IELTS Academic Reading Practice Test 147

Light-emitting diodes, or LEDs, are now so ordinary that they rarely attract attention. They sit behind the power button of a television, illuminate the screen of a mobile phone and, as bright white bulbs, light homes and offices. Yet the technology has a history stretching back more than a century, and for most of that time it was regarded as a scientific curiosity of little practical value.

The first report of light from a semiconductor came in 1907 from Henry Joseph Round, an experimenter working for Marconi on radio detectors. Round was studying crystals of silicon carbide, or carborundum, then used in the 'cat's whisker' receivers of early radio sets. Passing a current through one of them, he noticed a faint yellow glow at a contact and published a short note about the effect. The note attracted almost no attention and was forgotten for decades.

Similar effects were observed independently in the 1920s, when Bernhard Gudden and Robert Pohl studied light emitted by zinc sulphide. In 1936 a French researcher, Georges Destriau, obtained a glow by passing a current through zinc sulphide powder suspended in oil, and the phenomenon acquired the name electroluminescence. None of these devices was useful: they required high voltages, produced very little light and gave no hint of how the effect might be made efficient.

An explanation had to wait for a proper understanding of semiconductors. In a crystal, electrons occupy permitted bands of energy, and a considerable gap separates the band in which they normally sit from the band in which they can move freely. When an electron crosses that gap it releases the difference as a photon. The width of the gap therefore fixes the colour, since a wide gap yields energetic blue photons and a narrow one low-energy red photons.

The material that first worked was gallium arsenide. In 1955 Rubin Braunstein, working at the Radio Corporation of America, reported that diodes made from this compound emitted infrared radiation when a current passed through them. Others soon obtained visible red light, but only faintly and only when the device was cooled with liquid nitrogen, an obvious obstacle to commercial use.

The breakthrough is usually credited to Nick Holonyak Jr, then working at General Electric in Syracuse, New York. In 1962 he demonstrated a diode made from gallium arsenide phosphide that glowed visibly red at room temperature, and he has often been described as the father of the light-emitting diode. General Electric announced that the new lamps would eventually replace the ordinary bulb; in reality the first devices were dim, fragile and expensive, and they served only as indicator lights.

Commercial production began in 1968, when Monsanto and Hewlett-Packard introduced red LEDs and put them into calculators. The early devices gave perhaps a thousandth of the light of a small incandescent bulb and cost far more, but they lasted almost indefinitely. A researcher at Monsanto, M. George Craford, produced the first yellow LED in 1972, while the emerging market for optical fibres drove the search for materials that emitted at longer wavelengths.

One colour remained out of reach. Green and blue light require a wide energy gap, and the obvious candidate, gallium nitride, was extraordinarily difficult to handle. Its crystals were full of defects and, more seriously, nobody could make the p-type version, the half of a diode that supplies positive charge carriers, without which no junction and therefore no light can exist. Experiments with zinc selenide produced blue light in the laboratory, but in devices that faded within hours.

The obstacle was overcome in Japan. At Nagoya University, Isamu Akasaki and Hiroshi Amano grew gallium nitride of far better quality than anyone had achieved before and in 1989 created the missing p-type material by bombarding it with electrons. Meanwhile Shuji Nakamura, at the small chemical company Nichia, devised different growth methods and in 1993 produced a blue LED bright enough to sell. The three shared the Nobel Prize in Physics in 2014.

Blue light transformed the technology. By the mid-1990s a white LED had been produced by coating a blue emitter with a phosphor that turns part of the blue light into a broad yellow glow. The same materials gave rise to the blue laser, which allowed far more data to be stored on an optical disc. Efficiency improved at a rate few had predicted: the earliest visible LEDs turned well under one per cent of their electrical input into light, whereas the best modern white devices convert more than half.

Since lighting is often estimated to consume around a fifth of the world's electricity, the consequences of that gain are hard to overstate. The faint yellow glow that Henry Round noticed in a crystal of carborundum has become the light by which much of the world now works and reads.

A fuel cell generates electricity by combining hydrogen with oxygen, without burning anything. The principle has been known since the nineteenth century, but only in the past few decades have cells become cheap enough and durable enough to sell. A fuel cell is not a battery. A battery holds a fixed store of energy and must be recharged, whereas a fuel cell produces electricity continuously for as long as fuel and air are supplied. Its attractions are that it is quiet, that it has no moving parts in the stack itself, and that its exhaust is water.

The most widely used design is the proton-exchange membrane cell. Two porous electrodes, the anode and the cathode, are separated by a thin polymer film. This film is the electrolyte, and its behaviour explains the whole device. It conducts protons — hydrogen ions carrying a positive charge — but it blocks electrons and it blocks the two gases. Hydrogen arriving at one side and oxygen at the other therefore cannot meet and burn; the only route between them is through the membrane for protons and around the outside for electrons.

On the anode side, hydrogen flows over a layer that contains platinum. The metal acts as a catalyst, splitting each hydrogen molecule into two protons and two electrons. The protons pass through the membrane towards the cathode. The electrons cannot, so they are forced to travel through an external circuit, and it is this flow of electrons that does useful work, lighting a lamp or turning a motor, before they reach the other electrode.

At the cathode, oxygen from the air combines with the protons that have crossed the membrane and with the electrons returning through the circuit. The product is water. The only exhaust from the cell is therefore water, together with heat, and the reaction is remarkably clean. Cells of this kind operate at about 80°C, far cooler than a combustion engine, which is why they start quickly from cold and need no heavy insulation. If water is not removed, it can flood the electrode and prevent oxygen from reaching the catalyst.

Fuel cells have found their first markets where batteries are awkward. Several cities operate fuel-cell buses, and manufacturers have built lorries and delivery vans for routes that return to a depot each night. Forklift trucks in warehouses and factories use them because refuelling takes only a few minutes and the only emission is water, which is acceptable indoors. The distance covered between refills can be greater than that of a battery vehicle of similar size. Fuel cells also provide backup power for buildings, mobile phone masts and data centres, where a supply that can run for hours without a noisy generator is valuable.

Most hydrogen is not made cleanly. The great majority is produced by steam methane reforming, in which natural gas reacts with steam at high temperature and pressure, releasing carbon dioxide. Hydrogen can also be made by electrolysis, the splitting of water by an electric current. Electrolysis is clean only when the electricity comes from a low-carbon source, and it is at present more expensive than the gas route, so green hydrogen remains a small share of total production.

Storing hydrogen is harder than storing petrol. It is the lightest of all gases, so a given mass occupies a large volume. To fit a useful amount into a vehicle, it is compressed to pressures of around 700 bar or cooled to minus 253°C, at which point it becomes liquid. Compression and liquefaction both consume energy, and the tanks that hold high-pressure gas are heavy, being wound from carbon fibre. Liquid hydrogen must be kept in an insulated tank, and some of it boils off even then, so the fuel cannot simply be left for months.

Two obstacles have slowed adoption: cost and durability. Platinum is scarce and expensive, and a fuel-cell stack needs more of it than a catalytic converter does. The membrane and the electrodes also degrade, and a stack may lose a noticeable share of its power after a few thousand hours of use. Hydrogen from electrolysis is costly, and the network of refuelling stations is sparse. A vehicle that cannot be refuelled is not much use, however clean its exhaust. The cost of the stack has fallen as production has risen, but it remains a barrier for private cars.

Research is therefore directed at several targets. One is to reduce the amount of platinum, by spreading it in thinner layers, alloying it with cheaper metals, or replacing it with catalysts based on iron and nitrogen. Another is to raise the operating temperature so that the cell tolerates impurities in the hydrogen. Recycling of platinum from old stacks, better membranes and cheaper methods of compression are all being studied. Solid oxide cells, which work at much higher temperatures, are being developed for stationary power rather than vehicles.

Fuel cells are unlikely to replace batteries everywhere, and the two technologies are better seen as partners. Batteries suit small cars and short journeys, where their efficiency is hard to beat; hydrogen suits heavy vehicles, long distances and sites where a grid connection is impossible. The eventual size of the hydrogen economy will depend less on the elegance of the chemistry than on the unglamorous questions of cost, storage and infrastructure. For all the enthusiasm, the number of hydrogen vehicles on the road remains small compared with the number powered by batteries.

In March 2018 three researchers at the Massachusetts Institute of Technology published an analysis of every rumour tracked on Twitter over eleven years. Soroush Vosoughi, Deb Roy and Sinan Aral examined about 126,000 stories, some true and some false, and found that the false ones spread further, faster and more widely. A false story reached 1,500 people in about a sixth of the time a true one took.

The result was widely reported as proof that the internet had broken the truth. The authors offered a more careful explanation. False stories were more novel, and novelty provokes the surprise and disgust that persuade people to pass a message on, whereas accurate reports of events tend to be sober. The researchers also checked whether automated accounts were to blame and concluded that they were not: humans, not bots, did most of the spreading. That matters, because it locates the problem in ordinary behaviour rather than in a conspiracy of machines.

The obvious remedy is to correct false claims. Here psychology offers a warning. Repeated statements feel truer than unfamiliar ones, a phenomenon known as the illusory truth effect, and it works even when the listener knows better. In one experiment Gordon Pennycook and his colleagues showed that a single exposure to a fake headline raised the chance that people would later judge it accurate. A correction that dwells on the false claim may therefore leave the reader with a stronger memory of the claim than of the refutation.

Practical guides have drawn the lesson. The Debunking Handbook, first published by Stephan Lewandowsky and John Cook and since revised with other authors, advises fact-checkers to lead with the correct fact, mention the myth only once, explain the technique by which it misleads, and offer an alternative account that fills the gap the myth occupied. This advice rests on a considerable body of experimental work, and newsrooms in many countries have adopted it.

Does this mean that corrections are futile? For a while that was the fashionable conclusion, largely because of a 2010 paper by Brendan Nyhan and Jason Reifler which found that corrections could sometimes strengthen a false belief, especially among people already committed to it. The backfire effect became famous. Later work has been less alarming. A set of experiments by Thomas Wood and Ethan Porter, published in 2019, tested corrections on 52 different issues and found that they generally moved beliefs in the right direction, and a meta-analysis by Nathan Walter and Sheila Murphy reached a similar conclusion, while noting that the effects tend to fade.

If people accept corrections when they read them, why does misinformation persist? Partly because much of it is never read carefully in the first place. Pennycook and David Rand have argued that sharing is driven less by ideology than by inattention. In their experiments, simply asking participants to judge whether a headline was accurate before deciding whether to share it improved the quality of what they chose. A small pause, in other words, may be a more promising intervention than a lecture.

A second line of work tries to protect people before they meet a falsehood. Inoculation, or prebunking, exposes audiences to a weakened example of a manipulation technique so that they recognise it later. Sander van der Linden, Jon Roozenbeek and others have shown that short videos, and a simple online game in which players build their own fake news empire, improve people's ability to spot manipulation. Such campaigns can be delivered to millions at modest cost, which makes them attractive to public bodies.

Platforms, meanwhile, have tried changes to design. Warning labels attached to disputed posts have modest effects; more important may be reducing the algorithmic amplification of material that is unverified but emotionally arousing, because the reward structure of a network determines what spreads. Community Notes, the crowd-sourced system used on X, displays a note only when users with different views have found it helpful, and studies suggest that this bridging requirement produces notes acceptable to a wide range of readers.

There is a serious objection to all of this, which is that somebody has to decide what counts as false. Centralised removal of content places enormous power in the hands of platforms or governments, both of which make mistakes, and the machinery can be turned against legitimate criticism or awkward journalism, a form of censorship that no company is well placed to police. My own view is that the least dangerous measures are those that change the environment in which people decide: friction before sharing, ranking that favours accuracy, and education that teaches the tricks of manipulation, rather than the adjudication of individual claims by an authority.

None of these measures will produce a dramatic victory. Misinformation is not a single problem with a single cause, and the technologies that generate it, including systems that now write plausible text and fabricate convincing images, are improving faster than the defences. But a modest reduction in careless sharing, repeated across millions of decisions, would be worth a great deal, and so would a habit of doubting claims that happen to be flattering. The aim should be to make the truth slightly easier to find rather than to declare it.