I was teaching a grad course on principles of Materials design, and was explaining the students how lattice anharmonicity in materials leads to seemingly unrelated properties such as coefficient of thermal expansion (CTE) and electrostriction. Around the same time, we had an ongoing collaborative activity in our groups at CeNSE to integrate epitaxial BaTiO3 onto Si. Sandeep Vura was the grad student leading that effort. On some of these samples, which were synthesized at "less-optimal" conditions for ferroelectric BaTiO3, we observed that their CTEs are quite large. These samples were also defective, and non-stoichiometric. Along with Subhajit, a post doc that recently joined my group, we thought, we should check these samples out for their electrostrictive behavior. Apply voltage, measure strain. Sounded simple. Available resource for us, thanks to our departments strong MEMS activity, is a laser doppler vibrometer. Problem at that time was, all over the world, no one reported electrostrain measurements from this tool. Subhajit modified the set up to cater to our samples that are on a substrate (and not MEMS devices), and showed us that such sensitive measurements can be done, first. 

Our BaTiO3 thin films are grown on Si through buffer layers. One needs to be vary of contributions of various layers in this complex structure to the measured properties. With an awareness of these complexities, Subhajit and Sandeep started measuring the electrostrain in these samples. Initially, they saw "butterfly" loops, typical of ferroelectric materials. Only that our defective Barium titanate is not ferroelectric. Enter Shubham, a first year grad student at that time, who was itching to do experiments! This was also at a time that both Subhajit and Sandeep's stint was ending in the lab. All of us realized that any second order response, should give butterfly hysteretic loops, if there is a phase difference between input and output. 

With this new found realization, there was indication that the effect that we see, if it is electrostriction, is HUGE (and anisotropic)!! Shubham took over, showed that this is a consistent effect over samples of different thicknesses, came up with data analysis protocols to filter out the first order, second order and higher order strain effects. He convincingly showed that we are indeed measuring large electrostrains at ultrasmall electric fields. 

The next challenge is to see if this large strain is being produced by the electric field itself, or due to some form of current-induced heating (which is also a second order effect). My good friend Saurabh Chandorkar and our joint student Upanya decided to help us with electrothermal simulations, and also some preliminary measurements of device temperature using IR thermometry. All this pointed out to the fact that indeed what we are seeing is a giant electrostriction, upto frequencies of ~10 kHz, which was not seen before. 

Shubham then went into a superman mode, and with help from several of the other co-authors (including Bart and Majid from Groningen) did systematic experiments to understand the possible reasons from structural, chemistry and electrical properties point of view that correlate with the giant electrostriction that we have run into. With all this ammunition, and really convincing ourselves of these defect injected effects, we decided to report these results. 

Submitted, waited for 4 months. Finally the peer reviews came in from Nature Communications. I must say, it is peer reviews (which are available to read now) like this that make me believe in the scientific process. In my limited experience of publishing papers, this is the best peer review we ever saw, and I am glad that this was some of my students first experience. Rigorous critique, but every question aimed at clarifications and suggestions to solidify our claims. Working on the peer review, and designing experiments asked of us by the reviewers was a great learning process for all of us. 

Up until now, we argued that our huge electrostrains at low electric fields are a result of large electrostriction, based on primarily eliminating all the other possibly mechanisms that can give rise to similar signals. The referees noticed this, and suggested that we do a converse measurement i.e. apply stress and look at the change in susceptibility (capacitance). Magic of thermodynamics means that the ratio of change in susceptibility (electric property) to change in stress (mechanical property) is the same electrostriction coefficient, which is the ratio of generated strain to the applied electric field. However, this converse measurement is not affected by all the other factors such as Joule heating etc.. that could obfuscate the interpretation of direct electrostrain measurement.

Shubham and Upanya set up these unconventional measurements in several ways: a) nanoindentation to apply stress on devices, while simultaneously measuring change in capacitance, and b) bending which applies homogenous stress on the films while simultaneously measuring change in capacitance (as a function of stress). To our awe, both these measurements yielded values of electrostriction which solidified the giant electrostriction claims from earlier measurements. An issue at this point could have been availability of samples to test out these new measurements, now that Sandeep had left to work with industry. Rama put up his hand and took the responsibility of generating  several copies of the samples within record times.

Over the next two reviews, all the reviewers believed in the existence of this effect, but we debated the exact mechanisms at play. What was quite satisfying is the quality of the scientific debate, the scientific temperament to agree when proved wrong and strongly making ones point, especially if it is corroborated by the data. In this ever stressful world, these processes make us happy that we are in a noble profession of doing science. 

After about 1 year of submission, and back and forth reviews, when we heard about the acceptance of the manuscript, we were very happy. Our work was baptized by fire, but came out stronger. We now understand what next, and have a systematic long-term research plan chalked out to develop on these interesting results. 

In any case, this is the backstory. One take home message about this work is that material defects are not useless entities that need to be eliminated. One can engineer them effectively to create unconventional properties in materials, which otherwise do not exist. Our non-stoichiometric Barium titanate, and its giant electrostriction is a poster child of this philosophy.

One last thing with which I will sign off is to ask the readers to look at the author list on this paper. Some of us jokingly remark that half of CeNSE (our department in IISc) is in this. You'll find MEMS experts, device people, materials scientists everyone contributing in this. I personally thank this wonderful ecosystem, that lets us do some nice work.