The Chemical Evolution from Older (323–318 Ma) towards Younger Highly Evolved Tin Granites (315–314 Ma)—Sources and Metal Enrichment in Variscan Granites of the Western Erzgebirge (Central European Variscides, Germany)

Round 1

Reviewer 1 Report

A very well-written manuscript, that presents an excellent data set, is supported by good figures and has the potential to generate significant interest in the (tin) granite-related mineral resources community, around and well beyond the Erzegibirge. I support publication, subject to minor revision.

Suggested revisions

The common homogeneity of many datasets between, and within, individual granite groups is remarkable and is used to discount certain source types. There needs to be a more robust justification in these instances, using references and/or or additional data. Line 593-595. The concluding statement "In contrast to previously published suggestions [6, 7], we can exclude a substantial role of intense sedimentary weathering as an important control factor for later Sn and W enrichment in granite related ores of the Western Erzgebirge." is central to the work. But, although challenging, it would be better if there were more proxy data available to allow direct comparsion with the Ordovician shales of the Frauenbach Group?
Line 154. Sedimentary vs quartzo-feldspathic is an ambiguous source classification. Is 'sedimentary' being used as a proxy for a particular type of sedimentary rock such as pelite or greywacke? Be more specific.

Line 315: published 18O analyses for the Ordovician shales of the Frauenbach Group? If so, they should be referenced here.
  Line 552-554. The statement "We suppose that the high enrichment in hydrothermal greisen fluids was possible due to a thick continental crust composed in large parts of Variscan granites that were already enriched in these elements." is not substantiated. What about the local role of melt:fluid partitioning?

Line 603-604. The statement "Efficient leaching by hydrothermal fluids led to a very strong enrichment (up to several orders) of Sn and W in the greisen ore bodies." is not substantiated. How do you test leaching vs a Sn+/-W enriched magmatic-hydrothermal fluid in which enrichment was a consequence of melt:fluid partitoning?

There might be too many data tables within the body text.

Very minor typographical annotations marked-up in the attached pdf.

Comments for author File: Comments.pdf

Author Response

Please ee attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

 

 

The paper presented by Tichomirowa et al. is an interesting work about the origin (sources and partial melting conditions) and melt differentiation as main causes of tin enrichment in Variscan granites of the Western Erzgebirge (Bohemian Massif, Germany). I agree with authors in such a combination of conditions, pre-enriched crustal protoliths and exhaustive crystal fractionation processes, as the main drivers in generating Sn-rich granitoids. The work is well constructed and supported by the data, summarizing and filling the previous data from literature.

In my opinion, the work in its current form is acceptable to be published in Minerals journal with only minor corrections.

My major concerns are as follows:

Some Tables are very large and make unnecessarily long the manuscript. Table 1 (Sample location), 4 (Lu-Hf-O isotopes on zircon) and 5 (Trace elements in zircons) could be addressed to Supplementary Material. Moreover, Table 2 (Bulk rock geochemistry) could be edited (or try at least) to be smaller.

 

Discussion (subchapter 5.1 Identification of source rocks). I agree with authors that metasedimentary protoliths at high-grade metamorphism (to generate significant fractions of granite magma batches) are suitable for generating studied peraluminous granites. I also agree with the different Sr-O isotopic ratios that those materials must have at lower crustal levels (as it happens in other sectors of the European Variscan Belt ( e.g., central Spain: Villaseca et al., 1999, Jour. Petrol. 40). So this argument is enough robust for me.

 

For me, the most debatable point is in subchapter 5.4.1 (source enrichment). It is a surprise for me that restitic granulite rocks were more enriched in Sn and W than metasedimentary precursors. It is very intriguing how to reach Sn and W enrichments in granulitic rocks (restite- and melt-rich fractions). To my knowledge, these elements can have some contrasting behavior during crustal melting as W remains in the restite-fraction (mostly in rutile) (up to 400 ppm in lower crustal granulite xenoliths –LCGX-from central Spain) whereas Sn is more concentrated in melt-rich fractions as it appears at very low concentrations in restitic granulite rocks (our unpublished data on LCGX is mostly below 2 ppm). Rutile must have Kd^W > Kd^Sn as deduced from literature (e.g., Luzzivoto et al., 2009, ChemGeol 261; Carocci et al., 2019, Minerals 9). The scarce data presented by Tichomirowa et al., 2018 (Lithos 302-303) means that they should be taken with caution.

 

Summarizing, I agree with authors that the combination of suitable metasedimentary sources with efficient crystal fractionation processes within fluxing-rich high-silica magmas, promoting syn-to-late magmatic-hydrothermal interactions, play a significant role in the genesis of Sn-W-ore mineralizations within and around some peraluminous granite bodies. We have published recently a similar model (from enriched metasedimentary sources-to-highly fractionated granites) to explain Li-rich pegmatite fields around Variscan peraluminous granites (Roda-Robles et al., 2018, Ore Geol. Rev. 95). So, crustal recycling of previously enriched metasedimentary sources seems to be a prerequisite for the generation of the Sn-W or Li-P ores related to some Variscan peraluminous granites.

 

Minor typographic mistakes and suggestions are:

Line 36: voluminous granite plutons intrude the European upper crust

Figure 1 References: introduce the numbering [29], [30] and [32] in (c) plot

239: The last sentence is very short, please, explains a little the meaning of “clearly different chemical compositions of the EIB-granites when compared to all OIC”

246 and 249: We have calculated temperatures… we have applied the Al2O3/TiO2 thermometer…

319: garnet- and muscovite-bearing restites

357: record a granulitic source (I suppose!)

402: A similar relationship is displayed by all porphyritic varieties…

Figure 10: put c2 in another part of the pattern to be more visible!

 

Author Response

Please see attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

General and major comments

This is a valuable scientific contribution to decipher the origin of the sources for the Erzgebirge granites, with high quality analytical data performed with up to date technologies, but the paper suffer of weaknesses concerning:

the limited number of samples anlyzed per granite bodies the interpretation of the data Concerning the number of samples, the selected ones do not cover the variations of chemical composition of the idividual granite plutons compared to the data presented by Förster et al., 1999. In consequence the degree of metal enrichment related to magmatic processes within the pluton is only partly assessed. the interpretations (see below)

p. 8: Harker diagrams you are using tend to align all the data for the different diagram for exemple for the one involving, Ba, Sr ...
There are better diagrams to discriminate the different granite types : A-B (multicationic diagrams of Debon and Lefort), Th-B diagrams, Th-Ta ... using mainly the less mobile elements. Some of these diagram will be attached to the comments.

p. 13 : You write that inherited zircons are scarce, but in peraluminous low temperature leucogranites Zr solubility is very low and thus inherited zircon are common especially in the Zr richest sampes. This is supported by the fact that the zircon saturation temperature estimations between the samples of the three peraluminous leucogranites differ of up to 150°C ! Whereas for the higher temperature weakly peraluminous Kirchberg high-K calcalkaline granite the difference is only of about 50°C ! I suggest that the lowest temperatures for the peraluminous leucogranites are the ones the closest to the real temperature of the magma

p. 14: Table 3:

The calculated mean Sri is not significant because strongly influenced by the highly altered sample Eibenstock 612, as shown by its very high peraluminous index due to strong plagioclase alteration (very low Ca and Na contents) and the percolation of hydrothermal fluids indicated by very high Rb [986 ppm], W, Sn [137 ppm], and Bi [1060 ppm] contents. This has led to an anomalously high Sri value of 0.7198, all others Sri are below 0.710. This samples has also the lowest Zr and REE contents, much lower that all other samples. This may have lead also to the anomalous value obatined for T DM value.

Also samples EIB 713 and KIB786 are significantly altered. All these samples should have been discarted from the study looking for primary signature of the granites.

Sample BER790 is also probably altered but to a lesser extend (high Rb, Sn, W. He has also extremely low Zr and REE contents, which may also explain the non realistic T DM value.

p. 22: The "slightly higher scatter for EIB, especially for zircons/zircon rims with black CL)" is due to the metamict state of these crystals which are very rich in uranium (up to several percent). Thus, this has no meaning for the interpretation of the source rocks !

You write "all of our samples have low CaO/Na2O ratios (≤0.3; Table 2)" but this is not related to the characteristics of the source but to the fact that these granites are highly fractionated, with fractionation of most of the anorthite componant as indicated by the diagram CaO versus B = Fe + Mg + Ti). Each granite body has its proper correlation trend, except Eibenstock because of more important alteration.

You write "garnet- und muscovite bearing restites left behind after granulite-facies melting have low CaO/Na2O ratios because they are poor in plagioclase" What type of rocks are you melting in granulite facies ? Restites have always higher CaO/Na2O ratios tha the protoliths because the anorthite componant if more refractory than the albite componant of plagioclase, same rule for Ca-rich minerals such as amphibole or pyroxene !
The speculation you are doing on the possible source is rather weak. No tight constrain.

p. 26: You write "K-feldspar loss due to fractional crystallization in the fine-grained evolved granites" This is a too simplistic approach of the geochemical variations the B-CaO diagram indicates that you have fractionation of
- plagioclase with decreasing CaO contents
- biotite with decreasing B parameter, its fractionation may explain at least part of Ba content decrease
with Th, Zr or REE versus B diagrams fractionation of:
- zircon and monazite (except for Kirchberg)
whereas, despite biotite fractionation K2O content remains either constant or increases (clearly in the Eibenstock pluton), with during progressive fractionation progressive increasing proportion of of muscovite crystallization instead of of K-feldspar.
In figure 2C, the decrease of the Ba and Sr contents are better explained by a combination of plagioclase + biotite fractionation rather than by that of K-feldspar.

You write: "defined Ca-rich rocks (our  undifferentiated rocks with phenocrysts) as “cumulative rocks” that represent melts together with a precipitating mineral" but Ca-rich rocks are rich in Ca-rich plagioclase, not in K-feldspar. These rocks are not richer in K-feldspar but in biotite !
You may propose unmixing but it will be unmixing of plagioclase + biotite, plus eventually some minor K-feldspar. However this does not explain the variations of mineralogical composition of the minerals. The plagioclase is becoming more and more albitic from the more cafemic granites to the more leucocratic and probably the Ba-content o the K-feldspar decreases, because crystallizing from a Ba-poorer melt due to biotite fractionation.

"The Ti content of bulk rocks was often used as an indicator for fractional crystallization" Yes, but the sum Fe+Ti+Mg is more relevant because Ti is in very low concentration on such leucocratic granites and thus the analytical precision on Ti as a major element may not be excellent.

You write: "temperatures. The granite samples from KIB show a good correlation between the Ba content and the calculated zircon saturation temperature (Fig. 2d). As explained above the variation of zircon saturation temperature in the weakly peraluminous high-K calc-alkaline granite of Kirchberg is of about 50°C only because there was no zircon oversaturation at the contrary of the peraluminous leucogranites which exibit a large scatter of zircon saturation temperature of 150°C indicating oversaturation and variable zircon accumulation.

"They record a temperature variation of ~100°C (850 – 750 °C)" no only a little more than 50°C !

You write: "while their fine-grained rock types without K-feldspar phenocrysts do not show a correlation of Ba contents with zircon saturation temperatures (Fig. 2d). There is no inter-correlation between Ba contents and calculated zircon saturation temperatures for EIB samples" despite you have also porphyritic facies in Eibenstock pluton and in the other ones. Therefore the porphyritic character cannot explain the variations you are refering to as explained above !

You write: "KIB granites tend to the highest temperatures followed by ASB. The latest intrusion (EIB) yielded similar or even lower temperatures compared to earlier plutons (ASB, BER, KIB)" despite not based on calculations, the results with the method of P (1980) seem to be the most relevant:
- Kirchberg high-K calcalkaline granite has the highest temperature in accordance with what is generally estimated for this type of magma, and its high accessory minerla content (highest average Zr, Th, REE and Ti contents)
- Eibenstock has the lowest temperature in accordance with its strongly peraluminous leucogranitic character, and its low accessory mineral contents (lowest average Zr, Th, REE and Ti contents), the high values reflecting possible crystal accumulation because of their low solubility in such low temperature highly peraluminous melts.

You write: "(porphyritic varieties may have higher melting degrees and melting temperatures)." If you are proposing unmixing, you should have only one magma submitted to various degrees of unmixing. Her you prose a sequence of partial melts formed at different temperature, this is a totally different model of genesis ! You have to choose between the two models or discuss them separately.

You write: "all samples display large within-sample-variations both for epsilon Hf and 18O values (Figure 8a) that should also indicate some changes in sources or mixing processes" Very interesting observation. Yes such peraluminous leucogranites result from low degree of partial melting of heterogeneous mid-crustal material which may comprise sedimenst and acidic volcanic and plutonic rocks with mutiple melt generation and injection, explaining the isotopic and trace element variations between and within the plutonic bodies.
I suggest a model similar to the one proposed by Farina et al., 2012, in Mineralogy and petrology, 106, 193-216
I would be interesting to have more samples covering the complete range of geochemical variations as obtained in the work of Förster at al., 1999. The sample number per plutonic body is very limited in your paper.

p. 29: You write "Restite layers left behind after granulite-facies melting consists mainly of muscovite," I am quite surprised to have so much muscovite in granulite facies rocks which are typically anhydrous ! Muscovite is not stable in granulite facies conditions, why this mineral is observed in restite of granulites.Granulites are typically highly depleted in many trace elements hosted in biotite and muscovite. A late hydrothermal process is necessay and may have modified trace elements concentrations and secondarily enriched them in Sn and W.
I have verified this problem in the paper of WILLNER et al. 1997. J. of Petrolgy and he clearly states : "
White mica generally grows around kyanite at the expense of potassic feldspar and kyanite. The degree of this retrograde white mica formation varies strongly. Therefore these sampes cannot be used as representing pristine composition of the granulite facies rocks. The problem is to determine when this retrograde crystallization has occurred relatively to the formation of the granites !
Moreover these rocks represent very small volumes in the present outcrops. They occur as lenses in lower grade metamorphic rocks. When they were representing larger unit, they probably have had another composition.

You write: "Consequently, the source rocks were already enriched in Sn and W, but less than 10 times compared to UCC" These rocks having retrogressed with the influx a a lrage amount of fluifs (Willner et al., 1997, J of Petrolgy) its depends when this enrichment has occured before, during or after granite formation.

p. 31 You write : "All samples from EIB are
Minerals 2019, 9, x FOR PEER REVIEW 10 of 38
enriched in Sn (by a mean factor of 11) and W (by a mean factor of 5 compared to UCC" If you do not take into account the hydrothermalized samples le level of enrichment in Sn and W is similar in Eibenstock and Bergen intrusion and only slightly higher compared to the other ones (except for W).

You write: "For the EIB samples, increasing Rb concentrations (correlated with increasing F (Figure 11c) and Cs concentrations, slightly increasing Nb and Ta concentrations, and decreasing TiO2, Ba, Sr and REE concentrations) do not correlate with Sn or W, i.e. differentiation within the EIB granite did not lead to further enrichment of Sn and W."

For the Eibenstock granite you have 19 samples plotted in the diagrams but only 8 in the geochemical table given in the text. From where are coming the other ones ?
Also your conclusions arise also beause you have a limited sampling of the Eibenstock granite, and compared to the one of Förster you have not the most fractionated granites (the most leucocratic), they are among the less leucocratic, and you have only weakly mineralized greisens.
In the data of Förster there are more leucocratic (more fractionated ?) Eibenstock granites with up to 100 ppm Sn without any evidence of greisenization

To discriminate magmatic enrichment versus hydrothermal enrichment you have to use a stable parameter of magma composition variation such as Ti or better the B parameter.
For exemple in the K/Rb versus B you can observe a nice decrease of the K/Rb ratio with decreasing B (biotite fractionation probably reflecting magmatic fractionation) for all the granites except Eibenstock (a scatter of the data is observed), probably because it has been strongly affected by late magmatic/hydrothermal fluids.
The paragraph in yellow is very confusing and do not discriminate clearly what is magmatic and whai is hydrothermal in these evolutions !
The Bergen granite is transitional between the Kirchberg and Eibenstock granite types.
The Aue-Schwaerzenberg has a typical peraluminous leucogranite trend in a Th versus B diagram, with decreasing Th content with decreasing mafic mineral content at the contrary of the Kirchberg pluton.

You write; "greisen samples have neither the highest F nor P concentrations (Figure 11c, d)" This conclusion derive from the fact that the number of samples in the diagram is much higher than the one in the table from your paper.
I we take your data the altered samples have the highest Rb and F contents compared to the rest of the granites (see attached figure)

p. 33you write "In summary, there is an enrichment of both Sn and W with time in the source(s) of granitic melts, 557 but the enrichment degree is relatively low" the enrichment in the source has not to be extremely important, but over a given threshold to have the metals available for enrichment in the silicate melts during partial melting (lower is the melting degree higher will be the concentration of the matls in the melts). Then, another enrichments tep occur during fractional crystallization and/or residual crystals unmixing.
The parrallel history of uranium fractionation in such granites is an illustration of these successive steps (eg: M Cuney 2014. Felsic magmatism and uranium deposits. Bulletin de la Société Géologique de France 185 (2), 75-92).

You write: "low Ca and Na together with high K concentrations were reached in restites by melting" It is quite strange Ca is hosted in refractory minerals (anorthite, amphibole or pyroxene which tends to remain in the restitic part, Na and K typically fractionate into the silicate during partial melting !
These granulites are quite strange !

You write: "this enrichment is usually not higher than 12 times 570 the UCC level (Figures 11, 12). Therefore, we conclude that the role of fractional crystallization was often overestimated" This is due to the fact that you have not analyzed the most fractionated rock of the Eibenstock pluton !

p. 34 you write: "7], we can exclude a substantial role of intense 593 sedimentary weathering as an important control factor for later Sn and W enrichment in granite" You cannot readily exclude this source because it seems that there is no isotopic data on the Ordovician shales.

 

Minor comments :

p. 5: You write: "the initial 87Sr/86Sr ratios calculated from bulk rock data from the evolved YIC granites are very questionable because they often resulted in too low"

It is the opposite, they are typically too elevated because of secondary Rb enrichment and Sr leaching during hydrothermal alteration as shown by your greisen sample EIB612.

p. 8: You define these granites as A-type, but not only A, but A2 better named high-K calc-alkaline granites. A1 granites are peralkaline granites.

You write: "Noteworthy, only the least evolved KIB samples show geochemical similarities to A-type granites." Why are you writing that ? KIB samples are all typical high-K calc-alkaline granites, especially visble with the trends in the A-B and Th-B diagrams.

p. 25 Have you any data on Paleozoic sedimentary rocks (shales, arkoses) ?

You write: "for the greisen sample (EIB 612), the high model age resulted from very low Nd concentrations and related high 147Sm/144Nd ratios that could be caused by the strong hydrothermal overprint" why have you selected such an altered sample in this discussion ?

p. 29. You write : "Such high enrichment factors cannot be achieved by a fractionated 482 new melt batch but indicate a F-rich fluid that was able to dissolve high concentrations of chemically 483 different elements." These variations mainly result from the metamict state of these zircons due to very high uranium concentrations (up to several % !)
THis paragraph does not add anything to the main objective of your paper.

 

Comments for author File: Comments.pdf

Author Response

please see attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

Thank you for having taken into consideration the comments of the reviewers. I really appreciate your rather unique contribution on the isotopic composition of zircons, but there are still a few points which need some clarifications in the revised version of your manuscript:

(i) I just would like to try for the last time to show you that the parameters of Debon and Lefort are the most efficient to follow magmatic evolution with important consequences in the interpretation of your data.

line 235-236, the trends in Ba, Sr, Rb diagrams do not correspond only to K-feldspar fractionation but mostly to plagioclase and biotite fractionation. moreover when considering the B vs K2O diagram, provided in my first review, at the contrary K2O content increase with fractionation (decrease of B parameter) for Eibenstok, and no significant variation for the other plutons ! Such an increase is also clearly observed in the trend defined in the Q-P diagram with the sampling of Förster et al. (Figure attached) In addition, Rb content increase with decreasing K content, (most intensively in the altered samples) and is not well correlated to fractionation (see B-Rb diagram) contrarily to the other plutons. Its increase is dominated by the magmatic-hydrothermal stage (pegmatitic-hydrothermal trend of Shaw 1968). Finally as Rb being very well correlated with Sn, Sn enrichment in the Eibenstock granite is not correlated with fractionation, but with to the magmatic-hydrothermal stage. You write "We suppose that the greisen fluid was not developed by fluid unmixing from EIB melt because the
considerably higher 87Sr/86Sri ratio for the greisen sample EIB 612", however such high Sri may result from an interaction of the magmatic-hydrothermal fluids with enclosing micaschists !

(ii) Your samples are not representative of the whole evolution of the granites. I have plotted the complete set of data of Förster et al just for the Eibenstock granite (diagram Q-P attached). You can see that your sampling mostly cover part of the most fractionated granites and the trend defined by your samples is perpendicular to the trend defined from Förster et al samples. Therefore you cannot be sure to have characterized the complete evolution for the enrichment of the granites in Sn, W in particular (for example in the sampling of Förster et al., W contents vary from 4 to 84 ppm and Sn from 15 to 98 ppm in the Eibenstock granite (contrarily to what you write lines 537-539) without evidence of greisen type alteration.

Besides the Q-P diagram clearly define the high-K calc-alkaline trend for the Kirchberg pluton.

(iii) My third major concern is with the presumed source I admit that your data on zircon point out to such a source, but other sources may exists at depth which has not been charcaterized because no rocks are available as outcrops or enclaves !

the low CaO/Na2O of the granites results first from their derivation from low degree of melting at relatively low temperature (except Kirchberg which have higher ratios up to 0.65 in Förster et al. data), second from An plagioclase fractionation, and then possibly from the source. "Restite layers were interpreted as resulting from multiple melt extractions": when you have melting most incompatible elements tend to fractionate in favour of the silicate liquid ! After multiple melting episodes you have still a source with high K, Rb, Cs, U, Sn, W which is very surprising. However a metasomatic event has reenriched the source in K, Rb, Cs, with new formation of muscovite. Why this enrichment do not also concern U, Sn and W ? You invoke Sn-W-rich Ti oxides as the host of these metals, but are their abundance sufficient to explain the whole rock content, but how to mobilize these elements during an ultimate phase of melting at relatively low temperature (especially for the Eibenstock pluton, the one which is the most enriched in these metals ?). Moreover, the highest Sn and W contents in Ti-oxides are found in crystals coming from W-Sn deposits not from granulite facies rocks.

 

Comments for author File: Comments.pdf

Author Response

Please se attachment

Author Response File: Author Response.pdf