Nanolytics Instruments' MWA detector features high radial and wavelength resolution. While similar specifications apply to other existing prototypes, its speed for data acquisition is unmatched, due to sophisticated software solutions.
When redesigning the detector arm, we entrusted an external photonics expert with computer modelling of the entire beamline. Optics were optimized, based on these calculations, to project a 10 µm event range in the measurement cell onto the 20 µm entrance slit of the spectrometer. Thus, a real optical resolution of 10 µm is achieved in the radial domain.
However, it cannot be circumvented that the optics is lens based, and that no achromats can be used as this would degrade UV intensity significantly. Chromatic aberration dictates that optics cannot be optimized for the entire UV-Vis range. Thus, optics are optimized for a reasonable average of 400 nm. This does not reduce optical resolution, but it causes different ranges of the spectrum to reach the spectrometer with individual quantum efficiency. Outer ranges in the spectrum will be registered with lower intensity than emitted by the Xenon light source. This does not represent a problem, as the Xe spectrum is very ragged by nature, and reference intensites are measured for calculation of absorbance.
Radial resolution is opposed by the depth of field inside the measurement cell. The beam does not enter the cell parallel, but gets narrower until it is focussed in the 2/3 plane of the cell. Afterwards, it diverges again until it is focussed by the ocular lens onto the spectrometer slit. Thus, the light will experience a slightly hourglass shaped path through the cell. This should be taken into account when relying on a real 10 µm resolution. It is still subject to research to determine what impact this error might have on the final results. However, this limitation may be considered state of the art and is not circumventable in any existing MWA design for the time being.
The Ocean Optics USB 2000+ spectrometer used in our MWA optics features 2048 channels for a wavelength range from 185 to 800 nm. This yields approximately 3 wavelengths per nanometer. Though this resolution is probably higher than the grating's precision, it does contribute to data redundancy and quality.
While greater ranges of this data may be discarded when no light is absorbed, a considerable number of datasets are available in the range of specific absorbance maxima. Also, it might be considered to average a number of channels e.g. to integer wavelengths and use the high number of datasets to reduce experimental noise.
For positioning the spectrometer slide, we apply a Zaber linear actuator with a positioning precision of 0.05 µm. The motor can move at a maximum of 8 mm/s. While scanning, the motor will be set into continuous motion rather than starting and stopping at every desired position. This reduces stress on the motor and increases scanning speed by almost one order of magnitude. More details on scanning modes are described below.
Light source and operation frequency
One of the fundamental differences in design in comparison to commercial AUCs is that in the MWA, the Xe flash lamp is located outside the vacuum chamber and does not form the vacuum seal. The disadvantage of the need for frequent lamp cleaning is eliminated.
The lamp is located in a sealed module with a parabolic mirror, providing a maximum yield of light transmitted into the fiber. Lamp intensity can be adjusted conveniently with a dial on the Omega Device. With a maximum flash frequency of 530 Hz, it is faster than the spectrometer. The duration of a pulse being approximately 700 ns, it will cover an angle range of 0.5 degrees at a maximum angular velocity of 60,000 rpm, short enough to strobe a single cell sector.
Single shot, multishot, hyposhot modes
In contrast to the prototype published by Coelfen, Nanolytics Instruments' MWA system provides enough lamp intensity (even at throttled lamp power) to saturate the spectrometer with a single flash. Consequently, there is no need to record more than one flash at a given radial position in order to collect a sufficient amount of intensity. This fact paved the way to a fundamentally different mode of operation: Rather than stopping and moving the spectrometer assembly hundreds of times during a scan, the motor is brought into continuous movement, while the spectrometer and light source are operating at maximum flash frequency. Thus, Nanolytics Instruments' MWA is operating at the maximal speed physically possible: one flash with every rotor revolution reduces the time required for a scan to seconds.
In single shot mode, one flash occurs at every rotor revolution. At 3000 rpm, the duration of one revolution is 20 ms, at 15000 rpm, it requires 4 ms. This means that for collecting 1000 datapoints (as for the range from 6 to 7 cm at 10 µm resolution), the system will require 20 seconds, or 4 seconds at 15000 rpm, respectively. Scanning could not possibly be faster, as every rotor revolution is used for a shot.
In hyposhot mode, every other or more revolution is skipped. This is necessary when rotor frequency exceeds the maximum spectrometer frequency. The spectrometer is specified to work with a maximum frequency of 333 Hz. To be on the safe side, we have limited the maximum frequency to 250 Hz. This corresponds to an angular velocity of 15,000 rpm. At a higher rate, e.g. 16,000 rpm, every second revolution will be skipped, and the spectrometer will work at a frequency of 133 Hz. This means that data acquisition will take longer at this speed (7.5 seconds) than at the lower speed of 15,000 rpm, again increasing with higher angular velocities. However, at yet higher speeds, no scan will take longer than 8 seconds.
At 30,000 and 45,000 rpm, the system will enter into higher hyposhot modes automatically.
Multishot mode has been implemented for peak performance of calibration procedures. For delay calibration, it is necessary to collect intensity data for the entire rotor at a fine angle resolution. For this task, the maximum frequency of the spectrometer is applied regardless of angular velocity. Thus, more than one shot is recorded for one revolution at speeds smaller than 15,000 rpm. Flash frequency is fine adjusted to match prime angles, and the data is resorted after 360 shots have been made. Thus, delay calibration at 1 degree resolution will always consume no more than 1.5 seconds. In practice, five series will be recorded, yielding 0.2 degrees resolution in 1800 datapoints.
The speed of measurement is unmatched and to be exceeded only with faster spectrometers. Please note that scanning times given above apply to the pure experiment of scanning a sector. An overhead of several seconds for programming the microcomputer and ca. 10 seconds (depending on resolution) for transferring scan data from the spectrometer to the computer will add to this time. Eventually, at most inconvenient conditions (high data resolution at lowest angular velocity), scanning two sectors of a cell will consume approximately 90 seconds, including the overhead. Delay calibration at 0.2 degrees resolution requires approximately 30 seconds.
The high speed of data acquisition is possible because the system is designed to require only one flash at each radial position for sufficient intensity, enabling continuous motor feed motion and exploiting every rotor revolution. The drawback is, as light flashes are not identical in intensity, a relatively high noise level. Other MWA designs require integration of multiple flashes, taking much longer but automatically reducing noise by averaging. However, the user has the liberty to average replicates at his own convenience, balancing his priority for speed of measurement with the signal/noise ratio. In any case, the noise is purely statistical and likely to be removed by subsequent fitting.
In summary, Nanolytics Instruments' MWA system is at least threefold faster than state of the art commercial instruments (recording only one wavelength rather than a complete spectrum) and ca. 8 times faster than other MWA prototypes.