Ultrafast laser sources offer extremely brief bursts of energy and correspondingly high peak optical powers. For example, commercially available sources can exceed a megawatt of peak power and produce pulses of duration less than 10 femtoseconds. As a point of reference, the duration of a few-femtosecond pulse is roughly 16 orders of magnitude less than a minute – approximately the same as the ratio between the age of the universe (about 13 billion years) and one minute. These remarkable properties of ultrafast sources have made them extremely powerful tools – particularly in non-linear optics and metrology – as evidenced by ultrafast-pulse-related Nobel Prize awards in 1999 and 2005.


Much of the utility of ultrafast pulses stems from their status as some of the briefest events we can produce and observe. However, this distinction raises significant difficulties in the measurement and characterization of the pulses themselves. Basically, with no shorter-duration events to use for comparisons, ultrashort pulses cannot be directly measured in a traditional manner. The only practical strategy is to use indirect nonlinear measurements generated by using the pulse itself as a reference. An early (and intuitive) approach is to measure the pulse autocorrelation function by superimposing two temporally offset copies of a pulse on a comparatively slow detector. In some cases the extent of the measured autocorrelation can be used to infer pulse duration, although in other cases (such as when the pulse has a 'chirped' profile) such interpretations can lead to highly misleading conclusions. Over the years a wide range of improved ultrafast measurement techniques have been proposed and vigorously (sometimes contentiously) debated. As the field has matured, measurement capabilities have been expanded and subtle but important experimental issues have been identified and addressed.


The paper considered here is the latest offering from a pioneering group in the field of ultrafast metrology. The method they describe exemplifies many of the best features of modern pulse measurement techniques: an elegant spectroscopic measurement scheme; robustness to experimental uncertainties; efficient and intuitive signal processing algorithms; high resolution measurements of both spatial and temporal field behavior; and the capacity for self-consistency checks between the measured data and the recovered field profile. The underlying measurement strategy is known as Spatial Encoded Arrangement for Temporal Analysis by Dispersing a Pair of Light E-fields (SEA TADPOLE). In this technique a spatially-varying ultrafast beam is measured via: 1) comparison to a previously characterized ultrafast reference; 2) spatial sampling through scanning of a collection fiber; and 3) interferometric spatial-spectral measurements. In this article the authors address small drifts in the optical properties of the instrumentation that can lead to significant errors in certain aspects of the data – a problem common to many ultrafast measurement techniques. By taking additional measurements and computationally enforcing known physical constraints, these errors can be estimated and compensated. The additional measurements effectively introduce a degree of data redundancy leading to a well-posed field estimation problem. The authors further exploit this information by showing the self-consistency of their results. By comparing measured and predicted data at different sensor planes, they demonstrate the ability to validate the measured pulse profile. This work convincingly describes a practical method for reliably measuring ultrafast fields at a femtosecond timescale and at a sub-micron spatial resolution. This is clearly an important and impressive capability.

---

---

 Add Comment

You must log in to add comments.