Chapter 164: The World Sees the Light
12 August 1974 — 15 September 1974Gorakhpur; Cambridge, Massachusetts; Eindhoven; Tokyo; Delhi; Bombay; London; Stanford, California
The Physical Review Letters paper appeared on August 12th.
Not dramatically. Not announced. It appeared in the journal the way papers appeared in journals: as one item among several in an issue, preceded in the table of contents by a paper on quantum chromodynamics and followed by one on neutron scattering in liquid helium. The journal was distributed to its institutional subscribers — the universities and research institutes and corporate laboratories in forty-three countries that maintained subscriptions — and the August 12th issue arrived in those institutions' libraries and physics department reading rooms with the ordinary velocity of the international postal service.
The paper's title: High-Efficiency Blue Light-Emitting Diodes in InGaN/GaN Multiple Quantum Well Structures with Solution to P-Type Dopant Activation.
Authors: R. Chandra, S. Iyer, A. Malhotra, V. Kumar, R. Mehta, P. Sharma, K. Shergill.
Institutional affiliation: Indian Semiconductor Manufacturing Corporation, Gorakhpur, Uttar Pradesh, India.
The paper was eight pages long with four figures and forty-two references. The methods section described the MOCVD growth conditions, the multi-quantum-well architecture, and — in the specific, technically precise language of a Physical Review Letters paper — the hydrogen passivation solution. The results section presented the efficiency data, the electroluminescence spectra, and the white LED characterisation. The discussion section was careful and conservative, as Physical Review Letters papers were required to be, making no claims beyond what the data directly supported.
The data directly supported quite a lot.
At MIT's Research Laboratory of Electronics, a postdoctoral researcher named Dr. James Chen arrived at seven-thirty in the morning on August 12th and found the Physical Review Letters issue in the department's mail. He flipped through the table of contents in the standard way of a researcher checking for papers relevant to his work.
He saw the ISMC paper title.
He stopped.
He had heard about the ISMC patent in June through the patent database alert system he had configured for GaN-related filings. He had read the claims. He had read the efficiency figure and had experienced the specific reaction that semiconductor researchers experienced when they read an efficiency figure that exceeded the state of the art by a factor of thirteen: first, the assumption of error; second, the careful reading for the mechanism; third, the dawning possibility that there was no error.
He had spent two weeks in June thinking about the hydrogen passivation mechanism. He had concluded, privately, that it was correct.
He had told no one.
He was telling no one because he was twenty-nine years old and was a postdoctoral researcher and telling people that an Indian company had solved the central problem in the field and moved semiconductor optoelectronics forward by fifteen years was the kind of statement that required either complete confidence or complete anonymity, and he had enough confidence to be telling people but not quite enough to be comfortable with the consequences of telling people.
Now the paper was in his hands.
He read it.
He read it in the specific way he read papers that were potentially significant: standing at the mail table, not yet gone to his office, with his coat still on. He read the abstract first. Then the methods. Then the results. Then the discussion.
At eight-fifteen, he walked to the office of Professor Robert Sato, his supervisor, and knocked.
Professor Sato was fifty-three years old, a Japanese-American physicist who had built one of the three or four best III-V semiconductor research groups in the world and who had the specific quality of a great scientist: the ability to receive unexpected information without defensive resistance, to hear something that challenged his understanding and to engage with the challenge rather than reject it.
He looked up.
"Jim," he said.
"Professor," Chen said. He put the paper on Sato's desk. "The ISMC paper. Physical Review Letters. August 12th."
Sato looked at the paper. He had seen the patent claims in June. He had spent considerably more than two weeks thinking about them.
"Sit down," he said.
Chen sat.
Sato read the paper. He read it the way he read papers of potential significance: completely, all the way through, without speaking. Chen waited. This took twenty minutes. Chen was accustomed to waiting. Sato read slowly because he read thoroughly.
When Sato finished, he looked at the ceiling for approximately thirty seconds.
Then he said: "The phosphor characterisation. Figure four."
"Yes," Chen said.
"The CRI above 75 at 5,600 Kelvin colour temperature. That's a commercially useful colour rendering index. That's not a laboratory curiosity."
"No," Chen said.
"An 86 lumens per watt white LED with CRI above 75 is — Jim, that's better than a compact fluorescent lamp."
"I know," Chen said.
"Compact fluorescents are the most efficient commercially available white light source in 1974," Sato said.
"I know," Chen said again.
Sato looked at figure three — the electroluminescence spectrum of the Wafer 847 device. The sharp, symmetric peak at 465 nanometres. The full-width at half-maximum of 22 nanometres. The specific profile of a well-fabricated quantum well device operating efficiently.
"The linewidth," he said. "Twenty-two nanometres FWHM. For a quantum well device, that's — the inhomogeneous broadening is low. Very low. Their growth conditions are producing good material uniformity."
"Which is why the efficiency is high," Chen said.
"Which is why the efficiency is high," Sato confirmed. He closed the paper. He looked at Chen. "We received a request last week. From ISMC's legal team. Through the patent process. They asked if MIT's Research Laboratory of Electronics would be willing to conduct independent testing of device samples when available."
"I know," Chen said. "You mentioned it at the group meeting."
"I said I was considering it," Sato said. "I have decided. We will test the samples."
Chen looked at him. "Even if the results confirm the claims?"
Sato looked at him for a moment.
"Jim," he said. "If the results confirm the claims, the results confirm the claims. That's what testing is for. The purpose of scientific measurement is not to confirm our assumptions about what is possible. It is to determine what is."
Chen was quiet.
"The samples will arrive when the paper's publication is confirmed to the legal team," Sato said. "I expect that's within the week. When they arrive, we test them with our best equipment. We apply every measurement standard we have. We report what we find."
"And if we find 10.4 percent EQE?" Chen said.
"Then we report 10.4 percent EQE," Sato said. "And then we write our own paper citing the ISMC work, confirming the results, and beginning to work on the understanding of why the solution works as well as it does." He paused. "The correct response to a significant result from a new source is not scepticism for its own sake. It's verification. And if verification confirms the result, it's extension."
He looked at Chen.
"Also," he said, "we should invite Chandra to give a seminar here."
Chen blinked. "Already?"
"The paper is eight pages," Sato said. "There is twenty times as much to discuss as fits in eight pages. The right forum for that discussion is a seminar. Yes, already."
At Eindhoven, in Philips Research Laboratories, Dr. Hans Vriesenga had been waiting for the PRL paper since the patent filing in June.
He was not a man given to impatience. He had been in semiconductor research for twenty-two years and had the settled patience of someone who understood that important results arrived when the science was ready and not before. But he had been waiting for this specific paper with the specific alertness of a researcher who had been told, by the patent claims, that something significant was about to be revealed, and who needed the full technical disclosure to understand what it was.
He read it on August 13th — the European postal service being one day slower than the American.
He read it in his office with the door closed.
He had three members of his research team read it simultaneously in the adjacent meeting room.
At ten-thirty he opened the connecting door.
"Well?" he said.
Dr. Annemarie de Vries, his senior device physicist, was the first to answer.
"The growth conditions," she said. "The reactor temperature profile during the InGaN active layer growth. They're operating at a lower temperature than what we've been using — about forty degrees lower. That's not how I would have designed the process."
"But?" Vriesenga said.
"At the lower temperature, indium incorporation into the InGaN alloy is more efficient," she said. "Which means you get a better compositional uniformity in the quantum wells. Which means lower inhomogeneous broadening. Which means the linewidth they're reporting — twenty-two nanometres — is actually explained by the growth conditions. It's not a measurement artifact. It's a consequence of how they're growing the material."
"And the hydrogen passivation solution," Vriesenga said.
"It's right," said Dr. Pieter Janssen, the materials scientist. He had said nothing yet and now said three words with the specific weight of a man who had spent two months hoping the patent claim was wrong and had just confirmed it was not.
"Explain," Vriesenga said.
"The Mg-H bond dissociation energy," Janssen said. "It's approximately 2.3 electron volts. At 700 degrees Celsius in thermal equilibrium, the hydrogen atoms have sufficient thermal energy to break the bonds and diffuse through the GaN crystal. At lower temperatures, the diffusion is too slow. At higher temperatures, you risk damaging the InGaN quantum wells. The 700 to 750 degree range in the patent — it's exactly right."
"You're saying the temperature range in the patent is not arbitrary."
"It's not arbitrary," Janssen said. "It's the correct range, and the fact that it's the correct range is not obvious. It requires knowing the bond energetics and the diffusion kinetics and the damage threshold of the InGaN layers and choosing the window that satisfies all three constraints simultaneously." He paused. "It's the work of someone who understood the problem completely before designing the solution."
Vriesenga sat down.
He had a habit, when receiving significant information, of sitting down before he spoke next. The sitting was the processing time.
"The samples," he said.
"We requested them in June," de Vries said.
"Through the patent process," Vriesenga said. "But ISMC said they'd provide samples after the paper was published." He looked at the paper on the table. "The paper is published."
"I'll call Vikram Sharma," de Vries said. Sharma was ISMC's legal representative in New York, who had been the contact for the verification process.
"Yes," Vriesenga said. "And when the samples arrive — we test them the same day. Not the next day. The same day they arrive."
"Of course," she said.
Vriesenga looked at the paper.
"Twenty-two years," he said. Not to them — to himself, or to the paper, or to the specific long career that the paper was summarising in its implications for everything that had come before. "Twenty-two years in this field. I've read thousands of papers. I've had two papers of my own that I considered significant. One of them is in this journal. Volume 28, 1969."
"I know the paper," de Vries said.
"When I read it for the first time — my own paper — I felt that I had contributed something real to the field," Vriesenga said. "Today I read someone else's paper and I feel something different. I feel that the field has moved past what I thought it was capable of. That someone has solved a problem I was working on by approaching it from a direction I hadn't considered."
He paused.
"That feeling," he said, "is either humbling or exciting depending on whether you are primarily invested in your own position or primarily invested in the field's progress."
De Vries and Janssen were quiet.
"I hope," Vriesenga said, "that I am primarily invested in the field's progress."
He picked up the paper.
"Call Sharma," he said. "Get the samples."
In Tokyo, the reaction moved differently.
Japan had no direct equivalent of the informal academic networks that distributed information in the United States and Europe — the phone calls between colleagues, the immediate circulation of preprints, the hallway conversations at conferences that constituted the informal scientific communication system of the Western research world. But Japan had something else: the specific internal communication networks of large corporations whose research laboratories were staffed by scientists who were also employees, and who reported upward to management structures that had a direct financial interest in understanding what a new technology meant.
Sony's optical research division head, Dr. Akira Matsumoto, had read the patent claims in June and had been following the situation since. He had the PRL paper on his desk on August 14th — two days after publication, delivered by the fastest available route because he had instructed his administrative assistant to flag the ISMC paper and obtain it immediately.
He read it.
He read it with the thoroughness of a man who had been waiting for it.
Then he walked to the office of Sony's Director of Advanced Technology, Dr. Yoshikazu Ando.
"I need thirty minutes," he said.
Ando looked up. He was sixty-one years old, a man whose career at Sony had encompassed the transistor radio and the Betamax recorder and who had therefore lived through two of the twentieth century's most significant consumer electronics transitions and who understood, from that lived experience, what a technological transition looked like from the inside.
"The LED paper?" Ando said.
"Yes," Matsumoto said.
"I read it this morning," Ando said.
"And?" Matsumoto said.
Ando was quiet for a moment.
"Sit down, Matsumoto san," he said.
Matsumoto sat.
"I need to tell you," Ando said, "about something that happened in 1955."
Matsumoto waited. He knew that when Ando began a conversation with something that happened in 1955, the conversation was going somewhere important.
"In 1955, Texas Instruments announced the first commercial silicon transistor," Ando said. "At the time, Sony was a small company making tape recorders. Our founder — Ibuka san — saw the Texas Instruments announcement in a technical journal. He immediately understood that the transistor would replace the vacuum tube and that the company that understood this earliest and moved fastest would define the next era of electronics." He paused. "Ibuka san flew to New York. He negotiated a transistor manufacturing licence from Western Electric. He came back and we built the transistor radio. The rest is the history of Sony."
"Yes," Matsumoto said. He knew this history.
"What Ibuka san saw in 1955," Ando said, "was not just a better component. He saw a transition. The end of one era and the beginning of another. The vacuum tube era ending. The semiconductor era beginning." He paused. "The ability to see transitions — to recognise them while they are still beginning — is the rarest and most valuable capability in the technology industry."
Matsumoto looked at him.
"The white LED," Ando said. "What do you see?"
"I see the end of the incandescent era," Matsumoto said. "And the beginning of the solid-state lighting era."
"Yes," Ando said. "That is what I see too." He paused. "And the company that controls the core patents for the transition is ISMC. Which is a subsidiary of Shergill Industries. Which is an Indian company that also makes fighter aircraft and petroleum and tanks."
"Yes," Matsumoto said.
"In 1955, Ibuka san flew to New York to license the transistor," Ando said. "In 1974, I believe the appropriate action is to fly to Gorakhpur."
Matsumoto was quiet for a moment. "You're suggesting we go ourselves."
"I'm saying that this technology is important enough that it should be discussed at the most senior level we can establish," Ando said. "Not through lawyers. Not through intermediaries. The relationship we build with ISMC now will determine our position in the LED transition. That relationship should be built personally."
"Shergill Industries," Matsumoto said. "Karan Shergill."
"Yes," Ando said.
"He is twenty-four years old," Matsumoto said.
"Ibuka san was forty-seven when he flew to New York," Ando said. "The age of the person you are meeting is irrelevant. The technology they control is what matters."
He looked at Matsumoto.
"When can you be in Gorakhpur?" he said.
"I would need to understand the licensing position before—"
"No," Ando said. "Not yet. Before the licensing discussion, I want to see the facility. I want to understand what they have built. The licensing discussion is a commercial conversation. Before the commercial conversation, I want the scientific and technical understanding." He paused. "Can you arrange a visit?"
"I'll contact Vikram Sharma," Matsumoto said.
"Today," Ando said.
At Stanford, the paper's arrival was — complicated.
Professor David Allen had known the paper was coming since the patent claims in June. He had known because Vikram Sharma's office had provided a courtesy notification — not required, but offered as a gesture of professional respect — that the PRL paper would be published in August. He had therefore been expecting the paper for two months.
He had spent those two months doing two things simultaneously.
The first was continuing his own research. He was the sort of scientist for whom the work was primary regardless of external circumstances, and the external circumstances — the ISMC patent, the gap in efficiency, the Stanford priority claim that his lawyers were managing — were external. The work was internal. He kept working.
The second was thinking carefully about what the ISMC result meant. Not for the patent dispute — that was his lawyers' domain and he had learned in forty years of academic life that the things he did not know should be delegated to people who did know them. What the result meant for the science.
When the paper arrived on August 12th, he read it immediately.
He had read it four times by midnight.
On August 13th, he arrived at his lab at six in the morning and convened an informal group meeting at seven with the four members of his research group who were present.
"You've all read the ISMC paper," he said.
Nods.
"I want to hear what you think," he said. "Not what you think we should do about the patent. What you think the paper says."
A graduate student named Tom Harrington — third year, working on GaN growth — spoke first. He was twenty-five years old and had the specific quality of a young researcher who was not yet invested in any particular position and who could therefore read new information without the protective filter that investment in a position created.
"The hydrogen passivation mechanism," Tom said. "I've been trying to find a problem with it since June. I can't find one. The physics is right."
"Tell me what you mean by the physics is right," Allen said.
"The Mg-H bonds," Tom said. "At the growth temperature, hydrogen is produced from the organometallic precursors in the MOCVD process. It's always present during growth. It bonds to the magnesium dopants. The annealing drives the hydrogen out. I looked up the bond energetics. The temperature range they specify — 700 to 750 Celsius — is where the kinetics work. Below that it's too slow. Above that you risk damage to the InGaN wells."
"You looked up the bond energetics," Allen said.
"Yes," Tom said.
"After reading the patent in June?" Allen said.
"The day after," Tom said. "I wanted to understand whether the temperature range was arbitrary or derived. It's derived. The 700 to 750 Celsius window is exactly where the physics requires it to be."
Allen looked at him.
"What does that tell you?" Allen said.
"It tells me," Tom said carefully, "that whoever designed the experiment knew what temperature window they needed before they ran the experiment. They weren't optimising empirically. They designed toward a known target."
The room was quiet.
"You're saying they knew the answer before they did the experiment," Allen said.
"I'm saying they had a very specific, theoretically derived prediction about what would work," Tom said. "And the experiment confirmed the prediction. Which is — that's not how most material science research works. Material science is usually more empirical. You try things and see what happens and gradually converge on what works."
"This wasn't empirical," Allen said.
"Not in the same way," Tom said. "This looks more like — they had a complete theoretical understanding of the problem and they designed the solution based on that understanding and then experimentally confirmed it."
Allen was quiet.
The room was quiet.
Allen's other graduate student — a young woman named Sara who was working on device characterisation — said: "Professor Allen. I've been thinking about something since June and I want to ask you about it."
"Ask," Allen said.
"Our device," she said. "The Stanford device. 0.8 percent efficiency. We're proud of it. It's a genuine result. It demonstrates the concept." She paused. "But the ISMC device is at 10.4 percent. With a working white LED at 86 lumens per watt." She paused again. "We proved that blue light emission from GaN was possible. They proved that it was useful. Those are both genuine contributions but they're not the same contribution. What we did matters. What they did matters more."
The room was very quiet.
Allen looked at her.
"Yes," he said. "That's right."
"I just wanted to — I wanted to say it," Sara said. "Because the patent dispute is about whether we have priority and the answer to that is complicated and maybe yes and maybe no. But the scientific question — who solved the problem — that answer is not complicated."
"No," Allen said. "It's not."
He looked at the paper on the conference room table.
"I want to be honest with all of you," he said. "I have spent two months managing the emotional response to this, which is — difficult. We have spent years working on this material system. We achieved a genuine breakthrough in April. We filed a patent that represents real scientific work. And then, sixty-eight days later, a facility in India filed a patent that demonstrated our breakthrough was a step on the way to something much larger than what we had achieved."
He paused.
"The temptation, when that happens, is to protect what you have by attacking what the other group has achieved. To question the results. To find technical problems. To argue for legal priority as a substitute for scientific priority." He paused again. "I want to be clear that we are not going to do that."
He looked at the group.
"We are going to request device samples from ISMC for independent testing," he said. "We are going to verify their results. If the results confirm the claims — and I believe they will — we are going to publish a verification study and cite their work as the foundational result it is. And we are going to refocus our own programme toward the questions that the ISMC work has opened rather than the ones it has closed."
"What questions has it opened?" Tom asked.
"Many," Allen said. "The optimisation of the quantum well design. The reliability of the devices over time. The scaling of the manufacturing process. The integration of the LED into real lighting products. The ISMC work has solved the central scientific problem. The engineering work — the translation of the science into something the world can actually use — that's still largely open."
He looked at the paper.
"Also," he said, "I want to write to Chandra. Directly. Not through lawyers. A letter from me to him acknowledging his group's achievement and proposing a scientific exchange."
Sara looked at him. "The lawyers—"
"The lawyers are managing the legal situation," Allen said. "I am managing the scientific relationship. Those are different things. The legal situation is about priority claims and patent rights. The scientific relationship is about the field moving forward. Both matter. But I am a scientist. The scientific relationship is my primary responsibility."
He stood.
"Tom," he said. "Contact ISMC and request the sample package. Use my name. Tell them Professor Allen at Stanford is formally requesting samples for independent verification and would welcome a scientific exchange." He paused. "Not through Sharma's office. Directly to Chandra."
"Yes, Professor," Tom said.
"And Sara," Allen said. "Draft the letter. I'll review it this afternoon."
He walked out of the conference room and went to his office.
He sat at his desk.
He had been in science for thirty years. He had read many papers that moved the field. He had written papers that moved the field. He understood, from that experience, the specific quality of a result that was not merely a step forward but a reorganisation — a result that changed the shape of the space the field operated in.
The ISMC paper was a reorganisation.
The field of semiconductor optoelectronics was different today than it had been on August 11th. Not incrementally different. Categorically different. The hydrogen passivation solution had been published. The multi-quantum-well architecture had been demonstrated at commercial efficiency. The white LED had been shown to work. These were not improvements on existing knowledge. They were additions to the field's fundamental toolkit.
And they had been done in Gorakhpur.
He thought about what Tom had said: they had a complete theoretical understanding of the problem and they designed the solution based on that understanding.
He thought about the efficiency of that approach. The directness of it. The absence of the long empirical search through solution space that characterised most material science research.
He thought about what it would mean to have that kind of foresight. To know what the solution was before you built the experiment to confirm it.
He thought about Karan Shergill, who was twenty-four years old.
He did not conclude what that meant.
He began drafting notes for the Stanford group's next research direction.
In New Delhi, on August 14th, the Prime Minister received the full scientific briefing that Professor Menon had prepared.
The briefing was comprehensive — twelve pages, with the technical content calibrated to be rigorous without requiring a physics background. Menon had spent three days writing it, which was longer than he normally spent on a briefing document, because the importance of the subject and the importance of the audience were both at the highest levels he worked at, and both required maximum care.
Indira Gandhi read the briefing the way she read all documents: completely, with marginal annotations, without interrupting the reading with questions that could be answered by reading further.
When she finished, she looked at Menon.
"The verification," she said. "MIT, Caltech, and NPL Teddington. When will we have results?"
"The samples were shipped from Gorakhpur this week," Menon said. "The testing will take — my estimate — two to three weeks for each institution to run a full measurement protocol. We should have preliminary results in September."
"And when those results confirm the claims," she said — not if, when, which Menon noted — "what happens?"
"The world's major lighting and electronics companies will seek licensing agreements," Menon said. "GE, Philips, Siemens, Sony, Toshiba, Panasonic — every major company that operates in the lighting and display sectors will need to license the core patents to produce commercially viable LED products. The licensing revenue will be substantial. The industrial implications will be more substantial."
"More substantial how?" she said.
"The licensing revenue is money," Menon said. "It will be significant — potentially very significant over the next decade. But the more important question is whether India develops an LED manufacturing industry, or whether India licenses the technology to others who develop the manufacturing industry."
She looked at him.
"That distinction matters to you," she said.
"Enormously," he said. "In the transistor era — the 1950s and 1960s — America held the core transistor patents through Bell Labs. But Japan licensed those patents and built a manufacturing industry around them that eventually dominated global consumer electronics. The patent holder was American. The industry was Japanese." He paused. "India now has the core LED patents. India needs to build the manufacturing industry, not just the patent revenue."
"What does that require?" she said.
"Capital investment in ISMC's scale-up," he said. "The Gorakhpur facility is a research and pilot production facility. Commercial scale LED manufacturing requires a different level of infrastructure — larger reactors, automated assembly, quality control systems, supply chain development. ISMC is planning a commercial production facility. The government's role is to ensure that planning proceeds without the obstacles that large-scale manufacturing in India typically encounters."
"Regulatory obstacles," she said.
"Regulatory, but also infrastructure," he said. "Power, water, transportation. The facility will need reliable industrial-grade power supply. The Gorakhpur region's grid—"
"The Shergill energy division stabilises the Gorakhpur grid," she said. It was not a question.
"Yes," Menon said, slightly surprised.
"I know what Shergill has built there," she said. "Continue."
"The workforce development," Menon said. "Commercial LED manufacturing requires trained technicians at scale. The ISMC programme is producing researchers. Production technicians are different — they need different training. The technical education infrastructure in the region—"
"Gorakhpur Technical University," she said.
"Yes," Menon said. "Which is also Shergill's. The ecosystem is integrated."
She looked at the briefing document.
"Professor Menon," she said. "You described this achievement in our June meeting as: the nuclear test announced India's power, this announces India's intelligence. Do you still hold that view?"
"More strongly than in June," he said. "The nuclear test — I want to be precise — demonstrated that India could apply existing knowledge to produce a known result. Many countries had tested nuclear devices. We demonstrated that India could join that group. That is an important achievement. It changes India's strategic position." He paused. "This achievement is different in kind. A commercially viable white LED did not exist anywhere in the world. ISMC did not replicate something that existed. They produced something that did not exist and that the world's best researchers had been unable to produce." He paused again. "The distinction matters. Catching up is valuable. Being first is in a different category."
"India was first," she said.
"India was first," he confirmed.
She looked at the window. Outside: the South Block courtyard, the trees, the specific Delhi August light.
"What do you want from me?" she said. "Specifically."
"Three things," he said. "First: a personal communication from you to Chandra and the research team. Not a government statement. A personal letter. The team has spent eighteen months on something extraordinary and the acknowledgement of the person they work for, and the country they live in, matters."
"Already drafted," she said. "Kaul sent it last week."
"Good," he said. "Second: a meeting with Karan Shergill. Not to congratulate him. To understand what the next phase of the programme requires and what the government can do to support it. The LED manufacturing scale-up will encounter obstacles. Some of them are in the government's domain to remove."
"Schedule it," she said.
"Third," Menon said. He paused. "This is the one I'm less certain about raising."
"Raise it," she said.
"The Nobel Prize committee," he said. "The Physics Nobel. The committee makes its nominations in October. The PRL paper is published in August. If the independent verifications confirm the results by September — which I expect they will — the work will be in the conversation for the Nobel. I believe — and this is a personal opinion, not a scientific committee position — I believe it is the most significant result in semiconductor optoelectronics in the past decade. Possibly in the history of the field."
She looked at him.
"What do you want me to do about the Nobel committee?" she said.
"Nothing," he said. "The Nobel committee is independent. Any government communication to the committee would be inappropriate." He paused. "What I want you to be aware of is that this conversation is going to happen. In the scientific community, in the press, internationally. And I want the Indian government to be prepared to respond to it correctly — not to claim the award, not to lobby for it, but to acknowledge that the work was done here by Indian researchers and that the acknowledgement of that work by international scientific institutions is something India welcomes without having sought it."
"That is a careful position," she said.
"It is the correct position," he said.
She looked at the briefing.
"One more question," she said. "The PRL paper lists seven authors. K. Shergill is one of them."
"Yes," Menon said.
"Karan Shergill," she said. " Is the authorship appropriate?"
Menon looked at her.
"I asked Chandra directly," he said. "His response was: Shergill identified the solution pathway before the programme began, directed the programme's experimental strategy throughout, and made specific technical contributions at several critical junctures including the identification of the optimal annealing temperature range." He paused. "By the standards of authorship in physics — which require intellectual contribution to the work, not just funding or direction — yes. The authorship is appropriate."
She was quiet.
"A twenty-Four-year-old Nobel Prize Winner" she said. Not with disbelief. With the specific quality of a woman who had been managing exceptional people her entire political career and who was recalibrating her model of what exceptional meant.
"Yes," Menon said.
She said nothing for a moment.
"Good," she said. "And Professor Menon." She looked at him. "The third thing — the Nobel conversation. When it happens, I want it to be about the team. Not about any individual. India's achievement is a team achievement."
"Chandra's team," Menon said.
"Yes," she said. "Chandra's team. Including Karan Shergill." She paused. "All of them."
The MIT confirmation arrived on September 3rd.
Dr. James Chen and Professor Sato had run the measurements over seven days — not seven continuous days, a week of careful work in which each measurement was checked against calibration standards and repeated multiple times and the results were verified against independent calibration sources before any number was recorded as confirmed.
The final numbers:
External quantum efficiency: 10.2 percent. (Slightly below ISMC's 10.4 — attributed by Chen, in the measurement notes, to minor handling-related variation in the samples during international shipping, consistent with the expected degradation of less than 5 percent.)
White LED efficacy: 84 lumens per watt.
Colour temperature: 5,800 Kelvin.
Colour rendering index: confirmed above 75.
Linewidth (electroluminescence FWHM): 23 nanometres.
The MIT report's conclusion was four sentences:
These results are confirmed. The device architecture and fabrication process described in the Physical Review Letters paper by Chandra et al. produce devices with the claimed performance. The efficiency figures represent a significant advance beyond the current state of the art and are consistent with the theoretical predictions for optimally designed InGaN/GaN quantum well structures. The white LED device represents the first demonstration of a commercially practical solid-state white light source.
Four sentences. Professor Sato had drafted them himself. He had revised them eleven times.
The eleventh draft was the one he filed.
He did not revise it a twelfth time because the eleventh draft said exactly what needed to be said without elaboration and without qualification and without the defensive hedging that scientists sometimes used when publishing results that made their own previous understanding look insufficient. The eleventh draft was direct. The results are confirmed. Here is what they mean. That is all.
He sent the report to ISMC's legal team through Vikram Sharma's office.
He also sent a copy to Physical Review Letters with a cover note suggesting that the journal might consider publishing MIT's independent confirmation as a Brief Report in the subsequent issue.
Then he called Chandra.
The call reached Gorakhpur at seven in the evening.
Chandra was in his office — not the lab, the office, because it was after six and the lab work was being managed by the overnight team and he was at his desk reading the accumulated correspondence that had come in since the paper's publication.
There had been a great deal of correspondence.
Letters from scientists at universities and research institutions across India and internationally, each one representing a version of the same response to the paper: I read your work. I am reaching out because I want to establish contact. This was the specific currency of scientific communication — not money, not prestige, but the willingness to reach. When a paper produced this volume of reaching, it was the field's way of saying: this matters, we have seen it, we are responding.
Chandra was reading the letters and composing responses with the thoroughness that he brought to correspondence as he brought to everything: completely.
His phone rang.
"Dr. Chandra," the voice said. "Robert Sato. MIT Research Laboratory of Electronics."
Chandra sat up slightly. He knew the name. Everyone working in III-V semiconductor devices knew the name.
"Professor Sato," he said. "Thank you for calling."
"I'm calling," Sato said, "to tell you personally before you receive the formal report that our testing is complete." He paused. "Dr. Chandra, your results are confirmed."
The lab. The white LED on the bench. Wafer 847. The eighteen months. The 847 wafers. Arun's call to his sister. The patent number. The paper. August 12th.
All of it converged into this sentence.
"Thank you, Professor," Chandra said.
"The efficiency numbers," Sato said. "Our measurement gave 10.2 percent — slightly below your 10.4. Our instrument analysis suggests the minor discrepancy is attributable to sample handling variation during shipping. The device performance is exactly as claimed."
"Yes," Chandra said. "We expected some handling variation."
"The white LED," Sato said. "84 lumens per watt on our calibrated system."
"Yes," Chandra said.
"Dr. Chandra," Sato said. "I want to say something to you as a scientist. Not as part of the formal verification process. Personally."
"Please," Chandra said.
"I have been in semiconductor research for twenty-five years," Sato said. "I have seen many significant results. I have contributed several results that I consider significant. I have never read a paper that reorganised the field the way yours has." He paused. "The hydrogen passivation solution — the specific mechanism, the temperature range, the nitrogen atmosphere — I have been working adjacent to this problem since 1968. I was approaching it from a different direction. The direction I was approaching it from would not have found the solution. Not the way you found it."
Chandra was quiet.
"What was your direction?" he asked.
"Growth chemistry modification," Sato said. "Trying to prevent hydrogen incorporation during growth rather than removing it afterward."
"That was also Professor Rao's direction at IISc," Chandra said.
"It's the intuitive direction," Sato said. "Prevent the problem rather than solve it afterward. The post-growth annealing approach is less intuitive because it requires accepting that the problem will occur during growth and planning to address it after. That's a different way of thinking about the sequence of the process."
"It came from a specific understanding of the growth environment," Chandra said. "The MOCVD process generates hydrogen throughout growth. Trying to prevent hydrogen during growth means fighting the process chemistry. Working with it — accepting the passivation and then reversing it — is less elegant but more practical."
"Whoever identified that approach," Sato said, "understood the system deeply."
"Yes," Chandra said. He did not say more.
"I've also submitted a cover note to Physical Review Letters," Sato said. "Suggesting they consider publishing our confirmation as a Brief Report. I wanted you to know in advance."
"That is generous," Chandra said.
"It's the correct scientific practice," Sato said. "Independent confirmation of a significant result should be published. The field needs the confirmation to be in the record." He paused. "And Dr. Chandra — one more thing. I'd like to invite you to give a seminar at MIT. At your convenience. The depth of the work — eight pages doesn't capture it. There's twenty times as much to discuss."
Chandra was quiet for a moment.
He thought about the programme. The eighteen months. The 847 wafers. The team — Iyer and Malhotra and Kumar and Mehta and Sharma and young Arun Mehta.
"I would be honoured," he said. "I'll bring three members of my team."
"Of course," Sato said. "We'll be in touch about the scheduling."
"Professor Sato," Chandra said. "May I ask — the Stanford situation. Have you been in contact with Professor Allen?"
A pause.
"We spoke yesterday," Sato said. "He is — handling this with considerable professional dignity. He has requested samples from your team for independent testing. He has instructed his group to treat the verification as a scientific obligation rather than a competitive exercise." He paused. "He will do the right thing."
"The Stanford work is genuine," Chandra said. "Their device represents real scientific achievement. The concept demonstration — proving that GaN could emit blue light at measurable efficiency — that was not trivial. They did important work."
"Yes," Sato said. "They did. And you completed it."
In India, the MIT confirmation arrived like a second wave of the original news.
The first wave — the June patent filing and the newspaper coverage — had been significant but contingent: significant because the numbers were extraordinary, contingent because the numbers had not yet been independently verified. The sceptics — Professor Mishra at Delhi University who had said the numbers couldn't be real, the unnamed voices in various institutions who had kept their doubts to themselves — had the cover of waiting for verification.
The MIT confirmation removed that cover.
The Hindustan Times ran the story on September 5th.
MIT CONFIRMS ISMC LED BREAKTHROUGH — RESULTS ARE REAL, SAYS AMERICA'S TOP SEMICONDUCTOR LAB
Below the headline, in the first paragraph: The Massachusetts Institute of Technology's Research Laboratory of Electronics has confirmed that the high-brightness LED devices developed by the Indian Semiconductor Manufacturing Corporation in Gorakhpur perform exactly as claimed in the company's Physical Review Letters paper published last month. The confirmation eliminates remaining scientific doubt about the magnitude of India's achievement.
The story ran above the fold.
It ran front page.
At IIT Delhi, Professor Mishra read the story at breakfast.
He was the physicist who had told Subroto Bhattacharya in June that the numbers couldn't be real. He had said this because the numbers were, by the standards of what he knew the field to be capable of, implausible. He had said it as a scientist exercising appropriate scepticism. He had said it accurately — appropriate scepticism was the correct response to an extraordinary claim.
The MIT confirmation told him his scepticism had been correct as a scientific posture and incorrect as a prediction.
He put the newspaper down.
He picked up his phone.
He called Subroto Bhattacharya.
"Mr. Bhattacharya," he said. "You will remember me. You interviewed me for the June story."
"Professor Mishra," Subroto said. "Of course."
"I told you the numbers couldn't be real," Mishra said.
"You did," Subroto said.
"I was wrong," Mishra said.
A pause.
"Professor," Subroto said carefully, "that's—"
"I want to say it clearly so there is no ambiguity," Mishra said. "I said those numbers were not physically possible. MIT has confirmed they are physically possible and have been achieved. I was wrong." He paused. "I was exercising appropriate scientific scepticism given what I knew. The appropriate scientific response to the confirmation of a result I doubted is to update my position. I am updating my position."
"Would you like me to include a statement from you in my next story?" Subroto said.
"Yes," Mishra said. "Include exactly what I just said. I was wrong. The results are confirmed. India has produced something extraordinary." He paused. "And include this: the scepticism I expressed in June was the correct scientific response to an extraordinary claim. It was proven wrong by the data. That is how science is supposed to work. The system worked correctly. I said the numbers couldn't be real. The MIT laboratory tested the device and said the numbers are real. I accept the MIT result. This is how science resolves disagreements."
"You want to be quoted saying you were wrong," Subroto said.
"A scientist who will not say he was wrong when the data shows he was wrong is not a scientist," Mishra said. "I am a scientist. Include the statement."
Subroto included the statement.
In Gorakhpur, in the engineering college hostel, a group of first-year students stayed up until one in the morning on September 5th reading everything they could find about the LED breakthrough.
They were seventeen, eighteen years old. They had arrived at engineering college three weeks earlier — fresh from the IIT JEE and the state engineering entrance examinations, fresh from the intense preparation of class eleven and twelve, arriving at the next stage of the education that was going to make them engineers.
Their names were Ravi, Suresh, Kavitha, Pradeep, and Ananya.
Ravi had read the June story. He had read the August paper — or tried to; he understood perhaps thirty percent of the content, which was more than most seventeen-year-olds and less than he would understand in four years. He had read the September confirmation story.
At eleven-thirty, after the lights-out rule had been technically observed by the corridor supervisor and practically ignored by the hostel floor, he said to the others: "The ISMC facility is seven kilometres from here."
No one said anything for a moment.
"I know," Suresh said.
"The people who built the white LED," Ravi said. "They are seven kilometres from where we are sleeping right now."
"I know," Suresh said again.
"When we graduate," Ravi said, "we could—"
"They'd never take us," Pradeep said. "ISMC takes IIT graduates. PhD level."
"They have an undergraduate attachment programme," Kavitha said. She had looked this up. She was the person in the group who looked things up. "Through the Gorakhpur Technical University. The engineering college has a memorandum of understanding."
Everyone looked at her.
"Since when?" Ananya said.
"Since April," Kavitha said. "Deshpande sir at IIT Bombay called them. They set up the MOU in April. Second-year students from affiliated engineering colleges can apply for a six-month attachment."
Silence.
"Second year," Ravi said.
"One year from now," Kavitha said.
Another silence.
"I'm applying," Ravi said.
"You have to be second year," Pradeep said.
"In one year I will be second year," Ravi said. "And then I'm applying."
He looked at the others.
"You're all applying," he said. "All of us."
In the ISMC lab at nine in the evening on September 5th, Chandra was writing the acknowledgements section of the paper he was beginning — the second paper, the one that would go into the growth conditions and the material characterisation in the detail that eight PRL pages could not contain.
He had been thinking about the acknowledgements section for two days.
There was a standard way to write acknowledgements in a physics paper. The funding sources, the institutions, the equipment providers, the colleagues who had read drafts. Chandra was drafting the standard acknowledgements and finding them insufficient for the specific situation.
He was thinking about what was insufficient.
The funding came from ISMC, which came from Karan Shergill. The institutional support was the ISMC facility, which Karan had built. The research direction had been provided, initially, by Karan's specification of the hydrogen passivation approach as the correct pathway. The reactor capability had been purchased with Karan's capital. The team's salaries had been paid from Karan's programme.
This was all in the standard category of funding acknowledgements, and standard funding acknowledgements were three sentences at the end of the paper.
But there was something else that was not in the standard category.
There was the fact that this work had happened at all.
The work had happened because in 1973, a twenty-two-year-old industrialist in Gorakhpur had decided that India should be first in LED technology, had built the facility required, had identified the correct solution pathway, had recruited the team capable of executing it, had provided the resources and the direction and the protection from bureaucratic friction that allowed the team to work without distraction for eighteen months.
None of that was in the standard acknowledgement format.
He sat with his pen above the paper for a long time.
Then he wrote:
The authors acknowledge the vision and support of Karan Shergill, without whose understanding of the problem's structure and whose commitment to creating the conditions for its solution this work would not have been possible. The acknowledgement here is not for funding, which is described separately. It is for the specific intellectual contribution of identifying the correct approach before the research began, and for providing the institutional environment in which eighteen months of sustained, focused experimental work could produce these results.
He read what he had written.
He kept it.
The Caltech confirmation came on September 5th.
EQE: 10.3 percent. White LED: 85 lumens per watt.
The NPL Teddington certification — the most rigorous measurement authority in the world for optical devices — came on September 8th.
EQE: 10.1 percent. White LED: 83 lumens per watt.
All three confirmations were within measurement uncertainty of ISMC's original claims.
All three were consistent with each other.
All three confirmed: the technology was real.
In the week after the three confirmations, something happened across India that was difficult to describe precisely because it was not a single event but an accumulation of small moments in many places simultaneously.
It happened in the engineering college in Gorakhpur where five first-year students stayed up until one in the morning talking about seven kilometres.
It happened in Professor Anand Mehta's class at VJTI Bombay when he walked in on September 9th and wrote on the blackboard: White LED: 86 lm/W. MIT confirmed: 84 lm/W. Caltech confirmed: 85 lm/W. NPL confirmed: 83 lm/W. Location: Gorakhpur, Uttar Pradesh. And then sat on the edge of his desk and said to his class: "Any questions?" and a student named Deepa Menon put her hand up and said: "When can we visit?"
It happened in the retired schoolteacher Harpal Singh's flat in Chandigarh where he read the September 5th confirmation story and called his son the Air Force pilot and said: "I wrote you a letter in June. I didn't know then that I was right about something bigger than I knew."
It happened in the laboratory at IISc Bangalore where Professor Malati Krishnaswami read the NPL Teddington certification on September 9th and sat alone in her office for twenty minutes before calling the two postdoctoral researchers who worked in her group and saying: "I want to change our research direction. I want to discuss it today."
It happened in the university reading rooms and the college hostels and the factory canteens and the tea stalls adjacent to engineering colleges and the specific conversations between parents and children where the story crossed from the technical world into the ordinary world through the translation that families performed, imperfect and essential.
It happened in different forms and with different content and with different levels of technical understanding. But at its core it was the same thing in all the places it happened:
The understanding that something had changed.
Not a small change. Not an incremental change. A change in the kind of thing that was possible here.
India had invented something.
Not improved something. Not built something according to someone else's design. Not manufactured something under licence. Invented. From the beginning, by Indian researchers, in an Indian facility, funded and directed by an Indian industrialist who was twenty-three years old.
Invented.
And the world — MIT, Caltech, NPL Teddington — had confirmed it.
In the Gorakhpur factory, on the evening of September 10th, Karan was at his desk at eleven at night.
The S-35 programme was in a good phase — the structural testing was proceeding well, Rathore had flagged a minor issue with the FBW in one specific flight regime that was being addressed, but the programme was on track. The Arjuna production order was being implemented. The Star Wars music was being built in Bombay. The petroleum was flowing. The Jiyo supplement was in 90,000 households.
He was reading the NPL Teddington certification report.
He had been reading it for an hour.
Not because it required an hour to read — it was three pages, concise and precise in the way of NPL documentation. He had been reading it because he was sitting with what it meant.
10.1 percent. 83 lumens per watt.
The MIT number was slightly different. The Caltech number was slightly different. The NPL number was slightly different. All of them different from ISMC's original measurement. All of them within the range that legitimate measurement variation across instruments and calibration standards explained. All of them confirming the same thing.
The device was what it was.
He set the report down.
He thought about March 21st, 1973. Not this year — last year. The day he had started the programme. Chandra's first briefing. The team of six researchers. The new MOCVD reactor that had arrived from Germany in January. The plan.
He had known, going into the programme, what the result would be. He had known the hydrogen passivation mechanism. He had known the temperature range. He had known the multi-quantum-well architecture. He had known the efficiency potential.
He had known because he came from a future where these things had been worked out, published, discussed, celebrated, and eventually implemented in the billions of LED devices that had lit the twenty-first century he had left.
He had brought that knowledge back.
He had used it to direct a programme that had arrived at the result eighteen months later.
This was the part of what he did that he thought about most carefully and least clearly.
He had not built the device. Chandra and the team had built the device. They had done the 847 wafers. They had optimised the growth conditions. They had characterised the devices. They had written the paper. They had done the work.
He had pointed them toward the answer.
Was that the same as knowing the answer?
He had been thinking about this distinction for three years, since the first programme he had directed based on knowledge from the future. The S-27. The petroleum. The LED. In each case: he knew what the result should be, he directed the programme toward it, Indian hands and Indian minds did the work of building it.
His contribution was the direction. His contribution was the knowledge that a direction was correct before the experiment confirmed it.
In science, knowing which direction to look before the experiment was — it was not nothing. It was, in fact, the thing that made the difference between programmes that took twenty years and programmes that took two.
But it was not the same as doing the experiment.
He sat with this for a long time.
Then he opened the notebook and wrote:
September 10, 1974. NPL Teddington confirmed. 10.1 percent. 83 lumens per watt.
MIT: 10.2 percent. 84 lumens per watt. Caltech: 10.3 percent. 85 lumens per watt.
The device is confirmed by three of the world's most rigorous measurement institutions.
I knew it would be confirmed. I have known since March 1973 that the device would work. The knowledge did not make the eighteen months unnecessary. The knowledge made the eighteen months focused on the right problem instead of the wrong one.
Chandra's team built it. I pointed them toward what to build.
That is the honest accounting of this.
He looked at what he had written.
Then he added:
The world is calling it India's achievement. It is India's achievement. The team is Indian. The facility is Indian. The work was done here.
The question I will be asked — am I already being asked it — is: how did you know what to build? The answer I give is: I understood the problem. The answer I cannot give is the complete answer.
I cannot say this. I say: I understood the problem.
That is true. It is incomplete. The incompleteness is the condition I live in.
He closed the notebook.
He turned off the desk lamp.
Outside, the factory floor was running the night shift. The S-35 assembly bay had its lights on. The Arjuna components section was working. The ISMC facility a kilometre east had its own lights — the reactors that ran continuously, the measurement equipment, the overnight team.
The white LED was on the bench.
It was always on the bench.
Running.
86 lumens per watt.
The world had confirmed it.
India had invented it.
Both things were true.
He sat in the dark office for a moment longer.
Then he went home.
End of Chapter 164
Independent Verification Record — ISMC Blue LED Programme
Physical Review Letters Publication: August 12, 1974
MIT Research Laboratory of Electronics (Professor Robert Sato, Dr. James Chen): Confirmation date: September 3, 1974 EQE measured: 10.2% (claimed: 10.4%) White LED efficacy: 84 lm/W (claimed: 86 lm/W) Conclusion: Results confirmed. First demonstration of commercially practical solid-state white light source.
California Institute of Technology (Division of Engineering and Applied Science): Confirmation date: September 5, 1974 EQE measured: 10.3% White LED efficacy: 85 lm/W Conclusion: Device performance as claimed. Efficiency figure represents significant advance over current state of the art.
National Physical Laboratory, Teddington (Primary measurement authority, United Kingdom): Certification date: September 8, 1974 EQE measured: 10.1% White LED efficacy: 83 lm/W Colour temperature: 5,750K CRI: confirmed above 75 Conclusion: Measurement certified. All performance parameters within stated specifications, accounting for measurement uncertainty across instruments.
Stanford University Independent Testing: Request received from Professor David Allen: August 13, 1974 Sample request submitted to ISMC: August 15, 1974 Testing commenced: September 1, 1974 Results: To be reported in a Brief Report to Physical Review Letters (forthcoming)
Licensing inquiries received as of September 15, 1974: Philips, GE, RCA, Sony, Siemens, Toshiba, Sharp, Panasonic. Licensing programme structure under development by Shergill legal team. No agreements concluded.
Nobel Prize committee: Informal inquiry to Chandra, September 14, 1974. ISMC LED programme under consideration for Physics Nobel nomination, 1975 cycle.
The device: Still on the bench in Dr. Chandra's laboratory, Gorakhpur. Still running. 86 lumens per watt. The world has seen it. The world has confirmed it. The world is changed by it.