'Squeezed' light exhibits smaller quantum fluctuations than no light at all. Localized squeezed regions have now been produced along an infrared light wave and probed with unprecedented time resolution. See Letter p.376
Heisenberg's uncertainty principle1, one of the pillars of quantum mechanics, states that certain pairs of physical quantities cannot be determined at the same time with arbitrarily high precision. For example, the velocity of an elementary particle becomes more uncertain the better its position is defined, and vice versa. Something similar happens to the oscillating electromagnetic fields that make up ordinary light. Instead of being sine waves that have perfectly defined amplitudes at each moment in time, Heisenberg's curse imposes a band of uncertainty around them. Although this band cannot be made thinner at all times, the amplitude of the field can be defined more precisely (a 'squeezed' band) in some parts of the wave, provided that it fluctuates more wildly (a 'swollen' band) in adjacent sections. On page 376, Riek et al.2 report the first direct observation of this effect by closely following the time evolution of such an uncertainty band to within single oscillations of an infrared light wave.
The electromagnetic field emitted by a stable laser source is the form of light that comes closest to exhibiting a perfect sine wave. However, even in this case, a constant level of quantum noise is always present when the amplitude of the field is being measured. These fluctuations are extremely small, and their effect is negligible for macroscopic light fields. But they become noticeable when very weak fields are used, and deleterious when measurements of extreme precision and sensitivity are attempted. Such a tiny band of quantum noise is present even in the absence of a net electromagnetic field — these 'vacuum fluctuations' limit the precision of highly sensitive measurement devices, such as gravitational-wave detectors3.
Fortunately, since the 1980s, physicists have learnt how to tame quantum noise4. Mostly thanks to the strong interaction of laser light with nonlinear crystals5, non-classical light states have been produced whose uncertainty bands develop alternating squeezed and swollen regions, instead of being equally wide over time. Studies of such phenomena have so far relied mostly on homodyne detection schemes6,7. In these, an intense light field overlaps with the one under investigation, and is used to extract information about the latter in a time-averaged fashion over many successive light-wave oscillations. However, if the pattern of noise is not repetitive and substantial variations exist between oscillations, a detection scheme of much higher temporal resolution is required. In particular, for accurate time sampling, a probe is needed whose duration is much shorter than the oscillation period being investigated.
Providing these is not easy. For visible light, a single oscillation lasts for little more than a femtosecond (10−15 s). Producing even shorter pulses to use as precise probes is currently impossible because, even though attosecond (10−18 s) pulses exist8, their generation cannot be controlled at the exceptional level of accuracy and stability required to detect and study quantum fluctuations.
Part of Riek and colleagues' solution to this problem was to move the investigated spectral region towards longer wavelengths. By shining a laser pulse into a nonlinear crystal, the authors produced a mid-infrared oscillating field and redistributed the vacuum fluctuations along different parts of the wave (Fig. 1). Because the oscillation period of mid-infrared light is more than ten times that of visible light, Riek et al. could then use another ultrashort laser pulse (of a now manageable duration of a few femtoseconds) as an extremely well-controlled probe of the 'slowly oscillating' mid-infrared field. Interactions of the two light waves in a second nonlinear crystal allowed the authors to measure the instantaneous amplitude of the slowly oscillating field by mapping it onto subtle changes in the ultrashort probe pulse.
Sampling the oscillating amplitude of a light field over time is a formidable feat in itself9, but in 2015, Riek and collaborators perfected the technique10. Thanks to new detection schemes and extraordinary control over all the possible sources of experimental noise and instabilities, the authors succeeded in accurately isolating the quantum fluctuations associated with such amplitude measurements. Now, by comparing the amount of quantum noise at different points on their mid-infrared light wave with vacuum fluctuations, Riek et al. are the first to directly follow the time evolution of a quantum uncertainty band, from squeezed to adjacent swollen regions.
The authors' fascinating results have twofold significance. First, they offer a unique time-domain perspective on the study of fundamental issues in quantum physics — for example, they relate changes in vacuum fluctuations to ultrafast variations in the speed of light in nonlinear media. Second, the research opens up exciting prospects for applications. Now that we know how to produce and precisely address them, such quiet moments in time could be exclusively selected for performing measurements (ranging from spectroscopic trace-gas detection to gravitational-wave interferometry) in a new regime of exceptional precision and sensitivity.