Origins of Batch-to-Batch Variation: Organoindium Reagents from Indium Metal

Abstract

Yields of organoindium reagents synthesized from indium metal were previously reported to be highly dependent on metal batch and supplier due to the presence or absence of anticaking agent. Here, single-particle fluorescence microscopy established that MgO, an additive in some batches nominally for anticaking, significantly increased the physisorption of small-molecule organics onto the surface of the resulting MgO-coated indium metal particles. An inert and relatively nonpolar boron dipyrromethene fluorophore with a hydrocarbon tail provided a sensitive probe for this surface physisorption. SEM images revealed markedly different surface properties of indium particles either with or without MgO, consistent with their different physisorption properties observed by fluorescence microscopy. We further documented incomplete commercial bottle labeling regarding the presence and composition of this anticaking agent, both within our laboratory and previously in the literature, which may complicate reproducibility between laboratories. Trimethylsilyl chloride pretreatment, a step employed in a subset of reported synthetic procedures, removed the anticaking agent and produced particles with similar physisorption properties as commercial batches of indium powder distributed without the anticaking agent. These data indicate the possibility of an additional substrate/catalyst physisorption mechanism by which the anticaking agent may be influencing synthetic procedures that generate organoindium reagents from indium metal, in addition to simple anticaking.

Graphical Abstract

The United States National Institutes of Health (NIH) recently implicated a reproducibility crisis in the biomedical sciences.^1,2 As part of addressing this crisis,³ the NIH called for improved authentication of key research resources.^1,2,4,5 An Organometallics roundtable similarly highlighted the need for improved experimental reproducibility when the source or supplier of reagents is changed.⁶ The (im)purity of reagents has a well-documented, sometimes-sordid/sometimes-lucky history in metal-mediated and -catalyzed reactions.^7–14 Here, we examine the origins of batch-to-batch variation in the synthesis of organoindium reagents from indium metal powder. Organoindium reagents are particularly useful in organic synthesis due to their ready participation in cross-coupling reactions and their higher functional group tolerance compared to alternative organometallic reagents.¹⁵ Generation from indium metal powder and organohalides offers the simplest and most atom-economical route to these organoindium reagents. In impressively careful studies, Minehan and Yoshikai separately noted that the specific batch of indium powder significantly influenced yields of organoindium reagents produced through this route.^16,17 Both researchers traced this difference to the presence or absence of anticaking agent: Batches of indium with anticaking agent produced higher yields of organoindium reagents than those without (Figure 1a).

Figure 1.

a. Reactions with yields dependent on indium batch. b. Acros indium bottle label stating 99.999% purity (no mention of anticaking agent additive) and of Aldrich indium stating 99.99% purity.

We became interested in if single-particle fluorescence microscopy and SEM data could aid in identifying differences in these batch-to-batch behaviors beyond simply the macroscale effect of reducing metal conglomeration. A combination of single-particle fluorescence microscopy and SEM herein establishes that the anticaking agent increases the physisorption of organic compounds to the surface of the indium metal powder. This effect provides an additional potential mode of impacting reactivity of indium metal towards organic and organometallic species during these reactions.

In addition, through examination of the literature reports, and of indium bottles from different suppliers in our own laboratory, we identified three factors problematic to reproducibility:

1. Some commercial bottles that contain the anticaking agent do not state so on the bottle label. For example, Figure 1b shows a photo of indium metal powder from Acros, produced in 2016, labeled as 99.999% indium. Although the presence of an anticaking agent is not listed on the bottle label, batch analysis information obtained from the supplier upon request states that it contains 2% MgO anticaking agent (certificate of analysis in SI). Further, this purity is reported on the bottle label as that for “trace metal analysis”, and yet the anticaking agent contains a metal (oxide), thus the phrasing of this purity statement is conceivably easy to misinterpret.

2. A parallel gap in commercial labeling has muddled the correct identification of the chemical composition of the anticaking agent. Both Minehan and Yoshikai referred to the anticaking agent as “Mg” in their reports, concluded from supplier information available at the time; indeed, certificates of analysis from Aldrich state this indium was distributed with “~1% Mg” as recently as 2017 (see SI for certificate). Here, a labeling gap appears to have arisen from acknowledgement of just the (trace) metal in the anticaking agent, rather than its chemical composition as the oxide. That assessment is consistent with certificate of analysis information that our laboratory obtained from the other supplier, Acros, and with website information,¹⁸ stating that the current anticaking agent is “MgO” in Acros batches of indium. Underscoring the scientific complications thus caused, Minehan attempted a control reaction wherein 1% magnesium metal was intentionally added to an indium metal sample that did not contain anticaking agent.¹⁹ This addition failed to reproduce the effect of the anticaking agent. We suggest retrospectively that this control experiment failed because the true anticaking agent was MgO and not Mg.

3. Which supplier’s indium batches contained or did not contain anticaking agent changed with time: Yoshikai noted that Aldrich was the supplier of a batch of indium containing anticaking agent.²⁰ Minehan also found that indium metal from Aldrich contained anticaking agent at the time and produced the highest yields. According to the current catalog, however, Sigma–Aldrich does not presently distribute indium with anticaking agent.²¹ In contrast, indium with anticaking agent is available currently through Acros.¹⁸ This observation underscores that the distributor/supplier information alone, as typically provided in experimental descriptions, is insufficient for robust reproducibility.

In order to explore differences in the surface properties of indium metal batches, we used a combined single-particle fluorescence microscopy and SEM approach. This approach enabled examination of the differences in surface composition and physisorption properties between two different batches of indium metal that both had bottle labels stating ≥99.99% pure (Figure 1b).

To perform the fluorescence microscopy experiments, the different batches of indium metal were treated with a solution of fluorescent probe 1 in THF (2 h, ambient temperature, Figure 2a). This sensitive probe enabled assessment of the surface physisorption properties of different samples of indium metal towards small-molecule organics. Probe 1 was selected for these studies because of its high quantum yield, chemical inertness, and solubility in the organic solvents^22,23 used by Minehan and Yoshikai (Figure 2b). Probe 1 features a boron dipyrromethene (BODIPY) core with an alkyl chain and serves as a model for the small-molecule organic substrates and/or catalyst–ligand complexes previously used in synthetic studies. The fluorescence intensity on the surface of indium particles correlates with the quantity of 1 on those surfaces. This relationship enables comparison of the degree of surface physisorption between different batches of indium metal: higher physisorption produces greater fluorescence.

Figure 2.

a. Preparation procedure for fluorescence microscopy imaging. b. Chemical structure of imaging agent 1. c–e. Fluorescence microscopy images of each representative particle of Aldrich indium and both TMSCl pretreated and untreated Acros indium. d–h. SEM images of each representative particle of Aldrich indium and both TMSCl and untreated Acros indium.

Figures 2c–e display the data from these fluorescence microscopy studies. A sample of indium metal from Aldrich, which did not contain anticaking agent, is completely dark (Figure 2c). This image does not show evidence for significant physisorption of 1. In contrast, a sample of indium from Acros, that did contain anticaking agent, showed substantial physisorption of 1 (Figure 2d). This physisorption is indicated by the widely distributed bright green “hot spots” of 1 visible on a representative single indium particle in the center; other similarly coated indium particles are visible elsewhere in the image. The heterogeneous distribution of fluorescence on single particles, wherein some areas are bright green and other areas are relatively dim, reflects the heterogeneous surface properties of single particles.

An additional variant of synthetic procedures was provided by Knochel in 2008, in a report on preparing organoindium reagents from indium metal.²⁴ This report listed ChemPur as the supplier for the indium metal, and did not mention the presence or absence of anticaking agent. Of particular interest to our laboratory, this report described pretreatment of the indium powder with trimethylsilyl chloride (TMSCl) and 1,2-dibromoethane. TMSCl has been previously shown to remove surface oxides from metal powders.²⁵ Indeed, when indium metal that contained anticaking agent was pretreated with TMSCl (~0.4 M solution in THF, 2 h), the powder exhibited markedly reduced physisorption of 1, instead showing fluorescence intensities similarly low as the batch supplied without anticaking agent (Figure 2e). This data bolsters the ideas that the anticaking agent is the cause of the increased physisorption, and that anticaking agent is removed upon TMSCl pretreatment.

Thus, the protocol employed by Knochel proceeds under conditions effectively without anticaking agent, because the pretreatment step would have removed it even if it had been present in the original sample. This reported procedure appears to contrast with the methods reported by Minehan and Yoshikai, in which optimal yields were obtained in the presence of anticaking agent. It therefore seems plausible that pretreatment steps in the Knochel protocol may also remove an indium oxide or carbonate layer in a step unrelated to anticaking agent. Such roles for pretreatment steps separate from anticaking agent remain speculative and are not addressed directly with our present study.

SEM analysis of the batches from different suppliers and with different pretreatments showed marked variation (Figure 2f–h), consistent with observations from fluorescence microscopy. SEM comparison between the sample from Aldrich that did not contain anticaking agent, and the batch from Acros that did, showed the following clear differences: The indium powder from Aldrich consisted of smooth, roughly spherical particles (Figure 2f). In contrast, the indium powder from Acros consisted of irregular particles substantially coated with a flaky material, assigned as MgO (Figure 2g). TMSCl pretreatment of this sample, however, removed the flaky coating, leaving behind a heavily etched indium surface (Figure 2h). These data established a correlation between the presence of anticaking agent in the indium samples as observed by SEM and increased surface physisorption of organic molecules as observed by fluorescence microscopy. This observation strengthens the assignment that the increased physisorption observed in the indium batch containing anticaking agent is caused by the anticaking agent.

To further investigate the assignment of increased physisorption as due to the properties of MgO, we carried out a control experiment with authentic MgO powder and probe 1 (Figure 3). In this control experiment, MgO powder was treated and analyzed by fluorescence microscopy in the same fashion as the indium powders had been previously, per the procedure shown in Figure 2a. The resulting powder showed very high fluorescence, such that individual particles of MgO could not be resolved (representative image, Figure 3). This observation is in line with prior reports of physisorption of small-molecule organics to MgO.^26,27 This high physisorption of probe 1 on an authentic sample of MgO powder is consistent with the assignment of the anticaking agent as the cause of the differential surface physisorption in the three samples of indium, which were different by batch or pretreatment step.

Figure 3.

Bright fluorescent microscopy image of a representative MgO sample, showing high physisorption of 1.

These physisorption results are briefly summarized in Figure 4a. Combined, these data establish that the surface properties of the indium powder are highly dependent on batch and on (pre)treatment. This dependency raises the possibility that the anticaking agent influences reactivity through physisorption of substrate or organometallic catalyst, in addition to its established role in preventing metal conglomeration on the bulk level (Figure 4b). Prior reports of electron transfer through thin coatings of MgO,²⁸ for example, suggest one such possible route for this coating to influence reactivity between the indium metal and physiosorbed substrates or catalysts, although establishing a specific route is not the thrust of the studies herein.

Figure 4.

a. Schematic summary of fluorescence microscopy experimental results. b. Visual representation of organic substrate and catalyst physisorption to indium vs. no physisorption.

In conclusion, robust reproducibility of organometallic reactions hinges on the documentation, characterization, and authentication of metal reagents. We here unravel origins of differences in batches of indium metal powders, influenced by supplier, year, and pretreatment step. Existing literature documentation, as well as examination of our own in-laboratory commercial indium batches, indicate that bottle labels and certificates of analysis detailing composition have been incomplete. Paths forward to address these reproducibility considerations include: more comprehensive labeling and composition information from distributors; inclusion of batch number in reported supporting information; increased use of materials characterization techniques during synthetic methods development to authenticate resources, understand compositions, and achieve the NIH mandate;⁵ and mechanistic and synthetic considerations of surface effects resulting from batch additives and subsequent pretreatment steps. In the present case, single-particle fluorescence microscopy and SEM data show that the presence of anticaking agent changes the surface physical features and physisorption properties of the indium powder, suggesting a possible additional influence of this additive beyond its simple role in anticaking.