Feynman Says Atomic-Scale Machines Face Design Limits, Not Physics

Muon Ray

Physical laws permit machines and data storage down to single-atom scales because apparent technical barriers stem from flawed macroscopic engineering assumptions rather than fundamental limits.

AI-assisted summary of Richard Feynman "Tiny Machines" Nanotechnology Lecture - aka "There's Plenty of Room at the Bottom" — original source. Check the original for full context.

Source published: 2012-08-23

Key ideas

Richard Feynman reviews the trajectory of nanotechnology from his 1959 lecture to 1984, noting that atomic-scale digital storage using clusters of roughly 100 atoms per bit could compress the text of all global libraries into a dust speck measuring 1/200th of an inch. (4:48)
Scaling machines down is not simple geometric shrinkage because physical forces shift unevenly; microscale combustion engines fail due to rapid thermal dissipation outracing mechanical expansion, while micro-swimmers require rotational flagella to navigate fluid that behaves like thick honey. (36:49)
Feynman argues that commercial engineering limits, such as electromigration or optical diffraction in photolithography, should not be mistaken for fundamental barriers, as they can be bypassed through alternative architectures and shorter wavelengths like X-rays. (23:56)
Distinguishing between fundamental and applied physics, Feynman asserts that applied science has virtually infinite scope for creating new technologies within verified physical laws, whereas pure fundamental physics may eventually reach completion. (60:13)

Detailed notes

## Background Published on YouTube by Muon Ray on August 23, 2012, this video captures Richard Feynman in 1984 revisiting and updating his visionary 1959 lecture 'There's Plenty of Room at the Bottom.' The description highlights how Feynman re-examines the feasibility of nanotechnology twenty-five years later, evaluating commercial photolithography, electron microscopy, atomic manipulation, and the fundamental boundaries imposed by the laws of physics.

## TL;DR Feynman explores the theoretical and practical frontiers of extreme miniaturization, showing how vast libraries of information can be compressed into a speck of dust using atomic-scale encoding. He examines modern semiconductor manufacturing, telemanipulators, and micro-engines, demonstrating why shrinking mechanisms requires completely rethinking designs to accommodate shifting scaling laws, such as extreme thermal dissipation and high effective fluid viscosity at low Reynolds numbers. Distinguishing temporary engineering hurdles from immutable natural limits like optical diffraction and the speed of light, he champions scientific imagination strictly bound by physical law. The session expands into reflections on subatomic particle probing, theoretical physics, and the psychological rush and humbling hazards of scientific discovery in his work on beta decay and superfluid helium.

## Atmosphere & style The presentation carries Feynman's signature blend of razor-sharp physical intuition, conversational warmth, and playful irreverence, making deep scientific principles instantly accessible through colorful metaphors like dancing Rockettes, ocean waves tossing corks, and desert jackrabbits knocking over mirrors. His delivery is energetic and spontaneous, punctuated by hearty chuckles, rapid-fire rhetorical challenges, and a strict no-nonsense approach to metaphysical speculation. Listening feels like attending an intimate, freewheeling masterclass that moves seamlessly from the mechanics of microscopic motors to the vulnerable, intoxicating mania of scientific breakthrough and the sobering domestic wisdom of waiting until tomorrow to see if an idea survives.

## Bulletin - [04:48](https://www.youtube.com/watch?v=4eRCygdW--c&t=288s) Richard Feynman reviews the trajectory of nanotechnology from his 1959 lecture to 1984, noting that atomic-scale digital storage using clusters of roughly 100 atoms per bit could compress the text of all global libraries into a dust speck measuring 1/200th of an inch. - [36:49](https://www.youtube.com/watch?v=4eRCygdW--c&t=2209s) Scaling machines down is not simple geometric shrinkage because physical forces shift unevenly; microscale combustion engines fail due to rapid thermal dissipation outracing mechanical expansion, while micro-swimmers require rotational flagella to navigate fluid that behaves like thick honey. - [23:56](https://www.youtube.com/watch?v=4eRCygdW--c&t=1436s) Feynman argues that commercial engineering limits, such as electromigration or optical diffraction in photolithography, should not be mistaken for fundamental barriers, as they can be bypassed through alternative architectures and shorter wavelengths like X-rays. - [01:00:13](https://www.youtube.com/watch?v=4eRCygdW--c&t=3613s) Distinguishing between fundamental and applied physics, Feynman asserts that applied science has virtually infinite scope for creating new technologies within verified physical laws, whereas pure fundamental physics may eventually reach completion.

## Key takeaways - [00:47](https://www.youtube.com/watch?v=4eRCygdW--c&t=47s) Miniaturization limits are governed by physical laws rather than manual dexterity, allowing digital information to be stored as arrangements of contrasting atoms. - [02:33](https://www.youtube.com/watch?v=4eRCygdW--c&t=153s) A 20,000-fold reduction in length and width allows the entire Encyclopædia Britannica to fit onto the head of a pin, and can be read back using electron microscopy. - [07:01](https://www.youtube.com/watch?v=4eRCygdW--c&t=421s) Writing at nanoscale resolutions can be accomplished conceptually by operating electron microscope focusing optics in reverse. - [23:56](https://www.youtube.com/watch?v=4eRCygdW--c&t=1436s) Engineering hurdles like electromigration resemble early beliefs that the sound barrier was impassable; they represent limitations of specific designs rather than physical laws. - [26:33](https://www.youtube.com/watch?v=4eRCygdW--c&t=1593s) Macroscale telemanipulators could be miniaturized recursively, with larger mechanical hands fabricating smaller tool sets in exponential manufacturing cascades. - [29:08](https://www.youtube.com/watch?v=4eRCygdW--c&t=1748s) Planar microfabrication with sacrificial etching enables moving components like electrostatic motors and micro-shutter arrays without manual assembly. - [36:49](https://www.youtube.com/watch?v=4eRCygdW--c&t=2209s) Miniature internal combustion engines cannot function because surface area-to-volume scaling causes heat to escape into cavity walls before gas can push a piston. - [39:49](https://www.youtube.com/watch?v=4eRCygdW--c&t=2389s) At low Reynolds numbers, water behaves like thick syrup, making reciprocating strokes ineffective and requiring rotational corkscrew flagella for microscale propulsion. - [45:35](https://www.youtube.com/watch?v=4eRCygdW--c&t=2735s) Scientific imagination must operate within a strict straitjacket defined by experimentally verified physical laws, ruling out speculative devices like anti-gravity machines. - [48:29](https://www.youtube.com/watch?v=4eRCygdW--c&t=2909s) Thermal vibrations in solid lattices prevent single-atom long-term data storage, but clusters of roughly 100 atoms remain stable against thermal disruption for billions of years. - [01:01:56](https://www.youtube.com/watch?v=4eRCygdW--c&t=3716s) Visible light cannot resolve or manipulate features smaller than its wavelength, necessitating ultraviolet or X-ray lithography for submicron fabrication. - [01:10:22](https://www.youtube.com/watch?v=4eRCygdW--c&t=4222s) Acoustic frequencies reach an absolute physical cutoff when the acoustic wavelength becomes equal to the interatomic lattice spacing of the medium. - [01:14:50](https://www.youtube.com/watch?v=4eRCygdW--c&t=4490s) Feynman identifies his derivation of the V−A theory of weak interactions and the theoretical explanation of frictionless liquid helium superfluidity as his most thrilling discoveries.

## Detailed walkthrough ### [00:47](https://www.youtube.com/watch?v=4eRCygdW--c&t=47s) The Ultimate Limits of Small Writing and Atomic Coding

Richard Feynman opens the lecture by posing a deceptively straightforward question: how small can machinery actually be made? He mentions overhearing people chatting in the bath, puzzled about his topic and wondering aloud what on earth he could mean by “tiny machines.” His answer to them was as literal as it gets: he is talking about very small machines. As Feynman puts it, “I am talking about very small machines.” But before diving directly into mechanical devices, he decides to lay the groundwork by examining the physical limits of writing and encoding numbers at minuscule scales.

When considering how small a number or mark can be written, Feynman emphasizes that the constraint is not human finger dexterity, but the fundamental laws of physics. Because writing simply consists of arranging contrasting matter—such as depositing dark atoms over light ones—one cannot write on an individual atom itself. To encode information at the absolute physical frontier, one must arrange distinct atomic patches, such as alternating sequences of gold and silver atoms to form binary dots and dashes. While a single atom could theoretically hold a state, setting aside a square patch measuring roughly one hundred atoms on a side provides a reliable, well-defined mark.

- Theoretical writing limits bottom out at the scale of atoms, since a mark is merely an arrangement of one type of atom deposited over another. - Information can be encoded digitally as contrasting physical materials, such as alternating dots and dashes of gold and silver. - A conservative, practical baseline for distinct atomic encoding is an area approximately 100 atoms on a side.

```mermaid flowchart TD accTitle: Information Scaling Hierarchy accDescr: Diagram showing Feynman's scaling comparison from standard print to atomic encoding. subgraph Macro["Macro Scale"] A["Encyclopædia Britannica (20,000 sq ft of print)"] end subgraph Micro["2D Micro Scale (20,000x Reduction)"] B["Pinhead Surface (40,000,000x area shrink)"] C["Single Library Card (Entire Caltech Library)"] end subgraph Atomic["3D Nano Scale (Atomic Arrays)"] D["Speck of Dust (1/200th inch cube)"] E["All Global Knowledge (5-atom cubes in gold/silver)"] end A -->|"optical reduction"| B B -->|"card equivalent"| C B -->|"volume encoding"| D D --- E ```

---

[02:33](https://www.youtube.com/watch?v=4eRCygdW--c&t=153s) Compressing Libraries and the Encyclopædia on a Pinhead

Feynman recalls how people often marvel that the Lord’s Prayer has been etched onto the head of a pin, but he proposes a far bolder benchmark: condensing the entire thirty-odd volumes of the *Encyclopædia Britannica* onto that same tiny surface. If all twenty to thirty thousand pages of the encyclopedia were laid out on the ground, they would cover roughly 20,000 square feet. To fit that massive surface onto a standard pinhead requires reducing the linear dimensions of every word, letter, and illustration by a factor of 20,000. Because reduction operates across two dimensions, this equates to shrinking the required surface area by a factor of 400 million, a reduction that remains well above fundamental atomic barriers. Feynman highlights the feasibility of the concept: “If a little figuring you'll figure it out yourself it's about 20,000 times reduced so if we could reduce the size of the words the letters the dots in the pictures the whole thing everything that's in the andik up here only by 20,000 times this way and of course this way too so it's 40 million times difference in area there's a lot of difference in area between the head of a pin and 20,000 Square ft uh then we could write the entire encyclopedia on a head of a pin and as I will indicate soon that's not too difficult.”

Extending this 20,000-fold two-dimensional reduction to entire academic collections illustrates the staggering capacity of miniature media. Feynman calculates that the entire Caltech Scientific Library could easily fit onto a single standard library card. In a playful scenario, he notes that if a major academic library in Brazil were ever destroyed by fire, Caltech could replace its entire institutional catalog simply by mailing them one card. Scaling up to the Library of Congress in Washington—vastly larger than Caltech's collection—Feynman estimates that all its volumes could comfortably fit inside the physical dimensions of a standard issue of *Time* magazine.

- *Encyclopædia Britannica* spans roughly 20,000 to 30,000 pages, taking up approximately 20,000 square feet of area if laid flat. - A 20,000-fold reduction in both length and width yields a 400,000,000-to-1 reduction in surface area. - At a 20,000-fold reduction, the entire Caltech scientific catalog fits on one library card. - The entire Library of Congress collection fits within the pages of a single *Time* magazine format.

---

[04:48](https://www.youtube.com/watch?v=4eRCygdW--c&t=288s) 3D Information Storage in a Speck of Dust

Feynman points out that putting pages onto the surface of a pinhead still only uses two dimensions, leaving the entire internal volume of the pin untouched. If instead information is packed into three-dimensional space by stacking bits through the depth of a material, storage density explodes. He envisions using little three-dimensional blocks measuring five atoms on a side—about one hundred atoms total per bit—encoded in Morse-style patterns of gold and silver atoms. Restricting the encoding strictly to text rather than images allows all written information across human history to be compressed beyond ordinary comprehension.

Feynman calculates that if one gathered every book from every library on Earth—Turkish, Hungarian, English, and all others—the aggregated text would fit inside a volume measuring just 1/200th of an inch on each edge. This microscopic volume corresponds to the barest speck of dust discernible to the naked human eye. Feynman underscores this dramatic contrast: “It turns out if you take all the books in all the libraries Turkish Hungarian everything all over the entire world and just take the information because I can't get the pictures at this scale then all this goes into a volume of material one two 200th of an inch on a side which is the smallest piece of dust that you can possibly see.”

Reflecting on this figure, Feynman quips that our everyday books are absurdly, unnecessarily oversized. Beyond convenience and portability, ultra-dense information storage offers a defense against historical catastrophes. Remembering the destruction of the ancient Library of Alexandria, where centuries of collective knowledge were permanently wiped out by fire, Feynman jokes that humanity should manufacture countless identical dust specks carrying the entirety of human knowledge and scatter them across the globe so that no civilization-ending disaster could ever destroy our collective heritage.

- Three-dimensional storage utilizes cubes roughly five atoms on an edge (~100 to 125 atoms per informational unit). - The text of every library book on Earth compresses into a single cube measuring 1/200 of an inch on a side. - A volume of 1/200th of an inch represents the limit of what can be seen by the naked human eye without magnification. - Dispersing duplicate dust particles offers absolute disaster insurance against cultural loss like the burning of Alexandria.

---

[07:01](https://www.youtube.com/watch?v=4eRCygdW--c&t=421s) Optical Limitations and Reverse Electron Microscopes

Having demonstrated how much information can be compressed, Feynman turns to the practical challenges of reading and writing at such microscopic scales. Reading an encyclopedia reduced by a factor of 20,000 cannot be accomplished using conventional optical microscopes. Because visible light possesses an inherent wave nature, optical instruments cannot resolve structures smaller than the wavelength of light, hitting an insurmountable optical limit around 2,000 magnifications. However, electron microscopes bypass this barrier entirely, routinely reaching magnifications from 20,000 up to 200,000 times. Resolving a 20,000-fold reduction is therefore well within the capabilities of existing electron optics. Feynman summarizes: “It would be very easy to read this book that we wrote on the head of the pin or this encyclopedia with an electron microscope.”

Writing at that resolution presents the reciprocal engineering challenge. Feynman suggests operating an electron microscope in reverse, much like turning a telescope around to look through the wrong end to make large objects appear tiny. By running the focusing lenses backwards, a broad beam containing the large-scale pattern of letters and pictures could be focused into an extraordinarily sharp electron beam to etch microscopic patterns. Feynman notes that when he originally proposed this concept twenty years earlier in his 1959 lecture, it was purely a theoretical prediction; while reverse electron beam systems remained difficult and underdeveloped, the fundamental physics was entirely sound.

- Standard light microscopes cannot resolve features beyond roughly 2,000 magnifications due to the wavelength of visible light. - Electron microscopes achieve magnifications between 20,000 and 200,000 times, making reading at a 20,000-fold reduction straightforward. - Demagnification for writing can be achieved conceptually by operating optical or electron lens systems in reverse. - The concept of the demagnifying electron beam writer originated in Feynman's initial 1959 presentation of the idea.

---

[08:45](https://www.youtube.com/watch?v=4eRCygdW--c&t=525s) Commercial Miniaturization and Computer Chip Fabrication

Feynman transitions from speculative records like etching the Lord’s Prayer to what modern industry was actively producing by the mid-1980s. Commercial incentives had shifted away from novelty engravings toward microelectronics and computer chips. Presenting a slide of an integrated circuit measuring roughly three millimeters across, he asks the audience to look past the familiar novelty of computing and appreciate the staggering mechanical achievement embodied in the silicon die. Feynman insists on acknowledging this milestone: “From a point of view of humankind and its development it's really quite an achievement to be able to manufacture something with such finess of detail.”

He reflects on how society frequently laments the decline of classic craftsmanship, praising old-world artisans who painstakingly hand-polished mechanical parts. Yet the geometric tolerances and alignment notches manufactured onto a modern computer chip far exceed the accuracy of any manual polishing in history. Industrial chip production routinely achieves linear reductions of 2,000 times using optical photolithography. By directing light through precision microscope lens systems operated in reverse, manufacturers project complex macroscopic circuit designs down onto silicon wafers coated with photosensitive material to mass-produce functional micro-machinery.

- Modern integrated circuit dies measure approximately 3 millimeters across. - Microelectronic manufacturing tolerances and alignment notches surpass the finest hand-polished mechanical craftsmanship in history. - Commercial photolithography routinely reduces complex circuit artwork by 2,000 times using inverted optical lens systems. - Optical projection remains the industrial standard for 2,000-fold reductions before reaching the physical limits of visible light.

[11:20](https://www.youtube.com/watch?v=4eRCygdW--c&t=680s) Photolithography and Chip Fabrication

Richard Feynman details the step-by-step chemical and optical mechanics used to construct integrated circuits on an industrial scale. The process begins with an exceptionally pure silicon substrate, chosen because at micro scales, the smallest speck of dirt or surface scratch resembles a “great big monster Boulder” that would ruin the entire structure. Inside a vacuum chamber, oxygen is introduced to form a thin surface layer of silicon dioxide—essentially quartz, sand, or a microscopic layer of glass. While silicon serves as a semiconductor/conductor, this glass layer functions as an insulator. Feynman enthusiastically grabs colored chalk to sketch out how the architecture builds upward layer by layer.

```mermaid flowchart TD accTitle: Photolithography Process Flow accDescr: Sequential steps in photolithography from pure silicon substrate to final metallic interconnection.

subgraph Prep ["1. Surface Preparation"] A["Pure Silicon Substrate"] --> B["Thermal Oxidation: SiO2 Glass Layer"] B --> C["Spin-Coat Photoresist Layer"] end

subgraph Pattern ["2. Optical Patterning"] C --> D["Project Patterned Light"] D --> E["Resist Exposed Areas to Solvent"] E --> F["Dissolve Unexposed Photoresist"] end

subgraph EtchBuild ["3. Etching & Layering"] F --> G["Hydrofluoric Acid Etch: SiO2 Dissolved"] G --> H["Strip Scaffolding Photoresist"] H --> I["Deposit Doped Silicon / Contacts"] I --> J["Evaporate Metal Interconnects"] end ```

On top of the insulating glass, a chemical layer known as photoresist is evaporated. Light is then projected onto it in a precise black-and-white pattern using an optical system. Where the light strikes, the photoresist hardens and becomes insoluble against subsequent chemical etching. The unexposed material is dissolved away, leaving protective caps over the silicon dioxide. Hydrofluoric acid is then applied to eat away the bare glass, carving out distinct insulating columns on top of the silicon base. Finally, the remaining photoresist scaffolding is chemically stripped.

Feynman explains that this cycle—applying photoresist, exposing a pattern, etching away material, and depositing doped silicon or metallic connectors—is repeated continuously across multiple strata to produce interconnected transistors. Looking down from above, the finished device represents a breathtaking union of industrial art and engineering: “It's an enormously complicated electric circuit they can make calculations all by itself with the design, the thing that determines the pattern is the original arrangement of light and I just wanted to show you how it is made.”

- Silicon acts as the base conductor/semiconductor, while silicon dioxide ($\text{SiO}_2$) acts as an insulating glass layer. - Hydrofluoric acid dissolves exposed glass regions while leaving photoresist-masked columns intact. - Current commercial lithography miniaturizes designs by a factor of roughly 2,000 times, hitting the physical resolution limit imposed by the wavelength of visible light.

---

[15:14](https://www.youtube.com/watch?v=4eRCygdW--c&t=914s) The World's Smallest Drawing

Feynman pivots from commercial chips—limited to a 2,000-fold reduction—to experimental research pushing toward a 20,000-fold or greater scale reduction. Because light cannot resolve details that small, alternative methods must be invented. Feynman cites cutting-edge experiments conducted at Cornell University's microfabrication laboratory (the National Research and Resource Facility for Submicron Structures). To illustrate this frontier, Feynman introduces his friend, the artist Tom Van Sant, whom he praises as a true Renaissance man: “He's a real man of the present era and not one artist who sneers at science and doesn't understand the world he's in, he loves the world he's in and he loves art.”

Van Sant collaborated with the Cornell researchers to execute the smallest drawing ever made in human history: a depiction of a human eye etched into a single salt crystal. Using a focused beam to bore holes and carve away the salt, the researchers created a pattern that could only be inspected afterward through an electron microscope. The completed drawing is reduced by a factor of 100,000 relative to a normal human eye. Across its entire span, the artwork measures approximately 100 atoms wide. Feynman notes with amusement that the researchers were actually disappointed because their beam could have resolved dots half that size and lines half as thick, but Van Sant specifically insisted on an exact 100,000:1 reduction. The image also features an accidental artistic sweep caused by a passing truck that vibrated the entire building, proving that at this atomic scale, the slightest environmental tremor becomes a massive disruption.

- Drawing location: Etched into a salt crystal at a Cornell University microfabrication facility. - Physical width: Approximately 100 atoms across. - Reduction scale: 100,000 times smaller than a natural human eye. - Imaging technique: Visualized exclusively using an electron microscope.

---

[18:20](https://www.youtube.com/watch?v=4eRCygdW--c&t=1100s) The World's Largest Drawing: Van Sant's Landsat Eye

Feynman reveals why Tom Van Sant insisted on the exact 100,000-to-one scale factor: Van Sant had previously created the largest drawing in human history, scaled to exactly 100,000 times *larger* than a human eye. Displaying slides showing abstract patches of color that expand outward to reveal eyebrows and orbital rings, Feynman zooms out until the entire Los Angeles metropolitan basin comes into view. The scale of the artwork is so vast—measuring 2.5 kilometers across—that it cannot be viewed from the ground; it can only be seen from 600 miles above the Earth via the Landsat satellite.

```mermaid flowchart LR accTitle: Van Sant Landsat Mirror Reflection accDescr: Path of sunlight reflecting off Van Sant's desert mirror array directly into Landsat satellite scanners.

subgraph Orbit ["Space (600 Miles Altitude)"] LS["Landsat Satellite Scanner"] end

Feynman explains the mechanics of how the satellite generates images. As Landsat flies overhead, its sensor beam sweeps back and forth across the terrain, calculating reflected light intensities in three distinct color bands moment by moment to assemble a grid of pixels. In Landsat's resolution, each individual pixel represents an entire acre of land. Rather than attempting to coat dozens of whole acres in white material, Van Sant utilized solar geometry. Collaborating with Landsat operators to obtain precise orbit times, he positioned an array of 24 mirrors throughout the desert. Each mirror was calibrated with spirit levels and transit angles so that at the exact second the satellite swept over that specific acre, the sun's reflection shot straight into the camera optics, completely saturating the sensor into a brilliant white pixel.

Feynman points out a solitary flaw in the desert satellite portrait where one pixel failed to light up properly. When Van Sant went back into the desert to discover what had gone wrong, he found that a wild jackrabbit had hopped across the mirror and knocked the alignment off by just enough to miss the satellite's aperture. Feynman chuckles at the detour: “I know that's not got to do with small things, it has to do with large things, but I couldn't resist after I showed you the smallest drawing ever made to show you that the same artist.”

- Canvas dimension: Spans roughly 2.5 kilometers across the desert outside Los Angeles. - Viewing platform: Visible only from orbit via the Landsat Earth observation satellite at 600 miles altitude. - Resolution equivalence: Each individual image pixel spans roughly one acre of ground area. - Mirror network: 24 precision-angled mirrors positioned across the desert floor to blind individual pixel sensors on flyby. - Flaw origin: A stray jackrabbit bumped one of the desert mirror mounts, knocking it out of alignment.

[22:47](https://www.youtube.com/watch?v=4eRCygdW--c&t=1367s) The Scale of the Small: From Eyelashes to Saturn

To convey the sheer miniaturization involved in micro-lithography, Feynman presents a thought experiment comparing the drawing of a child's eye to everyday objects and cosmic bodies. If one takes an ordinary human eyelash—a single hair—and magnifies it until it reaches the size of an entire eye, the original eye expands to the width of the lecture hall. Magnifying an eyelash of *that* massive eye across the room once again causes the eye itself to stretch to two and a half kilometers across.

Stepping up yet another factor of 100,000 creates an astronomical scale. Looking for something 100,000 times larger than that two-and-a-half-kilometer eye reveals a striking celestial counterpart: the planet Saturn ringed in space. Feynman points out that this exact ratio describes the gulf between our macroscopic world and micro-scale lithography: “As 2 and 1/2 kilometers is to your eye, so that your eye is to that original first drawing, which is so very tiny.”

- Scaling factor: Sequential jumps of $100{,}000\times$. - First jump: Eyelash magnified to eye size stretches the human eye across an entire room; repeating the step expands the eye to $2.5\text{ km}$. - Cosmic comparison: Another $100{,}000\times$ step matches the visual diameter of Saturn and its rings.

---

[23:56](https://www.youtube.com/watch?v=4eRCygdW--c&t=1436s) The Sound Barrier Analogy and Atomic-Scale Computing

Feynman shifts his focus from present industrial realities to physical limits: what is possible in principle versus what is merely practical today. When engineers attempt to shrink electrical traces down to a half or a third of their present size, they encounter electromigration—the phenomenon where the flow of electric current physically displaces the metal atoms until the wires pull themselves apart. While engineers view this failure mode as an insurmountable wall, Feynman compares it directly to the early twentieth-century myth of the sound barrier. Early aeronautical engineers assumed air was incompressible because slower propeller aircraft operated where air did not squeeze tightly together. When approaching Mach 1, those specific propellers and wing designs failed, prompting claims that humans could never fly faster than sound.

```mermaid flowchart TD accTitle: Sound Barrier Analogy for Nanotechnology accDescr: Flowchart illustrating how design limits differ from fundamental laws of physics. subgraph SoundBarrier["The Sound Barrier Precedent"] A["Incompressible Aerodynamics Design"] --> B["Propellers Fail Near Mach 1"] B --> C["Assumed 'Barrier' (Design Limit)"] C --> D["New Design (Jets/Supersonic Airfoils) Breaks Speed of Sound"] end subgraph Nanotech["The Atomic Wire Barrier"] E["Macroscopic Metal Wire Design"] --> F["Electromigration Destroys Thin Wires"] F --> G["Assumed 'Barrier' (Design Limit)"] G --> H["Quantum Atomic Architecture Stores 1 Bit per Atom"] end ```

Feynman stresses that conventional engineering barriers are often mistaken for limits of nature: “These limitations that you hear about mean only that you can't go on with the same design, but it doesn't mean that it's not possible in principle.” Bypassing wires entirely, Feynman asks how small a computer could theoretically be if built directly out of individual atoms interacting through quantum mechanical states. While such nanoscale devices must account for Heisenberg's uncertainty principle, his own theoretical investigations confirm that the basic laws of physics comfortably allow a computer to store and process a full bit of information within a single atom.

- Electromigration: Current-induced mechanical displacement of atoms tears extremely narrow wires apart. - Sound barrier parallel: Early aviation theories broke down because they treated air as strictly incompressible, mistaking a propeller limitation for an immutable law of physics. - Quantum-scale limit: Physics permits digital computing architecture down to a density of one atom per bit of information.

---

[26:33](https://www.youtube.com/watch?v=4eRCygdW--c&t=1593s) Tiny Machines with Movable Parts: The Master-Slave Manipulator Scheme

Turning from computing hardware to mechanical engineering, Feynman pivots to machines with actual moving parts. Anticipating the immediate skepticism of his audience, he remarks on the collective expression asking "what for?" His cheerful response is simply that exploring extreme miniaturization serves as pure "mental entertainment" and fun, pointing out that curiosity does not cost anything and practical applications usually follow later.

Feynman discusses the physical mismatch between human anatomy and microscopic mechanisms: “You see today we have great big hands and they're this fat and lumpy, and we can make tiny little watches, but that's nothing compared—I don't mean a computer watch, cause that's not got moving parts, but a machine with moving parts inside that is extremely small.” To bridge this gap, he revisits the master-slave telemanipulators used in radioactive hot-cells to handle hazardous materials remotely. By scaling down the electrical linkages, full-sized operator hands could manipulate a pair of miniature mechanical hands equipped with tiny wrenches. Those small hands could then assemble an even smaller set of hands, repeating the reduction recursively. To avoid ending up with only a single microscopic tool after an enormous expenditure of effort, each stage would fabricate multiple copies—two hands building four, four building eight—yielding tens of thousands of microscopic manipulators executing the same master control signals in parallel.

```mermaid flowchart LR accTitle: Recursive Master-Slave Manipulator Hierarchy accDescr: Exponential scaling of telemanipulator hands from macroscopic controls to nanoscale swarms. A["Operator Hands (Macro Controls)"] --> B["1st Generation Small Hands"] B --> C["2nd Generation Smaller Hands (2x Replication)"] C --> D["3rd Generation Micro-Hands (4x Replication)"] D --> E["Thousands of Nanoscale Slave Hands Operating in Parallel"] ```

---

[29:08](https://www.youtube.com/watch?v=4eRCygdW--c&t=1748s) Sacrificial Etching and Electrostatic Motors

Acknowledging that the telemanipulator strategy had seen limited experimental success, Feynman champions a far more viable method: adapting planar semiconductor manufacturing to produce free-moving mechanical parts. Rather than building a solid, monolithic block of material, a fabricator can deposit alternating layers of silicon, silicon dioxide, conductors, and a distinct sacrificial material (which he color-codes as "blue"). By selectively etching or chemically dissolving away the blue substance while leaving the surrounding structural material intact, engineers could create mechanically unconstrained, loose components resting inside guided channels.

Using the existing lithographic wiring techniques, electrical voltages can be routed to fixed conductive plates surrounding the loose piece. Electrostatic forces can then push and pull the component across the substrate. Feynman explains that an array of electrodes could actuate a series of parts in sequence, forming microscopic conveyors or tiny vehicles that transport materials across the silicon chip. He explains: “Therefore the second way of manufacturing small machines would be the evaporating, etch-freeing pieces of object, etc.—you can move them with electrical forces.”

- Structural layers: Silicon, silicon dioxide, and standard electrical conductors. - Sacrificial release: Depositing a selective material that dissolves completely in chemical etchant without damaging surrounding structural layers. - Actuation: Electrostatic attraction and repulsion controlled via microfabricated surface traces.

---

[31:22](https://www.youtube.com/watch?v=4eRCygdW--c&t=1882s) Applications to Optical Shutters and Limits of Light

While insisting he was not primarily concerned with immediate utility, Feynman offers an application for these electrostatic micro-actuators: high-intensity optical shutters. By positioning tiny movable plates over arrays of microscopic holes, electrostatic charges could open and shut individual apertures across a flat plane to modulate light. Unlike conventional cathode-ray tube television screens—which produce weak, diffuse light from a phosphor screen in a box—an array of microscopic light shutters could have an extremely bright external light source beamed straight through it. By opening and closing individual shutters electronically, it could project brilliant, high-contrast moving images onto large surfaces.

An audience member interrupts with a question regarding optical resolution limits: “Sir, what I'm confused about: you said something can only be magnified 2,000 times optically with optics problem. If you could reduce the light, as you're talking about small shutters or small cameras, would it be possible then?” Feynman clarifies the distinction between electron microscopy and optical limits. Light cannot resolve details smaller than its own wavelength; the shutter system he is describing operates near the 2,000-diameter optical boundary where light still passes cleanly through physical apertures. Emphasizing the physics of the wave limit, Feynman replies: “No, you can't make the light any smaller than the size of the wavelength of light.”

- Optical shutter display: A matrix of electrostatically actuated micro-shutters modulating a high-intensity background lamp. - Contrast with CRTs: Solves brightness limits of phosphor screens by acting as a high-throughput light valve for large-scale projection. - Diffraction barrier: Apertures and shutters cannot shrink below the wavelength of visible light if they are intended to modulate optical rays.

[32:49](https://www.youtube.com/watch?v=4eRCygdW--c&t=1969s) Precision Chemical Placement and Molecular Manufacturing

Feynman considers alternative techniques to circumvent the tedious photolithographic methods used to fabricate microstructures. Under the existing regime, aligning consecutive exposures and masks requires painstaking workmanship to ensure the second picture sits precisely over the first. To streamline this, he envisions an active chemical deposition device: a wide, flat structural framework fitted with an array of tiny nozzles featuring movable apertures and built-in micro-valves. By actuating these valves, chemical reagents could simply be squirted exactly where intended on demand. Because this approach removes optical lithography and light wavelengths from the fabrication equation, scaling limitations vanish, enabling manufacturing to become progressively smaller.

```mermaid flowchart TD accTitle: Comparison of Chemical Assembly Methods accDescr: Contrasts conventional lithography, active nozzle arrays, and direct atomic placement.

subgraph Macro["Conventional Methods"] A["Optical Lithography"] -->|"Tedious multi-mask alignment"| B["Imperfect Layering"] end

subgraph Micro["Mechanical Deposition"] C["Flat Nozzle Array"] -->|"Movable micro-valves"| D["Precision Jetting of Reagents"] end

subgraph Nano["Atomic Limits"] E["Direct Atomic Positioning"] -->|"Place individual atoms"| F["Synthesized Molecules without Wet Chemistry"] end ```

He then expands this vision toward replicating complex biological and chemical systems. In living organisms, specialized enzyme catalysts embedded within lipid membranes execute delicate, localized reaction control. While human engineers can currently assemble crude membranes with random chemical distribution, they lack the dexterity to position specific enzymes in predetermined layouts. Feynman notes that having microscopic tools capable of pushing individual components would fundamentally revolutionize chemical synthesis. Ultimately, at the absolute physical limit, synthetic chemistry would no longer rely on indirect laboratory manipulations:

> “All that a chemist does is rearrange make a new molecule with a different arrangement of atoms so if you went down to a small enough scale you just put the atoms down where you want them to make the molecule no hocus pocus with bottles and colored liquids actually that's really magical how they do that.”

- A flat matrix of micro-nozzles with mechanical shut-off valves could replace optical etching by squirting chemicals directly in place. - Eliminating light diffraction removes the physical resolution limits imposed by photolithography. - Nature controls metabolic pathways by anchoring specific protein enzymes onto membrane sheets. - Direct mechanical manipulation of matter at the atomic scale would replace bulk liquid-phase reaction chemistry.

---

[35:10](https://www.youtube.com/watch?v=4eRCygdW--c&t=2110s) Micromachines: Microscopic Gladiators and Swallowing the Surgeon

Turning to unconstrained, free-swimming mechanisms, Feynman speculates about untethered devices operating at microscopic dimensions. Reflecting on the commercial adoption of early computers, he notes that novel technologies often gain initial traction through games and entertainment. He imagines a microscopic aquatic gladiator operated via external electrical signals, equipped with a tiny sword so an operator peering through a standard optical microscope can battle single-celled paramecia in real time.

He then recalls an audacious medical concept proposed two decades earlier by his colleague Albert R. Hibbs. Hibbs proposed building miniature motorized surgical mechanisms capable of operating directly inside the body. The patient would ingest the micro-device, which would then navigate into the circulatory system under external control, travel into diseased blood vessels, and excise plaque from arterial walls:

> “A friend of mine named Al Hibbs suggested something else that's going to shock you and the idea is to swallow the surgeon the surgeon is a machine with knives that's very tiny they can go down around and be controlled and go into the blood vessels and down in here and carve away the plaque or whatever heck that has to do and come okay you swallow yes fantastic you swallow the surgeon.”

Feynman admits with self-deprecating humor that while the conceptual core of this lecture dates back twenty years to conversations with Hibbs, modern experimental imaging and microfabrication photographs prove that the technology is finally beginning to catch up with the theory.

- First conceptual consumer application envisioned: an operator using an electrically guided micro-swordsman to duel live paramecia under a microscope lens. - Al Hibbs' concept: "swallow the surgeon"—a remote-controlled, miniature automated surgical tool capable of arterial navigation and plaque excision. - Feynman acknowledges that while he originally voiced these ideas decades prior, emerging microfabrication techniques were finally validating the premise.

---

[36:38](https://www.youtube.com/watch?v=4eRCygdW--c&t=2198s) Dimensional Scaling: Heat Dissipation and the Failed Micro-Engine

Feynman warns that scaling down macroscopic machines is not simply a matter of geometrically shrinking components. Physical forces change drastically across dimensional scales; as an object shrinks, ratios between weight, tensile strength, air resistance, and electrical resistance shift unevenly. In an electric motor, simply scaling down the copper wiring causes electrical resistance to rise so high relative to power output that the device fails, necessitating radical design adjustments like superconductivity. He compares this paradigm shift to the replacement of propeller-driven aircraft with jet propulsion when scaling barriers emerged.

To illustrate how traditional physics rules break down when scaled down naively, Feynman presents a thought experiment: an internal combustion engine scaled down to microscopic size. In a full-sized engine, the chemical combustion of gasoline releases heat, the gas expands, and the piston drives mechanical work. Although heat leaks through the cylinder walls into the cooling jacket, this represents an insignificant fraction of total energy. However, scaling the engine down causes surface-area-to-volume ratios to derail the process entirely:

| Dimension / Property | Scaling Factor ($L \to 0.1L$) | Physical Implication | |:--- |:--- |:--- | | **Linear Dimension** | $10\times\text{ reduction}$ | Length, bore, and stroke shrink by a factor of 10 | | **Combustion Volume** | $1000\times\text{ reduction}$ ($10^3$) | Fuel payload and generated thermal energy decrease by 1,000 | | **Surface Area** | $100\times\text{ reduction}$ ($10^2$) | Cylinder wall surface area decreases by only 100 | | **Heat Loss Rate** | $10\times\text{ relative increase}$ | Heat leaks out into the cylinder walls ten times faster |

If scaled down 100 times, the mismatch becomes insurmountable:

> “As soon as you would explode this infinitesimal amount of gasoline in this tiny tiny cavity then the heat of it wouldn't be before it had any chance to push anything it would leak out immediately into the container so you have to redesign.”

- Scaling dimensions down by $10\times$ reduces cylinder volume and total combustion heat by $1,000\times$, while surface area drops by only $100\times$. - Surface area per unit volume expands tenfold, draining thermal energy straight into the casing before expanding gas can push the piston. - Scaling an internal combustion engine down 100 times renders it completely non-functional due to instantaneous thermal dissipation.

---

[39:49](https://www.youtube.com/watch?v=4eRCygdW--c&t=2389s) Low Reynolds Swimming: The Physics of Bacterial Propulsion

Feynman addresses fluid mechanics, an area where intuition from the human macro scale fails completely. When macroscopic observers imagine a microscopic submarine cruising through water, they fail to recognize how scale affects fluid drag. If an entity operating at micro-scale dimensions were blown up to human size, the surrounding water would feel as viscous as extremely thick honey. In this low-Reynolds-number regime, traditional locomotion mechanisms fail; a micro-scale fish flapping its tail fins cannot propel itself forward via fluid inertia, as it merely gets glued into the surrounding "goo."

```mermaid flowchart LR accTitle: Viscous Regime Locomotion Mechanics accDescr: Shows the distinction between macroscopic fin propulsion and microscopic corkscrew flagellar motion.

subgraph MacroScale["Macro-Fluid Regime"] A["Flapping Fins / Paddles"] -->|"Relies on water inertia"| B["High-speed forward motion"] end

subgraph MicroScale["Micro-Fluid / High Viscosity Regime"] C["Low Reynolds Environment"] -->|"Fluid behaves like thick honey"| D["Rotating Flagellar Screw"] D -->|"Internal rigid core turns inside sleeve"| E["Screws vehicle forward through viscous media"] end ```

To demonstrate that micro-locomotion is physically viable despite these viscous forces, Feynman appeals to biology:

> “How do I know then that I can make very tiny machines that it's possible according to the laws of physics because I'm a physicist and I check the laws of physics and it's okay I'm telling you but I got another way to tell you living things have already done it bacteria swim they swim through the water at the scale that corresponds to thick goop.”

Bacteria bypass the inertia problem using flagella configured like rigid rotating corkscrews. A central helical filament turns inside an outer sleeve, threading through the viscous medium much like a physical screw biting into cork or thick honey. While rotating a corkscrew in open water would be useless for a human swimmer, in the microscopic regime it is an optimal mode of propulsion. Feynman concludes that navigating these physical changes makes miniaturization an exhilarating engineering puzzle. Closing the lecture, he quips that practical applications will undoubtedly emerge in twenty or thirty years, joking that he will return then to give another talk explaining what they are finally used for, before opening the floor to questions.

- In microfluidic realms, water exhibits the effective viscosity of heavy honey; fin-based flapping fails because inertia is negligible. - Bacterial flagella operate on non-inertial locomotion: a rigid corkscrew shaft rotates inside a sheath, screwing through the surrounding fluid. - Macro-scale screw propulsion in water is inefficient without high rotational speeds, whereas low-Reynolds-number fluid dynamics makes it ideal. - Feynman predicts that practical nanotechnological applications will emerge within 25 to 30 years, concluding the session and inviting audience questions.

[43:23](https://www.youtube.com/watch?v=4eRCygdW--c&t=2603s) Scaling Up vs. Scaling Down: Viscosity and Aerodynamics

An audience member asks whether scaling up works in reverse: if fluids appear extremely viscous to microscopic entities, would a gigantic being moving through the air experience the atmosphere as thick and soupy enough to support its flight? Feynman corrects this intuition, explaining that scaling up does not make air behave like honey. At macroscopic scales, the "ickiness" or true viscosity of air is already relatively negligible. Propellers and wings initiate dynamic flows where the governing factor is inertia, not fluid thickness. An airplane stays airborne by pushing down a mass of air; the recoil from that downward momentum counterbalances gravity.

```mermaid flowchart TD accTitle: Scaling Effects on Aircraft Flight accDescr: Diagram illustrating how scaling an aircraft up changes the balance between wing lift and mass.

subgraph ScaleUp["Scale-Up Factor: L × 100"] Area["Wing Area: Scales by L² (10,000×)"] Volume["Aircraft Weight: Scales by L³ (1,000,000×)"] end

subgraph Outcome["Flight Mechanics"] Lift["Dynamic Lift (Air Momentum Recoil)"] Deficit["Deficit: Weight Outpaces Lift by 100×"] Solution["Engineering Fix: Make Structure Hollow"] end

Area --> Lift Volume --> Deficit Lift -.->|"Insufficient to balance"| Deficit Deficit --> Solution ```

Scaling an airplane up directly, however, encounters a different mathematical barrier: the square-cube law. If an identical airplane design were simply enlarged 100 times in linear dimensions, it would be unable to leave the ground. Its lifting surface area increases only by the square (a factor of $100^2$ or 10,000), while its weight increases by the cube (a factor of $100^3$ or 1,000,000). To compensate, actual large airplanes must be engineered with vast hollow spaces rather than solid bodies to prevent dead weight from outracing wing area. Feynman underscores that different physical properties dominate at different scales: “You're going the wrong way; viscous—you got to get effective viscosity going for smaller scale.”

- Dynamic resistance dominates over viscosity for large airborne objects. - Lift relies on Newton's third law: downward acceleration of air mass provides the upward recoil force. - Structural weight scales as length cubed ($L^3$), whereas wing lift scales as length squared ($L^2$). - True high-viscosity behavior only becomes dominant when shrinking down to micro scales.

---

[45:35](https://www.youtube.com/watch?v=4eRCygdW--c&t=2735s) Anti-Gravity and "Imagination in a Tight Straitjacket"

When an attendee asks how an original thinker like Feynman would design a miniaturized anti-gravity machine, Feynman answers without hesitation: “I don't know how to make any anti-gravity machine.” Even if such an invention could solve monumental human problems, wishing for an outcome cannot overturn known physical realities. Feynman articulates his personal philosophy of scientific inquiry, describing it as a strict intellectual exercise: “The game I play is a very interesting one: it's imagination in a tight straitjacket, which is this—that it has to agree with the known laws of physics.”

He distinguishes between mere fantasy and disciplined creativity. While anyone can design impossible technologies by inventing arbitrary new laws of physics, real science requires pushing the boundaries of what is possible strictly within the established framework of verified nature. Once a concept is imagined, it must immediately be subjected to rigorous analysis to see if it violates observed principles. Under current gravitational theory, there is no physical mechanism to cancel or invert gravity. He quips with deadpan humor that the only workable anti-gravity machines are mundane support structures: “The only anti-gravity machine—that is, things which oppose gravity, which are very effective—are things like you're using now: a pillow or a floor under your behind that is an anti-gravity machine, and will support you in space above the Earth a few feet, in this case, for reasonably unlimited time.”

---

[47:20](https://www.youtube.com/watch?v=4eRCygdW--c&t=2840s) The Scope and Certainty of Physical Laws

In response to a follow-up question regarding how absolute the known laws of physics truly are, Feynman maps out the boundaries of human knowledge. Science always has an unsettled edge where discoveries continue to emerge. However, within specific experimental realms that have been tested exhaustively over generations, physical laws are settled with tremendous certainty. The validity of physical predictions depends on the regime of the variables being investigated.

If one ventures into speculative realms—such as distances a million times a million times a million times smaller than a centimeter (sub-nuclear or Planckian scales)—physicists do not yet possess definitive answers and must remain humble. But in the domain discussed in his lecture, stretching from standard macroscopic dimensions down to atomic interactions, the underlying physics is exceedingly well established. Feynman emphasizes that within this intermediate realm, “the laws that I need for the scales that I've been using, down to atomic scales and so on, are in extremely good control; it's not very likely that I'm making any mistake there.”

---

[48:29](https://www.youtube.com/watch?v=4eRCygdW--c&t=2909s) Atomic Jiggling, Thermal Fluctuations, and Information Storage

A participant raises a challenge from earlier in the talk: if matter is engineered at the scale of 5 to 100 atoms, how can one prevent the individual atoms from constantly moving around? Feynman clarifies that at ordinary temperatures, atoms are never entirely stationary; they undergo constant thermal vibrations. However, in solid materials such as gold or silver, the chemical binding forces holding neighboring atoms together are strong enough to trap them in a crystal lattice. Each atom jiggles, but it remains locked in place by its neighbors.

Feynman uses the analogy of a marching band: individual musicians may sway or move erratically, but the collective formation remains coherent because each person constantly readjusts relative to their peers. If the thermal energy rises significantly, or if the interatomic forces are too weak, the atoms lose their positions and roll past one another, transitioning from a solid into a liquid state. This explains why Feynman specified data storage clusters of roughly 100 atoms ($5 \times 5 \times 5$ atoms) rather than single atoms. If a memory bit relied on only a few atoms, random thermal fluctuations would eventually knock an atom loose; over years, the entire mark would degrade. Calculating the stability of a 5-atom-thick cube of gold reveals that thermal agitation would require billions of years to disrupt the data, ensuring the preservation of human records across deep time.

When asked whether focused lasers could be utilized to manipulate or write on that tiny atomic scale, Feynman explains that light faces a hard physical boundary set by its wavelength: - Visible light cannot be focused down to individual atoms due to the diffraction limit. - Lasers achieve reductions on the order of 2,000 times smaller than macroscopic scales (sub-micron), but cannot reach the atomic scale. - Manipulating individual atoms with light is akin to attempting to pick up minuscule specks using giant, clumsy tools much larger than the target itself.

---

[51:33](https://www.youtube.com/watch?v=4eRCygdW--c&t=3093s) The Speed of Light as an Impassable Barrier

A listener draws an analogy between historical sound-barrier breakthroughs and the speed of light, asking whether the laws of physics permit matter to travel faster than light. Feynman states unequivocally that they do not. While humans originally failed to understand what would happen at Mach 1 due to dynamic pressure challenges, special relativity represents a fundamentally different, fundamental boundary.

According to relativistic mechanics, any object initially traveling slower than light cannot be accelerated to or past the speed of light ($c$). Adding kinetic energy increases the object's velocity, but it also increases its effective relativistic inertia. Pushing it harder requires progressively massive amounts of energy for ever-diminishing gains in speed. As an empirical example, Feynman points to high-energy particle accelerators like the Stanford Linear Accelerator Center (SLAC). In such machines, electrons are pumped with immense quantities of electrical energy until they fall short of light speed by only an imperceptible fraction—roughly 1 part in 160 million. At these extreme velocities, the relativistic mass of each electron swells to roughly 40,000 times its resting mass, making a single electron around 20 times heavier than a stationary proton.

As the excerpt ends, an audience member raises a question about antimatter, wondering whether hypothetical entities in a spiritual realm beyond normal human vision could be composed of matter vibrating at velocities exceeding the speed of light.

[54:06](https://www.youtube.com/watch?v=4eRCygdW--c&t=3246s) Observability, Antimatter, and the Nature of Science

A member of the audience presses Richard Feynman on whether unseeable phenomena might operate under different laws. Feynman responds with characteristic scientific pragmatism, insisting that physics is fundamentally bound to empirical measurement. If someone asks about entities that cannot be detected by any instrument whatsoever, he notes, those entities could be doing anything at all, and science simply has no method to verify them. He emphasizes:

> “The only thing I can tell you are the things that I can make observations about if you ask me questions about something that I cannot see with any instruments whatsoever they can be doing anything whatever I have no way to tell.”

When the questioner pivots to antimatter and whether it represents something spiritual or vibratory, Feynman rejects the mysticism outright. He clarifies that antimatter is fully accessible to laboratory experimentation; physicists routinely create it, measure it, and understand its governing laws. It is ordinary physical substance just like conventional matter. When asked whether it vibrates, Feynman points out that matter’s vibration depends strictly on circumstances—just as ordinary matter vibrates when you strike a bell with a hammer. When another audience member asks, “What makes the world work?”, Feynman quips with a laugh about how easy so-called "small questions" are to answer, deflecting the vast metaphysical riddle with humor.

[55:50](https://www.youtube.com/watch?v=4eRCygdW--c&t=3350s) Writing Books on Wires and Fundamental vs. Engineering Problems

Asked what area of research he is currently pursuing, Feynman explains that the tiny machines concepts he is presenting are not his formal daily research, but rather an ongoing hobby he has played with on and off since around 1950. He recounts sitting on a beach daydreaming about the absolute minimum space required to record an entire book. His theoretical solution was to evaporate layers of materials one after another across a large area—allocating roughly five atoms per bit of information. By his calculation, a stack about one inch thick over a one-inch square would store the entire volume; slicing that block into micro-wires yielded thousands of copies of the identical text, each preserved on a wire merely two inches long. He remembers running around the beach excitedly asking strangers if they realized what was possible, joking that he "didn’t sell any bananas."

```mermaid flowchart TD accTitle: Book Miniaturization by Layer Deposition accDescr: Process showing how evaporating layers allows writing thousands of duplicate books on micro-wires. subgraph Deposition ["Layer-by-Layer Fabrication"] A["Evaporate thin material layers across 1-inch square base"] --> B["Encode data at ~5 atoms per bit"] B --> C["Accumulate layers to ~1-inch total thickness"] end subgraph Slicing ["Parallel Duplication"] C --> D["Cut block longitudinally into thin micro-wires"] D --> E["Yields thousands of wires (~2 inches long), each containing a full copy of the book"] end ```

When an audience member asks him to name the most difficult problem in physics he can visualize, Feynman emphasizes the difference between basic physics and applied engineering. He characterizes miniaturization and micro-machinery as an engineering puzzle:

> “This is a a game engineering problem because the sizes I've used are well within the realm where there's not much uncertainty and there's technical problems of making it which are interesting but not the most important problem in physics.”

For Feynman, fundamental physics involves identifying the domains where current laws fail, rather than cleverly exploiting well-established laws.

[58:17](https://www.youtube.com/watch?v=4eRCygdW--c&t=3497s) Theorists, Data Droughts, and the Future of Discovery

An attendee brings up Feynman's classic book *The Character of Physical Law*, noting Feynman’s twenty-year-old prediction that the excitement in physics might eventually diminish as discoveries saturate, comparing it to tourists overrunning and spoiling a pristine landscape. Feynman reflects on his past prediction and admits he made a miscalculation about human nature:

> “I didn't realize a propensity of human beings to speculate and so as a matter of fact when less data comes in more theorists are working with more crazy theories than before and I hadn't contemplated that because my day things were more vigorous when we got a theory we could check it.”

While experimental data has indeed slowed down because experiments have become progressively larger, costlier, and more arduous, theorists have responded by having "much more fun" inventing wilder conjectures without the immediate discipline of experimental refutation.

Noticing audience members beginning to shuffle toward the exits for other evening activities, Feynman graciously suggests letting those who need to leave depart, while offering to stay behind to answer questions from whoever wants to remain.

- In his earlier career, theoretical ideas were tethered strictly to rapid, vigorous experimental testing. - Slower empirical feedback loops ironically inflate theoretical speculation rather than subduing it.

[01:00:13](https://www.youtube.com/watch?v=4eRCygdW--c&t=3613s) The End of Physics vs. Open-Ended Applied Science

Addressing a question about whether science will eventually discover all laws and run out of fundamental work, Feynman lays out two sharply divergent paths for fundamental research versus applied science. For fundamental laws, he considers it possible that physicists might successfully unify every law so that every experiment fits, concluding the quest. Alternatively, fundamental physics might stagnate if experiments become excessively complex, abstract, and unaffordable, leading society to withdraw support from the field's "high priests."

Applied physics, on the other hand, faces no such ceiling:

> “We're talking about two things finding new laws that is discovering new things about nature and the other is finding new ways to use what we know the second one will go on much longer than the first and I cannot envision how it stops.”

Feynman sees the permutations of applying known principles to technology, materials, and computation as virtually infinite.

[01:01:56](https://www.youtube.com/watch?v=4eRCygdW--c&t=3716s) Thermodynamic Cycles at the Microscale and the Limits of Light

Returning to the mechanics of micro-devices, a participant asks how heat engines could possibly run at minuscule scales given the severe constraints of cooling rates and thermodynamics. Feynman points out that heat engines do not need to rely on macroscopic pistons or molecules bouncing in a gas chamber. Heat represents random motion, but that randomness can manifest in various degrees of freedom—such as spin systems, nuclear magnetism, or magnetic dipoles flipping back and forth while the atoms remain fixed in place. He notes that living biological organisms obtain work from chemical energy without resembling internal combustion engines, demonstrating that engineers should not stay locked into conventional mechanical layouts.

After humorously brushing off a comment about whether physics establishments burn original thinkers at the stake, Feynman turns to a lingering question about why optical lithography cannot be reduced indefinitely. He explains the physical limitation of light using the analogy of ocean waves and corks floating along a shoreline. If waves possess a fixed wavelength, it is physically impossible to focus or manipulate them down to an area smaller than that characteristic length:

> “It's impossible to figure out how to get waves at that this normal wavelength to focus any shorter than about this size you can by bringing the waves from different sides from a much bigger Dimension bring things in focusing the waves but you can't get them any shorter than this characteristic length of the weight and so the sloshing on the shore will move corks over about this much area it's impossible to pick one cork.”

Diffraction dictates that wide waves inevitably disturb an entire swath of corks across their characteristic dimension, preventing light from isolating or resolving features smaller than its own wavelength.

[01:04:55](https://www.youtube.com/watch?v=4eRCygdW--c&t=3895s) Overcoming Lithography Limits: X-rays and the Electromagnetic Spectrum

Responding to an audience question about overcoming the fundamental resolution limits imposed by the wavelength of light, Richard Feynman clarifies that using different wavelengths is precisely the path forward. Shorter wavelengths are simply called ultraviolet light or X-rays. While moving to X-rays requires completely abandoning traditional silver-based photographic emulsions and inventing alternative lithographic techniques to expose photoresists, Feynman acknowledges that many laboratories are actively pursuing X-ray lithography to shrink features far below standard optical limits. Admitting his earlier slip, Feynman says: “I'm sorry I answered your question wrong before saying there was a limitation because it's light I proposed I was assuming it was light.”

```mermaid flowchart LR accTitle: The Electromagnetic Spectrum accDescr: Continuity of the electromagnetic spectrum from long low-frequency waves to short high-frequency radiation.

subgraph LowFreq["Long Wavelength / Low Frequency"] direction TB A["Radio Waves"] --> B["Television Waves"] B --> C["Infrared / Heat Waves"] end

subgraph Optical["Optical Window"] direction TB D["Red Light"] --> E["Blue Light"] end

subgraph HighFreq["Short Wavelength / High Frequency"] direction TB F["Ultraviolet Light"] --> G["Soft / Hard X-rays"] G --> H["Gamma Rays"] end

LowFreq --> Optical Optical --> HighFreq ```

Feynman emphasizes that the entire span—from radio waves to gamma rays—is unified under a single physical framework: - Ultraviolet light cuts feature sizes roughly in half compared to visible light. - Progressing further into extreme ultraviolet, soft X-rays, and hard X-rays allows wavelengths to shrink smaller than an individual atom, smaller even than an atomic nucleus. - In the opposite direction, longer wavelengths stretch through red light, infrared, far infrared, heat waves, television broadcast frequencies, and radio waves. - From a fundamental physics standpoint, all of these phenomena are identical: harmonic oscillations of electromagnetic radiation operating across an enormous continuous scale, differing only in their oscillation frequencies and wavelengths.

---

[01:07:21](https://www.youtube.com/watch?v=4eRCygdW--c&t=4041s) Detecting Subatomic Particles and Nuclear Probing

An audience member asks how anyone can possibly detect and measure objects that are smaller than atoms when our tools seem to be made of atoms themselves. Feynman teases the questioner with characteristic humor before diving into an intuitive mechanical analogy: “I have such trouble with you you're always confused how can you detect something that's smaller than atoms what can we have to detect them with that takes a great deal of ingenuity.”

He proposes imagining a wall composed of unknown targets, with only tennis balls available to throw at it: - If you throw tennis balls at random against the wall, the fraction that bounce back versus those that pass cleanly through directly reveals the effective cross-sectional area (the size) of the targets. - A target the size of a pin will register far fewer reflections than a target the size of a tennis ball. - Beyond simple collision rates, physicists leverage the relationship between collision energy and wavelength: bombarding metal targets with energetic electrons generates radiation whose wavelength decreases as particle energy increases. - Even when wavelengths become too minuscule to observe directly, the underlying physical laws hold, allowing researchers to calculate dimensions that correlate across extensive internal experimental checks.

Feynman sketches the staggering emptiness inside matter: if an atom were enlarged to the size of an entire lecture room, its nucleus would be a tiny speck—roughly 100,000 times smaller, like the printed period at the end of a sentence. Modern high-energy accelerators push probe wavelengths down to one-thousandth the diameter of the smallest nucleus, enabling physicists to map the interior structure of protons. As Feynman remarks wryly: “That's what your taxpayer money has been able to produce.”

---

[01:10:22](https://www.youtube.com/watch?v=4eRCygdW--c&t=4222s) Acoustic Waves, Ultrasound, and the Atomic Cutoff

A participant asks whether sound waves could be used instead of light to fabricate or probe micro-architectures. Feynman initially hesitates, noting that ordinary audible sound has wavelengths far too long for millionths-of-an-inch manufacturing, and that thermal agitation inside matter causes severe attenuation and absorption. However, as the questioner presses on ultrasound, Feynman reconsiders: “That's a nice idea I'm going to think about it some more no I'm not fooling because it's not so easy to get to the dimension of X-rays is rather complicated but to get sound that has the size for the X-rays is if I'm not mistaken not that difficult although I'm not sure I have to check the numbers and stuff okay but it's not a bad idea I think tomorrow I'll see it's a bad idea but right now...”

```mermaid flowchart TD accTitle: The Physical Cutoff of Acoustic Waves accDescr: Progression of acoustic frequency until wavelength matches atomic spacing.

A["Audible Sound: Long wavelengths, macroscopic vibrations"] --> B["Ultrasound: High frequency, sub-micron wavelengths"] B --> C["Interatomic Limit: Wavelength equals distance between adjacent atoms"] C --> D["Fundamental Cutoff: Collective wave motion ceases; sound loses physical meaning"] ```

When asked what happens if acoustic frequency is driven indefinitely higher, Feynman details the hard physical boundary where sound ceases to exist: - Sound is not an abstract field; it is organized kinetic motion traveling through discrete matter. - As the frequency increases, the wavelength eventually approaches the interatomic spacing of the lattice carrier. - To illustrate why higher pitches cannot exist, Feynman invokes a chorus line: “You get this wave that they like to make you know one one end does the next next next next so you see this wave of beautiful legs rising okay over a distance but it's perfectly clear that if I would ask you how with those Rockettes can you make a wave where the distance between the ups and the downs is less than the distance between the Rockettes how can I answer there is no such a way.” - Once adjacent atoms are oscillating in complete opposition (one up while its immediate neighbor is down), no shorter acoustic wave can be constructed; the concept of sound loses physical meaning entirely.

---

[01:14:00](https://www.youtube.com/watch?v=4eRCygdW--c&t=4440s) Relativistic Electrons and Physical Dimensions

The discussion shifts to high-speed particle dynamics when an audience member asks about accelerating electrons toward the speed of light: does an electron remain an electron as its relativistic mass increases, and does its diameter change? Feynman responds directly: “Oh yes it's still an electron just a electron moving very close to speed of light and has the correct properties for such an object.”

Regarding electron size, Feynman cautions that an elementary electron possesses no experimentally measured classical diameter, making the question ill-posed without a rigorous quantum field definition. However, if one considers a composite object with an unambiguous physical size, such as an entire atom, accelerating it to near-light speed leaves its transverse diameter unaltered in directions perpendicular to motion, while relativistic length contraction compresses it along its line of flight.

[01:14:50](https://www.youtube.com/watch?v=4eRCygdW--c&t=4490s) The Most Exciting Discoveries: Weak Interactions and Beta Decay

Richard Feynman finishes an aside regarding special relativity and how high-speed objects become distorted. He remarks that perpendicular dimensions remain unchanged while the moving object appears increasingly flattened into an ellipsoid, ultimately resembling a thin disc as its velocity approaches the speed of light. Chuckling, he tells the audience that he is not actually giving a lecture for his idiosyncratic thinking course on relativity, joking that he merely tossed it out as an advertisement for the upcoming year before opening the floor to the audience for remaining questions.

An audience member poses an interview-style question, asking Feynman what single thing he would consider the most exciting discovery, problem, or concept he has played with across his entire scientific career. Feynman pauses thoughtfully, noting that picking a single moment is an accident of personal history rather than an absolute judgment of profound scientific importance. He clarifies that what makes a moment truly exhilarating is when the solution arrives suddenly after prolonged effort. While these might not be his objectively most famous discoveries, two specific instances stand out in his memory for the sheer fun and excitement they brought him.

The first was his work formulating the fundamental laws governing beta decay and the weak interaction. Feynman recalls working through guesses for the algebraic structure of the weak force. While he acknowledges that his formulation was not the final word, he emphasizes that it was significantly closer to the truth than any existing theory at the time. His formulation held up remarkably well for years until a subsequent generation of physicists expanded and refined it into the modern, complete electroweak framework. As Feynman reflected on looking back at his theoretical breakthroughs, “It’s not the most important discovery, but it’s the one that was fun because, well, there were two like that, so it’s a little hard.”

- In relativistic kinematics, an object moving near light speed flattens along its direction of motion into an ellipsoid and eventually appears disc-like, while transverse dimensions remain unaffected. - Feynman's first highlighted thrill was deriving the V−A structure of the weak interaction governing beta decay, outperforming prevailing models before being incorporated into broader standard theory.

[01:16:30](https://www.youtube.com/watch?v=4eRCygdW--c&t=4590s) Cracking the Quantum Mystery of Liquid Helium

Feynman identifies his second great thrill as the two years he spent deciphering the bizarre physics of superfluid liquid helium. When helium gas is liquefied and cooled to sufficiently low temperatures below the lambda transition, it completely sheds its viscosity and flows without friction. To illustrate this frictionless behavior, Feynman describes a classic laboratory apparatus: a platinum wire sealed tightly into a glass rod. Because platinum and glass possess slightly different coefficients of thermal expansion, dropping the apparatus to cryogenic temperatures causes them to contract at unequal rates, opening an infinitesimal crack along the seal.

When this hairline fissure connects two vessels of liquid helium at different heights, the superfluid flows effortlessly through the opening to equalize the levels. Because it encounters zero viscous drag, the fluid's momentum causes it to overshoot the equilibrium level, coasting upward in the second container before oscillating back and forth indefinitely. Although physicists knew the basic laws of quantum mechanics ought to explain this behavior, the presence of billions upon billions of strongly interacting helium atoms made the equations impossible to solve directly by brute force. Unraveling the phenomenon required visual and physical imagination rather than raw computation.

```mermaid flowchart TD accTitle: Platinum-Glass Differential Contraction Mechanism accDescr: Diagram showing how cooling glass and platinum wire opens a microscopic fissure that allows superfluid liquid helium to flow frictionless. subgraph Preparation["Apparatus Setup"] A["Platinum wire sealed in glass rod"] B["Cryogenic cooling"] A --> B end

subgraph MechanicalEffect["Differential Contraction"] C["Unequal thermal expansion rates"] D["Infinitesimal crack forms along seal"] B --> C C --> D end

subgraph SuperfluidFlow["Superfluid Transport"] E["Liquid helium enters fissure"] F["Zero viscosity flow"] G["Overshoots equilibrium level and oscillates"] D --> E E --> F F --> G end ```

The breakthrough arrived in a sudden flash. Feynman had conceived what he suspected was the correct theoretical mechanism, but he immediately talked himself out of it with a counterargument based on an assumed smooth experimental curve. He found himself trapped in intellectual frustration until he suddenly remembered data from another experiment proving that the curve was not smooth after all. In an instant, the contradiction vanished and the theoretical model fell into place. Recalling the relief of that sudden realization, Feynman explains: “I had the idea and I had talked myself—I couldn’t see what was the matter because I had this wrong idea about a certain curve from experiment, and I forgot, and then I suddenly remembered this fact which would have saved the explanation, so it was like suddenly discovering it opened up.”

- Superfluid liquid helium exhibits zero viscosity at low temperatures, permitting perpetual flow through microscopic apertures. - Differential thermal contraction between platinum and glass creates controlled sub-microscopic channels for testing frictionless flow. - Because helium atoms in the superfluid state lack resistance, fluid flowing between connected columns coasts past the point of gravitational equilibrium, oscillating back and forth without viscous damping.

[01:18:23](https://www.youtube.com/watch?v=4eRCygdW--c&t=4703s) The Euphoria of Discovery and "Wait Till Tomorrow"

Feynman elaborates on the psychological rush of scientific insight, explaining that the true joy of research occurs when years of stalled labor give way to an explosive moment of total clarity. In those moments, a researcher feels they have grasped the master key: ideas begin pouring out, connections form across multiple problems, and one can sprint through detailed calculations to see how far the insight reaches. Yet he cautions that this intoxicating euphoria is accompanied by the constant danger that one has overlooked an elementary mathematical error or misunderstood empirical facts.

With self-deprecating humor, Feynman reveals that his domestic life often serves as a grounding reality check against these theoretical highs. Whenever he believed he had made an earth-shattering breakthrough, he would rush downstairs to tell his wife that he had finally cracked the problem. Recounting her customary, skeptical reaction to his exuberance, Feynman mimicked her pragmatic reply: “I come down, ‘I’ve gotten it! I got it, I got it!’ You know, ‘I’ve heard it before. Wait till tomorrow.’”

He admits that her skepticism was almost always justified. More often than not, the light of the following morning revealed either a subtle flaw in his reasoning or the deflating discovery that someone else had beaten him to the punch long ago. He quips about the running academic trope of finding out that “the Russians have already discovered it 200 years ago—like tennis!” With the room laughing at the joke, Feynman brings the question-and-answer period and the evening lecture to an end, thanking the host with a lively sign-off: “Thank you, Mr. President. Okay. Help, help, I was okay!”

- The emotional experience of scientific discovery is characterized by a rapid cascade where a single unifying principle suddenly unlocks multiple peripheral details. - Feynman's wife served as a skeptical counterbalance to his theoretical excitement, advising him to sleep on claims before celebrating. - Morning evaluations frequently exposed faulty mathematical steps or revealed that Soviet theorists had published the identical finding years earlier.

## Notable quotes > [45:35](https://www.youtube.com/watch?v=4eRCygdW--c&t=2735s) The game I play is a very interesting one: it's imagination in a tight straitjacket, which is this—that it has to agree with the known laws of physics. — Richard Feynman

## Mentioned Richard Feynman, Tom Van Sant, Albert R. Hibbs, Encyclopædia Britannica, California Institute of Technology (Caltech), Caltech Scientific Library, Library of Congress, Time magazine, Library of Alexandria, Cornell University, National Research and Resource Facility for Submicron Structures, Landsat, Werner Heisenberg, Stanford Linear Accelerator Center (SLAC), The Character of Physical Law, Rockettes

## Video description Donate and Support this Channel: https://www.paypal.com/donate/?cmd=_s-xclick&hosted_button_id=BLJ283JMTMT7S (Original Transcript Available Here:http://muonray.blogspot.ie/2012/12/richard-feynman-theres-plenty-of-room.html) Richard Feynman gave his famous talk "There's Plenty of Room at the Bottom" on December 29th 1959 at the annual meeting of the American Physical Society at the California Institute of Technology (Caltech) as his vision on how physics and engineering could move in the direction that could eventually create nanotechnology.

Really good ideas and strokes of genius are often manifest in the right questions being asked: How small can information be encoded? How can information be written? How can information it be read? All of these important "Hows" were asked by Feynman in a time when computers had to be put in large rooms and when the impending space race was forcing engineers to do some serious strategic thinking in making technology small enough to be lifted by rockets into space to function as serious tools in scientific exploration and defence.

Feynman himself may not have invented the technology we see in the development and continuity of the computer age, but the fact that even in the early 1960's nanotechnology was being considered as a serious field of study was definitely a factor contributing to the boom in computer technology seen in the late 20th century and continues to reach more spectacular levels of sophistication in the 21st century.

Jump 25 years forward into the year 1984, when Feynman tries to retell his 1959 lecture from a more modern perspective in that many aspects of his vision have been fulfilled, particularly with the invention of the electron microscope, the atomic force microscope and experimental manipulation of the atomic scale of matter. Also discussed is the current practical field of photolithography for the manufacture of bipolar transistors and junctions used in computer chips done on an industrial scale and how this process continues with ever decreasing wavelength capabilities of lasers from UV to X-rays. Feynman also discusses the boundaries of miniaturization and how the scale differences affect the function of certain aspects of technology as well as in nature.

In the true spirit of Feynman, the discussion goes into the colorful details and gives diagrammatic examples of how this field had progressed from 1959 to 1984. We can only imagine how Feynman would have felt about the modern developments in nanotechnology in the 21st century where entirely exotic principles of physics may begin to become technologically significant, including vacuum fluctuations and quantum entanglements. Without a doubt he would have found our developments exciting but always within the realms of understanding by studying the most fundamental language of nature, quantum mechanics, to bring our macroscopic brains into visualizing, however abstractly, the tiny machinery of nature.

[Source](https://www.youtube.com/watch?v=4eRCygdW--c)

中文简报

费曼称原子尺度机器面临的是设计限制而非物理限制物理定律允许将机器与数据存储缩小至单原子尺度，因为显见的技术障碍源于有缺陷的宏观工程假设，而非基本极限。费曼探讨了极致微型化的理论与实践前沿，展示了如何利用原子尺度编码将浩瀚的图书馆信息压缩至一粒微尘大小。他审视了现代半导体制造、遥控操作手和微型发动机，论证了为何缩小机械结构需要彻底重新设计，以适应不断变化的尺度规律，例如极端的散热效应以及在低雷诺数下流体所呈现的高有效黏度。他区分了暂时的工程障碍与光学衍射、光速等不可改变的自然极限，倡导在严格遵循物理定律的前提下发挥科学想象力。此外，该议题还延伸到对亚原子粒子探测、理论物理学的思考，以及他在β衰变和超流氦研究中所经历的科学发现的心理激荡与令人自省的风险。理查德·费曼回顾了从他1959年演讲到1984年间纳米技术的发展轨迹，指出利用每比特约100个原子的集群进行原子尺度数字存储，可将全球所有图书馆的文本压缩至直径仅1/200英寸的微尘中。 (4:48) 缩小机器并非简单的几何等比缩小，因为物理作用力的变化并不均匀；微型内燃机会因极快的散热速度超过机械膨胀速度而失效，而微型游动机械则需要旋转鞭毛才能在表现得如浓稠蜂蜜一般的流体中穿行。 (36:49) 费曼指出，商业工程上的限制（如光刻中的电迁移或光学衍射）不应被误认为是基本屏障，因为它们可以通过替代架构以及采用X射线等更短波长来绕过。 (23:56) 在区分基础物理与应用物理时，费曼断言，在已验证的物理定律范围内，应用科学在创造新技术方面拥有几乎无限的空间，而纯粹的基础物理学最终可能会走向完备终结。 (60:13)

Reading summary

Richard Feynman argues that the fundamental laws of physics permit vastly denser information storage and functional machinery built down to atomic scales, provided engineers adapt their designs to the shifting physical forces that govern micro-dimensions rather than naively shrinking macroscopic concepts. Physical Limits of Atomic Data Storage (0:47) Feynman demonstrates that information storage boundaries bottom out at the scale of atoms, where writing consists of depositing contrasting matter such as alternating patches of gold and silver. A two-dimensional reduction of linear dimensions by a factor of 20,000 shrinks surface area 400-million-fold, allowing all thirty-odd volumes of the Encyclopædia Britannica to fit on the head of a pin, or the Caltech Scientific Library onto a single library card. By expanding data packing into three dimensions using blocks measuring roughly five atoms on a side (around 100 atoms per bit), Feynman calculates that all written text across human history would compress into a volume measuring 1/200th of an inch on each edge—a speck of dust barely visible to the naked human eye. Addressing concerns about atomic thermal motion, he explains that while individual atoms in a solid lattice jiggle constantly, a five-atom-thick cluster of gold remains bound by neighboring crystal forces, requiring billions of years before thermal agitation could disrupt the recorded data. Microfabrication and Electrostatic Mechanical Devices (29:08) To manufacture actual movable machinery at microscopic scales, Feynman champions adapting planar semiconductor photolithography rather than relying on human dexterity or recursive telemanipulator hands. By depositing alternating layers of silicon, silicon dioxide, and sacrificial materials, fabricators can chemically etch away intermediate scaffolding to release unconstrained components within guided channels. These liberated parts can then be propelled and sequenced using electrostatic forces delivered through microfabricated surface electrodes, creating microscopic conveyors, motors, or high-intensity optical shutter displays. Because optical lithography hits a physical barrier around 2,000 magnifications due to visible light wavelengths, Feynman points to alternative methods such as reverse electron microscopes, X-ray lithography, and arrays of precision-valved chemical deposition nozzles that bypass optical diffraction limits entirely. Dimensional Scaling and the Non-Intuitive Micro-Regime (36:38) Feynman emphasizes that shrinking machines is not a matter of pure geometric scaling, because physical forces alter dramatically across dimensions. A miniature internal combustion engine fails because scaling down linear dimensions tenfold reduces combustion volume and fuel energy by 1,000-fold, while cylinder wall surface area drops by only 100-fold, causing combustion heat to leak out instantly before driving the piston. Electrical resistance also surges relative to output power in scaled-down copper windings, requiring radical design adjustments. Similarly, macroscopic fluid intuition collapses at micro scales: in low-Reynolds regimes, water behaves as viscous as thick honey, rendering macroscopic swimming motions useless. Microscopic organisms such as bacteria resolve this not by flapping fins against fluid inertia, but by rotating rigid helical flagella like corkscrews to drill through the viscous medium. Distinguishing Engineering Hurdles from Laws of Physics (23:56) Feynman contrasts transient engineering barriers with inviolable laws of physics, comparing perceived microelectronic bottlenecks like electromigration to the early twentieth-century myth of an impassable aeronautical sound barrier. While engineers often interpret contemporary design failures as natural limits, genuine physical boundaries—such as the speed of light or diffraction limits tied to radiation wavelength—remain strictly impassable. Defining his methodology as 'imagination in a tight straitjacket,' Feynman insists that scientific speculation must remain anchored strictly within verified empirical laws. While fundamental research might eventually stall if experiments become prohibitively expensive or if all natural laws are unified, Feynman maintains that applied science faces an inexhaustible horizon, offering endless permutations for rearranging known physical principles down to individual atoms. While Feynman acknowledged that his proposals began as theoretical exploration, mid-1980s advances in microchip fabrication, submicron drawings etched onto salt crystals, and biological precedents proved that atomic-scale engineering is physically viable. The ultimate realization of nanotechnology hinges not on overturning physical laws, but on developing fabrication tools capable of operating within quantum and atomic regimes.

精炼阅读摘要

理查德·费曼（Richard Feynman）在讲座中指出，微型机械与极度微缩的信息系统在物理原理上完全可行，限制发展的并非自然法则而是传统工程思维。他通过原子编码、微纳制造工艺与尺度缩放效应等系统论证，阐明了在已知物理学定律约束下拓展微观机械与应用科学的广阔前景。原子尺度编码与极端信息压缩 (0:47) 费曼指出，信息记录的物理极限在于原子尺度，写入本质上是将一种原子排布在另一种原子之上（如金、银原子交替排列构成二进制位）。若将平面线性尺度缩小20,000倍（面积缩减4亿倍），即可将全套三十余卷《大英百科全书》刻在针头上；照此比例，加州理工学院科学图书馆的所有文献也可容纳于单张标准借书卡。若进一步向三维纵深堆叠，以边长5个原子（约100个原子/比特）的微块仅存储文本，整个人类历史上所有图书馆的书籍信息，均可压缩进边长仅1/200英寸的立方体中——相当于肉眼勉强可见的一粒尘埃。针对热运动质疑，费曼解释固体中化学键束缚力如同紧密排布的军乐队，能将原子锁定在晶格内；5个原子厚度的金晶体依靠晶格力，在常温热扰动下可稳定存在数十亿年，足以安全保存人类文明记录。超越光学极限的微纳加工与成像 (7:01) 20世纪80年代微电子工业已在3毫米见方的硅芯片上实现约2,000倍缩微，通过纯硅基底氧化生成二氧化硅绝缘层、涂布光刻胶、紫外或可见光曝光及酸蚀刻，其几何公差大幅超越传统工匠手工研磨精度。但由于光的波动性，光学显微镜与投影系统受制于波长限制，放大与解析极限约在2,000倍，无法继续向下微缩。为突破光波屏障，费曼提出反向运行电子显微镜聚焦透镜系统以电子束进行微观刻蚀；康奈尔大学实验室随后的实验证实了类似构想——艺术家Tom Van Sant与科研人员合作，在单晶盐颗粒上刻出直径缩小100,000倍的人眼图案，线宽仅约100个原子。费曼指出，进一步超越该界限需使用波长极短的X射线光刻，或采用微喷嘴阵列实现主动化学沉积与直接原子级组装。微型机械的牺牲层制造与静电驱动 (29:08) 针对活动机械部件，费曼早年设想通过主从遥控手递归复制出成千上万副微型机械手并行制造，但他指出更可行的工业路径是借鉴平面半导体工艺。通过交替沉积结构材料与特定的“牺牲层”材料，在最后阶段用化学试剂将牺牲层选择性溶解剥离，即可在微米尺度释放出置于导轨内、自由活动的零部件。这些悬浮或滑动的微部件可由集成电极产生的静电场力推拉操控，构成微型传送带或小车。费曼列举其实际用途，例如制造由微小静电活门构成的高亮度光学快门阵列，用于投射高对比度动态图像，克服传统阴极射线管显示器的发光弱点。此外，他还推介了阿尔伯特·希布斯（Albert Hibbs）提出的“吞下外科医生”构想，即制造可巡航于人体血管内、受外控切除血管斑块的微型医疗机械。尺度效应与流体热力学重构 (36:38) 费曼警示，微观机械绝非宏观物体的简单等比例缩放，平方-立方律会导致表面积与体积之比发生剧烈畸变。以微型内燃机为例，线性尺寸缩小10倍，燃料容积与热能释放骤降至1/1,000，但缸壁表面积仅缩小至1/100，相对散热速度激增10倍，导致热量在做功前即泄漏殆尽；微电机导线截面剧减也会引发严重的电迁移与过热。流体力学在微观低雷诺数下同样反常：水在微米尺度黏滞如浓稠蜂蜜，传统依靠流体惯性的游禽摆尾机制完全失效，只会被黏液困死；细菌唯有利用刚性纤毛以螺旋开塞钻方式转动才能推进。类似地，放大工程亦受尺度约束：飞机几何放大100倍时，升力机翼面积仅扩大10,000倍，重量却剧增至1,000,000倍，若不进行大幅掏空优化，机器将因重力超载而无法离地。紧身衣中的想象力与科学边界 (45:35) 面对超光速与反重力等提问，费曼强调科学探索是“身穿紧身衣的想象力”——一切构想必须经受已知物理定律的严格检视。他严格区分了工程设计障碍与物理底层极限：历史上被视作飞跃极限的“音障”以及微电路中的电迁移现象，均属于特定设计在接近临界态时的工程失效，能够通过喷气发动机或全新量子架构突破；然而狭义相对论规定的光速限制则是不可跨越的客观定律，在斯坦福直线加速器中电子被加速至相对论质量达静止质量40,000倍的实证即证明了这一点。费曼指出，基础物理探索可能因定律被完全掌握或实验成本过于高昂而收敛，但基于已知规律排列组合的应用物理与微纳创造，其探索空间将永无止境。费曼的理论推演为现代纳米技术与微机电系统奠定了思想框架，确立了原子尺度机械制造的物理合宪性。但他亦强调，微观世界并非无约束的玄学乐园，材料热涨落、声波原子截止频率及辐射损伤等物理边界依然严苛；前沿技术的实现取决于对物理定律的敬畏与创造性工程实践，而非脱离实证规律的理论狂想。

Detailed notes — zh

## Background 该视频由 YouTube 频道 Muon Ray 于 2012 年 8 月 23 日发布。视频记录了理查德·费曼（Richard Feynman）在 1984 年重温其 1959 年著名演讲《底层有大把空间》（There's Plenty of Room at the Bottom）时的精彩讲座，从更现代的视角回顾纳米科技、电子显微镜、光刻工艺的发展，并深入探讨微型化与原子级操纵的物理边界。 ## TL;DR 费曼在演讲中探讨了信息微缩与制造微型机械的底层物理极限，阐明将全人类知识编码进一粒微尘或单原子计算机在物理法则上完全可行。他深入浅出地拆解了芯片光刻流程与超越光学极限的前沿探索，并以艺术家的微缩与大地画作为引，揭示了微观与宏观尺度的惊人对比。随后，费曼详细剖析了尺度变化对物理机制的颠覆——例如微型内燃机的散热死局以及低雷诺数下细菌如开瓶器般的螺旋推进。在漫长而活跃的问答环节中，他强调科学探索是“戴着紧身衣的想象力”，必须严格恪守已知物理定律，并回顾了自己在β衰变与液氦超流理论中体验到的极致思维顿悟。 ## Atmosphere & style 整场讲座充满了费曼标志性的幽默、机智与纯粹的求知狂热，现场笑声与惊叹声此起彼伏。他擅长用生活化的奇妙比喻（如高中管乐团、拔木塞的螺旋钻、海浪与软木塞）化解高深的理论物理，将复杂的工程难题讲得妙趣横生。听众与费曼之间的问答互动毫无距离感，既有严谨的物理辨析，也不乏费曼调侃妻子让他“等明天再说”的自嘲温情，带给观众一种在智力巅峰漫步却又极其亲切鲜活的沉浸体验。 ## Bulletin - [04:45](https://www.youtube.com/watch?v=4eRCygdW--c&t=285s) 微缩信息存储的物理极限在于原子尺度：利用三维原子阵列（约100个原子编码一个字符），全人类所有图书馆的文字信息理论上可压缩至一粒肉眼勉强可见的尘埃（边长仅 1/200 英寸）中。 - [15:14](https://www.youtube.com/watch?v=4eRCygdW--c&t=914s) 传统光刻技术受可见光波长衍射极限约束只能缩小约 2,000 倍，通过改用电子束、深紫外及 X 射线等更短波长的辐射，人类可大幅突破光学制造瓶颈并向单原子操纵迈进。 - [36:45](https://www.youtube.com/watch?v=4eRCygdW--c&t=2205s) 微观机械绝不能简单等比例缩小：由于表面积与体积比剧增、流体在微观下呈现类似高黏度蜂蜜的低雷诺数特性，微型机械必须推翻宏观设计，改用静电驱动、细菌鞭毛螺旋推进等新机制。 - [45:33](https://www.youtube.com/watch?v=4eRCygdW--c&t=2733s) 费曼倡导在严格遵守已知物理定律的前提下发挥“戴着紧身衣的想象力”，断言物理原理完全支持单原子比特计算与原子逐个组装的终极化学，否定了反重力等违背物理规律的伪科学设想。 ## Key takeaways - [00:08](https://www.youtube.com/watch?v=4eRCygdW--c&t=8s) 书写与信息存储的物理底层极限是原子，信息无法被压缩到比原子更小的尺度，但可以通过原子阵列编码实现极端微缩。 - [02:26](https://www.youtube.com/watch?v=4eRCygdW--c&t=146s) 只需将线性尺寸缩小 20,000 倍，整套《大英百科全书》即可写入普通针尖大小的二维表面上，且能借助电子显微镜轻松读取。 - [04:45](https://www.youtube.com/watch?v=4eRCygdW--c&t=285s) 若利用三维空间并以 5×5×5（约 100 个原子）的阵列编码一个字符，全人类图书馆的文字信息均可装入一粒边长 1/200 英寸的微尘中，且足以抵抗数十亿年的热涨落干扰。 - [11:20](https://www.youtube.com/watch?v=4eRCygdW--c&t=680s) 半导体芯片制造采用沉积、光刻胶曝光、化学蚀刻与剥离的循环工艺，其最终物理几何结构完全由投射到光刻胶上的光学明暗图案决定。 - [15:14](https://www.youtube.com/watch?v=4eRCygdW--c&t=914s) 可见光衍射极限限制了传统光刻只能实现约 2,000 倍的缩放，而利用聚焦电子束可在盐晶体上雕刻出仅约 100 个原子宽度的超微细线条。 - [23:56](https://www.youtube.com/watch?v=4eRCygdW--c&t=1436s) 微型导线面临电迁移断裂的技术瓶颈并不代表微型计算的终结，正如音障被打破一样，量子力学原理完全支持制造以单个原子作为一个比特的计算机系统。 - [26:33](https://www.youtube.com/watch?v=4eRCygdW--c&t=1593s) 通过级联制造的主从微型机械手或利用自复制架构，成千上万只并联的微型机械手可以在微观尺度上协同组装零部件。 - [29:11](https://www.youtube.com/watch?v=4eRCygdW--c&t=1751s) 利用半导体牺牲层腐蚀技术可释放出微米级的悬空可动部件，结合静电驱动可制成微型光快门阵列，用于实现高亮度动态投影。 - [32:49](https://www.youtube.com/watch?v=4eRCygdW--c&t=1969s) 终极的化学合成可以直接在微观尺度将原子精确放置在指定位置来构建目标分子，从而彻底替代传统试剂反应手段。 - [35:10](https://www.youtube.com/watch?v=4eRCygdW--c&t=2110s) 微型机器人在医学和娱乐领域潜力巨大，包括微型草履虫搏斗装置以及能够吞入体内、沿着血管游走清除动脉斑块的微型外科机器人。 - [36:45](https://www.youtube.com/watch?v=4eRCygdW--c&t=2205s) 内燃机等宏观机械在微观尺度下完全失效：气缸容积缩减至千分之一时表面积仅缩小至百分之一，热量会在推动活塞前瞬间散逸殆尽。 - [39:45](https://www.youtube.com/watch?v=4eRCygdW--c&t=2385s) 在低雷诺数的微观流体中，惯性几乎消失而黏滞阻力占据绝对统治，微型机器需要模仿细菌鞭毛的开瓶器螺旋旋转机制才能有效游动。 - [43:23](https://www.youtube.com/watch?v=4eRCygdW--c&t=2603s) 飞机的几何尺寸不能随意放大，因为重量正比于尺寸的三次方增长，而升力仅正比于面积的二次方增长，超大飞行器必须采用空心轻量化结构。 - [45:33](https://www.youtube.com/watch?v=4eRCygdW--c&t=2733s) 科学构思应当是“戴着紧身衣的想象力”，所有突破性构想都必须在已知且经严格验证的物理定律边界内进行推导与检验。 - [51:33](https://www.youtube.com/watch?v=4eRCygdW--c&t=3093s) 相对论明确限制有质量物体的运动速度无法达到或超越光速，向粒子注入极高能量只会使其动质量急剧膨胀（如高能电子质量膨胀数万倍）。 - [01:10:22](https://www.youtube.com/watch?v=4eRCygdW--c&t=4222s) 高频声波在微观介质中存在物理上限，当波长缩短到相邻原子间距时，离散原子只能做反向振动，“波”的物理描述便彻底失去意义。 - [01:16:09](https://www.youtube.com/watch?v=4eRCygdW--c&t=4569s) 费曼回顾其学术生涯中最令人激动的顿悟时刻——放射性弱相互作用猜想与液氦超流物理机制的攻克，并指出了科学灵感后反复验错纠错的客观规律。 ## Detailed walkthrough ### [00:08](https://www.youtube.com/watch?v=4eRCygdW--c&t=8s) 微型机器与微缩书写的物理极限费曼在演讲开场时幽默地提到，他在洗澡时听到周围的人在议论“微型机器”（tiny machines），纷纷困惑他到底在说些什么。费曼对他们说，他指的就是“非常微小的机器”。在正式讨论机器之前，费曼决定先探讨一个更基础的问题：微缩书写。我们究竟能把文字或数字写到多小？这里的限制并不是指人类手指在没有辅助工具时的书写极限，而是指利用特殊机械所能达到的物理极限。费曼指出，数字本身无论写得大还是写得小，其代表的信息都是一样的。然而，书写的最终物理极限是原子。我们无法在单个原子上写字，因为所谓的“笔迹”本质上就是将一种原子分布在另一种原子之上（例如在白原子上铺上黑原子）。因此，若想在原子尺度上写字，就必须使用某种代码。例如，可以用一小片金原子代表“点”，一小片银原子代表“划”，通过类似摩斯密码的编码来表示数字。费曼保守估计，一个编码单元在单边上大约需要 100 个原子（尽管理论上可能缩减到单个原子）。为了让大家对这个尺度有更直观的感受，他决定用一些具体的例子来做对比。 > “最终的极限是，你当然无法让一个数字无限变小——无论写得大还是写得小，数字本身都是一样的，你可以说任何尺寸都行，但你无法让它比原子还要小，你无法在原子上写字。” —— 理查德·费曼 --- ### [02:26](https://www.youtube.com/watch?v=4eRCygdW--c&t=146s) 将《大英百科全书》写在针尖上人们常常津津乐道于将《主祷文》（The Lord's Prayer）刻在针尖上的手艺。费曼以此启发听众：如果我们想把整套《大英百科全书》（Encyclopedia Britannica）都写在针尖上，这可行吗？《大英百科全书》大约有两三万页，如果把每一页纸都平铺在地上，会占满整整 20,000 平方英尺的巨大面积。要将这么大的面积缩减到针尖大小，需要将线性尺寸缩小 20,000 倍。这意味着在面积上要缩小 400 万倍。费曼解释说，这种程度的微缩在物理上其实并不困难，而且在这个尺度上，我们甚至根本不需要动用原子层面的技术。如果能实现 20,000 倍的微缩，整个加州理工学院（Caltech）科学图书馆的所有藏书信息都可以装在一张小小的图书卡上。费曼开玩笑说，如果巴西的科学图书馆不幸遭遇火灾被烧毁，我们只需要给他们邮寄一张图书卡，里面就包含了加州理工图书馆的所有知识。而规模更大的华盛顿国会图书馆，在同等微缩尺度下，也只需要一本《时代》（Time）杂志大小的介质就能全部装下。 > “如果我们能把书里的文字、字母、图像中的网点，也就是百科全书里的所有一切，在这个方向上缩小两万倍，当然在另一个方向上也缩小两万倍……那么我们就能把整部百科全书写在一个针尖上。” —— 理查德·费曼 --- ### [04:45](https://www.youtube.com/watch?v=4eRCygdW--c&t=285s) 三维空间与“尘埃”中的人类知识库到目前为止，费曼所讨论的还仅仅是在针尖的二维表面上进行书写，并没有利用针尖内部的三维空间。真实的《大英百科全书》是有厚度的，一页叠着一页。如果我们不仅在表面书写，而是深入到物质的内部，并利用原子进行三维书写，情况会怎样？假设我们使用 5 个原子见方的小立方体（即 100 个金、银等原子组成的阵列）来代表一个字符，并用类似摩斯密码的编码来代替图像和文字。经过计算，费曼得出了一个惊人的结论：如果只保留文字信息（因为在这个尺度下无法保留图像），全世界所有图书馆里的所有藏书——无论是土耳其语、匈牙利语还是其他任何语言——其全部信息都可以装进一个边长仅为 1/200 英寸的立方体中。这个体积相当于人类肉眼勉强能看清的一粒微小尘埃。费曼调侃道，如果我们担心人类文明会像当年亚历山大图书馆被付之一炬那样走向崩溃，我们应该赶紧把所有图书馆的副本制成这样的小尘埃，并撒得到处都是，因为“他们不可能清除掉所有的尘埃”。 > “……所有这些信息都可以装进一个边长仅为二百分之一英寸的体积中，这正是人类肉眼所能看到的最小的尘埃颗粒。这就是人类所有信息排列的最终结晶。” —— 理查德·费曼 ```mermaid flowchart TD accTitle: 费曼设想的信息微缩尺度演进 accDescr: 展示从传统书籍到针尖、再到三维原子尘埃尺度的信息存储空间压缩过程。 A["传统图书馆 (如 Caltech 图书馆) 占地数万平方英尺"] -->|"线性缩小 20,000 倍"| B["针尖表面 (2D 写入) 可容纳整部《大英百科全书》"] B -->|"利用单张图书卡空间"| C["单张图书卡 可容纳整个 Caltech 图书馆"] C -->|"利用三维空间 (3D 写入) 以 100 个原子 (5x5x5 纳米立方) 代表一个字符"| D["一粒微小的尘埃 (边长 1/200 英寸) 可容纳全人类所有图书馆的文字信息"] ``` --- ### [06:50](https://www.youtube.com/watch?v=4eRCygdW--c&t=410s) 如何读取与书写微缩信息面对如此微小的书写，人们自然会问：我们该如何阅读它？如果使用普通的显微镜，由于光的波动性（即光的波长限制），其放大倍数无法超过 2,000 倍，因此无法看清。但是，电子显微镜的放大倍数可以达到 200,000 倍。而我们只需要 20,000 倍的放大率，这仅比光学显微镜的极限高出 10 倍，因此用电子显微镜阅读写在针尖上的百科全书实际上非常容易。那么，我们又该如何书写它呢？费曼提出，未来可以通过一种“反向电子显微镜”来实现。就像我们可以把望远镜倒过来拿，从大的一端往里看会让所有东西变小一样，我们也可以反向使用显微镜的透镜系统，将大尺寸的文字投影并聚焦成一束极细的电子束，从而在极小的介质上刻印出微缩图像。费曼提到，他第一次发表这个观点的部分内容是在 20 年前（1959 年），他很惊讶当时人们还没有把这个设想付诸实践，但现在（1984 年）他已经可以向大家证明，这完全是可行的。 > “……未来某天，我们有可能通过一种类似于‘反向电子显微镜’的装置来书写它，即把大尺寸的文字输入，反向使用透镜来控制一束非常细微的光束……” —— 理查德·费曼 --- ### [08:44](https://www.youtube.com/watch?v=4eRCygdW--c&t=524s) 商业微缩技术的现状与电脑芯片随后，费曼将话题转向了当时（1984 年）工业界实际应用的微缩技术。我们现在并没有真的在针尖上刻写《主祷文》，因为商业利益和关注点已经发生了转移。相反，我们正在大规模制造的是电脑芯片。费曼展示了一张电脑芯片的幻灯片，该芯片的实际尺寸只有大约 3 毫米宽。虽然在投影屏幕上很难看清其极其微细的结构，但从人类技术发展的角度来看，能够以如此高的精细度进行批量制造，已经是一项了不起的成就。费曼赞叹了这种“令人愉悦的精密工艺”。他指出，过去人们引以为傲的精密抛光技术，其误差甚至比这些芯片上最微小的凹槽还要大。目前，工业界已经能够利用光学透镜系统（即反向显微镜原理），在 2,000 倍的微缩尺度下（即光学极限范围内）制造出这些复杂的芯片结构。 > “……从人类及其发展的角度来看，能够制造出细节如此精细的东西，确实是一项了不起的成就……” —— 理查德·费曼 - **《大英百科全书》微缩数据**：约 20,000 至 30,000 页，平铺面积约 20,000 平方英尺，线性缩小 20,000 倍后可完整写在针尖上。 - **三维原子存储极限**：以 100 个原子（5x5x5 阵列）代表一个字符，全球所有图书馆的文字信息可装入边长 1/200 英寸的立方体（一粒肉眼可见的最小尘埃）。 - **显微镜放大极限**：光学显微镜受限于光波长，极限约为 2,000 倍；电子显微镜可达 200,000 倍。 - **展示芯片尺寸**：实际宽度约为 3 毫米，采用约 2,000 倍微缩的光学曝光技术制造。 ### [11:20](https://www.youtube.com/watch?v=4eRCygdW--c&t=680s) 光刻技术与集成电路制造流程理查德·费曼（Richard Feynman）在黑板上用彩色粉笔生动地拆解了半导体芯片的制造机制。他强调，底层必须是一块极其纯净的硅片（Silicon），不能有哪怕最微小的瑕疵或尘埃，因为“在这个尺度下，任何污垢、划痕或缺陷都是一块巨大的怪兽级巨石（great big monster Boulder）”。在真空环境中通入氧气后，硅表面会氧化生成一层极薄的二氧化硅（Silicon dioxide）。费曼用通俗的语言解释道，这其实就是石英、沙子或一层薄玻璃，它是优良的绝缘体，而下层的硅则是导电体。随后，工艺人员会在其表面蒸镀一层被称为“光刻胶”（Photoresist）的感光化学材料。此时利用光学投影系统，将预先设计的黑白图案像投影片一样照射在光刻胶上。被光照射到的区域会发生化学交联，从而具有耐溶解或耐蚀刻的特性；未受光照的区域则会被化学溶剂洗掉。接着，使用氢氟酸等能够腐蚀玻璃的试剂侵蚀暴露的二氧化硅，被未溶解光刻胶遮蔽的二氧化硅则留存下来，形成微小的绝缘柱结构。最后，作为临时脚手架的光刻胶被另一种试剂彻底剥除。费曼指出，这种“沉积材料—涂胶—曝光—显影蚀刻—去胶”的过程会层层循环推进。工程师会在绝缘层之间再次蒸镀掺杂硅（用于构建晶体管结）或金属连线，逐步构建出复杂的双极型晶体管与导电引线网络。费曼在白板上完成绘制后总结道，从上方俯瞰，这既是令人惊叹的工业制造奇迹与艺术品，也是一个完全能够自主完成复杂运算的电路网络，而决定这一切微观几何图案的，“正是最初那一束光线的空间排布”。 ```mermaid flowchart TD accTitle: 半导体光刻工艺基本循环流程 accDescr: 展示从硅基底、氧化层制备、涂覆光刻胶、曝光蚀刻到形成金属互连的多层制造流程。 subgraph Step1 ["第一阶段：基底与绝缘层构建"] A["高纯度硅片基底 (导体)"] --> B["真空充氧氧化"] B --> C["生成极薄二氧化硅层 (绝缘体/玻璃)"] end subgraph Step2 ["第二阶段：光刻图案转移"] C --> D["表面蒸镀光刻胶 (Photoresist)"] D --> E["光学图案投影曝光"] E --> F["洗去未曝光光刻胶"] end subgraph Step3 ["第三阶段：蚀刻与结构成型"] F --> G["酸液腐蚀裸露的二氧化硅"] G --> H["剥离残留光刻胶 (去除脚手架)"] H --> I["沉积晶体管半导体层与金属连线"] end ``` - **纯度要求**：纳米至微米级制造对尘埃零容忍，微观颗粒对电路而言等同于巨石。 - **材料特性**：硅作为导电基底，二氧化硅（石英/玻璃）作为层间电气绝缘介质。 - **结构确定机制**：电路的最终物理拓扑结构完全由投射到光刻胶上的光学明暗图案决定。 --- ### [15:14](https://www.youtube.com/watch?v=4eRCygdW--c&t=914s) 超越光学极限：康奈尔实验室与世界上最小的画费曼指出，常规工业光刻受限于可见光的波长，微缩能力大约只能停留在 2,000 倍左右的缩小极限，而他在 1959 年演讲中提出的“缩小 20,000 倍”则属于截然不同的物理前沿。“我们无法用光做到缩小 20,000 倍，因为光只能达到大约 2,000 倍的缩放，所以如果我们真要实现它，就必须改用其他方法。”目前这类研究属于非商业的前沿探索，比如康奈尔大学的微加工实验室就在尝试挑战微观制造的物理极限。费曼随后介绍了一位跨界艺术家朋友汤姆·范·桑特（Tom Van Sant）。费曼热情地称赞他是一位真正热爱科学、艺术与商业交融的现代人物，“而不是那种蔑视科学、完全不理解自己身处何种世界的艺术家；他深爱他所处的世界，也热爱艺术”。范·桑特向康奈尔实验室提出请求，希望他们能协助将自己设计的一幅图案制作出来，由此诞生了世界上尺寸最小的一幅画作。幻灯片展示了这幅画在透射电子显微镜下的形貌：这是一只人类眼睛的图案。它的制造方式是在盐晶体上利用精细的电子束进行偏转扫描，像微型挖掘机一样轰击剥离盐分子来勾勒出线条。相比于真实的人眼，这幅画的尺寸缩小了整整 100,000 倍，其线条横跨的物理宽度仅仅只有大约 100 个原子左右。画作中带有某种特殊的波浪状质感，费曼风趣地揭秘道，那根本不是刻意的艺术笔触，而是实验室外刚好有一辆卡车开过，引起的微小机械震颤直接传导至电子束上而留下的抖动痕迹。实验室人员曾遗憾地表示，如果技术调整得当，原本可以把线条厚度和图案整体再缩小一半，但范·桑特坚决拒绝了进一步缩小，因为他必须将其严格维持在人眼尺寸的十万分之一。 - **光学瓶颈**：受衍射极限限制，传统可见光光刻技术只能实现约 2,000 倍的微缩。 - **微观加工载体**：康奈尔实验室利用盐晶体（Salt crystal）作为衬底，使用聚焦粒子束剥蚀表面刻制图案。 - **物理尺度指标**：画作整体宽度仅跨越约 100 个原子的间距，比人类肉眼真实尺寸缩小了 100,000 倍。 - **环境振动干扰**：在 100 个原子的尺度下，地面上一辆过路卡车引发的微震就足以在加工轨迹中留下显著的波纹。 --- ### [18:20](https://www.youtube.com/watch?v=4eRCygdW--c&t=1100s) 荒漠中的镜阵：地球上最大的画与野兔意外为什么范·桑特一定要把盐晶体上的眼睛做成真实人眼的“十万分之一”？费曼在切换幻灯片后揭晓了答案：因为这位艺术家此前曾创作过另一幅呼应之作——一幅比真实人眼“放大 100,000 倍”、被誉为世界最大的人造图像。屏幕上首先展示了一幅极具现代艺术风格的眼部色块图，随后画面逐渐拉远，露出了眉毛和眼袋的整体轮廓；当幻灯片缩放到最宏观尺度时，观众席赫然发现背景竟是整座洛杉矶城及周边荒漠，那只“眼睛”正横卧在大地上。费曼解释道，这幅巨型大地画作的跨度达到了 2.5 公里，肉眼从地面根本无法识读，只能依靠在约 600 英里（约 965 公里）高空运行的陆地卫星（Landsat）从太空俯瞰。Landsat 卫星通过扫描传感器对地面进行往返扫掠，实时测算并记录三色光谱反射光，最终在数字图像上生成一个个代表特定方向的像元（Pixel）。在卫星图像的尺度下，每一个像素点对应地表大约 1 英亩（约 4047 平方米）的面积。范·桑特并没有动用浩大的工程去在干旱沙漠中粉刷上百英亩的土地，而是巧妙地利用了几何光学原理：他在 2.5 公里的范围内精准布设了 24 面反射镜。根据 Landsat 运行轨道和精确过境时间，他计算好每面镜子的俯仰与方位角。当卫星扫描传感器掠过对应那一英亩土地的瞬间，太阳光恰好直射镜面并反射进卫星相机的光电倍增管，造成瞬间的强烈强光过曝与饱和，使得整整一英亩的像素点在数字图像上全部显示为耀眼的白色。费曼最后调侃地指出，最终的卫星成像图上出现了一个显眼的黑点，某个特定像素没有亮起来。范·桑特回到现场排查故障后发现，原来是一只野兔（Jackrabbit）在卫星过境前从镜子上方踩踏奔跑过去，撞偏了镜架的角度，导致阳光未能准确射入卫星镜头。费曼笑着向听众自嘲，这幅巨画虽然与极小尺度的微观物理无关，但他实在无法抗拒在展示完世界上最小的画之后，紧接着与大家分享同一位艺术家利用卫星与镜子创作出世界最大画作的奇妙联系。 - **宏观缩放比例**：大地艺术画作尺寸达到 2.5 公里宽，精确设定为真实人眼的 100,000 倍放大。 - **观测方式**：必须通过约 600 英里高空的 Landsat 陆地卫星多光谱扫描传感器在轨合成。 - **光学致盲技巧**：用 24 面小型镜子在特定时刻将日光反射至传感器，利用饱和溢出效应点亮单颗覆盖 1 英亩的地面像元。 - **野外干扰因素**：沙漠野兔奔跑导致的物理碰触使镜面偏转角度失准，造成最终图像缺失一个关键白色像素。 ### [22:47](https://www.youtube.com/watch?v=4eRCygdW--c&t=1367s) 极小与极大的尺度对比：从睫毛到土星环 Feynman 通过形象的尺度放大，帮助观众直观理解微观世界到底有多么微小。他以前文提到的一幅画作为对照基准，将其中的细节与人眼做层层递进的比较。如果取普通人眼上的一根睫毛，将这根睫毛放大到整只眼睛那么大，那么此时整只眼睛的尺寸就会膨胀到跨越整个讲堂的宽度；但这仅仅走完了一半的尺度阶梯。接下来，若再取那只巨大眼睛上的一根睫毛，同样将其放大到讲堂大小，那么原本的眼睛尺寸将暴涨到 2.5 公里宽。为了让听众更震撼地感受这种百倍千倍的跨度，Feynman 提议再向上放大一个同样的量级——即再放大 100,000 倍。若要在宏观世界中绘制一个比那只 2.5 公里大的眼睛还要大上 10 万倍的“天体之眼”，那恰好是天空中壮丽的天体——带有光环的土星。Feynman 由此总结道，从 2.5 公里的巨眼到真实人眼的比例，正如人眼与那幅极其微小的原始微缩图画之间的比例，这种对称的几何放大揭示了微观世界的极深跨度。 ```mermaid flowchart TD accTitle: 尺度放大层级对比 accDescr: 展示从微观图画到宏观土星环的跨尺度放大过程 subgraph MicroScale ["微观世界"] A["微缩图画中的眼睛"] B["真实人类眼球"] end subgraph MidScale ["讲堂与公里尺度"] C["放大睫毛至眼球大小: 眼球约讲堂宽"] D["二次放大睫毛: 眼球达 2.5 公里宽"] end subgraph CosmicScale ["宇宙尺度"] E["再放大 100,000 倍: 土星及其光环"] end A -->|"100,000 倍递进"| B B -->|"第一级放大"| C C -->|"第二级放大"| D D -->|"再放大 100,000 倍"| E ``` ### [23:56](https://www.youtube.com/watch?v=4eRCygdW--c&t=1436s) 导线的局限与打破“音障”：单原子计算机的物理原理在讨论完现有技术能达到的尺寸后，Feynman 将话题引向计算机进一步微型化的本质极限。他强调，这不再是探讨“今天工程上是否可行”，而是探究“物理原理上是否可行”。如果在现有设计下将导线粗细缩减到二分之一或三分之一，就会遭遇严重的物理阻碍：当电流长期流经极细导线时，物质本身会发生迁移（即电迁移效应），在宏观导线中这微不足道，但在极细导线中则会直接导致导线断裂。许多工程师因此感到悲观，但 Feynman 直言这就像当年航空界所面临的“音障”一样。在过去，人们基于空气不可压缩的假设来设计螺旋桨和机翼，得出飞机绝不可能超越音速的结论；一旦将空气的可压缩性纳入考量，旧式的螺旋桨设计确实彻底失效。然而，物理定律从未禁止超音速飞行，它需要的仅仅是全新的飞行器设计范式。在微型化计算上同样如此，现有的导线瓶颈只意味着“用老方法走不通”，并不代表物理学宣告了死刑。关于原理上的突破，Richard Feynman 明确指出： > “It just requires a completely different design so these limitations that you hear about mean only that you can't go on with the same design, but it doesn't mean that it's not possible in principle.” 摆脱传统导线束缚后，Feynman 构想直接利用原子状态和原子间相互作用来构建计算机。在这种尺度下，经典物理失效，必须引入量子力学和测不准原理。Feynman 兴奋地分享了自己推导出的结论： > “Just let me tell you that it turns out to be possible according to the laws of physics in principle to build a computer in which each bit or each little piece of information is one atom large and there's no problem in that.” - 极细导线失效原因：长期导电引发物质迁移，导致微导线物理断裂。 - 航空“音障”类比：证明技术瓶颈往往源于旧范式（如假设空气不可压缩），而非自然法则的不可逾越。 - 终极物理极限：量子力学原理完全允许制造以单个原子作为一个比特（bit）的计算系统。 ### [26:33](https://www.youtube.com/watch?v=4eRCygdW--c&t=1593s) 制造微型可动机械：主从机械手与自复制层级构想 Feynman 将演讲重心从纯粹的信息处理设备转向更传统的机械设备——即拥有可动部件的物理机器。他幽默地注意到听众脸上的疑惑表情，仿佛都在问“造这些到底有什么用？”。Feynman 洒脱地表示这纯粹是“脑力娱乐”（mental entertainment），不需要一开始就背负功利的目的，即便没有实际用途，“缺乏应用也不会消耗你一分钱”，思考本身就是莫大的乐趣。当今人类制造精密微型手表的双手显得庞大而笨拙，要想打造具备微观可动零件的机械，必须探索颠覆性的制造思路。他提到的第一种设想借鉴了核电站处理放射性物质所用的“主从机械手”（slave hands）。操作员在防辐射室外扳动控制杆，通过电气线路操控内部的从属机械爪进行扭螺丝、倒试剂等精细作业。同样地，外部的大型机械手可以通过电路驱动比自身小得多的精密小机械扳手；而这套更小的机械手造出来后，并不是停在原地，而是用它去制造下一代尺寸更小的第三代机械手。为了解决最终只能用一对微型手制造一台机器的低效问题，Feynman 构想了一种几何级数递增的自复制架构：第一代机械手制造时，直接制造出两套或更多套更小的机械手；随后每一代都呈倍数制造下一代。最终，将有成千上万只极微小的机械手并联接入相同的控制导线，步调一致地在微观世界中并行组装微型设备与机械零件。 ### [29:11](https://www.youtube.com/watch?v=4eRCygdW--c&t=1751s) 表面微加工与微机电构想：牺牲层与光开关应用随后，Feynman 提出了第二种更加实用的制造方案：利用芯片制造中的多层沉积分层光刻工艺。他反问为什么芯片结构必须全是实心固态，我们完全可以改变材料布局。例如，在硅基底上沉积二氧化硅等绝缘材料、导电层，并在特定结构之间填入一种被他称为“蓝色材料”的物质。这种“蓝色材料”具有独特的化学特性——它可以被某种溶剂轻松溶解剥离，而周围的硅和棕色结构（如二氧化硅）却完好无损。一旦将这种中间材料溶蚀洗净，原本被包裹的微小部件就会脱离拘束，成为悬空、可活动的自由部件（即现代 MEMS 中的牺牲层腐蚀技术）。 ```mermaid flowchart LR accTitle: 微机械牺牲层制造流程 accDescr: 展示利用牺牲层溶解制造微型可活动零件的工艺步骤 subgraph Litho ["薄膜沉积与图形化"] A["硅基底与电极层"] --> B["沉积易溶材料 (牺牲层)"] B --> C["沉积结构材料 (如硅/绝缘层)"] end subgraph Release ["化学释放与驱动"] C --> D["特异性溶剂溶解易溶层"] D --> E["释放活动部件 (形成自由微机械)"] E --> F["施加微电压实现静电位移驱动"] end ``` 这些加工出来的微活动部件与周围的电气走线相连，工程师可以在其两侧施加不同的电压，利用微观静电荷驱动零件来回移动。Feynman 发挥想象力提出，若是设计一系列连续排列的微型活动件，就可以像输送带或微型小车一样，在硅片表面逐级传递微小物体。对于这套系统的实际应用，Feynman 顺手提出了一种极具前瞻性的方案——光学快门阵列： - **微型光快门阵列**：在一个开孔阵列上方覆盖微型静电可动挡板，通过施加电压独立控制每个微孔的开合，从而调制光线的通断。 - **高亮度动态投影应用**：传统的阴极射线管或荧光屏投影电视受限于荧光粉的发光强度，画面暗淡且尺寸受限；而微型快门阵列可以直接透射高强度的外部独立光源，通过电控快速开关各个快门点形成高动态、超高亮度的投影图像（该设想正是后来德州仪器 DLP 数字微镜技术的前身）。在演讲的最后，一位听众现场提出疑问：前文曾提到光学显微镜受限于光的波长，最多只能放大 2000 倍，那么这种微型快门是否会受制于减小光束尺度的物理障碍？Feynman 当即澄清纠正：光的波长确实存在下限，人们无法将光线聚焦到小于其波长的尺度；但此时他所讨论的可动快门结构，尺度恰好就在 2000 倍放大的光学尺度（微米级）左右。这种尺寸不仅完全能够让可见光顺畅通过，而且在机械制造领域，已经远远小于人类此前手工制造过的任何微型机械。 ### [32:49](https://www.youtube.com/watch?v=4eRCygdW--c&t=1969s) 微尺度喷涂制造与操纵原子的“终极化学” 理查德·费曼（Richard Feynman）开始探讨微型机械的其他潜在用途。他指出，当前芯片制造以及将不同物质精确置于特定位置的传统工艺极其繁琐冗长；为了让第二层图案与第一层图案精准对齐，所需的技术精度要求高得令人抓狂。费曼提出了一种替代思路：为什么不直接把需要的物质喷射到指定位置？可以设想一个平整的大型框架结构，上面排布着大量带有可移动微孔的喷嘴，化学物质从这些微孔中喷射出来。但这需要微型阀门或机械移动部件，而此类结构恰恰可以利用前文探讨的微型机械来制造。一旦这种微型喷涂阀门阵列就位，光刻中光的衍射极限便不再成为障碍，微型化进程将迎来全新的技术路径。更进一步，费曼设想模仿生物体内复杂的化学反应。生命系统依靠细胞膜，将特异性酶或特定分子嵌入固定位置，充当极其敏感高效的催化剂，精巧地调控生命活动。目前人类逐渐理解了这些多层膜结构的运作机制，但受限于技术尺度，我们还无法像细胞那样直接制造并随意排布酶分子。为了自由组装各种化学反应体系，人类需要能在微观尺度上推移分子的微型工具。费曼随即推导出了这一逻辑的终极形式：如果微观操作达到原子级别，甚至不再需要传统的化学家。正如费曼所言：“All that a chemist does is rearrange make a new molecule with a different arrangement of atoms so if you went down to a small enough scale you just put the atoms down where you want them to make the molecule no Hocus Pocus with bottles and colored liquids actually that's really magical how they do that.” 尽管费曼开玩笑称化学家倒腾五颜六色试剂瓶的操作像“变戏法”，但他坦言化学家能用宏观手段合成复杂分子确实令人惊叹。 ```mermaid flowchart TD accTitle: 从传统制造到原子级操纵的演进 accDescr: 流程图展示了从传统光刻对齐问题到喷嘴阵列构想，再到仿生膜与终极原子操纵的发展阶段。 subgraph SG1["制造方式演进"] A["传统光刻制造 （多层图形对齐繁琐/光波长限制）"] --> B["微型喷嘴阵列直接喷涂 （绕过光学极限/微型阀门控制）"] end subgraph SG2["分子与原子操纵"] B --> C["仿生特异性酶与细胞膜 （受限工具精度，无法精确直接排布）"] C --> D["终极原子操纵 （直接将原子摆放到指定位置，合成目标分子）"] end ``` --- ### [35:10](https://www.youtube.com/watch?v=4eRCygdW--c&t=2110s) 自主游动微机、草履虫决斗与“吞下外科医生” 在讨论了固定在基底上的机械构件之后，费曼转向了自由游动的微型机械。他幽默地提出，任何新技术在初期最容易流行并商业化的往往是游戏，微型机械也是如此。人们完全可以制造一个微小的机械体，由外部发射电信号进行远程遥控；这个小东西配备了一柄微型的“剑”，操作员可以在显微镜下一边观察，一边操控自己的微型角斗士去和草履虫（Paramecia）决斗。不仅如此，费曼提到了他的朋友阿尔·希布斯（Albert Hibbs，简称 Al Hibbs）构想的一个更为激进的医疗应用：把微型外科医生“吞下去”。希布斯构想制造一个极其微小、配备手术刀具的受控机器，让患者将其吞入腹中；它可以穿过组织进入血管，沿着血管游走到病灶区域，精准刮除血管壁上的动脉粥样硬化斑块。费曼回顾道，希布斯早在20年前就向他提出了这个构想。尽管这篇演讲在概念起源上有着悠久的历史，但结合近年最新研发出来的微结构图像，这些构想正一步步展现出现实可行性。 - 微型受控机器概念：自带微型执行机构（如微型刀具、探针），可通过外部电信号感知和遥控。 - 娱乐与医疗两大方向：显微镜下与草履虫搏斗的微型竞技玩具；进入人体血管内切除病变沉积物的外科微型机器人。 --- ### [36:38](https://www.youtube.com/watch?v=4eRCygdW--c&t=2198s) 尺寸效应与微型内燃机的热力学死局费曼严肃地指出，研制微型机械绝非简单地把现有宏观设备等比例缩小。如果在磁力或机械结构上盲目按比例微缩电动机，它根本无法运转。当物体的特征尺寸缩小或放大时，重力、结构强度、空气阻力和电阻等物理量随尺度的变化比例完全不同。例如，普通导线的电阻在极小尺度下会不成比例地剧增，带来致命的发热问题；工程师必须利用超导材料或全新架构来绕过这些瓶颈。为了直观展示简单等比例微缩为何不可行，费曼以微型内燃机为例进行了详细的热力学推导。在宏观内燃机中，火花塞引爆汽油产生高温膨胀气体推动活塞，气缸壁虽会带走一部分热量，但大部分热量仍用于对外做功。若将尺寸缩小10倍： - 气缸在长、宽、高三个维度各缩小为原来的 $1/10$，气缸容积以及燃烧产生的热量相应暴跌至原来的千分之一（$1/1000$）； - 然而，气缸的散热表面积仅仅缩小到原来的百分之一（$1/100$）； - 表面积与容积之比相对扩大了10倍，意味着热量向缸壁的泄漏速度提高了整整10倍。如果将尺寸缩小100倍，这种设计在物理上将彻底破产。费曼阐述道：“As soon as you would explode this infinitesimal amount of gasoline in this tiny tiny cavity then the heat of it wouldn't be before it had any chance to push anything it would leak out immediately into the container so you have to redesign.” 极微量的汽油刚一发生微爆，热量还没来得及推动活塞，就会瞬间散逸到周围的缸壁金属中。因此，微型机械必须推翻宏观旧范式，依据微观主导的物理规律从头设计。 --- ### [39:45](https://www.youtube.com/watch?v=4eRCygdW--c&t=2385s) 低雷诺数流体世界：从黏稠蜂蜜到细菌的螺旋推进在微观尺度下，流体流动规律与宏观世界天差地别。许多人质疑微型机械在流体中的活动能力，费曼解释说，当把微小物体在水中的相对运动放大到人类尺度时，水不再像我们感觉中的水，微小物体面对的环境相当于在“极度黏稠的蜂蜜”（thick honey）中游泳。在这样的低雷诺数环境中，流体的惯性阻力不再主导，黏滞阻力完全掌控一切。宏观鱼类依靠左右摆动鱼鳍、利用水的惯性推力向前游动的策略彻底失效；微型机器鱼如果只靠拍打鳍片，只会像陷在黏胶中一样动弹不得。费曼强调，物理定律绝不禁止微型物体在黏液中运动，而最直接的铁证就是自然界中的微生物： - 细菌在尺度上完全处于这种“黏稠浆糊”的世界中，但它们却能自如游动。 - 细菌采用的完全是另一种游动机制：其鞭毛（flagella）本质上是一根微型的“开瓶螺旋塞”（cork screw）。 - 鞭毛由包裹在套管（sleeve）内的坚硬螺旋轴芯构成。轴芯在套管内旋转，犹如开瓶器咬进软木塞一般，依靠剪切和螺纹推力将菌体向前硬生生“拧”过去。费曼指出，这种开瓶器螺旋式的推进方式在宏观水体中低效至极，除非旋转极其迅速才能产生足够的惯性推力（宏观下鱼鳍摆动要高效得多）。正是因为不同物理效应在不同尺度下的占比发生了剧烈颠覆，重新设计微型机械才变得极为有趣。在演讲尾声，费曼展望道，在微观物理规律允许的范围内重新审视并设计所有人类熟悉的机械，将是一场引人入胜的探索。他预言：“When you allow for the fact that when you go to different scales different things at different proportions become important you have a delightful time in redesigning all kinds of familiar Machinery to see if you can do it and give what take 25 30 years there'll be some practical use for this.” 尽管他目前还无法确切预知这项技术的具体终极形态，但他笑称或许再过二三十年，自己会再做一场演讲，向大家汇报“这些该死的小玩意儿”究竟有了什么实际用处。随后，费曼在全场掌声中结束讲座，进入观众问答环节。 ### [43:23](https://www.youtube.com/watch?v=4eRCygdW--c&t=2603s) 尺度放大与空气动力学：为什么大飞机不能只靠几何放大一名听众提问：如果在极小尺度下液体会变得异常黏稠，那是否意味着如果建造一个极其庞大的生物或机器，空气对于它而言也会变得足够“厚重”，从而能够像穿过浓稠液体一样支撑它移动？理查德·费曼（Richard Feynman）首先反驳了这个假设，指出尺寸增加并不会使空气变得黏稠。他解释说，在宏观尺度上，空气的黏滞性（即像蜂蜜一样的黏稠阻力）早已变得微不足道。飞行器之所以能飞，依靠的是动态阻力与惯性——机翼在向前推进时将空气向下推，而向下排开空气的反作用力将飞机托举在空中。即使把飞机放大很多倍，这一动力学原理依然适用。然而，费曼指出，简单的等比例放大存在严重的物理极限。如果将一架普通飞机的尺寸等比例放大 100 倍，它根本无法起飞。这是因为飞机的自重随着尺寸的三次方（体积）增长，而机翼产生的升力只随着尺寸的二次方（表面积）增长。为了解决这个问题，大型飞行器必须做成空心的。费曼强调，在不同的尺度下，起决定性作用的物理效应是完全不同的：“你走错方向了，要获得有效的黏度，你得走向更小的尺度。” - **升力与重量的尺度法则**：重量正比于尺寸的立方（$L^3$），升力（依赖机翼面积）正比于尺寸的平方（$L^2$）。 - **大型飞机的工程对策**：必须采用空心薄壁结构以削减自重，否则自重增长过快将彻底压垮升力。 - **黏滞效应的方向性**：只有在极微小的尺度下，流体的黏滞效应（有效黏度）才会占据绝对主导；在巨大尺度下，惯性与动态冲量才是主导因素。 --- ### [45:33](https://www.youtube.com/watch?v=4eRCygdW--c&t=2733s) 带着紧身衣的想象力与反重力机一名听众请教费曼，作为一位极具独创性的思考者，会如何着手设计一台微型的“反重力机器”（anti-gravity machine）。费曼毫不犹豫地回答自己根本造不出来，并阐明了他在科学思考中所坚守的哲学底线。费曼描述自己的思维方式是“戴着紧身衣的想象力”：“我玩的游戏非常有趣，那是被紧身衣紧紧束缚的想象力——它必须与已知的物理定律相符合。”他解释说，自己从不假设物理定律已经改变，而是在人类已知物理法则的严苛边界内，竭尽全力运用想象力去探索什么是真正可行的。这需要先发挥想象构思可能性，然后再回过头进行严格的物理分析与校验。谈及反重力时，费曼直言自己会立刻放弃，因为在现有的引力物理学中，反重力没有任何理论依据。他幽默地指出，目前唯一存在且极其有效的“反重力机器”，就是听众屁股底下坐着的椅子、垫子或地板——这些结构提供了对抗重力的支持力，让人类能够稳定悬浮在地表上方几英尺处，而且能支撑极长的时间。当听众追问已知的物理定律究竟有多“绝对”时，费曼解释说，科学的前沿总有未知领域，定律在某些极端尺度下可能会失效（比如比厘米小上成千上万亿倍的超微观距离）；但在人类日常经验的厘米级尺度，以及本场讲座讨论的原子级尺度，现存的物理定律已经经过了无数次深入且精确的检验，其可靠性极高，几乎不可能出现根本性差错。 ```mermaid flowchart TD accTitle: 费曼的物理构想与验证流程 accDescr: 描述费曼如何在已知物理定律的框架内发挥想象力并检验设计可行性 subgraph S1["构想阶段 (Imagination)"] A["提出创新设想与设备形态"] end subgraph S2["束缚与检验 (Straight Jacket)"] B{"是否符合已知物理定律？"} C["判定物理上行不通 (如：反重力机器)"] D["判定物理上可行 (如：原子尺度组装)"] end subgraph S3["执行/放弃 (Outcome)"] E["果断放弃该构想"] F["细化工程与尺度物理设计"] end A --> B B -- "不符合" --> C B -- "符合" --> D C --> E D --> F ``` --- ### [48:29](https://www.youtube.com/watch?v=4eRCygdW--c&t=2909s) 热抖动中的原子固态与激光的光学极限另一位听众提出疑问：如果不打破物理定律，如何在原子尺度上制造零件？那些只有 1 个或 5 个原子大小的零件难道不会自行散架吗？如何让这些原子乖乖停留在原地？费曼解释道，在常温下原子确实从来不是静止的，它们永远处于剧烈的高频热抖动（jiggling）之中。但在金或银等固体物质中，原子之间的结合键力相当强韧，将它们牢牢约束在晶格位置上。费曼形象地打了个比方：固态中的原子就像是一支正在行进的高中管乐团，“就像高中乐队在游行，没人真正知道该做什么，大家都在摇摇晃晃，但这没关系，队伍依然保持在一起。”如果温度过高，或者材料内部的相互作用力太弱，原子就会滚出原位、脱离队伍，固体便熔化成了液体。费曼解释了他在演讲中为何设定每个信息比特需要由一个 $5 \times 5 \times 5$（共 125 个原子，约 100 个原子量级）的微小立方体块来存储。如果只使用 5 个原子组成的微团，热抖动的统计涨落会随着时间推移，偶然导致一个原子逃逸脱落，随后接二连三流失，几年后人类记录在上面的全部知识信息就会荡然无存；而采用 5 个原子边长的微团，经热力学计算，其被热抖动彻底破坏并导致信息遗失所需的时间长达数十亿年。紧接着有听众询问是否可以用激光来实现原子尺度的精密加工。费曼澄清说，激光虽然是相干且高度可控的高品质光束，但它本质上依然是光，受到光波波长的衍射极限限制。激光加工的微缩极限大约在 2000 倍左右，无法触及单个原子的尺寸。他试图打比方解释这种尺度错位：这就像试图用比目标物体巨大得多的工具去夹取细小颗粒一样，哪怕工具再锋利也是不可能办到的。 - **原子热运动与固态稳定性**：固体并非静止不动，而是原子在强束缚势阱中的热振动。 - **热涨落与存储寿命**：原子微团的自发热扩散概率随尺寸减小呈指数上升；$5 \times 5 \times 5$ 原子团可确保信息稳定留存数十亿年。 - **激光的光学波长极限**：常规激光受限于微米/亚微米量级的波长衍射极限，无法直接用于单原子尺度的微观操纵。 --- ### [51:33](https://www.youtube.com/watch?v=4eRCygdW--c&t=3093s) 光速壁垒、狭义相对论与物质质量膨胀一位听众联想到人类打破音障的历史，并指出人眼只能感知可见光，进而提问物理定律是否允许物质运动超越光速。费曼给出了坚决的否定回答，并现场普及了狭义相对论的基本物理现实。费曼指出，在现存物理定律和相对论框架下，任何初速度低于光速的有质量物体，都不可能被加速到甚至超越光速。当你持续对物体施加巨大的推力并注入越来越多能量时，它的速度确实会越来越快，但其惯性质量也会急剧攀升。随着物体变得越来越沉重，进一步提升哪怕一点点速度都需要极其庞大的能量输入，使其只能无限逼近光速，却永远无法达到或逾越。费曼以斯坦福直线加速器中心（SLAC）等高能粒子加速器中的真实实验为例。在这些庞大的加速装置中，物理学家向电子注入了巨量的电磁能量，将电子加速到极其逼近光速的程度——其速度与光速的差距仅剩下千万分之几甚至亿分之几。在这一极高能态下，根据相对论质量效应，原本极其轻微的电子，其动态质量可以膨胀到静止质量的 40,000 倍，甚至比一个静止质子还要重上 20 倍。当该听众随后转向反物质，并试图借此引申讨论通灵与精神世界中是否存在超光速振动的实体时，费曼始终立足于实验可验证的实证物理框架，坚决排除伪科学与超自然遐想的干扰。 - **相对论光速极限**：“根据目前的物理定律和相对论，任何当前运行速度低于光速的物体，都无法被加速到超过光速。” - **相对论质量增加效应**：注入物体的动能转化为惯性质量，导致加速难度呈非线性剧增。 - **高能电子的极端物性**：在大型直线加速器（如 SLAC）中，电子速度与光速相差不到数千万分之一，其动质量膨胀为其静止质量的约 40,000 倍。 ### [54:06](https://www.youtube.com/watch?v=4eRCygdW--c&t=3246s) 实验可观测性与反物质的本质理查德·费曼（Richard Feynman）在回答观众席的提问时，重申了物理学的根本原则：科学家只能探讨仪器能够观测到的现象。费曼明确表示：“关于任何我用任何仪器都看不到的东西，如果你问我，它们可以是任何状态、做任何事情，我根本无从知晓，我没有任何办法去判断。”但只要提问者能够提供哪怕一条线索，指明某种可以进行观测的途径，他就能通过设计测试来给出回答；否则在缺乏经验事实的前提下，自己只能保持“空无所知”（empty-minded）。随后，有观众提及反物质并试图探讨其是否具有某种“精神性”（spiritual）。费曼果断打消了这种玄学猜想，强调科学界在实验室里对反物质了如指掌：物理学家不仅能在实验室中直接观察到它，还能针对它开展各种实验。他指出，反物质并非任何超自然事物，“它和普通物质属于同一类东西，同样可以进行实验研究，我们对其行为规律掌握得非常透彻”。当被追问反物质是否也是一种“实体”（substance）以及它是否会振动时，费曼解释道，反物质就像普通物质一样是实实在在的物质；至于振动，普通物质在受到撞击时（例如用锤子敲击一口钟）当然会振动，反物质也遵循对应的物理规律。接着，一位听众开玩笑地问自己的伴侣总问他“是什么让世界运转”，费曼幽默地调侃道，面对这种宏大问题，“那是你的难题，不是我的难题”。 - **可证伪与仪器检验**：无法被任何实验仪器触及的假设不在物理学讨论范畴之内。 - **反物质的物理属性**：反物质并非神秘的精神产物，而是与普通物质遵循同等明确物理法则的实体，可测、可控、可知。 --- ### [55:55](https://www.youtube.com/watch?v=4eRCygdW--c&t=3355s) 海滩上的狂想：把整本书蒸镀到一根两英寸的细线上有观众询问微纳制造是否为费曼目前的主要研究方向，费曼纠正道，这并不是他的主业，而更像是一种消遣和游戏，从 1950 年左右开始他就断断续续地以此自娱自乐。他回忆起当年自己在海滩上闲坐时的一次思维漫游：那时他开始琢磨，要记录一整本书的内容，最小需要多大的空间？ ```mermaid flowchart TD accTitle: 海滩微型化存储构想 accDescr: 费曼构思的分层蒸镀与切线多副本制造流程 subgraph Evaporation ["大面积层叠蒸镀"] A["大面积基底 (1平方英寸)"] -->|"逐层蒸镀原子 (5个原子代表1比特)"| B["累积至约 1 英寸厚度"] B -->|"完成全书信息存储"| C["1立方英寸的完整信息块"] end subgraph Slicing ["切片分割为线材"] C -->|"横向/纵向均质切割"| D["数以千计截面相同的细线 (Wires)"] D -->|"每根长 2 英寸"| E["每根细线均独立包含一整本书的副本"] end ``` 在 1950 年代的条件下，费曼构想了一种简单可行的方法：在大面积基底上逐层蒸镀材料。他当时设定用 5 个原子来编码 1 个比特（bit），经过测算发现，只要在一块 1 英寸见方的面积上蒸镀到大约 1 英寸的厚度，就能把一整本书的全部信息完全记录进去。更绝妙的是，由于大面积蒸镀在横向上是完全均匀一致的，如果把这块 1 立方英寸的材料纵向切成成千上万根极细的细线，那么每一根截面相同的细线都将完整包含这本书的一份副本。费曼当时被这个结果深深震撼了——仅仅一根 2 英寸长的细线，竟然就能容纳一整本书的全部信息。他激动得在海滩上到处跑，拽住陌生人喊道：“嘿，听着，你们知道这在物理上是可能的吗？！”然而海滩上的游客完全无动于衷，费曼戏谑地自嘲道：“当时一个‘香蕉’也没卖出去（根本没人买账）。” - **存储尺寸推算**：在 1950 年前后，费曼基于单比特 5 个原子的尺度，算出 1 平方英寸基底蒸镀至 1 英寸厚即可写下一整本书。 - **批量复本机制**：大面积均质沉积后再切割成线缆，可在微观尺度上一次性复制出数千份书本信息。 - **科研定位**：费曼强调自己只是偶尔因对小型计算机、微型机械或某种电阻特性的好奇而涉足微纳技术，这属于工程游戏，而非他的根本理论物理研究。 --- ### [57:05](https://www.youtube.com/watch?v=4eRCygdW--c&t=3425s) 最困难的物理学问题与纯理论家的“疯狂狂欢” 当有听众追问“在你能想象的范围内，全物理学中最困难的问题是什么”时，费曼直截了当地指出，前面讲的微型机械绝不是物理学最重要的问题，因为所涉及的尺度完全在已知物理定律确定性的适用范围之内，虽然存在很多有趣的技术与工艺挑战，但归根结底属于工程技术问题。物理学真正重要的课题永远是探寻那些“我们尚不知晓定律的领域”。一位观众提到了费曼 20 年前出版的著作（《物理定律的本性》，*The Character of Physical Law*），书中费曼曾做过一个比喻：物理学的黄金探索期可能会像游客蜂拥而至的美丽风景区一样，随着景致被完全探明而逐渐失去新奇与活力；提问者询问这一预言是否正在变成现实。费曼坦诚承认自己当年的预测出了偏差：“事实证明，当进来的实验数据变少时，理论物理学家反而拿着比以往更疯狂的理论在四处折腾，这是我当年完全没料到的。” 费曼解释道，在自己的青年时代，物理学处于极度严谨求实的阶段，一旦有人提出假说，立刻就有实验数据进行验证和淘汰；而如今，高能物理实验越来越庞大、耗时且昂贵，数据产出变得十分缓慢，理论家们失去了强有力的实验约束，反而有大把空间凭借人类擅长猜测的天性，肆意构思各种光怪陆离的理论来自娱自乐。 - **根本物理 vs 工程技术**：纳米尺度构件受已知经典与量子力学规律支配，其难点在于工艺制造，而非定律缺失。 - **数据匮乏导致的理论膨胀**：实验进展减缓不仅没有抑制理论研究，反而由于缺乏实验证伪的门槛，诱发了更多缺乏检验的纯数学推演与猜测。 --- ### [59:34](https://www.youtube.com/watch?v=4eRCygdW--c&t=3574s) 物理学终局的两种可能：定律探究的边界与应用物理的无穷讲座接近尾声，部分听众准备离席，费曼体贴地建议赶时间的听众先行离场去参加其他活动，自己留下来继续解答其余问题。针对一位女听众关于“物理学是否终将把所有碎片拼合完毕并终结”的疑问，费曼系统地分析了基础物理学走向终局的可能路径。费曼清晰地区分了物理学的两个不同维度：一是发现自然界的新基本定律；二是为已知规律寻找全新的应用方式。对于前者，费曼认为存在几种可能性： 1. **彻底解明**：物理学家找到了所有基础定律，经过反复检验无懈可击，探索自然画上句号。 2. **倦怠放弃**：基础探索越来越远离日常生活，实验设施变得极其昂贵，研究蜕变为小圈子“祭司阶层”的自说自话，最终失去社会资助，研究者逐渐流失。 3. **陷入永恒无解的拉锯**：人类永远无法得到最终答案，但在成百上千年中持续狂热求索；或者彻底绝望而放弃。但对于第二种维度——应用物理学（Applied Physics），费曼态度截然不同：“应用物理学的发展将比基础物理定律的探寻持续久远得多，因为我们对已知定律的应用方式还有无数种根本没有想到；我完全无法想象它会如何终结。” 随后，有听众提出在微观尺度下制造热机是否可行、需要依托怎样的物理循环。费曼回应称，热的本质是分子的无序运动，但在微观尺度，这种无序运动不必局限于普通热机中分子的碰撞回弹，它完全可以体现在自旋（如磁矩的随机翻转或核磁效应）上。人类可以利用自旋的随机涨落构建磁循环热机，也可以从生物界汲取灵感——活体生物从化学反应中获取动力，它们的工作方式远比宏观的内燃机更巧妙多维。 --- ### [01:03:51](https://www.youtube.com/watch?v=4eRCygdW--c&t=3831s) 光学极限的本质：波长与聚焦的物理障碍最后，一名听众对微缩光刻的物理极限提出疑问：在微观尺度下，到底是什么物理机制使得光无法被无限缩小并聚集到更微小的点上？费曼用海浪拍岸的直观物理图像来解释衍射极限（Diffraction Limit）： > “想象一下海浪拍打海岸的情景，岸边散落着大量的软木塞。如果你想在极其狭小的空间内单独移动某个软木塞，而你唯一获准使用的海浪具有特定的固定波长，那么你根本不可能把这种波长的波浪聚焦到比其自身特征波长还要小的尺寸范围内。” 费曼进一步拆解了这个动力学过程：哪怕你从极其宏大的外部尺度，把来自各个方向的波浪向中心汇聚，干涉与波动力学规律也决定了你无法将扰动局限在小于其特征波长的空间之内。最终，岸边海水的涌浪必然会同时冲刷并移动覆盖该波长尺度内的一大片软木塞。在光学微缩与光刻技术中也是同理，单色光的波长决定了其能够解析并聚焦的绝对物理下限，人类根本无法利用特定波长的光束去单独操控小于该波长尺度的单颗“软木塞”（微观原子或分子）。 - **海浪与软木塞类比**：水波无法精准只晃动一颗软木塞而不扰动同一波长范围内的相邻物体。 - **波长决定分辨率极限**：聚焦光束的最小截面受限于其自身特征波长，这是波动光学不可逾越的底层物理限制，也是推动光刻技术不断迈向更短波长（紫外线、X射线）的根本原因。 ### [01:04:55](https://www.youtube.com/watch?v=4eRCygdW--c&t=3895s) 突破可见光极限：光刻中的电磁波谱与 X 射线面对听众关于能否突破波长限制的提问，理查德·费曼（Richard Feynman）首先纠正了自己早先略显狭隘的回答。他承认，如果仅局限于所谓的“光”，确实会受到波长的硬性约束，但物理学中并不受制于此。若改用波长更短的辐射，例如紫外线甚至 X 射线，制造更微小结构的设想在原理上是完全可行的，并且当时各大实验室已在全力研究利用 X 射线使光刻胶（resist）感光的工艺。尽管传统的感光银盐技术无法直接套用，必须开发全新的工艺流程，但这正是缩小微电子元件的关键前沿方向。 ```mermaid flowchart LR accTitle: 电磁波谱能量与波长关系 accDescr: 展示从长波到短波的电磁波演变，频率与能量递增，波长递减。 subgraph LongWaves ["长波长 / 低能量"] A["无线电波 & 电视信号"] --> B["红外线 & 热辐射"] end subgraph VisibleRange ["可见光区"] B --> C["可见光 (红到蓝)"] end subgraph ShortWaves ["短波长 / 高能量"] C --> D["紫外线 / 远紫外"] D --> E["软 / 硬 X 射线"] E --> F["硬伽马射线"] end ``` 费曼借此向听众梳理了广袤的电磁波谱：从长波端的无线电波、电视传输信号、热辐射和红外线，到中间狭窄的人类可见光区，再延伸至远紫外线、软 X 射线、硬 X 射线乃至能量极高的硬伽马射线。他指出，日常语言按不同用途赋予了它们五花八门的名称，但在物理学家的眼中，它们在本质上是完全统一的现象，只是振动频率从慢到快的调和振动（harmonic oscillation）。关于这种物理本质的统一性，理查德·费曼解释道：“It's just the same damn thing along a whole scale from one enormous range okay.” - 紫外线可将图案尺寸缩小约一半，而更进一步的超远紫外（extreme UV）和 X 射线则能进一步压低极限。 - 硬 X 射线与伽马射线的波长能够远小于原子尺寸，甚至达到原子核的尺度。 - 在物理本质上，所有这些辐射均属于电磁辐射（electromagnetic radiation），只是宏观名字随波长范围发生改变。 --- ### [01:07:27](https://www.youtube.com/watch?v=4eRCygdW--c&t=4047s) 如何探测小于原子的尺度：碰撞截面与高能物理现场一位听众坦言自己被搞糊涂了，费曼幽默地调侃对方“总是很困惑”，随后切入核心问题：当我们手中的所有工具都是由原子构成的时，我们究竟该如何探测比原子还要小的东西？费曼用了一个网球的直观比喻来解释散射截面的原理。假设手头只有网球，若把网球随机扔向一堵挂着目标的墙，通过统计反弹回来的球数与直接穿过去的球数比例，就能推算出目标的几何尺寸大小。目标越小（比如从球缩小成一枚大头针），发生正面碰撞反弹的几率就越低，这就构成了测量微小截面的基础。费曼进一步解释了现代高能物理如何通过加速粒子来获得更短的探测“探针”。通过加速电子等带电粒子使其撞击靶材，注入的能量越高，产生的电磁波波长就越短。即使这一波长短到无法直接用尺子衡量，科学家也可以根据已被严格检验的物理定律，通过高能碰撞产生的短波来解析极其微观的世界。为了让大家直观感受原子内部的空旷，费曼打比方说，如果整个原子有一间房间那么大，那么原子核就只相当于房间里一个微小的尘埃或标点符号中的句点，其尺度比原子整整小了 10 万倍（100,000 times smaller）。谈及现代对质子内部结构的探索成果，理查德·费曼风趣地感慨道：“We're down on the insides of protons and things at the moment that's what your taxpayer money has been able to produce.” --- ### [01:10:22](https://www.youtube.com/watch?v=4eRCygdW--c&t=4222s) 声波的微型化极限：超声波与“声波的尽头” 一位女士举手发问，询问能否将光线换成声波来处理这种微小尺寸。费曼最初指出日常声波的波长在微米尺度上实在太长了，但提问者提醒他注意超声波。费曼大度地承认自己方才说得太绝对，并迅速展开讨论。在介质中传播的高频超声波波长确实可以变得极短，但在极微观的尺度下，声波存在一个致命弱点：光波可以在真空中传播，而声波是物质内部粒子运动的机械波，由于原子在常温下始终处于杂乱无章的布朗热运动中，极高频率、极短波长的声波会在物质中遭遇极强的吸收与衰减，难以维持长距离的相干波列。不过费曼称赞这是一个十分有趣的构想，甚至打趣说自己今晚要仔细算算数据，也许明天就会发现这主意不行，但当下它值得深入琢磨。随后听众追问：如果不断提高声波的频率，音调一直上升，最终会发生什么？费曼明确指出，声波存在物理上的绝对上限。当声波波长缩短到与介质中相邻原子之间的间距相当时，“波”的概念就彻底失效了。费曼生动地拿百老汇无线电城火箭女郎（The Rockettes）踢腿排舞来做类比：一排演员依次踢腿，可以形成一道漂亮的行进波浪；但如果要求波峰和波谷之间的距离比两个相邻演员之间的间距还要窄，这在逻辑上根本无法实现。因此，在原子构成的离散介质中，当相邻原子只能做反向振动时，系统就达到了振动模式的终点。关于无法无限制提高声波频率的本质，理查德·费曼总结道：“And so you can't make a sound wave where the vibration of the successive atoms is in a pattern which is what a sound wave is it's a pattern like the Rockettes but which each atom next to each other is any faster than opposite there just isn't any way to describe it so that's what I mean by there's no more sound it doesn't mean anything anymore.” --- ### [01:14:00](https://www.youtube.com/watch?v=4eRCygdW--c&t=4440s) 接近光速下的粒子性质：电子与相对论讲座进入尾声的提问环节，一名听众询问，在粒子加速器中电子的速度被加速到极为接近光速时，电子的质量会增加，那它“是否仍然是一个电子”？尺寸是否会变大？费曼毫不犹豫地肯定回答，它毫无疑问依然是正规的电子，只是其运动行为严格遵循狭义相对论所描述的高速运动规律。关于“电子直径”是否会发生变化，费曼指出在现代量子电动力学和粒子物理中，点状的基本粒子（如电子）并没有一个可以直接测量的经典刚性“直径”，因此讨论其直径变化需要预先严格定义。但如果换作具有明确物理边界的复合实体（例如一个完整的原子），当它被加速到接近光速时，在垂直于运动方向的维度上，其直径将保持不变，而在运动方向上则会发生相对论性的尺缩效应。 ### [01:14:50](https://www.youtube.com/watch?v=4eRCygdW--c&t=4490s) 狭义相对论插曲与问答环节的展开费曼顺着前文的话题打趣地提到狭义相对论效应：当物体的运动速度非常接近光速时，它在运动方向上会显得越来越扁平，像一个椭球体，直到最后几乎变成一个薄盘。他笑着对听众说：“我确信你们了解相对论这一点，但我并不是在给我古怪思维课（idiosyncratic thinking class）的学生讲相对论，我只是想让大家知道，这是为明年的课程做个预告，我们今天不讨论这个。”随后，他宣布结束正式演讲，进入自由提问环节。台下一位听众举手提出了一个典型的“采访式问题”，询问费曼在漫长而辉煌的科研生涯中，觉得哪一件事是他发现、思考或探索过的“最令人兴奋的事情”。费曼略微思索后回应道，科学发现的兴奋感往往是特定个体在历史长河中的一次偶然经历，带有极度个人化的色彩，因此很难给出一个非黑即白的标准答案。他进一步澄清，最让人兴奋的瞬间往往不是“最重要”的发现，而是答案突然降临的瞬间。 - 费曼打趣预告明年的相对论课程，声明今天的讲座不展开深究。 - 听众提出经典采访问题：费曼科研生涯中最令人兴奋或玩得最开心的一个发现是什么。 --- ### [01:16:09](https://www.youtube.com/watch?v=4eRCygdW--c&t=4569s) 两大最令人兴奋的科学顿悟：弱相互作用与液氦超流费曼回顾了自己的学术历程，挑出了两个带来最极致愉悦感的时刻。第一个是他对放射性衰变中弱相互作用（β衰变）规律的猜测。虽然他最初的猜测没有做到百分之百的绝对精确，但比前人猜得要准得多，这一理论在相当长的一段时间内都是正确的，直到后来被新一代物理学家发展成了更完备的形式。第二个令他痴迷两年的重大问题是液氦（Liquid Helium）的超流现象。在极低温度下，液氦会变成一种毫无阻力流动的液体，其黏度彻底降为零。费曼生动地描绘了一个实验设计：把一根玻璃棒和一根白金细丝封在一起，降温时由于两者的热膨胀系数不同，玻璃与白金之间会拉扯出一条极其微小的裂缝（infinitesimal crack）。用这条极细微的裂隙连接两个高度不同的氦容器，液氦会毫无阻碍地流过去，甚至会因为惯性滑行而冲过头（overshoot），然后再荡回来往复振荡，完全感受不到任何流动阻力。 ```mermaid flowchart TD accTitle: 费曼攻克液氦超流物理机制的顿悟过程 accDescr: 展示费曼从面对海量原子方程的困境，到陷入错误假设，再到灵光一现解开超流之谜的思维推演路径。 subgraph S1["理论困境"] A["量子力学基础定律明确"] --> B["体系包含太多原子 无法直接解析薛定谔方程"] end subgraph S2["思维碰壁"] B --> C["需要极大的物理想象力"] C --> D["构建假说但误以为某条实验曲线是平滑的"] D --> E["自我反驳，陷入理论死胡同"] end subgraph S3["顿悟解惑"] E --> F["猛然想起真实实验曲线并非平滑"] F --> G["灵感爆发，推导全线贯通 完全解开液氦超流机制"] end ``` 当时的物理学界深知量子力学定律是正确的，但面对液氦中天文数字般的原子数量，根本无法直接通过解方程得出宏观的超流特质。费曼苦苦思索了两年，直到某一天脑海中突然闪过正确的图像。一开始，他用一个论据否定了自己的想法，因为他误以为某条来自实验的曲线是平滑的；但就在片刻之后，他猛然记起真实实验里那条曲线根本不是平滑的，那个关键事实瞬间挽救了他的理论。费曼描述道：“当时就像‘砰’的一声，一切问题都迎刃而解了，就像突然推开了一扇大门，答案瞬间涌现。” - **β衰变定律**：费曼的弱相互作用猜想极大地推进了当时的粒子物理认知，成为后续更完备理论的基石。 - **液氦超流性**：在极低温下黏度消失，通过白金与玻璃收缩形成的微观缝隙无阻力流动，形成无阻尼振荡。 - **攻克历程**：历经两年苦思，从原本被错误实验记忆卡住的死局中突然顿悟，彻底解开量子力学如何在宏观流体中体现的机制。 --- ### [01:18:23](https://www.youtube.com/watch?v=4eRCygdW--c&t=4703s) 兴奋过后的“等明天”与幽默收尾费曼感叹道，当一个人为了弄懂某件事苦熬数年，突然间一切豁然开朗时，那种感觉不仅极度刺激，而且充满了纯粹的快乐。在那一刻，“整件事情如同咆哮着奔涌而出”，你可以紧握这把钥匙推演所有细节，同时在心底暗暗祈祷千万别再犯蠢搞错。然而，这种科学探索中的欣喜若狂往往伴随着现实的幽默。费曼笑着调侃自己的私生活，说他的妻子早就对他这种狂热见怪不怪了。每当他在楼上研究出成果，兴高采烈地冲下楼冲着妻子大喊“我搞明白了！我搞明白了！”时，妻子总是波澜不惊地回敬一句：“我以前听你说过这话了，等明天再说吧（Wait till tomorrow）。”费曼坦白地承认：“她几乎总是对的，几乎每一次都是对的，一次又一次，‘等明天吧’，结果到了明天我果然发现自己犯了个错误。” 讲座在全场的笑声中接近尾声。观众席和费曼互相开玩笑打趣，费曼戏谑道，很多时候你自以为做出的新发现，回过头一看，“苏联人两百年前就已经发现了，就像打网球一样”。在轻松愉悦的气氛中，主持人用一句“谢谢你，总统先生（Thank you, Mr. President）”向费曼致谢，费曼在一片善意的掌声与笑声中打趣喊着“救命，救命啊”，正式结束了这场内容丰富、妙趣横生的讲演。 > “在经历了数年的苦思冥想之后，一切突然变得清晰起来，或者说哪怕还没有彻底明朗，但你确信自己已经拿到了钥匙，整件事情如山洪般倾泻而出。” —— 理查德·费曼（Richard Feynman） - 费曼妻子的经典箴言“等明天再说”，精准戳穿了科学家灵感上头时常有的盲目乐观。 - 探索科学往往经历“兴奋—自省—纠错”的反复循环，即使顶尖物理学家也不例外。 - 讲座在关于“苏联人早就发现过了”的调侃与听众热烈的欢呼互动中画下句号。 ## Notable quotes > [45:33](https://www.youtube.com/watch?v=4eRCygdW--c&t=2733s) 我玩的游戏非常有趣，那是被紧身衣紧紧束缚的想象力——它必须与已知的物理定律相符合。 — 理查德·费曼 > [59:34](https://www.youtube.com/watch?v=4eRCygdW--c&t=3574s) 应用物理学的发展将比基础物理定律的探寻持续久远得多，因为我们对已知定律的应用方式还有无数种根本没有想到；我完全无法想象它会如何终结。 — 理查德·费曼 > [00:08](https://www.youtube.com/watch?v=4eRCygdW--c&t=8s) 最终的极限是，你当然无法让一个数字无限变小——无论写得大还是写得小，数字本身都是一样的，你可以说任何尺寸都行，但你无法让它比原子还要小，你无法在原子上写字。 — 理查德·费曼 > [54:06](https://www.youtube.com/watch?v=4eRCygdW--c&t=3246s) 关于任何我用任何仪器都看不到的东西，如果你问我，它们可以是任何状态、做任何事情，我根本无从知晓，我没有任何办法去判断。 — 理查德·费曼 > [01:18:23](https://www.youtube.com/watch?v=4eRCygdW--c&t=4703s) 在经历了数年的苦思冥想之后，一切突然变得清晰起来，或者说哪怕还没有彻底明朗，但你确信自己已经拿到了钥匙，整件事情如山洪般倾泻而出。 — 理查德·费曼 ## Mentioned 理查德·费曼 (Richard Feynman), 加州理工学院 (Caltech), 大英百科全书 (Encyclopedia Britannica), 康奈尔大学微加工实验室, 汤姆·范·桑特 (Tom Van Sant), 陆地卫星 (Landsat), 阿尔·希布斯 (Albert Hibbs), 斯坦福直线加速器中心 (SLAC), 无线电城火箭女郎 (The Rockettes), 《物理定律的本性》 (The Character of Physical Law) ## Video description Donate and Support this Channel: https://www.paypal.com/donate/?cmd=_s-xclick&hosted_button_id=BLJ283JMTMT7S (Original Transcript Available Here:http://muonray.blogspot.ie/2012/12/richard-feynman-theres-plenty-of-room.html) Richard Feynman gave his famous talk "There's Plenty of Room at the Bottom" on December 29th 1959 at the annual meeting of the American Physical Society at the California Institute of Technology (Caltech) as his vision on how physics and engineering could move in the direction that could eventually create nanotechnology. Really good ideas and strokes of genius are often manifest in the right questions being asked: How small can information be encoded? How can information be written? How can information it be read? All of these important "Hows" were asked by Feynman in a time when computers had to be put in large rooms and when the impending space race was forcing engineers to do some serious strategic thinking in making technology small enough to be lifted by rockets into space to function as serious tools in scientific exploration and defence. Feynman himself may not have invented the technology we see in the development and continuity of the computer age, but the fact that even in the early 1960's nanotechnology was being considered as a serious field of study was definitely a factor contributing to the boom in computer technology seen in the late 20th century and continues to reach more spectacular levels of sophistication in the 21st century. Jump 25 years forward into the year 1984, when Feynman tries to retell his 1959 lecture from a more modern perspective in that many aspects of his vision have been fulfilled, particularly with the invention of the electron microscope, the atomic force microscope and experimental manipulation of the atomic scale of matter. Also discussed is the current practical field of photolithography for the manufacture of bipolar transistors and junctions used in computer chips done on an industrial scale and how this process continues with ever decreasing wavelength capabilities of lasers from UV to X-rays. Feynman also discusses the boundaries of miniaturization and how the scale differences affect the function of certain aspects of technology as well as in nature. In the true spirit of Feynman, the discussion goes into the colorful details and gives diagrammatic examples of how this field had progressed from 1959 to 1984. We can only imagine how Feynman would have felt about the modern developments in nanotechnology in the 21st century where entirely exotic principles of physics may begin to become technologically significant, including vacuum fluctuations and quantum entanglements. Without a doubt he would have found our developments exciting but always within the realms of understanding by studying the most fundamental language of nature, quantum mechanics, to bring our macroscopic brains into visualizing, however abstractly, the tiny machinery of nature. [Source](https://www.youtube.com/watch?v=4eRCygdW--c)

Source quotes

[45:33] 理查德·费曼: 我玩的游戏非常有趣，那是被紧身衣紧紧束缚的想象力——它必须与已知的物理定律相符合。 [59:34] 理查德·费曼: 应用物理学的发展将比基础物理定律的探寻持续久远得多，因为我们对已知定律的应用方式还有无数种根本没有想到；我完全无法想象它会如何终结。 [0:08] 理查德·费曼: 最终的极限是，你当然无法让一个数字无限变小——无论写得大还是写得小，数字本身都是一样的，你可以说任何尺寸都行，但你无法让它比原子还要小，你无法在原子上写字。 [54:06] 理查德·费曼: 关于任何我用任何仪器都看不到的东西，如果你问我，它们可以是任何状态、做任何事情，我根本无从知晓，我没有任何办法去判断。 [78:23] 理查德·费曼: 在经历了数年的苦思冥想之后，一切突然变得清晰起来，或者说哪怕还没有彻底明朗，但你确信自己已经拿到了钥匙，整件事情如山洪般倾泻而出。

Source quotes

[45:35] Richard Feynman: The game I play is a very interesting one: it's imagination in a tight straitjacket, which is this—that it has to agree with the known laws of physics.

Original transcript

[0:08] [Music] [0:28] e [0:30] [Music] [0:47] how small can you make [0:50] Machinery okay that's the subject and [0:53] because I've heard people around in the [0:54] bath saying tiny machines what is he [0:56] talking about tiny machines and I said [0:58] to him you know very small machines and [1:00] it doesn't work I am talking about very [1:03] small machines okay now before I start [1:08] on machines I'd like to talk about very [1:10] small writing [1:12] first okay how small can we make writing [1:16] or say numbers how big does a number [1:19] have to be if you want to write numbers [1:22] down what's the smallest you could [1:24] possibly make them I don't mean how high [1:27] if you're very delicate with your finger [1:28] how small you can make them but with [1:30] special machinery and so on what is the [1:32] ultimate limit well the ultimate limit [1:34] is that you can't make a number of [1:36] course the number is just as good if you [1:37] write it bigger or smaller you say any [1:39] size any size but you can't make it [1:41] smaller than atoms you can't write on an [1:44] atom you can't Mark the atom number one [1:46] because marks you see are just more [1:48] atoms spread over other atoms black [1:50] atoms on top of white ones or whatever [1:52] so the way that you have to do to write [1:55] something would be let's say to have a [1:56] little patch of gold followed by a [1:58] little patch of silver by another silver [2:00] by a gold and so in some sort of code [2:02] like dots and dashes to represent [2:04] numbers and though you could say the [2:07] smallest we could go would be say [2:09] perhaps 100 atoms probably can go down [2:12] to one atom but if you want to make it [2:14] nice 100 atoms on a side and I'd like to [2:16] talk about how small that really is [2:19] because you don't quite appreciate how [2:21] much you could write if you could write [2:23] that way so I start out further back and [2:27] talk about [2:28] uh [2:31] uh things that have done in the past [2:33] people have pointed out to me that the [2:34] Lord's Prayer has been written on the [2:36] head of a pin all right now let's see [2:39] what we would have to do if we wanted to [2:41] write the entire encyclopedia britanica [2:43] on the head of a pin can we do that well [2:47] first of all the encyclopedia Botanica [2:49] has something 20 30 some odd thousand [2:51] pages and you imagine that each page is [2:54] on a special piece of paper and you put [2:56] them all [2:57] over the ground and you get a a great [3:00] big area you know 20,000 Square ft and [3:03] that has to be shrunk down to the size [3:05] of a head of a pin and if a little [3:07] figuring you'll figure it out yourself [3:09] it's about 20,000 times reduced so if we [3:13] could reduce the size of the words the [3:17] letters the dots in the pictures the [3:18] whole thing everything that's in the [3:19] andik up here only by 20,000 times this [3:22] way and of course this way too so it's [3:25] 40 million times difference in area [3:27] there's a lot of difference in area [3:28] between the head of a pin and 20,000 [3:32] Square ft uh then we could write the [3:35] entire encyclopedia on a head of a pin [3:37] and as I will indicate soon that's not [3:40] too difficult to give you some idea of [3:44] the scale in which that to which that [3:46] corresponds an entire Library like the [3:48] Caltech scientific Library can all be [3:51] put on one library card we could send [3:55] all the [3:56] information that's in the library on one [3:58] library card say to Brazil if the [4:00] library in scientific library in Brazil [4:03] Burns that out and we just send them a [4:05] one library card which contains all the [4:07] information in all the books in the [4:09] Caltech Library the Congressional [4:11] library of Washington is larger and [4:13] requires something like Time Magazine to [4:16] have all that information at the scale [4:19] and so we see that if we could go down [4:21] to 20,000 times smaller and I'll show [4:24] tell you in a moment right away that [4:26] that's nowhere near limits of atoms are [4:29] not coming in at all at that scale [4:31] there's no problem in going 20,000 times [4:34] and that's the kind of scale that that [4:35] would be possible so just a moment while [4:39] I look at my [4:45] notes if [4:48] uh we made it in the three-dimensional [4:50] manner you see all that I've done so far [4:53] is writing on the flat of the pin I [4:55] haven't used the guts of the head of the [4:56] pin the encyclopedia britanica actually [4:59] uses a volume it has page after page if [5:01] we could write deeper you see not just [5:04] on the surface of the pin but in the [5:05] interior we could ask ourselves let's do [5:07] it now with the atoms that's much [5:09] further reduced to get five atoms on a [5:11] side little cubes that's 100 atom of [5:13] gold and silver and so on and now I'm [5:15] not doing the pictures but I'm just [5:17] putting all the words in some Call of [5:18] code like moss code with dots and dashes [5:20] gold and silver how much could we put [5:24] into what [5:26] space all right now it turns out if you [5:28] take all the books in all the libraries [5:31] Turkish Hungarian everything all over [5:35] the entire world and just take the [5:37] information because I can't get the [5:39] pictures at this scale then all this [5:42] goes into a volume of [5:44] material one two 200th of an inch on a [5:48] side which is the smallest piece of dust [5:50] that you can possibly [5:52] see that's the net result of all [5:55] Mankind's arrangement of information but [5:57] all that information could Be Remembered [6:00] in a piece of dust that [6:03] size and that gives us some idea about [6:05] the fact that there's plenty of room to [6:08] make things very much smaller than we [6:10] ever made them before our books are [6:12] obviously too big what's the sense of [6:14] having all that stuff in this big [6:16] Library when you can put it all in one [6:17] card oh it's convenient to have the [6:19] books in your hand but for some kind of [6:21] a summary of all the information and to [6:23] transmit the sub information from one [6:25] place the other or send it or suppose [6:27] that you're afraid that all of [6:28] civilization is going to collapse and [6:30] you would like to leave copies of the [6:32] libraries cuz you say everything in the [6:33] Alexandria library was in one library [6:35] and that got smoked out and that was the [6:37] end of that knowledge it would be guide [6:39] good hurry up we should make copies okay [6:41] so we have all this dust you see little [6:43] pieces of dust that have all have copies [6:44] of all over and they can't get rid of [6:47] all the dust you see anyway that's what [6:50] it amounts to now you might ask if we [6:53] had something that small I the five [6:57] atoms on a side is a little well we [6:59] could still read that but let's go back [7:01] to the 20,000 times reduced encyclopedia [7:03] which isn't as small as you can get but [7:05] pretty dramatic and good enough right [7:08] how can you read [7:09] it well if you try an ordinary [7:12] microscope to look at it you can't see [7:14] you can't manfy more than 2,000 times [7:16] because light has a structure and you [7:18] can't see any closer than the structure [7:20] called the wavelength of light but you [7:22] can use electron microscopes which don't [7:24] stop at 2,000 go up to 200,000 well we [7:27] only need 20,000 it's only 10 times [7:30] better than light it's rather easy [7:32] actually to see 20,000 times reduced [7:35] pictures with an electron microscope it [7:37] would be very easy to read this book [7:39] that we wrote on the head of the pin or [7:41] this encyclopedia with an electron [7:44] microscope the next question is how [7:47] would we write [7:49] it well it it's possible to write [7:52] it someday by using a kind of a thing [7:56] like an electron microscope in Reverse [7:58] in which to take the large scale writing [8:00] and use the lenses backwards to control [8:02] a beam which is so F which is very fine [8:05] instead of running the mic you know you [8:06] can run a telescope backward you look [8:08] through the wrong end you've probably [8:09] done this and everything looks small you [8:11] can do the same with a microscope and [8:13] you can do that with an electron [8:14] microscope so you can make the pictures [8:16] very tiny and imprint them easily 20 not [8:19] very easily this turns out at the [8:20] present time to be very difficult at the [8:23] moment that reverse electron microscope [8:25] has not been developed very far but I'm [8:27] going to tell you what the situ [8:29] situation is today the first time I ever [8:31] gave this part of the speech was 20 [8:33] years ago and you just said that [8:37] someday I'm surprised we haven't done it [8:39] yet and someday it'll be done now I can [8:41] show that it can be [8:42] done but before I do that I would like [8:45] to talk about what we are now actually [8:47] doing commercially in making things how [8:50] small do we make things how delicately [8:52] can we make them are we writing the [8:54] Lord's Prayer on the head of a pin no [8:57] we're not actually writing the Lord's [8:58] Prayer on the head head of a pin we're [8:59] doing a much smaller scale than that but [9:02] we're not writing the Lord's Prayer [9:04] because in the meantime interests have [9:06] changed somewhat and uh I'll show you on [9:09] the first slide if it's available please [9:12] uh something that youve all heard of [9:13] which is a computer [9:15] chip which is made uh what happened to [9:19] the slide oh I see [9:27] okay this is only uh [9:31] they're 20,000 times reduced and it's [9:33] very difficult to see it because it's so [9:37] fine is that we have that's it good and [9:41] it has to be focused it's really quite [9:43] difficult to see that there's a very [9:45] fine structure and you can see some of [9:47] the structure but you can't quite see [9:50] how fine it really is this whole thing [9:52] is about a 3 mm across and you've seen [9:55] these things in magazines and you say [9:58] well it's just Compu computers and so on [10:00] but from a point of view of humankind [10:02] and its development it's really quite an [10:04] achievement to be able to manufacture [10:06] something with such finess of detail the [10:10] patterns look rather beautiful when [10:11] they're worked out and it does appear as [10:14] an artistic thing too but the beautiful [10:18] thing about this is the delightfully [10:20] accurate workmanship we always talking [10:22] about workmanship don't do anything like [10:24] we used to do we used to polish things [10:26] down the accuracy which they polish [10:28] things is less than one of these little [10:30] notches in here now we can even make [10:32] something in that much detail and it's [10:34] made and used as is an example of a [10:37] computer chip I'm sorry to bother you to [10:39] put the lights back again but it'll be a [10:42] minute or two before I get to the next [10:44] slide is okay it's a kind of a [10:47] complicated [10:49] Arrangement because I wanted to explain [10:52] how such a thing is made this is made I [10:56] that was magnified about 20,000 times we [10:58] can make things [11:00] at 2,000 times I mean that was 2,000 [11:04] 20,000 times is very much harder because [11:06] 2,000 times we can use light and the way [11:09] it's done is to use a lens system a [11:12] microscope [11:14] backwards what we do is uh you take some [11:17] material in fact in this particular case [11:19] it's [11:20] silicon and there's a layer very [11:23] beautifully made very pure silicon the [11:26] reason is any piece of dirt or Scrat or [11:28] anything that's wrong with it it's a [11:30] great big monster Boulder at this scale [11:32] and you don't want any dirt so you get [11:34] very pure silicon and then in a vacuum [11:37] you let in oxygen and then what forms on [11:39] the surface of this is a layer of a [11:42] compound silicon dioxide which is simply [11:46] quartz or sand or or like glass it's a [11:49] thin layer of glass which is an [11:50] insulator silicon is a conductor see [11:53] we're going to build this thing up now [11:55] on top of this the next layer we put on [11:58] some uh oh I got lots of colors that's [12:02] great another chemical which is called [12:05] is evaporated on in the thin layer which [12:07] is called a photo [12:09] resist and [12:11] then light is shine shown on [12:15] here light comes [12:22] down let's say like this in a pattern no [12:26] light here and only light here and here [12:28] because it's an optical system it's a [12:30] picture in other words a picture is [12:31] projected black and white picture is [12:34] shown on here and what happens to the [12:36] light what the light does is make this [12:38] material resist etching [12:41] later or rather dissolving it off excuse [12:44] me so it gets to be resistant here where [12:47] the light shown I've got a simple [12:49] pattern but you might have a little [12:50] section in here and so forth so you can [12:53] make the shapes that you want by using [12:55] the light backwards to make this thing [12:58] so that it doesn't dissolve and then [13:00] what happens you dissolve this material [13:03] away [13:05] right and just have [13:07] these then you attack the silicon [13:11] dioxide to glass with hydrochloric acid [13:14] which dissolves glass and this part is [13:17] erased so that this [13:21] disappears i w let's say the silicon [13:24] dioxide is white and oh oh brown [13:27] brown Brown and so we get things like [13:31] this you see little Columns of brown on [13:34] top of the [13:36] Silicon [13:41] right and then the red stuff that I [13:43] showed there isolved by another chemical [13:45] because it was only a tool and a [13:47] scaffolding and so we get to this [13:49] picture remember that these are [13:51] insulators and this is a [13:54] conductor so the next step is to shine [13:58] some is to evaporate some silicon again [14:00] and to make a contact it goes on and on [14:03] and I'm not going to go on and on but [14:04] the next level say silicon or transistor [14:07] material which is silicon with something [14:08] dissolved in it let's say is L laid down [14:11] here in a layer all over filling all [14:16] this but then again with a resist and so [14:19] forth pieces of it are dissolved away so [14:22] they get little caps and so [14:24] on and make more insulator in various [14:28] layers each time deciding where the [14:30] stuff goes by using this photo resist [14:33] trick and there ultimately building up [14:36] say a metal connector finally with a [14:38] similar device the same general idea [14:41] let's say metal is put across the media [14:44] so now we have two transistors connected [14:47] together and that slide that I showed [14:50] showed the result looking down on this [14:52] of such a work of art and Commercial [14:56] effort industry to manufacture [14:59] a particular design of an electric [15:00] circuit it's an enormously complicated [15:02] electric circuit they can make [15:04] calculations all by itself with the [15:06] design the thing that determines the [15:08] pattern is the original arrangement of [15:09] light and I just wanted to show you how [15:12] it is [15:14] made that's only 2,000 times and we want [15:17] to now talk about 20,000 times small and [15:21] uh that is new and different it hasn't [15:22] been made yet because we can't do it [15:25] because light can only go down 2,000 [15:26] times so we have to do it some other way [15:28] if we're ever going to do I spoke about [15:30] writing at 20,000 times reduced and what [15:33] I want to show [15:34] you now is writing 20,000 times reduced [15:38] this is a non-commercial process this is [15:40] a process that is being worked out in [15:42] the frontiers of this business trying to [15:45] see how small we can make things and how [15:47] small we can write things and there's a [15:49] particular [15:50] example there a laboratory at Cornell [15:54] that makes these that's doing this [15:56] particular research micro something or [15:58] other laboratory right very small making [16:01] very things very small and I have a [16:03] friend Tom vant who's an artist who [16:07] loves Art and Science and commerce and [16:09] everything else together he's a real man [16:11] of the present era and not one artist [16:14] who snear at science and doesn't [16:16] understand the world he's in he loves [16:18] the world he's in and he loves art and [16:20] so he was asked them if he they would [16:24] make a drawing for him he designed the [16:27] paint the drawing and he made a [16:29] drawing which is the smallest drawing [16:32] ever made by anybody in the world okay [16:35] and in the next slide I want to show you [16:36] the smallest drawing in the world [16:43] okay this is supposed to be a drawing of [16:46] an eye it's an artistic work okay what [16:50] it really is was a salt crystal and a [16:53] beam that was moved around to make whole [16:57] to dig away the salt so as to make this [17:00] image and then the image is looked at in [17:02] an electron [17:04] microscope this thing compared to the [17:07] normal human eye is 100,000 times [17:11] reduced that's more than I've been [17:12] talking about before 100,000 times [17:15] smaller than the human eye is the actual [17:17] drawing and of course this is a [17:18] magnification back again so you can see [17:22] it the artist of course has the right [17:25] quality all art has this quality of of a [17:28] kind of you don't really care it's sort [17:30] of you know it's very [17:32] beautiful this was caused by a truck [17:35] that came by shaking the beam you know [17:37] shaking the apparat because even the [17:39] tiniest vibration at this scale is a big [17:41] movement and that produce that which [17:43] makes a very beautiful picture to get [17:45] some idea of the scale of this the cross [17:48] the distance across this thing is [17:50] approximately 100 [17:52] atoms which is a as small as anything [17:55] has been made yet the a the uh Tom had a [17:59] definite idea he wanted 100,000 times [18:02] less than an eye for a reason that I'll [18:04] you'll understand in a moment and they [18:06] were people there were disappointed [18:08] because they can make the dot about half [18:09] as big and the lines about half as thick [18:11] and a whole drawing about half size but [18:13] we insisted he and Tom insisted that [18:16] this be the [18:17] side because he had another drawing of [18:20] an eye which I would like to show you [18:21] the same artist had previously made [18:24] another drawing of an eye which is in [18:26] the next [18:27] slide [18:29] and that of course is the true kind of [18:30] art with different kinds of Patches at [18:33] different colors and so on which [18:35] represents Modern Art right that's a [18:37] beautiful picture of an eye I really you [18:40] know if you made an eye you'd make a [18:42] circle and a line but they make it such [18:44] beautiful colors and everything else to [18:47] see more about this picture I want to [18:49] show it at a different scale which is in [18:52] the next slide because he has included [18:55] here the eyebrows of the eye and [18:58] something Rings under the eyee you see [19:00] and there's the eye to get still a [19:03] better idea of the true scale of the [19:06] canvas for this picture we look at the [19:08] next [19:09] slide this is the city of Los Angeles [19:13] and there is the eye all right and this [19:16] is the largest picture in the world that [19:18] has ever been drawn okay okay it it is [19:22] 100,000 times larger than a normal [19:27] eye now I would like to tell you a [19:29] little bit about how the picture was [19:32] made how did that this of course is a [19:35] picture from the lat satellite this [19:37] drawing is so big you can't just look at [19:39] it you got to go up there 600 miles into [19:41] the sky and look down with the landat [19:44] satellite to see it is it possible go [19:46] backwards with the slides to two slides [19:48] back from [19:50] this yes thank [19:52] you what was actually done is what [19:54] happens with the lat satellite is this [19:57] it has a beam and it looks at the ground [19:59] goes back and forth like this as it [20:01] sails over the ground and the beam goes [20:03] back and forth and it's computed the [20:06] light that comes at any moment into the [20:08] cell moment after moment is computed and [20:11] an image a spot a little square is made [20:13] for that direction at that moment there [20:16] are three different colors information [20:18] that's taken in and that's added [20:19] together to get the color each one of [20:22] these is called a pixel because it's all [20:24] calculated from the thing and you don't [20:25] see that when you look at the full [20:27] picture just just like when you look at [20:28] a complete picture in a magazine you [20:31] know it's made of a lot tiny dots well [20:33] the dots they're too small to see on the [20:35] normal picture but here they are big [20:37] enough to see now each one of these [20:39] covers about an [20:41] acre and we'd like to know how Tom [20:44] vant had the energy to cover all these [20:47] Acres with white over a distance of 2 [20:50] and 1 12 kilom this drawing this picture [20:52] corresponds to a drawing 2 and A2 [20:54] kilometers wide and this is the way he [20:57] did it [20:59] he set up a [21:01] mirror at different stations there are [21:04] 24 mirrors in this picture each mirror [21:07] is set somewhere and the angle exactly [21:11] calculated because the the pro the [21:13] timing for the lat was well known was [21:16] given to him by the lat people and he [21:18] calculated the angle to set a mirror so [21:21] that at the moment that the beam went to [21:22] look at that [21:24] acre the angle was just right that the [21:26] sun went right up into the right into [21:29] the camera and you know saturated the [21:32] camera it was a flash of white so that [21:34] that particular pixel looked like the [21:36] whole thing was white it's just as [21:39] though you see you don't have to have a [21:41] whole acre so he had [21:43] 24 uh mirrors set up in the desert each [21:47] one calculated very accurately could we [21:50] just see that just one moment uh can I [21:52] see the picture just one moment because [21:54] there's one slight error in [21:56] it which I want to call your attention [21:58] to this pixel didn't come out quite [22:00] right there's an error there now I come [22:02] to the next to the next slide I mean the [22:05] final Slide the seventh one yeah that's [22:09] right that shows them setting up this is [22:11] Tom vant and there he is with the angles [22:15] you see and the levels and everything [22:16] else setting up his mirror to the [22:19] correct angle in order to uh reflect the [22:22] light into the lat in order to produce [22:24] this picture for us what happened was it [22:27] turned out after he saw the picture and [22:28] he went back to discover what happened a [22:32] jack rabbit had run over the mirror and [22:34] changed the angle so it didn't [22:37] work I know that's not got to do with [22:40] small things it has to do with large [22:41] things but I couldn't resist after I [22:44] showed you the smallest drawing ever [22:45] made to show you that the same artist [22:47] has also made the largest drawing ever [22:49] made and the comparison between the the [22:53] eye of [22:54] the child in this side can be seen [22:56] something like this [22:58] if you think of the eyelash which is a [23:00] hair on the on the ordinary eye and [23:03] magnify that hair until it's about the [23:06] size of an eye then an eye gets to be [23:08] across this room then you're halfway [23:10] there then you take an eyelash or that [23:12] eye and magnify till it's across this [23:15] room and the big eye is 2 and 12 km [23:17] another way of expressing the size is to [23:20] say let us go up another scale of the [23:22] same amount again of another 100,000 [23:25] what and try to draw an eye where would [23:27] we have to draw what is [23:29] it 100,000 times bigger than the 2 and2 [23:31] km eye it turns out as a beautiful eye [23:34] in the heavens namely Saturn with her [23:36] rings so that gives you some enor idea [23:39] of the enormous scale both up and all [23:42] the way down how tiny that draw that [23:45] drawing actually is as 2 and 1 12 kilm [23:47] is to your eye so that your eye is to [23:50] that original first drawing which is so [23:52] very [23:56] tiny then next question is what about [23:59] making computers still smaller and now I [24:02] launch onto another thing not a question [24:05] of what's practical today but what's in [24:07] principle [24:08] practical it's already noticed for [24:10] instance that if we were to try to make [24:12] the wires about half as big or a third [24:14] as big as the wires in that particular [24:17] design it's a strange thing that when [24:20] wires carry electricity tool on they [24:22] move the matter moves it doesn't move [24:24] much it doesn't make any difference [24:25] there but when the wires are very much [24:27] thinner it tears the wires apart so you [24:30] can't have a problem with wires and [24:32] everybody's worrying about that but [24:34] that's like the sound barrier you may [24:37] have be all too young to have heard of [24:40] the sound barrier but there was once a [24:41] time when it was said that no airplane [24:43] can travel faster than the speed of [24:45] sound and the reason was of course that [24:48] airplanes were designed considering for [24:50] assuming that the air does is not [24:52] essentially not very compressible it's [24:55] not easy to squeeze it into a small [24:57] space because the airplanes went small [24:59] slow enough that there wasn't much force [25:01] to squeeze it into a small space so it [25:03] always expanded it was always not [25:04] squeezed and the theory and the analysis [25:07] and the experiments all dead with air [25:09] that was not [25:10] compressed it was then realized that if [25:12] you took into account the compression [25:14] that airplane wouldn't work with its [25:15] propeller and everything wouldn't work [25:17] the propeller go too fast and it [25:18] wouldn't pull so there was a barrier how [25:20] are we going to go fast in the speed of [25:22] s it just means there's no but there's [25:24] no law of physics that says that you [25:26] can't go FAS than the speed of s [25:28] it just requires a completely different [25:30] design so these limitations that you [25:32] hear about mean only that you can't go [25:35] on with the same design but it doesn't [25:38] mean that it's not possible in principle [25:40] and therefore I have considered the [25:42] question never mind about making it with [25:45] wires let's make it a different way make [25:48] computers a different way with atoms [25:50] with interactions with certain kinds of [25:52] connections how small can it be made can [25:54] it be made so that each one atom in it [25:58] the state of one atom tells you yes or [26:00] no instead of gold and silver and the [26:02] answer is of course that the laws of [26:04] physics that you have to use are very [26:05] different than the laws here they're [26:06] called quantum mechanical laws the scale [26:09] is so small we have to worry about the [26:10] uncertainty principle and everything [26:12] else but just let me tell you that it [26:14] turns out to be [26:15] possible according to the laws of [26:17] physics in principle to build a computer [26:20] in which each bit or each little piece [26:24] of information is one atom large and [26:28] there's no problem in that that was [26:30] something I had some fun working [26:33] out but I said that this talk was about [26:37] machines and of course a computer is a [26:40] kind of machine but the machine you [26:42] usually think of when you think of [26:43] machines as machines with movable [26:46] parts now let us talk about the [26:48] possibility of making machines of [26:49] movable [26:51] Parts which are very tiny the immediate [26:54] look in all the faces what [26:56] for [26:58] mental [27:00] entertainment maybe someday they'll find [27:02] a use for it okay how small can we make [27:06] machines it's just thinking for the fun [27:08] of it okay don't worry about it that [27:10] hasn't any application it doesn't cost [27:13] you anything not to have an [27:15] application it's just fun okay so we're [27:20] not going to worry about how we use [27:21] these dumb things we're just going to [27:23] try to make them how could we make [27:24] movable matter Little Machines tiny [27:26] machines that would operate that are [27:29] very very small you see today we have [27:31] great big hands and they're this fat and [27:34] lumpy and we can make tiny little [27:36] watches but that's nothing compared I [27:39] don't mean a computer watch cuz that's [27:41] not got moving Parts but a machine with [27:42] moving parts inside that is extremely [27:45] small how can we do that one way that [27:48] has been suggested is this you know that [27:51] the power the radioactive plants and so [27:54] on have to manipulate bottles and turn n [27:56] nuts and so so forth and they have slave [27:59] hands which are operated through [28:01] electrical connections with bigger hands [28:03] you move these levers and that controls [28:06] the hands on the other end of the wire [28:09] well there's no reason why the hands on [28:10] the other end of the wire that is [28:11] wrenches and whatever they are you know [28:13] they were operated from out here need to [28:14] be the same size as the thing so you [28:17] have the wires run to control very tiny [28:19] wrenches to make small things and what [28:21] do you make with that you make another [28:23] pair of hands that are still smaller so [28:26] that now you can Rec connect your wires [28:28] to a smaller set of hands in which you [28:30] work for a while manufacturing things at [28:32] a smaller scale and then with those [28:34] making things at a smaller scale and so [28:36] on of course when you got all through [28:38] with this you'd say h I've only got one [28:40] pair of hands to make one tiny little [28:41] machine but actually the right way to do [28:44] it naturally is that when you made the [28:46] first pair you make two [28:48] pair and then they make two pair and [28:50] they make two pair so you have really [28:52] thousands and thousands and thousands of [28:54] very tiny pairs all following the same [28:56] wires and manufacturing small devices [28:59] machinery and otherwise that's one way [29:01] that's been suggested I don't that was [29:04] probably not a very practical way people [29:06] had tried a little bit but it didn't go [29:07] very far but how about using this scheme [29:11] how it's all solid well why does it have [29:13] to be all solid why can't we do the same [29:16] layer scheme and put in a piece I want [29:20] going to go into the details of laying [29:22] down the different layers but it's the [29:23] same idea you have the silicon and then [29:25] you have some other substance perhaps [29:27] not silicon oxide this time let's say a [29:29] substance we call [29:31] Blue of this shape let's [29:36] say it's no harm at manufacturing [29:38] there's a piece of silicon in [29:42] here I guess I should fill out the blue [29:45] there's a piece of silicon in here and [29:47] perhaps uh just to make an interesting [29:50] silicon dioxide i c most Brown last time [29:53] so to keep the things more or less [29:55] similar other substances here yeah [29:57] insulators or [30:01] conductors and other layers let's [30:04] simplify it you can make all kinds of [30:07] things but something like that and then [30:08] there's the blue material here but the [30:11] blue material is a substance that's [30:13] easily dissolved the way but not the [30:15] brown if you dissolved the away that's a [30:18] loose [30:20] piece okay now these things are [30:23] electrical they have all kinds of wiring [30:24] you can bring electrical voltages and [30:26] stuff all over these devices by using [30:28] the old other technique so you can put [30:30] electrical voltages on this the sides [30:32] and move this back and forth with an [30:33] electrical charge so you can move things [30:36] around now we've never done anything [30:38] like that but it's possible to make [30:40] devices in which we have loose pieces [30:42] that's another way of that's a very [30:44] simple Elementary fact that we can move [30:45] one piece you say that's nothing yeah I [30:47] know it's nothing but with a little more [30:49] imagination you put a sequence of pieces [30:50] or a sequence of to and move it in [30:52] succession from place to place and carry [30:54] objects over the surface of the Silicon [30:56] like cars and so on it's kind of fun I [30:59] haven't got the time to follow the [31:01] amusing time I had working on all these [31:04] things but you can start out and find [31:06] out that you can move things about with [31:08] Electrical uh devices and so forth and [31:11] therefore the second way of [31:12] manufacturing small machines would be [31:14] the evaporating edge freeing pieces of [31:16] object [31:18] Etc you can move them with electrical [31:20] forces for example what would the uses [31:23] be I told you I wasn't going to consider [31:24] it but just for fun the first use of [31:26] this thing [31:27] would be to make an optical shutter you [31:31] could have this thing over a hole and [31:33] when you move it the hole opens they're [31:35] very very tiny and they're all over an [31:37] area that's a light control you can make [31:40] pictures because you have electrical [31:41] control of which ones you open and which [31:43] ones you shut you say we can always [31:45] already make pictures but we can make [31:47] very powerful pictures that is TV [31:49] pictures are projected only in a small [31:50] box a big projection is rather poor [31:53] because you can Shine the Light and it's [31:55] kind of too weak for the fluorescent [31:56] screen but you can shine a strong light [31:59] directly through this thing because [32:00] there nothing but shutters and open them [32:01] and close them in different places and [32:03] make pictures move and so on with [32:05] electrical control sir what I'm confused [32:08] about you said something can only be [32:10] magnified 2000 times optically with [32:14] Optics problem if you could reduce the [32:18] light as you're talking about small [32:21] shutters or small cameras would it be [32:22] possible then no you can't make the [32:25] light any smaller than the size of the w [32:27] length of light these pictures that I [32:28] showed you of the with an electron [32:31] microscope which can go to a smaller [32:32] scale I'm back here at the scale 2000 I [32:35] should have been more clear it's much [32:36] smaller than any machine has ever been [32:38] made let me assure you already very much [32:41] but this would still be at the scale of [32:42] light and it would open and close holes [32:44] that light could go through yes I not [32:47] talking about making it smaller than [32:49] that I mean we haven't done anything [32:50] like [32:51] that magne make Genera you could think [32:55] of doing things like that yes [32:57] uh the question is other uses of the [32:59] stuff I thought of a few things the [33:01] present way of manufacturing these [33:03] things and putting different substances [33:04] in different places is a long and [33:06] tedious thing the workmanship needed to [33:08] get those second pictures to be at the [33:10] right place relative to the first ones [33:11] and so on it's a hell of a job suppose [33:14] instead we could squirt the stuff where [33:16] we wanted it how by having a big flat [33:19] framework with a whole lot of nozzles [33:21] with movable holes with the chemicals [33:23] coming through with but you have to have [33:24] a vve you have to move something so we [33:26] make it out of this kind of stuff once [33:27] you see then we bring It Go and now we [33:30] have all of the stuff in the right place [33:32] then of course there's no limit on the [33:33] scale light is not involved anymore and [33:35] we could gradually make things smaller [33:37] so there's another direction another [33:40] thing is to imitate a large number of [33:43] important chemical reactions and [33:44] biological reaction uh we have membranes [33:47] with special enzymes or special [33:49] molecules in special places they act [33:51] like special catalysts they control [33:53] reactions very delicately the kind of [33:55] delicate control that's in the human [33:57] body or in any living thing and these [33:59] controls we beginning to understand how [34:02] they're made they're layers of membrane [34:05] in which these protein molecules have [34:06] been set in certain particular places [34:09] but the scale is so tiny compared to [34:11] where we are and we can't think to [34:13] manufacture them directly and put the [34:15] stuff exactly where we want we may be [34:17] Try by putting something and maybe it [34:19] lands in the right place and by some [34:20] technique or other we can make a [34:22] membrane or another about right but we [34:25] can't make all the variety and all the [34:27] different possibilities for different [34:29] types of reactions by simply placing the [34:31] enzymes into the layers where we want [34:34] them but we need tools we need tiny [34:36] tools to push things around to make [34:38] things go as a matter of fact if you [34:40] went to the full limit we wouldn't need [34:42] chemists anymore all that a chemist does [34:45] is rearrange make a new molecule with a [34:47] different arrangement of atoms so if you [34:49] went down to a small enough scale you [34:51] just put the atoms down where you want [34:53] them to make the [34:55] molecule no h Hocus Pocus with bottles [34:57] and colored [35:00] liquids actually that's really magical [35:02] how they do [35:04] that well to [35:10] uh other things have been thought of the [35:13] question of making piece of Machinery [35:14] not that's connected like this in a hole [35:17] but a free [35:19] swimming mechanical thing you see [35:22] naturally the first application for [35:24] things just like it was for computers [35:25] that would be most popular would be [35:27] games and the game this time is that you [35:29] have this little thing that goes you [35:31] have a control from the outside it's [35:32] electrically sensitive and you can send [35:34] it signals it has a kind of [35:37] Sword see and it fights [35:40] paramia you watch under the microscope [35:43] and you have this thing and your [35:44] controler and you're fighting the [35:47] paramas with your little device a friend [35:50] of mine named Al hi suggested something [35:53] else that's going to shock you and the [35:55] idea is to swallow follow the [36:01] surgeon the surgeon is a machine with [36:03] knives that's very tiny they can go down [36:05] around and be controlled and go into the [36:07] blood vessels and down in here and carve [36:09] away the plaque or whatever heck that [36:11] has to do and come okay you SW yes [36:14] fantastic you swallow the surgeon [36:17] yes [36:20] Fant well I'm a little late with my idea [36:23] but the problem is that I manufacture [36:26] this hi gave me the idea 20 years ago [36:28] and this is a rather in some respects an [36:30] old [36:32] talk but not in all respects because of [36:34] these new [36:36] pictures I would like just to finish [36:38] this off though to say that there's [36:39] quite a considerable amount of u things [36:42] that you have to think about if you try [36:44] to make machines smaller because it [36:46] isn't merely a matter like you say of [36:48] magnetically making the same thing [36:49] smaller you take an electric motor and [36:51] made every part smaller it won't [36:54] work various things that come out in in [36:56] different proportions you see like the [36:58] weight and the strength and the air [37:00] resistance and ordinary thing like [37:03] electrical resistance in the other scale [37:05] is enormously too high relative and [37:07] makes a serious problem but we know how [37:09] to get around that with superconductors [37:10] and again the trick other design when [37:14] some proportions get very enough [37:17] different so that it's impossible to use [37:19] the old design then you can use some [37:22] other thing that's bigger to make a new [37:23] design just like you did with jet planes [37:25] there's no real barrier [37:27] but I would like to explain the idea [37:29] that scale doesn't really work very well [37:32] by imagining a rather silly thing to try [37:34] to do is to make a very tiny internal [37:36] combustion engine very very tiny because [37:41] it works the internal combustion engine [37:43] works like this you have a cylinder and [37:45] a you know spark plug and the gas goes [37:48] in the spark plug goes off nothing wrong [37:50] with making a smaller spark plug that's [37:52] all right the gasoline explodes it gives [37:54] a lot of Heat and the heat expands the [37:55] gas pushes the the piston and then [37:57] something happens or other and there you [37:59] go okay now some of the heat of the [38:03] gasoline is lost by heating up the [38:06] walls of the cylinder you know that [38:09] because you have to cool the engine I [38:10] mean that gets warm but it's rather an [38:13] insufficient in not very significant [38:15] amount of the heat most of the heat is [38:18] still used to expand the gas now if we [38:20] made the thing let us say 10 times [38:24] smaller that's you still would work [38:26] actually but you get the idea from 10 [38:28] let's we want to make it 100 times [38:29] smaller you'll see it's absolutely [38:31] hopeless let's take 10 times smaller [38:34] then because the cylinder is 10 times [38:36] shorter 10 times narrow and 10 times [38:39] each way the volume in the cylinder the [38:42] amount of gas that you made is a [38:44] thousand times smaller and the heat [38:46] that's been generated is a thousand [38:47] times smaller but the surface of the [38:50] cylinder to where the heat loss is going [38:52] to go if after a little bit you see it's [38:55] only 100 times smaller it's Square in [38:58] and if you contract just two ways to [39:02] make the Surface smaller it's only 100 [39:03] times smaller so you have a thousand [39:05] times less heat for 100 times less [39:08] surface that is to say the surface [39:10] relative to the volume that you're [39:11] heating is 10 times more it leaks faster [39:14] 10 times faster into the casing and if [39:17] you try to think of making the machine [39:19] the 100 times smaller which is really [39:22] not a very tiny machine from the point [39:23] of view that I'm taking of E wheaty [39:27] Machin you see that it's hopeless as [39:31] soon as you would explode this [39:32] infinitesimal amount of gasoline in this [39:34] tiny tiny cavity then the heat of it [39:37] wouldn't be before it had any chance to [39:40] push any anything it would leak out [39:42] immediately into the Container so you [39:45] have to redesign turns out following the [39:49] laws of physics there's no thing in [39:51] against it in principle every time I [39:53] mention this somebody always pulls up [39:55] some difficulty about electrical [39:56] resistance or something like this the [39:58] most interesting one is when I talked [40:00] about the object that was sailing along [40:03] in the water [40:04] right flowing moving through fluids is [40:07] entirely different things like [40:08] resistance it turns out if you look in [40:10] proportion is different if I would to [40:12] take a very tiny thing in water and [40:15] expand it to our Dimensions the [40:18] analogous situation is that it's [40:20] swimming in extremely thick [40:23] honey and the ordinary ways of flapping [40:27] let's say a fish you want to make a tiny [40:28] fish of flapping the the fins and so on [40:32] to push it nicely through the Water by [40:33] the inertia of the water doesn't work [40:35] because they're just ick ick in the icky [40:38] stuff okay and if the fish is stuck in [40:41] the in the [40:44] goo but how do I know then that I can [40:47] make very tiny machines that it's [40:48] possible according to the laws of [40:50] physics because I'm a physicist and I [40:51] check the laws of physics and it's okay [40:53] I'm telling you but I got another way to [40:55] tell you [40:56] living things have already done it [40:59] bacteria swim they swim through the [41:01] water at the scale that corresponds the [41:04] thick goop and how do they do it it [41:07] would be analogous to the following you [41:09] have a kind of a cork screw a piece of [41:12] iron or something this is more magnifi [41:15] inside a sleeve like a leather sleeve or [41:17] a rubber sleeve and you simply turn this [41:20] thing okay so the cor screw turns inside [41:24] the sleeve you got to Picture This you [41:25] see the sleeve is there so this cor [41:28] screw is turning the surface is the same [41:30] and it screws its way through the fluid [41:33] the fella of a bacterium is built that [41:36] way it's a cork screw of a piece on the [41:39] inside which is solid which turn and the [41:41] motion that's produced by this is really [41:44] like working with a cork screw going [41:46] through the stuff you're pulling [41:49] yourself forward it's as if you were in [41:51] sufficiently tough like cork you want to [41:53] move through cork you take a cork screw [41:56] go and push you know of course the [41:58] Cork's got to give a little because it's [41:59] not cork it's slightly thick honey so [42:01] you keep going and pull yourself forward [42:04] the only problem is what happens to the [42:05] cork screw that's the trick of having [42:07] the solid thing inside of the sleeve so [42:09] that you can carry it along with you it [42:11] doesn't get lost in the back and that's [42:13] the way a bacteria moves by a completely [42:16] different scheme that scheme is utterly [42:18] inefficient in real water turning this C [42:20] screw would have very little effect [42:22] unless you turned it very fast to get [42:24] some uh inertial resistance which is [42:26] better at the flap of flap like the fins [42:28] of a fish you see so when you allow for [42:31] the fact that when you go to different [42:33] scales different things at different [42:36] proportions become important you have a [42:38] delightful time in redesigning all kinds [42:41] of familiar Machinery to see if you can [42:43] do it and give what take 25 30 years [42:49] there'll be some practical use for this [42:51] what it is I don't know and maybe in 20 [42:54] or 30 years I'll think about that and [42:56] tell you give you another lecture on how [42:58] these damn things could be used thanks [43:01] very [43:12] much yeah I can [43:16] answer [43:18] well okay questions [43:23] any do that you mentioned when you go [43:26] very small in scale the liquid becomes [43:30] very viscous does that imply that if you [43:32] could build a very large being it could [43:35] move through the air in the same way and [43:38] that the air would then become thick [43:40] enough to support [43:43] it if you up the scale you have big [43:47] airplanes it work fine yeah I know but [43:49] supposing you up to 10,000 times or [43:51] something with nothing much happens [43:53] because already the uh [43:56] the resistance of the air I mean it's a [43:59] dynamic resistance now the the ickiness [44:01] of air what we call a viscosity the [44:03] analog of the honey is long since very [44:06] unimportant the Motions that you produce [44:09] with air propellers and so on start a [44:11] flow which it is true that the last [44:14] motions as it dies out has something to [44:15] do with the viscosity but the initial [44:17] motions are the ones that are important [44:19] they're Dynamic the airplane goes by the [44:22] inertia of pushing air down not [44:25] everything to do with resistance you see [44:27] by thickness it's just the [44:30] mass that moves that's pushed down by [44:32] the wing as it goes through and The [44:34] Recoil from the air going down holds the [44:36] plane up and that would be just as good [44:39] for a larger plane however it is true [44:42] that if you took the same kind of an [44:43] airplane and simply increased it by 100 [44:47] times it won't [44:49] fly the reason this time is that weight [44:53] has increased as the cube and the [44:55] lifting power from the area of the Wings [44:58] only by the square you know the the same [45:00] thing backwards so you have to have it [45:02] Hollow notice our big aanes are Hollow [45:05] because the weight if the airplane [45:07] really had the corresponding weight as [45:09] you increase the size it would get more [45:10] and more difficult to make it [45:12] fly it goes like that I mean everything [45:15] that you think of there this changes and [45:18] that changes and there it's true that [45:19] are different scales different things [45:21] are important so at a 10,000 times as [45:24] big the Air does not become this no the [45:26] air doesn't become viscous no you're [45:27] going the wrong way viscous you got to [45:29] get effective viscosity go going for [45:31] smaller [45:33] Sky yes [45:35] sir well let's see if I can phas this [45:38] right you're an original thinker I would [45:41] like to ask [45:42] you how would you go about designing [45:46] maybe a miniature somewhat smaller the [45:49] grand [45:50] Ki an anti-gravity machine I [45:54] can't you could use for no I can't I [45:59] don't know how to make any anti-gravity [46:00] machine you would lick the most problem [46:03] it doesn't make any difference I still [46:05] don't know how [46:07] to uh what the game I play is a very [46:10] interesting one it's [46:13] imagination in a tight straight jacket [46:16] which is this that it has to agree with [46:19] the known laws of physics I'm not going [46:21] to assume that maybe the laws of physics [46:23] have changed then I can design something [46:25] but I try supposing it's everything that [46:27] we know is true as we think it is as if [46:30] we do if we're wrong of course we can [46:32] red design something with the new laws [46:34] later but the game is to try to figure [46:36] out with what we know what's possible so [46:39] it requires imagination to think of [46:41] what's possible then it requires an [46:43] analysis back a checking to see whether [46:47] it fits it's allowed according to what [46:49] is known okay and in the case of [46:52] anti-gravity machine I immediately give [46:54] up because my understanding of the laws [46:56] of gravity are such that I don't see any [46:59] way it doesn't make sense for [47:00] anti-gravity okay the only anti-gravity [47:03] machine that is things which oppose [47:04] gravity which are very effective are [47:06] things like you're using now a pillow or [47:09] a floor under your behind that is an [47:11] anti-gravity machine and will support [47:13] you in space above the earth a few feet [47:16] in this case for un resonably unlimited [47:20] time yes how absolute are the known laws [47:24] of physics [47:26] they always we find more things all the [47:29] time uh and there are things that are [47:31] not known so there's an edge that's [47:33] unknown and there's a certain amount [47:36] that of behavior that's been studied [47:38] over and over in a certain realm you see [47:41] there are variables you change like the [47:42] size or the dimensions or something like [47:45] that when you get too far [47:47] to too small if I start to talk about [47:49] distances that are a million time a [47:52] million time a million time a million [47:54] times smaller than a ctim then I don't [47:56] know what the right laws are okay but in [47:59] the realm of a few centimeters for [48:01] example in ordinary Behavior they're [48:02] pretty well gone over there doesn't seem [48:05] to be anything wrong with them and the [48:07] laws that I need for the scales that [48:09] I've been using down to Atomic scales [48:11] and so on are in extremely good control [48:14] it's not very likely that I'm making any [48:16] mistake [48:17] there okay in the realm that I've been [48:20] talking about in this lecture if I were [48:23] talking about much smaller than that [48:25] then I would be more more [48:29] humble yes got one from way back when [48:32] you early started that was way high I [48:37] know slow you know it's it's it's about [48:41] making things with with parts that are [48:43] one atom big or even five atoms big I [48:47] got I use 100 until you said you were [48:50] not going to break the laws of physics I [48:51] wasn't going to Heckle you about this [48:53] but I it it has been concerning me from [48:55] well now how you going to make these [48:57] atoms stand still or do they do that the [49:01] atoms don't stand still at an ordinary [49:02] temperature they're always [49:04] jiggling but if you take a substance [49:06] like gold or silver the forces between [49:08] them is rather strong and they stay next [49:11] to each other they just jiggle in place [49:13] atoms which jiggle in place are solids [49:16] things that are solid are made of atoms [49:18] which although they're jiggling they [49:20] never get out of place if you took one [49:22] away the others in the right place it [49:23] comes it pulls them back you see it's a [49:25] perpet [49:26] check with your friend you okay yes I'm [49:28] it's like the people Marching In A it's [49:31] it's like the high school band March [49:33] okay nobody really knows what to do [49:35] they're going like this and it's okay it [49:37] holds together okay but if there's too [49:41] much jiggling and too many people are [49:43] out you can imagine if this line gets [49:45] broken up because he looks at him and [49:47] he's far out so he tries to get out and [49:48] he gets and so all tumbles around each [49:51] other so if you make too much [49:53] temperature or use a substance where the [49:55] forces are too weak they roll over and [49:57] they lose the strength and that's a [49:59] liquid okay so I don't want to make this [50:02] out of liquid I wanted I said gold and [50:03] silver and I meant ordinary temperatures [50:06] and that's why I used those five atoms [50:08] in a lump because it's if I had used [50:10] only I mean 5x 5 by five was 100 atoms [50:13] if I used only about five atoms the [50:15] jiggling would from time to time by [50:17] accident let one atom move away I would [50:19] get lost and then another and another [50:22] and after some years I would have lost [50:24] my [50:26] the description of the human Race's [50:29] knowledge and I want the seed to last [50:31] longer so I figured five atoms and [50:33] calculated that the time for that is [50:35] billions of years before that gets [50:38] disturbed by these [50:40] jiggling sorry oh does the [50:43] lasers give you a chance to work in that [50:46] small no the lasers are light and they [50:49] are used in these devices because [50:51] they're accurate light they're very [50:53] helpful in controlling the light and [50:55] focusing and everything else but they [50:56] can't get any smaller than a scale of [50:58] about 2,000 reduction because the light [51:02] wavelength there's a wave of light is uh [51:04] that big I mean it's millions of an inch [51:07] but uh they can't get any smaller than [51:09] that scale light has a scale and you [51:11] can't it's like using a [51:16] uh Chopsticks right to pick up [51:19] something well it's not very good [51:21] because you can still make it with [51:22] Chopstick because they have sharp edges [51:23] so I haven't got a very good analog yet [51:26] but it's using some kind of blobs to [51:28] pick something up that are much bigger [51:29] than the thing you're trying to pick up [51:31] and you can't do it yes sir one thing [51:33] that's always fascinated me kind of [51:35] brought it up today is the fact that at [51:37] one point in time man didn't know what [51:39] would happen if we went faster than the [51:40] speed of sound yes obviously we designed [51:44] something we got there yeah well the [51:47] human eye can really only detect light [51:50] that's true when we know the speed of [51:51] light when it's using in the ordinary [51:53] way it also can detect the suck and the [51:55] jaw suck on the side of the head as a [51:57] flash too but the normal uses were like [52:00] now my question is does physic the law [52:03] of physics include the fact that uh [52:08] matter can actually go faster than the [52:10] speed of no now any object which is [52:13] going slow according to the present laws [52:15] of physics and relativity any object [52:17] which is now going less than the speed [52:18] of light cannot be accelerated beyond [52:21] the speed of light if you keep on [52:23] pushing it harder and harder using up [52:25] more more energy to give it energy it [52:27] goes faster and faster and faster but [52:30] what happens is it gets heavier so to [52:32] give it a little bit more speed it's [52:34] more inertia right so it's harder and [52:36] harder and you put more and more energy [52:38] and put more and more energy just gets [52:39] heavier and heavier and gets closer and [52:41] closer and cl but never quite gets to [52:44] the speed of light and certainly never [52:45] surpasses it in the machines in which we [52:48] accelerate electrons for instance and [52:50] you go around like in slack here or [52:52] wherever we get them going so that the [52:54] man we've put so much energy in that [52:57] they become they come to something like [52:59] one and I been doing this quick so my [53:01] number may be slightly off 160 [53:04] billionth that may be slightly of the [53:07] speed of light in other words it's not [53:08] the speed of light it's short by [53:12] about 160 millionth okay sorry 160 [53:16] millionth of the speed of light and uh [53:19] 1,600 milon never mind very small [53:21] fraction it's very close and the mass of [53:24] these things has gotten to 40,000 times [53:27] the mass they have when they're standing [53:28] still an electron becomes 40 20 times [53:32] heavier than a proton what about [53:35] antimatter when you're dealing with [53:37] antimatter wouldn't it appear and what I [53:40] bring the question up uh it's because [53:42] this a likely community that to bring [53:45] this up is the fact that in uh in the [53:48] spiritual World which a lot of [53:51] people delve in it's possible that [53:54] things exist exist and live beyond our [53:56] ability to see [53:57] them and if this be the case is it [54:01] possible that they are vibrating at [54:04] speeds which are greater than the speed [54:06] of light therefore have a different law [54:09] the only thing I can tell you are the [54:10] things that I can make observations [54:12] about if you ask me questions about [54:14] something that I cannot see with any [54:15] instruments whatsoever they can be doing [54:18] anything [54:19] whatever I have no way to tell I have no [54:22] you can have what you [54:24] want long as I can't see it with any [54:27] instrument okay as long as I can't see [54:30] with any instrument but if you give me [54:31] one clue way you see I can make some [54:34] tests and i' be able to answer your [54:35] question otherwise I'm I'm empty minded [54:38] I Su anti no we know all about [54:41] antimatter we can see it in the [54:42] laboratory we do experiments with it [54:44] there's nothing spiritual about [54:46] it I mean excuse me because the whole [54:50] world might be completely spiritual and [54:51] all that stuff that's fine but it's the [54:53] same kind of stuff as matter it's just [54:55] is available to experiment we know we [54:57] know a large amount about the laws of [54:59] the way it behaves does it vibrate is it [55:02] substance yes it's substance in the same [55:04] way that any other substance is I mean [55:06] normal matter is substance and whether [55:09] normal matter vibrates or not is [55:11] something we'd have to discuss what [55:12] circumstan of course the normal matter [55:14] vibrates when you hit a bell with a [55:15] hammer it [55:17] vibrates yes a small question my lady is [55:20] asking me constantly what makes the [55:23] world work [55:28] Isn't it nice that small questions are [55:30] so easy to [55:31] answer are you have problems with your [55:34] lady there was a your [55:37] problem not my [55:50] problem is there any other yes yes what [55:55] next if if this is the area you're now [55:57] doing res searching no no I I play [56:01] sometimes this I've been doing over [56:03] since about 1950 on and off you know I [56:06] played on a beach one day I began to [56:09] think of how small a space I need to [56:11] write a book and I thought the way to do [56:14] it in those days an easy way would be to [56:16] evaporate layers one after the other [56:19] across on a big area and that's easy to [56:20] do and so and I had five atoms again for [56:23] one bit and so forth and found that the [56:26] thickness I'd have to go is about 1 in [56:28] in order to write the whole information [56:30] in the book but it's a 1 in square and [56:33] every inch is the same so if I sliced [56:35] that into wires each one would be the [56:38] same and I have thousands and thousands [56:40] of wires which a copy of the same book [56:42] and I was absolutely amazed that a book [56:44] could be put on a wire 2 in long [56:47] nowadays that's relatively easy but at [56:50] the time I was amazed I went around to [56:51] all the people on the beach and said Hey [56:53] listen do you know it's possible [57:00] didn't sell any [57:02] bananas yes what would you say is the [57:05] most difficult problem I'm sorry I [57:07] didn't finish but I I know I I work on [57:10] and off and so it's not real work [57:12] sometimes I get interested in the small [57:13] computer or the machine or some question [57:17] about resistance or something excuse me [57:19] what is would you say is the most [57:21] difficult problem in all of physics that [57:22] you can [57:23] visualize [57:25] what you ready to work on [57:29] it we have a very large number of [57:32] problems in physics that are not this [57:34] kind of problem this is a a game [57:36] engineering problem because the sizes [57:38] I've used are well within the realm [57:40] where there's not much uncertainty and [57:42] there's technical problems of making it [57:44] which are interesting but not the most [57:46] important problem in physics if you mean [57:48] by physics fundamental physics or the [57:50] laws of physics then the more important [57:52] problems have to do with disc covering [57:55] the places where we don't know the laws [57:57] okay which is a detailed thing which is [57:59] outside of the subject would take me [58:00] some time to explain I will be happy to [58:02] do so at another occasion [58:05] sir okay see my group is a tough group [58:09] yes these are members yes um I get one [58:13] that's like Pistons you know and se's [58:15] done he'll think of something [58:17] else yes at the end at the end of your [58:20] book you mentioned your book character [58:24] of the [58:26] one autograph for me [58:31] oh you you mentioned that you thought [58:33] there would come a time when the [58:34] excitement of what was going on in [58:36] physics would would slow down because [58:39] things have been discovered and you your [58:40] analogy was tourist discovering an [58:43] otherwise beautiful area and it becomes [58:45] less interesting and less nice as more [58:47] people know that book was written 20 [58:49] years ago do you think that's happening [58:51] no the well it's true that experiments [58:54] are going more slow slowly and each [58:55] experiment is more difficult than the [58:56] one before and information is coming in [58:58] slower nevertheless I made a mistake I [59:01] thought therefore that the theorists [59:02] would have less to work on and therefore [59:05] would they wouldn't be quite as [59:07] interesting but uh I didn't realize a [59:10] propensity of human beings to speculate [59:12] and so as a matter of fact when less [59:14] data comes in more theorists are [59:18] working with more crazy theories than [59:21] before and I hadn't contemplated that [59:23] because my day things were more vigorous [59:26] when we got a theory we could check it [59:28] nowadays since the data comes in slower [59:30] we have much more fun making large [59:32] amounts I see everybody's getting ready [59:34] see them sneaking out [59:37] then so I think we should uh let's do it [59:40] this way let everybody who wants to go [59:42] go and if someone wants to stay I'll [59:45] answer some more questions for them but [59:47] those who have something to do cuz there [59:49] are other interesting activities in this [59:52] wonderful wonderful place of vessel [59:55] so forget I'll come back to it in a [59:57] minute but wait let's the others leave [59:59] who wish to leave and then we'll come [60:01] back to those who have some questions [60:13] yet oh my God so many [60:16] questions the front oh she had a [60:19] question I ought to answer hers first [60:21] well it's something about the time [60:22] that's coming when all the things going [60:24] to f together and we're going to have [60:25] all the answers and we won't fin we [60:28] don't know whether that's true it's [60:30] possible but we don't know if it's true [60:32] it's [60:33] possible it's see either thing is [60:36] possible either that we find all the [60:37] laws that is to say we find laws and as [60:39] we check and check and check everything [60:41] fits everything fits everything fits and [60:42] it looks like we're done and maybe we [60:44] would be or it's possible that the the [60:48] data comes in slower and slower like [60:49] when you fall asleep or something like [60:51] that not really it doesn't go slow but [60:53] uh people get tired of it there's too [60:55] many new things and they not have any [60:57] application and it's too far out and [61:00] it's too expensive and the high priests [61:01] were involved in it and not supported it [61:04] much and less and less people doing it [61:06] that's another way that the fundamental [61:08] physics can stop but Applied Physics [61:10] like discussions of matters of this kind [61:13] will go much longer than that because [61:15] there's all kinds of applications of the [61:17] laws we haven't thought of yet we're [61:19] talking about two things finding new [61:21] laws that is discovering new things [61:22] about nature and the other is finding [61:25] new ways to use what we know the second [61:28] one will go on much longer than the [61:31] first and I cannot Invision how it stops [61:34] but the first one I could imagine either [61:36] that we get the answers or that we don't [61:39] get the answers and we don't get the [61:41] answers but we're frantic All Forever [61:43] thousands of years or we don't get the [61:45] answers and we give up or we get the [61:48] answers that's all obvious I can't see [61:50] any other [61:52] possibilities it's obvious some other [61:54] possibilities is what's going to happen [61:56] yes you say as the laws of physics we [61:58] need talked about the heat engine and [62:00] because the cooling rate that's all at [62:02] the scale that we know the laws well [62:04] what if you wanted to make a heat engine [62:06] what other mechanism would you use [62:08] because you've got to follow some sort [62:09] of cycle yes but there are many cycles [62:12] of magnetism of I [62:15] magnetized the nuclear magnetism for [62:18] example in the atoms and so on there's [62:20] many other things the heat is at random [62:23] motion but the random motion what it [62:25] doesn't have to be the motion of [62:26] molecules bouncing like an ordinary heat [62:28] in a engine but it could be the heat of [62:31] spin that is magnets swishing up and [62:33] down although the atoms that the [62:34] correspond were staying in place and so [62:36] on so there are different scales at [62:39] which we can get the heat operating [62:41] systems and reversible Cycles if you [62:44] want but I use the internal combustion [62:46] engine to illustrate that a particular [62:48] type of engine internal or heat engine [62:51] is not necessarily the way to go we [62:53] might get a clue from living creatures [62:56] which do get their power from chemical [62:58] reactions and you might call them heat [63:00] engines or not they are in a way [63:03] but uh actually so it's hard to answer [63:07] exactly I don't know exactly what you [63:08] want but I don't need to follow any [63:11] particular type of engine you see I use [63:13] different combinations of the [63:15] laws yes sir I put you down once let's [63:19] see if we can do it [63:21] again I [63:23] wonder power to be in physics power to [63:27] be in [63:29] physics will a will put a new original [63:32] thinker to the state and burn him alive [63:36] because the Earth is [63:38] flat after [63:40] all well I think not but we perhaps I'll [63:44] be the guy that gets burnt because the [63:45] Earth is f I'll take the [63:48] risk [63:51] right I want to back to something [63:54] drawing it yes wasn't clear no it wasn't [63:57] very the one thing that I'm not clear [63:59] about is that that scale yes what is it [64:03] that physically makes light impossible [64:06] to be reduced any further [64:09] light a light is like a let us [64:14] consider ocean waves or something [64:17] okay and imagine a large number of corks [64:21] on the shore and I want to move Corks [64:24] And a small space and the only waves I'm [64:26] allowed to use are this long it's [64:29] impossible to figure out how to get [64:30] waves at that this normal wavelength to [64:33] focus any shorter than about this size [64:37] you can by bringing the waves from [64:39] different sides from a much bigger [64:41] Dimension bring things in focusing the [64:43] waves but you can't get them any shorter [64:46] than this characteristic length of the [64:48] weight and so the sloshing on the shore [64:50] will move corks over about this much [64:52] area it's impossible to pick one cork [64:55] out with waves that are this long no [64:57] matter how you combine them [65:01] okay well by taking different waves yes [65:04] yes oh yes indeed because now what [65:07] different waves are not called light we [65:08] call them ultraviolet light or x-rays [65:12] for instance x-rays are in fact exactly [65:14] that they're light but with a shorter [65:17] wavelength than normal light now in [65:19] principle you could do and people are [65:21] trying to do this kind of a thing in [65:23] which the xray affect the resist you're [65:26] absolutely right they have to go through [65:28] other technique the photographic [65:29] technique for making the pictures can't [65:31] be silver and so forth but you do it a [65:33] different way and there's a lot of work [65:35] and x-rays is in fact one way that is [65:39] being worked on in various Laboratories [65:40] to reduce the size of this yes that's a [65:43] good idea I'm sorry I answered your [65:44] question wrong before saying there was a [65:46] limitation because it's light I propos I [65:49] was assuming it was [65:50] light yeah but it could be [65:52] x-rays uh or ultraviolet light that's [65:56] going a little bit ultraviolet light [65:57] about half size okay but beyond that we [66:00] call it super far ultraviolet or X rays [66:04] then there's soft x-rays hard x-rays so [66:06] on the size is going down and down and [66:08] down and that size can get smaller than [66:10] an atom it can get way much much smaller [66:13] in size of a nucleus Z hard gamma rays [66:16] and so on it's all the same scale but [66:19] the other direction in scale of longer [66:21] waves is called red light infrared light [66:24] far infrared light heat waves radio wave [66:28] or television transmission waves and [66:30] radio waves there's all one enormous [66:32] scale from one end to the other which [66:34] from a physics point of view properly is [66:37] the same thing and should be called by [66:39] the same name if we call it light you [66:41] may get a mistake because you think you [66:43] mean visible light okay we call it [66:45] electromagnetic radiation which is a [66:47] great big word but it's supposed to [66:49] Encompass this this phenomenon which can [66:52] have a scale like the ocean waves but [66:55] depending on what the scale is the [66:56] normal name of the thing is different [66:59] it's radio waves when one day harmonic [67:02] yes harmonic oscillation slow for radio [67:05] weight faster for television weight [67:08] faster for heat WS faster for and very [67:11] fast for red light and even faster for [67:13] blue light and then ultraviolet and and [67:15] so on it's just the same damn thing [67:18] along a whole scale from one enormous [67:21] range okay yeah I'm confused about I [67:27] have such trouble with you you're always [67:30] confused how can you detect something [67:33] that's smaller than atams what can we [67:35] have to detect them with that takes a [67:38] great deal of [67:39] Ingenuity let me explain to you how you [67:42] might be able to do [67:44] that suppose that you wanted to detect [67:47] the something smaller than tennis balls [67:49] and all you have is tennis [67:52] balls okay now if you hit a tennis ball [67:55] against a tennis ball it's unlikely to [67:57] bounce directly back well excuse me [67:59] that's too complicated I got to better [68:02] suppose you throw tennis balls at a wall [68:05] of these things that you want to test if [68:07] they're smaller then by seeing that the [68:09] number how many tennis balls come [68:11] through and how many bounce back you can [68:13] see how big the targets [68:17] are if one tennis ball has to occupy an [68:20] area size of a tennis ball so if you if [68:23] you had one tennis ball on the wall and [68:25] you threw them at random at it the [68:27] number that would bounce back would be [68:28] depending on the size of that ball but [68:31] if that wasn't a ball but a pin you'd [68:33] have to have better [68:34] aim yeah something like that it's [68:36] something like that so we measure that's [68:38] one way of going down in size okay uh [68:42] another way of going down in size is to [68:44] discover some [68:46] relationships for example when the ways [68:48] of making light one of the ways of [68:50] making light is to let electrons hit [68:53] something with a certain energy hit a [68:55] metal as you increase the energy when [68:57] you can still measure this wavelength [68:59] cuz it's still bigger than an atom you [69:01] measure it and you find the wavelengths [69:03] getting shorter as you increase the [69:05] energy of the [69:07] Collision now you get to a point where [69:09] you can't measured directly anymore but [69:11] the the way you understand the laws you [69:13] just use more energy you'll get shorter [69:15] wavelengths and that's the way we go to [69:18] very much shorter wavelengths and then [69:19] after we make measurements everything [69:20] fits together and there's lots of [69:22] internal checks and so on it's not very [69:24] easy to describe in a few words smaller [69:27] than atom not smaller than an electron I [69:29] guess gu smaller than well it's hard to [69:32] say how big an electron is but it'll be [69:35] pretty damn much smaller than an atom in [69:37] the following sense an atom say is the [69:40] size of the room okay okay then the [69:43] nucleus of the [69:44] atom is a tiny Spectre that takes me a [69:47] minute 100,000 times smaller 100,000 [69:51] times smaller so you have to think [69:53] exactly it's it's a little tiny dot like [69:56] a period in a sent which is much smaller [69:59] and that we can measure too but today we [70:01] have by using this method of using high [70:04] energy collisions with the high energy [70:06] corresponds to short wavelength we can [70:08] go down to a thousandth of the size of a [70:11] nucleus of the smallest nucleus so we're [70:14] down on the insides of protons and [70:16] things at the moment that's what your [70:19] taxpayer money has been able to produce [70:22] yes ma'am can you use sound the way you [70:24] use light could those little arrows be [70:27] sound they could in principle but sound [70:30] wavelengths are very long on this scale [70:33] these are these little things and the [70:35] distances I'm drawing there are [70:36] millionths of an inch but aren't there [70:38] new kinds of sound ultrasound ultrasound [70:41] same sort of yes ultrasound is smaller [70:44] and shorter and you [70:45] could indeed your right yes you could [70:48] use this very super frequency sound not [70:50] ordinary sound not sounds you can hear [70:52] with the ear no no I don't [70:54] supersonic no [70:56] not yes you're right I made a mistake [70:58] I'm sorry yes not auditory sound but you [71:01] could do this with short I mean is there [71:05] any it's so easy with the light and to [71:08] get at these small distances it's rather [71:10] difficult with the sound because the [71:11] sound is a wave in matter of [71:14] motion and the matter is disorganized [71:16] the matter is shaking all the time and [71:18] it's hard to keep organized waves going [71:20] any distance so it's kind of strong [71:22] absorption when you get the such short [71:24] wavelength so it's but it's not a bad [71:26] idea as a matter of fact and something [71:28] that must ought to be considered more a [71:31] sound waves can be made short rather [71:33] easily and that's a nice idea I'm going [71:36] to think about it some more no I'm not [71:39] fooling because it's not so e see to get [71:42] to the dimension of x-rays is rather [71:44] complicated but to get sound that has [71:46] the size for the [71:47] x-rays is if I'm not mistaken not that [71:52] difficult although I'm not sure I have [71:54] to check the numbers and stuff okay but [71:55] it's not a bad [71:57] idea I think tomorrow I'll see it's a [72:00] bad idea but right [72:01] now what happens to sound as you [72:04] increase the frequencies and go up goes [72:06] up and Pitch no no I mean Way Beyond [72:08] that upper limits Way Beyond That Way [72:11] Beyond it it goes to the point where the [72:14] wavelength is the distance between the [72:15] atoms and the substance that's trying to [72:17] carry the sound and then the sound can't [72:19] be a wave anymore and then what does it [72:21] become it's no long you can't make [72:23] sounds of any higher [72:26] pitch there's [72:28] no there isn't any okay yeah it's [72:36] uh as though you let's see if I can get [72:38] an easy example you have a whole lot of [72:40] people standing in a row and you want to [72:43] make a wave this way you tell this one [72:46] to raise his hand and when the other one [72:47] sees that hand going up to move his hand [72:49] and so on that's the rule everybody [72:50] looks at the other and so you can make a [72:52] wave in which the hands go up in a [72:55] succession along the stream right and [72:58] there's a certain distance it depends on [73:00] how fast people respond to who's moving [73:02] their hands up you know what I mean [73:04] there's this this the Ser like the [73:05] Rockets only not they're all doing it at [73:08] the same time but you get this wave that [73:10] they like to make you know one one end [73:11] does the next next next next so you see [73:14] this wave of beautiful legs Rising okay [73:18] over a distance but it's perfectly clear [73:20] that if I would ask you how with those [73:22] rocket [73:24] can you make a wave where the distance [73:26] between the ups and the Downs is less [73:28] than the distance between the Rockets [73:30] how can I answer there is no such a way [73:32] you can make the ups and downs of the [73:34] legs you know this leg is up and that [73:36] one's down quicker than one rocket [73:39] difference and so you can't make a sound [73:41] wave where the vibration of the [73:43] successive atoms is in a pattern which [73:45] is what a sound wave is it's a pattern [73:47] like the rocket but which each atom next [73:50] to each other is any faster than [73:51] opposite there just isn't any way to [73:54] describe it so that's what I mean by [73:56] there's no more sound it doesn't mean [73:58] anything anymore [74:00] okay just a moment we'll get to you yes [74:03] sir you mentioned in the electron [74:05] accelerator say that it approaches a a [74:07] speed of light close it uh the mass [74:11] increases in the electron what is is it [74:14] still an electron or oh yes it's still [74:16] an electron just a electron moving very [74:19] close to speed of light and has the [74:20] correct properties for such an object [74:23] does it get heavier you have to about [74:27] relativity does the diamer [74:30] increase uh there isn't the particular [74:32] diameter for an electron so until you [74:34] define what you mean by the diameter of [74:36] electron I would have a discussion have [74:38] to make a discussion however if you take [74:39] an object that has an evident diameter [74:42] such as an atom a complete atom and [74:45] accelerate it very fast the diameter [74:47] stays the same in the direction at right [74:50] angles to the motion but the object [74:52] appears to be more and more flattened [74:53] like an ellipsoid until it becomes [74:56] almost like a disc as you get very close [74:58] to speed of light I'm sure you know this [75:00] about [75:02] relativity I not giving lectures to my [75:06] idiosyncratic thinking class on [75:08] relativity I just want everybody know [75:09] that as an advertisement for next year [75:12] we're not discussing but I can't keep [75:15] them down okay I can't keep that so we [75:18] have any other [75:19] questions yes sort of a interview type [75:22] question [75:24] what would you say is the most [75:25] exciting exciting thing if there is one [75:28] thing that you have discovered or been [75:30] thinking or playing with over your [75:34] career well that's an accident of [75:36] history of a particular individual that [75:37] has got a profound [75:42] sense um but nevertheless [75:47] um it's not very [75:49] easy to [75:52] answer [75:54] it's exciting in the sense that the [75:55] answer came quick [75:57] suddenly that's it's not the most [76:00] important Discovery but it's the one [76:01] that was fun because well there were two [76:04] like that so it's a little [76:09] hard well I was a guessing at the laws [76:12] for the disintegration what's called [76:14] beta Decay weak in Direction laws I [76:18] didn't guess them right but I guessed [76:20] them much wrer than they had been [76:21] guessed before and it was right for [76:23] quite a while and then it was improved [76:25] by another generation of physicists in a [76:27] more complete [76:29] form that was fun and the other time [76:31] I've been working for two years trying [76:33] to understand why liquid [76:36] helium liquid helium is a helium when [76:38] it's liquefy for at sufficiently low [76:41] temperature flows without any resistance [76:44] that is no more viscosity it just goes [76:47] if you have a long thin crack for [76:49] example if you take a glass rod and put [76:51] a piece of platinum why in it and seal [76:54] it together then when you lower the [76:55] temperature because of the change in [76:57] expansion there's an infinite desmal [76:58] crack between the glass and the Platinum [77:01] or then you can use this crack and [77:03] connect two vessels of helium one higher [77:06] than the other and they'll flow through [77:08] like this and they'll overshoot by [77:10] coasting and then go back the other way [77:12] there's just absolutely no resistance to [77:15] flow helium had also a large number of [77:17] other properties that were strange and [77:20] unusual but we thought we had the corre [77:23] direct laws of physics which are in the [77:24] form of what's called quantum mechanics [77:27] but exactly how the quantum mechanics [77:30] behaved the formulas would produce these [77:32] results was not at all clear because [77:35] there were so many atoms involved that [77:36] it was too hard to solve the equations [77:38] directly and it took a lot of [77:39] imagination to see how the thing really [77:41] worked and uh I got a Big Thrill of that [77:45] because uh that sort of came to me [77:47] suddenly I had the right idea and I gave [77:50] myself an argument which showed it was [77:53] wrong because I supposed a certain curve [77:54] was smooth and I suddenly remembered [77:57] that that curve isn't smooth because we [77:58] have another [78:00] experiment and then it sort of went bang [78:04] everything was okay cuz I was in trouble [78:05] see I had the idea and I had talked [78:07] myself I couldn't see what was the [78:09] matter because I had this wrong idea [78:11] about a certain curve from experiment [78:13] and I forgot and then I suddenly [78:15] remembered this fact which would have [78:17] saved the explanation so it was like [78:20] suddenly discovering it opened up [78:22] Suddenly [78:23] it's always fun when it's sudden and uh [78:27] of course it's very exciting and [78:28] delightful when something like that [78:30] happens in you work for several years [78:31] trying to understand something then all [78:33] of a sudden it becomes sudden rather all [78:36] of a sudden it's clear okay or it's not [78:39] completely clear but it's clear that you [78:42] got the key that the whole thing is now [78:44] pouring out I mean everything is just [78:45] roaring and now you can do all kinds of [78:47] things and look at the details and keep [78:49] on going and so on and hope that you [78:51] don't get stuck that you didn't make [78:53] another mistake which happens very often [78:55] I have been very often excited my wife [78:58] says all the time I have a lady too she [79:01] says I come down I've gotten it I got it [79:03] I got it you know I've heard it before [79:07] wait till tomorrow she's almost always [79:10] right almost always and again and again [79:13] and again wait till [79:17] tomorrow well I find out that would they [79:19] made a mistake you [79:21] know okay the Russians discovered like [79:25] tennis you know H the Russians have [79:28] already discovered it 200 years ago like [79:30] tennis like tennis right thank you Mr [79:37] President [79:40] okay [79:42] help help I was okay