At the end of last year, a Japanese team discovered that there was more than meets the eye to a mechanochemical cross-coupling reactions catalysed by nickel salts. The paper revealed that abrasion of stainless steel during ball milling activated nickel catalysts – without the need for mechanoredox catalysts, piezoelectric materials or common reductants – challenging the assumption that the materials used for ball-milling, such as stainless steel and zirconia, are inert.
‘In nickel catalysed C–N cross coupling, you usually need a reductant,’ explains Julia Khusnutdinova, a coordination and catalysis chemist at Okinawa Institute of Science and Technology in Japan, who led the research team. ‘I started discussing with my student what could be the reductant – he proposed some crazy ideas, like it could be hydroxyl radicals and I said: “No, no, you have stainless steel – it could just be stainless steel”.’
So, her student set out to investigate; he measured abrasion and carried out detailed studies using transmission electron microscopy, as well as x-ray diffraction and a variety of spectroscopy techniques, to find out the exact cause.
They found that the mechanical abrasion was facilitated by additives commonly used in mechanochemistry, such as Celite or barium titanate. Their analysis also revealed the formation of iron- and chromium-coated particles during ball milling that likely serve as reductants that activate the nickel salts catalysing C–N cross-coupling or cross-electrophile C–C coupling reactions.
Khusnutdinova and her team say their findings serve as a ‘cautionary tale’ for the community, showing that the stainless steel vessel and balls often used in mechanochemistry are not necessarily bystanders in any reaction.
‘Comparison of different materials should become much more widespread, because even now, I see some papers where the reaction is only performed in stainless steel, but other materials are not explored,’ she says.
Unexpected reaction outcomes owing to seemingly ‘inert’ reaction materials, or rogue contaminants, are not a new phenomenon. Over the years Chemistry World has reported on glassware turbocharging the Katritzky reaction and promoting reactions in the Miller–Urey ‘primordial soup’ experiment; plasticware leaching chemicals into biological studies; and contaminated stirrer bars leading to ‘phantom’ catalyst-free reactions.
Metal-free Suzuki-coupling?
In the early 2000s, Nicholas Leadbeater, a synthetic chemist at the University of Connecticut in the US, had his own, somewhat heartbreaking, experience while working on microwave promoted reactions, specifically Suzuki couplings.
‘We’d done a lot of palladium catalysis of these reactions, and we discovered that we could use our microwave to heat our reaction mixture, safely and effectively, to 150°C,’ he recalls. ‘We found that we could do these reactions in water, which was really exciting. We could use just a little bit of the palladium catalyst, but you still got palladium in the reaction mixture.’
Leadbeater said he and his colleagues then wondered if there was a cheaper way they could do this using copper salts. ‘Lo and behold, we were able to do these coupling reactions and obtain really good yields of the product. And this amazed us … we did that with a whole load of substrates and wrote it up ready for publication.’
However, at the last minute they decided to do one final run.
‘I went into the lab and asked the student to do that … but in their haste to get this last reaction done, they forgot to add the copper and the reaction worked. We got an identical yield.’
They thought that they had discovered a metal-free Suzuki coupling reaction.
‘We thought … “this is groundbreaking”, because you don’t need any metal salts,’ says Leadbeater. ‘That has advantages in that it’s cheaper, but also, if you’re working in the pharmaceutical industry, one of the things you’ve got to do is get the metal out of your product before you turn it into a drug molecule to give to people.’
After they published the paper, Leadbeater was invited to speak at several high-level conferences, and big names in chemistry started coming up with possible mechanisms to explain what they had found.
It was at that point that the chemistry department in the UK which Leadbeater worked in closed and he headed to the US. However, when he got there, he found they were unable to reproduce any of the metal-free reactions.
‘Nothing worked. And this set me into a state of absolute panic … our fundamental discovery wasn’t working!’ says Leadbeater.
After some detective work they discovered that in the UK the sodium carbonate base used in the reactions comes with around 50 parts per billion of palladium, whereas in the US, it doesn’t.
‘So, what we’d actually discovered in the UK was a Suzuki coupling reaction, but one that you do with ultra-low levels of palladium – it was a contaminant in the sodium carbonate that was doing the reaction,’ explains Leadbeater. They rushed to publish what they had found before anyone could beat them to it and Leadbeater says he was convinced he would lose his job and that no one in the chemistry community would want anything more to do with him.
Thankfully that wasn’t the case. ‘People were very understanding, and then other people started coming out of the woodwork with things that they discovered with very low levels of metal,’ he says.
The experience fundamentally changed the way Leadbeater worked, as well as his mindset. He and his colleagues tried to spearhead an effort to highlight what had happened and advise other chemists to be careful to complete all the necessary tests before positing that they have discovered something new.
‘If it’s too good to be true, and it is probably too good to be true, you’ve got to make sure that you dot every i and cross every t, and even then you’ve got to look at it initially with a sceptical view,’ he says. ‘We now analyse absolutely everything.’
‘Another thing you could do is you could find friends in different places around the world, give them your procedure and say, “Hey, do this”. If we’d done that, then maybe we would have been saved the heartache.’
Of course, in chemistry not all unexpected outcomes caused by impurities are a dead end – some can lead to significant breakthroughs. In the case of the discovery of Ziegler polymerisation, contamination was the key.
An accidental breakthrough
‘[Karl] Ziegler was working on aluminium alkyls, and he was studying what is called the Aufbau reaction … adding ethylene to aluminium alcohols, and then later it would oxidatively be cleaved, and you get aluminium hydroxide and fatty alcohols – both interesting products,’ explains Ferdi Schüth, an expert in heterogeneous catalysis at the Max Planck Institute for Coal Research in Germany. ‘They were systematically looking into these reactions, and then they discovered in one batch that they didn’t get any higher aluminium alkyls, they got butene out of it, which was totally unexpected.’
After a strenuous investigation, the researchers found trace amounts of nickel in cracks in the steel autoclave – likely from a previous experiment or cleaning process. At this point, many scientists might have resorted to cleaning the autoclave before continuing their work, but Ziegler took a different path.
‘Ziegler then initiated a systematic study of metal impurities in this reaction,’ says Schüth. ‘He explored titanium, and that then was one of the really first experiments, from all accounts, where they got a solid block of polyethylene in the reactor.’
Schüth says that researchers should always be on their toes. ‘I often tell my students that the expected outcome of an experiment is boring – the unexpected outcome is more interesting … it can be turned towards discovery, or sometimes it may be a discovery, but it’s unpredictable, and it takes the keen eye to see if it’s worth looking more deeply into it or not. Like Ziegler did.’
Vessel effects kept under wraps
For Michael Martin Nielsen, a chemist at the University of Copenhagen, his fascination with the effects of vessel materials, such as glassware, started as an undergraduate when he became absorbed in the scientific literature. He came across a paper published by Teruaki Mukaiyama and colleagues in the 1980s, who had done a lot of work on glycosyl fluorides as electrophiles.
‘As I read these papers, I realised they never noted how they quenched or scavenged the hydrogen fluoride that’s obviously developing during this reaction and I figured that it was probably so obvious that no one bothered to describe it.’
He eventually went to his supervisor, concerned he was missing something, but he just thought it was something that people hadn’t really thought of before.
‘This is crazy; these are reactions that have been run thousands of times. And what makes it even more interesting is that it’s really well known that the yields of these reactions are highly dependent on the addition of large amounts of molecular sieves, zeolites which are used to scavenge water from organic solvents.’
Nielsen decided to carry out his own set of parallel experiments in both glass and Teflon and found that, to his surprise, the reactions worked much better in Teflon. ‘In regular glassware, without molecular sieves, I hardly observed any product, I just got hydrolysis. Whereas in Teflon, I actually got pretty nice yields, but in certain cases, I also observed significant decomposition.’
‘I then later found out that the hydrolysis observed in glass was caused by hydrogen fluoride reacting rapidly with the glass surface, essentially dissolving it, to form boron fluorides, silicon fluorides and water,’ Nielsen says. ‘On the other hand, the reactions in Teflon would eventually contain highly concentrated hydrogen fluoride, as well as the Lewis acid used to initiate the reaction, which caused some substrates to completely decompose.’
He ended up looking at another reaction type involved C–F-bond activation using benzylic fluorides and found that again, they had fewer issues in Teflon than they did in glass. ‘Fundamentally, what made it interesting for me was not the immediate impact … it was more that I was just surprised by something so obvious being ignored.’
What should the chemistry community do?
Nielsen ended up putting together a review, with his colleague Christian Marcus Pedersen, in which they explored ‘vessel effects’ in organic chemistry and made suggestions on when and how to investigate them.
‘The main reflection for me from writing this review, beyond that fluoride obviously reacts with glassware, is that we really have to be wary of what we use our glass for … in all cases, we are essentially bringing in some unknown factor to our reactions … there’s a lot to be discovered here.’
Nielsen and Pedersen do not believe that fluorinated or non-fluorinated plastic polymers generally provide a superior vessel material to borosilicate glassware or other glass types, however they do suggest implementing a variety of controls during reaction development.
These include running parallel experiments in hydrogen fluoride-resistant polymer materials during reaction development to assess vessel effects, storing organofluorine compounds in polymer-based containers and controlling for vessel effects during reactions using low catalyst loadings of either acid or base catalysts, by running parallel experiments in a plastic polymer reaction vessel.
‘I don’t think our recommendations are as useful as the general observation that the vessel material is an unknown … even if we know it’s glass, or even if we know it’s plastic – we should be wary of hydrogen fluoride; we still don’t know what we’re introducing, especially if it’s a non-pristine material.’
Even now, Nielsen says there’s no clear paper trail of people citing each other for finding vessel effects and that the phenomenon has likely been discovered 20 times in isolation, described and then forgotten about. ‘I referee papers on this where people are completely unaware that similar stuff has been described before – I think it’s really interesting.’
For Leadbeater, he feels it is important for the chemistry community to help ‘police’ its research output. ‘There’s now a proliferation of new journals and … work is sometimes getting submitted to some of these journals and then accepted and published that really isn’t done to the level that you need to do – it sort of pollutes the literature.’
Following her abrasion discovery, Khusnutdinova feels strongly that, rather than trying to avoid situations where contaminants may arise, chemists need to be more critical of their own results. ‘We should teach students to carefully think about the design of control experiments in such a way that they actually provide the answer, even if this answer is not what we expect,’ she says. ‘[We need to] consider other factors that we don’t expect to contribute – thinking even beyond your own original idea – because it’s chemistry, and the best results we get unexpectedly, by serendipity.’
No comments yet