) (b) Melts of pyroxenite
Melts Melts of peridotite
1.0 of pyroxenite Melts of CO2-peridotite
Melting of MARID
Melting of Phl-clinopyroxenite
0.5 Lamproites
15 Aillikites

CaO (wt.%)
Nephenites
FC3MS

0.0

-0.5 Melts of
peridotite
10
-1.0

-1.5 Melts of carbonated

peridotite
-2.0 5
5 10 15 20 25 0 5 10 15
MgO (wt.%) K2O (wt.%)
 Figure 5

1200 1150 1100 1050 1000 °C

1.2 1.0
(a) (b)
Melting reaction:
0.98RA+0.02Phl = 0.6Melt +0.24Cpx+0.15Opx
1.0

δ44/40Ca(‰)
103lnβ44/40Ca

0.8 0.8
Initial
value
Cpx 7/16
0.6 Calcite
Dolomite Lamproitic melt (6.6 GPa)
Magnesite 1/6
Magnesite 1/36 Residue (6.6 GPa)
Richterite
0.4 Apatite 0.6
0.4 0.5 0.6 0.7 0 0.1 0.2 0.3
106/T2 (K-1) Melting degree (F)
1.2
(c) (d) Melting reaction:
0.9 Initial value 6.6 GPa: 3Cpx + 2Gt + 2Mst = 4Opx + 3Melt

1.0
Initial
value
δ44/40Ca(‰)

δ44/40Ca(‰)

0.7
0.8
Cratonic geotherms
at 5GPa

0.5
0.6

30%Richterite Carbonatite melt (6.6 GPa)

Residue (6.6 GPa)
30%Richterite+5%Cpx
0.3 0.4
700 900 1100 1300 1500 1700 0 0.02 0.04 0.06
Temperature (oC) Melting degree (F)
 Figure 6

50 12
(a) Liquid immusibility (b)
Decarbonation
40 Fractional crystalization
of damtjernite
8
30

Al2O3 (%)
CaO (%)

20
Aillikites 4
Mela-aillikites
10 Damtjernites
Carbonatites Primary
Global carbonatites magma
0 0
0 10 20 30 40 0 10 20 30 40
SiO2 (wt.%) SiO2 (wt.%)
1.4 1.4
(c) (d)
1.2 1.2

1.0 1.0
δ44/40Ca (‰)
δ44/40Ca (‰)

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2
0 10 20 30 40 0 5 10
SiO2 (wt.%) Al2O3 (wt.%)
 Figure 7

1.2 Aillikites Banerjee et al. 2021

(a) (b) Primary-Sun et al. 2021
(c) Kimberlites
Mela-aillikites
Carbonated peridotites Nonprimary-Sun et al. 2021
1.0 Carbonatites in this study
ST203
δ44/40Ca (‰)

0.8

0.6
Carbonate-rich melts in
the mantle
0.4

0.2
0 10 20 0.702 0.707 0.712 -8 -6 -4 -2 0 2
SiO2 (wt.%) 87Sr/86Sr δ13C
 Figure 8

(a) (b)
1.0 1.0

MORBs MORBs
δ44/40Ca (‰)

δ44/40Ca (‰)
0.8 0.8

0.6 0.6
OIBs CSS basalts
Kimberlites CSU basalts
Aillikites Mela-aillikites
Nephelinites
0.4 0.4
20 30 40 50 60 0.2 2
SiO2 (wt.%) CaO/Al2O3

(c) (d) Lamproites

1.0 1.0

MORBs
δ44/40Ca (‰)

δ44/40Ca (‰)

MORBs
0.8 0.8

0.6 0.6

0.4 0.4
0 500 1000 1500 0 2 4 6 8 10
La/Lu Na2O+K2O (wt.%)
 Figure 9
(a)

1.4
Refractory cratonic
mantle lithsophere
1.2

1.0
Ambient upper
δ44/40Ca (‰)

mantle MORBs
0.8 Carbonated
mantle
0.6 MARID-rich mantle

0.4 The base of lithospheric

mantle under the NAC

0.2
2000 1500 1000 500 0
A ge (Ma)

(b) Mesoproterozoic (c) Late Neoproterozoic (d) Cretaceous

Lamproites Ultramafic lamprophyres Nephelinites

Crust Labrador Greenland

Lithosphere
Metasomatism Neoproterozoic
during metasomatic δ44/40Ca:
magma ascent mantle 0.75‰

Deviatoric Deviatoric
stress stress
Carbonated δ44/40Ca:
mantle 0.72-0.82‰
MARID veins δ44/40Ca: 0.6‰

Infiltration of hydrous
Hydrous silicate melt carbonatitic melt Asthenosphere flow caused by
localized lithosphere thinning
Asthenosphere
Table 1 Click here to access/download;Table;Table 1.xlsx

Table 1. Major element and Sr-Ca-C-O isotopic compositions and melting temperatures (T) and pressures (P) of the alkaline rocks at Aillik Bay and reference materials.
Samples SiO2a TiO2 Al2O3 FeO MnO MgO CaO Na2O K2O P2O5 H2O CO2 Mg# 87Sr/86Sri δ44/42Ca 2sd δ43/42Ca 2sd δ44/40Cab δ44/40CaRadc 2sd 2se n δ C δ O
13 18
Td P
wt.% wt.% wt.% wt.% wt.% wt.% wt.% wt.% wt.% wt.% wt.% wt.% ‰ ‰ ‰ ‰ ‰ ‰ °C GPa
Mesoproterozoic olivine lamproites (1374 Ma)
ST115a 45.1 4.05 8.78 12.4 0.12 8.88 7.2 1.3 6.37 1.1 2.0 1.2 56.1 0.7048 0.30 0.01 0.16 0.07 0.60 0.47 0.03 0.02 3 1429 4.8
ST208 44.6 3.943 8.16 11.61 0.12 8.71 6.3 1.4 6.1 0.87 2.0 5.1 59.8 0.7049 0.28 0.01 0.16 0.03 0.58 0.43 0.03 0.02 3 1436 5.1
ST223 45.7 3.47 9.76 11.2 0.13 8.05 7.8 2.4 5.07 0.86 1.3 3.5 56.2 0.7047 0.32 0.01 0.15 0.04 0.66 0.56 0.03 0.01 3 1372 4.1
ST237 47.0 4.772 8.27 12.34 0.33 4.91 6.5 1.4 7.33 1.2 1.7 3.3 44.1 0.7056 0.28 0.05 0.14 0.06 0.58 0.41 0.11 0.06 3 1514 5.6
Late Neoproterozoic aillikites (590-555 Ma)
ST164 24.1 3.3 3.3 13.2 0.20 18.6 18.1 0.20 2.0 3.0 2.5 10.3 73.7 0.7039 0.36 0.06 0.19 0.05 0.75 0.74 0.13 0.08 3 -5.2 9.4 1426 5.3
ST198a 26.8 3.8 3.5 13.7 0.20 17.3 14.3 0.70 1.7 1.7 1.7 13.3 71.4 0.7037 0.34 0.03 0.18 0.01 0.70 0.69 0.05 0.03 3 1502 6.4
ST225 26.4 3.3 2.5 13.2 0.20 20.9 16.4 0.20 1.4 2.0 2.2 10.1 75.8 0.7038 0.35 0.06 0.17 0.04 0.71 0.71 0.12 0.07 3 1444 6.2
ST250a 22.5 2.5 3.1 12.2 0.20 20.4 15.6 0.20 2.1 1.5 4.7 13.8 76.9 0.7039 0.35 0.05 0.16 0.10 0.72 0.71 0.10 0.06 3 -5.0 11.0 1427 5.0
Late Neoproterozoic mela-aillikites (590-555 Ma)
ST147b 31.6 5.8 4.0 17.5 0.20 16.8 11.0 0.60 1.9 1.3 3.4 4.6 65.4 0.7038 0.33 0.01 0.15 0.08 0.67 0.66 0.03 0.02 3
ST196 35.9 4.4 4.8 14.7 0.20 22.2 8.2 0.50 1.7 0.6 3.9 2.0 75.0 0.7033 0.35 0.04 0.16 0.10 0.71 0.70 0.09 0.05 3
ST244b 32.1 5.5 4.0 14.7 0.20 20.2 10.3 0.60 1.7 1.0 3.0 5.4 73.1 0.7046 0.35 0.02 0.15 0.02 0.72 0.71 0.05 0.03 3
Late Neoproterozoic damtjernites (590-555 Ma)
ST140 31.5 6.1 6.3 13.8 0.20 9.6 21.2 0.50 1.4 3.4 2.5 1.9 58.1 0.7036 0.33 0.03 0.15 0.01 0.68 0.68 0.05 0.03 3
ST174 33.1 6.1 8.5 15.8 0.30 5.9 18.3 2.2 2.4 2.3 3.6 0.2 42.4 0.7040 0.28 0.03 0.14 0.04 0.56 0.56 0.07 0.04 3
ST188a 30.7 4.8 7.4 14.4 0.30 6.4 19.8 4.3 2.6 2.9 2.9 2.3 46.7 0.7036 0.34 0.02 0.19 0.06 0.70 0.69 0.05 0.03 3 -7.0 11.4
ST206a 30.6 6.9 7 15.7 0.20 7.8 18 1.1 2 2.8 1.8 4.8 49.6 0.7038 0.29 0.05 0.15 0.10 0.60 0.59 0.11 0.06 3 -4.7 9.9
ST224b 33.5 4.9 5.4 15.8 0.20 15.9 10.3 1.7 0.90 1.2 1.8 7.3 66.5 0.7040 0.35 0.07 0.15 0.13 0.72 0.72 0.14 0.08 3
ST226 36.5 5.9 9.9 14.8 0.30 6.3 13.9 2.1 2.3 1.7 3.0 2.6 45.7 0.7039 0.31 0.03 0.15 0.06 0.62 0.62 0.11 0.07 3 -5.9 10.4
ST114 30.2 5.6 5.9 15.3 0.20 9.8 17.2 0.90 2.6 2.0 1.1 8.1 55.9 0.34 0.07 0.16 0.02 0.70 0.69 0.13 0.08 3
ST170 38.0 5.7 7.9 14.3 0.20 6.8 17 1.3 2.1 1.8 2.5 1.1 48.5 0.29 0.01 0.14 0.02 0.59 0.59 0.14 0.08 3
Replicate 0.32 0.03 0.18 0.06 0.65 0.03 0.02 3
ST256 32.7 5.343 6.5 14.35 0.25 11.08 14.013 1.0818 2.502 1.774 3.4 2.0 57.9 0.7050 0.35 0.04 0.19 0.01 0.71 0.70 0.07 0.04 3
Late Neoproterozoic carbonatites (590-555 Ma)
ST203 10.1 0.3 0.2 8.6 0.50 12.9 22.8 1.1 0.2 1.9 0.2 32.3 74.9 0.7047 0.44 0.03 0.23 0.06 0.90 0.90 0.06 0.03 3 -2.8 10.8
ST126 17.6 2.3 2.0 8.4 0.30 10.7 26.7 0.2 1.4 0.4 2.7 26.0 71.6 0.7058 0.40 0.02 0.19 0.07 0.82 0.82 0.03 0.02 3 -3.3 10.2
ST193a 9.1 1.2 1.8 5.3 0.30 7.2 38.4 0.2 1 2 1.6 30.5 72.8 0.7039 0.35 0.03 0.20 0.02 0.71 0.71 0.06 0.04 3
Replicate 0.35 0.01 0.16 0.06 0.71 0.03 0.02 3
ST198c 11.7 2.1 2.1 10.4 0.30 8.3 30.8 1.4 0.8 4.5 0.6 25.7 61.3 0.7039 0.35 0.03 0.17 0.07 0.71 0.71 0.06 0.03 3 -3.7 10.8
ST199 13.9 2 2.1 8.3 0.30 8.1 31.8 1.1 1.5 4 1.5 24.2 65.9 0.7039 0.37 0.04 0.16 0.04 0.76 0.76 0.07 0.04 3 -4.8 10.0
Cretaceous nephelinites (141 Ma)
ST100 38.1 2.5 12.1 11.5 0.15 10.6 13.7 0.74 1.7 1.5 5.5 1.1 64.8 0.7047 0.38 0.03 0.20 0.03 0.78 0.78 0.05 0.03 3 1252 2.5
ST102 44.3 1.7 13.2 10.5 0.21 8.1 10.3 2.3 1.9 0.73 3.9 0.6 60.5 0.7049 0.35 0.03 0.17 0.03 0.72 0.72 0.06 0.03 3 1288 2.7
ST103 34.7 2.2 10.7 11.8 0.27 9.7 12.1 3.0 1.64 2.5 6.4 3.7 59.5 0.7062 0.36 0.03 0.16 0.07 0.74 0.74 0.04 0.02 3 1309 2.8
Replicate 0.38 0.01 0.17 0.07 0.77 0.06 0.03 3
ST253 38.8 2.6 12.7 12.13 0.21 8.1 12.9 3.0 1.9 1.0 1.7 3.6 57.0 0.7056 0.37 0.03 0.17 0.06 0.76 0.76 0.06 0.04 3 1322 2.9
ST254 37.2 2.2 11.3 11.6 0.19 11.0 14.6 3.0 1.5 1.15 1.6 3.8 62.8 0.7052 0.37 0.04 0.18 0.05 0.76 0.75 0.09 0.05 3 1270 2.7
Reference materials
COQ-1 0.36 0.03 0.18 0.06 0.74 0.06 0.04 3
BHVO-2 0.38 0.04 0.17 0.02 0.77 0.07 0.04 3
BCR-2 0.40 0.04 0.18 0.07 0.83 0.08 0.05 3
Replicate 0.39 0.03 0.16 0.05 0.79 0.06 0.03 3
a: The major and trace elements, Sr isotopes, C-O isotopes of the alkaline and carbonatite rocks are from Tappe et al. (2006) and Tappe et al. (2007).
b: δ44/40Ca obtained using the mass-dependent fractionation law: δ 44/40Ca = 2.048 × δ44/42Ca (Heuser et al., 2016).
c: δ44/40CaRadiogenic is calculated by the addition of 40Ca ingrowth via the decay of 40K to 40Ca, details see Fu et al. (2022).
d: Estimate based on thermobarometry of silica-undersaturated melts developed by Sun and Dasgupta (2020).
Supplementary information

Click here to access/download

Supplementary material for online publication only
Supplementary information.docx
Supplementary tables

Click here to access/download

Supplementary material for online publication only
Supplementary tables.xlsx