Systemic issues in academic technology research: Case of lithium-ion batteries

Enough is enough. Having sifted through countless academic publications over the last few years, we have some observations and comments we want to share with the research community. 

Academic research is invaluable for society (that’s why our whole enterprise is dedicated to disseminating its outputs), but in order to thrive, researchers need to continuously reflect on its structure, evolution, and purpose. While such discussion is driven mostly by voices from within academia, feedback from “the outside” is equally important. This analysis, from an outsider’s perspective, will make some researchers cheer and make others uncomfortable, but our basic goal is to make everyone think.

Who are we speaking to?

The target audience is technology researchers. If you work on batteries, great! If you work on photovoltaics, sensors, fuel cells, or any other technology, the trends in your field are likely to be similar (if not the same). Most examples in this article relate to chemistry and materials science, since these disciplines produce the vast majority of publications.

Who are we to say?

For almost 3 years we’ve been running the Keeping Up with Batteries literature service to disseminate the latest battery science and engineering research. Each month, we examine over 100 journals and evaluate over 1,000 articles. Covering everything from fundamental battery operation to deployment of electric vehicles, we have a uniquely broad perspective on the field.

Are we biased?

We are neither from academia nor from industry, and have no stakes in any particular field or technology. We care about research with the potential to make an impact. Our main goal is to disseminate academic literature to engineers, consultants, policy makers, as well as academics themselves. We simply believe that technology research should be useful and used.

Race between quantity and quality

Most researchers would agree that quality beats quantity. Unfortunately, only some internalize this principle. When it comes to research and publications, much of academia has gone down the path of quantity – embracing impact factors, article counts, citations, and h-indexes. These are numbers that can be easily measured and compared. Moreover, many choose to believe they equate to scientific impact (even reward them through various financial or career-promoting incentives). The love of numbers is inherent to science, and so is the love of research areas where these numbers are on the rise. Although such areas can be urgent and important, like COVID-19 or clean energy, the amount of information they generate is expanding beyond what many researchers find manageable.

We have just witnessed an explosion of COVID-19 research, which, only a few months into the pandemic, was already producing thousands of new publications every week. Naturally, discussions about quality and quality followed. Battery research is different in magnitude but similar in nature. The rise of electric vehicles and grid energy storage has placed the lithium-ion technology at the frontline of the fight against climate change, sending reverberations through industry and academia. For most battery topics, the number of publications has approximately doubled in the past five years, and for some, it has tripled. Considering standard publication trajectories, these same topics will likely grow another 50-100% by 2025 before slowing significantly or giving way to newer ones. To be specific, there were almost ~28,000 Dimension.ai (and ~38,000 Google Scholar) new references with the term “lithium ion batteries” published only in 2019. That is THOUSANDS a month.

Are you a good reader? The famous novel To Kill a Mockingbird has around 300 pages and ~100,000 words. Scientific articles have between 2,000 and 10,000 words. This means that reading just 20 regular-sized (5,000 words) articles is like reading a full novel. For context, there were roughly 1,750 articles that discussed NMC (or NCM) cathode chemistry only in 2019 (as indexed in Dimensions). We invite you to do a small conversion exercise for all research topics you need to follow.

What is the message? First – researchers struggle to keep up with this amount of literature. According to our internal survey, an average scientist or engineer (academic and non-academic) reads around 20 new articles each month, spending an estimated 10 hours searching and reading new literature. That often includes patents, reports, and news. Unless one cuts on experiments, teaching, writing, or grant applications, the time dedicated to reading is unlikely to increase no matter how many more articles are being published.

Second – authors should consciously prioritize quality over quantity. We not only see countless sloppy publications with typos in the title or fragmented sentences in the abstract, but also countless permutation or multiplication of ideas (please read a pointed critique by Pumera et al. to get the picture). The overlap in content grows with the increasing number of publications. Maybe one needs to reconsider what counts as a meaningful contribution and what is actually worth his/her time, effort, and funding money. Maybe it’s time to raise the bar.

“Nanoflowers” is trending

Scientists, like all humans, are attracted to new trends. Trendy research topics capture imagination and earn immediate citations. (And who doesn’t like a quick “profit”?) Such topics either have enough vision and develop into a new area, or more often, they don’t have enough vision yet self-perpetuate due to the sheer momentum of their supporters. To avoid the latter, each scientist and engineer should periodically hit pause and self-reflect. Aristotle once said, “Knowledge of the fact differs from knowledge of the reason for the fact.” In other words, can one justify their research agenda? Is there actually an essential question one is answering or a practical problem one is solving? Or is he/she doing it just because everybody else is?

Here is a case study to illustrate the phenomenon. We compared the number of articles that contained the word “understand” (in any form) with the number of articles that contained the word “graphene”.* Results: there were fewer scientific publications using the word understand than those mentioning graphene. To rephrase it – the word that denotes basic research (even on graphene) and drives the whole scientific discourse can’t even match up to the popularity of a single trendy topic. (Note: it’s not our goal to single out or criticize any specific area of research. This example merely illustrates the hype phenomena that are common in science and are sometimes scrutinized by researchers themselves.)

// *Methodology: Li-ion battery articles published in 2019 and indexed in ScienceDirect.

Another case compares established cathode chemistry “LFP” (or LiFePO4) with a newly trending term “nanoflowers” (or any nano-, micro-, and flower variation). What would start as an unusual term to describe a specific morphology (flower) has soon become a phenomenon of its own. (Yes, someone can object that this search methodology is rather simplistic, but whatever the approach, the resulting statistics are not far from the unsettling reality.) Unsurprisingly, there are now more articles discussing nanoflowers and microflowers than those discussing one of the most instrumental battery chemistries science has produced. (And the number of original works on LFP is likely overestimated since the material is predominantly used for reference purposes.) Isn’t it ironic that industrial researchers at Tesla and CATL still see enormous potential where academia sees almost none?

Nevertheless, we want to highlight an important observation – doing unfashionable, hard, deliberate, and tedious research is rewarding. Although such research might not attract as much short-term academic attention as trendy research does, it might lead to a more significant long-term impact. Let’s compare scholarly and patent citations of an average article that has the words “nanosheet(s)” vs. “understanding OR mechanism” vs. “LFP” (or LiFePO4) in their title.*

// *Methodology: Li-ion battery articles published between 2010-15 and indexed in Lens.org. The keywords were chosen to form categories with sufficient article counts ≳250. The earlier 2010-15 time period was chosen to allow >5 years for articles to produce patent citations. The scholarly and patent citations were averaged across the 2010-20 period

The average article focusing on LFP attracted approximately double the patent citations than the average one focusing on nanosheets. Moreover, the LFP articles led to mainly granted patents with large patent families, indicating an intent to commercialize, while the nanosheet articles led to mainly pending patent applications (or patents with small patent families), indicating an intent to solely protect the intellectual property. Both trendy and non-trendy topics fared well – the former in citations, the latter in innovation impact.

Finally, there are less obvious trends that sway academic technologty research. For instance, scientists and engineers like to embrace complexity. Most journals are filled with articles that introduce overly complicated designs, promote extremely toxic and scarce materials, rely on arrays of convoluted methods, and develop, rather unconsciously, a plethora of solutions that have almost no chance of being considered for real-world applications. The novelty appears to get a priority over everything else. Advice to those striving for impact: keep it simple (here is an idea). To quote Da Vinci, “Simplicity is the ultimate sophistication.”

Image courtesy of Springer Nature (Schultz et al. 2020).

As another example, researchers are prone to think in a “bubble”. By following the latest research trends, reading the work of their peers, and attending academic workshops and conferences, they can shape their ideas without any consideration for the rest of the innovation chain. In fact, it’s not unusual for academic communities to set performance goals and directions that don’t correlate with practical demands. How to avoid this? One can engage with stakeholders outside academia, and read literature that goes far beyond his/her research domain (e.g. a battery chemist peeking into manufacturing projects, techno-economic analyses and market adoption studies, or relevant policy papers).

Academics are not industrial experts

Academics don’t collaborate enough with industry. Despite the benefits, such as funding diversification, exposure to business practices, or even job opportunities for students and postdocs, only a very few university researchers establish these connections. To be fair, corporations and start-ups are not always keen. There are numerous compromises and limitations that all sides need to negotiate (especially scientific independence, publication freedom, and intellectual property), but a well-structured partnership is often a win-win situation for everyone. (To get a better insight into corporate–academic relationships, we recommend reading this collection of articles by Nature.)

Battery researchers could have tapped the recent explosion of commercial interest in Li-ion battery technologies, yet based on a publication search (in Lens.org), only ~1.5% of battery articles between 2015-2019 were co-authored by an industrial partner. Scholarly publications might not be the only output of such collaborations; nevertheless, they do indicate the level of industrial involvement in creation of knowledge. And this level of involvement seems almost negligible (in 2019 alone, it was unimpressive ~1.3%).

The critical aspect of academic-industrial collaborations is mutual learning. For example, businesses gain access to state-of-the-art expertise, and in return, academics receive (for lack of a better word) a reality check. From our observations, articles co-authored with industrial partners are highly relevant, informed, focused, and subsequently well read and cited. These articles are as frequently fundamental as they are practical. They represent a unique voice and should be welcome in any engineering-centered field.

One of the many consequences of vanishing industrial involvement in publishing is that academics regularly step in and claim the authority to judge what is feasible, scalable, low-cost, or commercially desirable. Unfortunately, such authority often rests on shaky foundations, ranging from “common sense” to a superficial and sometimes misunderstood experience of the private sector. This weakens academic credibility. As much as a scientist wouldn’t promote unfounded speculations about their data, he/she shouldn’t take the liberty to speculate elsewhere. Science doesn’t necessarily conform to superficial common sense, so why should the business world?

This issue is not a fringe one. Academics making claims for the industry is a phenomenon that is enormously widespread and troubling. It skews scientific discussions, influences funding decisions, confuses the public, and alters the fate of whole research areas. (And you are correct to make a connection with the discussion above on hyped and trendy topics in academia). This bad practice applies to everyone who makes commercial claims about their material, approach, device, or algorithm without having the depth of relevant business knowledge or having consulted with a relevant industrial expert.

For instance, one can’t claim that his/her new material is “low-cost” unless they have estimated its total production cost and compared it with industrial standards. Are they aware that even “abundant” nickel has long been flagged in commodity markets and faces various supply risks? Do they know they need to do a proper analysis before promoting their organic compound as “cheap”? Have they ever conducted a scale up of a manufacturing process, preferably at an industrial level, to understand the word “scalable” before featuring it in the title of an article?

Free information means wider impact

Free information goes further than restricted information. As many are aware, there is currently a full-blown war that universities and funding agencies wage against publishers in order to liberate science from behind expensive paywalls. And unless one has spent their entire career at a few elite institutions in a few privileged countries, it is clear how much is at stake. It’s not only the non-elite universities that struggle with access to journals; it is government institutes, companies of all sizes (including tech start-ups), nonprofit organizations, think tanks, policy makers, consultants, journalists, and many others. All these people are unlikely to read an article that is not accessible.

Take a researcher at a small tech company – he/she finds a recent study relevant to their project and wants to glance through it. Unfortunately, it won’t be a click-and-read procedure. It will be any combination of filling out purchase forms, talking to a supervisor or a colleague, contacting administrative staff, waiting for approvals and/or signatures, going through a complicated procurement software, dealing with available funds and budget codes, only to proceed to the purchase and finally the reading. This process often involves different people, long waiting times, or significant disruptions to one’s daily work (not to mention that many organizations won’t even be willing or able to pay $50 per article). Another option is to send the author an email, which can go unanswered for weeks.

The simplest way for authors to have an impact outside academia is to enable people to read their work. One approach is to publish open access. There are well-documented benefits for choosing this option, yet only a fraction of researchers choose to do so. We looked specifically at the Li-ion battery literature between 2010-2015 and found that just 16% of the articles were published under an open access license (the percentage is comparable to general fields of chemistry and engineering – i.e. some of the lowest in all academic disciplines). However, these open access articles later attracted considerably more patent and scholarly citations than those behind paywalls (averaged across the 2010-2020 period). That’s a good deal.

This call is for everyone: Make your research accessible – either as open access, or archived in a preprint/postprint repository. If some publishers make it overly expensive to choose open access or make it difficult to use repositories, publish somewhere else. (Make sure to consult this list of potential predatory journals and publishers.) There are probably a handful of alternatives for each specific paper, including decent publishers who support access to information. You don’t have to compromise. Just remember that the next time someone comes across your paywalled article, your hard work might not be the one to help the next research project, technology invention, or government policy.

Key takeaways

Here are a few items each technology researcher should reflect on:

  • Publish quality over quantity (read),
  • focus on important scientific questions or engineering problems (read),
  • collaborate with industry, if possible (read),
  • and make results accessible (read).

Comments?

We are open to constructive discussion. Please send any feedback or comments to info@researchinterfaces.com. We will be happy to update the article if necessary.

Acknowledgement

This article went through an informal peer-review process. We thank the reviewers for their time and valuable feedback.


Keeping Up with Batteries

Monthly snapshots of the latest research, news, and analytics on Li-ion batteries.