Article pubs.acs.org/JPCA
Multimass Velocity-Map Imaging with the Pixel Imaging Mass Spectrometry (PImMS) Sensor: An Ultra-Fast Event-Triggered Camera for Particle Imaging Andrew T. Clark,† Jamie P. Crooks,† Iain Sedgwick,† Renato Turchetta,† Jason W. L. Lee,‡ Jaya John John,∥ Edward S. Wilman,‡ Laura Hill,∥ Edward Halford,§ Craig S. Slater,§ Benjamin Winter,§ Wei Hao Yuen,§ Sara H. Gardiner,‡ M. Laura Lipciuc,‡ Mark Brouard,§ Andrei Nomerotski,∥ and Claire Vallance*,‡ †
Rutherford Appleton Laboratory, Chilton, Didcot OX11 0QX, U.K. Department of Chemistry, University of Oxford, Chemistry Research Laboratory, 12 Mansfield Road, Oxford OX1 3TA, U.K. § Department of Chemistry, University of Oxford, Physical and Theoretical Chemistry Laboratory, South Parks Road, Oxford OX1 3QZ, U.K. ∥ Department of Physics, University of Oxford, Denys Wilkinson Building, Keble Road, Oxford OX1 3RH, U.K. ‡
S Supporting Information *
ABSTRACT: We present the first multimass velocity-map imaging data acquired using a new ultrafast camera designed for time-resolved particle imaging. The PImMS (Pixel Imaging Mass Spectrometry) sensor allows particle events to be imaged with time resolution as high as 25 ns over data acquisition times of more than 100 μs. In photofragment imaging studies, this allows velocity-map images to be acquired for multiple fragment masses on each time-of-flight cycle. We describe the sensor architecture and present bench-testing data and multimass velocity-map images for photofragments formed in the UV photolysis of two test molecules: Br2 and N,N-dimethylformamide.
1. INTRODUCTION
useful probe of bond strengths and internal excitation as well as a rapid means to distinguish between parent and daughter ions. Imaging studies on larger molecules present a different set of challenges to those on small molecules. The denser energy level structure of larger fragments means that resolving individual quantum states in the radial structure of the images is rare, and thus achieving the ultimate in velocity resolution is less important. However, such molecules often have multiple fragmentation pathways and understanding the competition between these pathways is an important aspect of probing the fragmentation dynamics. In a small-molecule experiment, it is usually sufficient to image a single fragment in order to obtain a complete picture of the fragmentation dynamics. However, for larger molecules, it becomes highly desirable to image multiple fragments on each time-of-flight cycle. The detectors used in most VMI experiments consist of a pair of microchannel plates (MCPs), which convert incoming ions into electron bursts, followed by a fast phosphor screen that creates an optical image of the electrons. The image on the phosphor is captured using a charged-coupled device (CCD)
1,2
Velocity-map imaging has been used with great success in the field of small-molecule reaction dynamics to study molecular photofragmentation events and other processes. The velocity distributions of fragment ions are highly sensitive to the detailed dynamics of the dissociation and yield information on the identity of the potential energy surface(s) involved, transition state geometries, bond strengths, and any product internal excitation. In recent years, the size of the molecular systems studied by velocity-map imaging (VMI) has steadily increased, and it has been shown that even relatively large molecules often yield structured, and therefore informationrich, images.3 VMI and related techniques are now routinely used to study the fragmentation of small to medium sized organic molecules in the gas phase, and in the longer term, such an approach has potential applications in mass spectrometric fragmentation studies. Tandem mass spectrometry (or MS/ MS) is becoming increasingly important in the study of biological molecules in the gas phase,4 and the ability to image the fragments as they fly apart from each other has the potential to add a new dimension to such studies. The velocity distributions of the fragments contain information on the energetics of the fragmentation process and could provide a © 2012 American Chemical Society
Received: October 5, 2012 Revised: October 22, 2012 Published: October 26, 2012 10897
dx.doi.org/10.1021/jp309860t | J. Phys. Chem. A 2012, 116, 10897−10903
The Journal of Physical Chemistry A
Article
Figure 1. (a) Schematic of the PImMS pixel, omitting trim DACs for simplicity; (b) photograph of the prototype PImMS sensor chip.
An alternative approach to time-resolved imaging measurements has been proposed by Zajfman and co-workers,11 who employ a pair of time-gated CCD cameras to determine accurate arrival times for ions arriving within the decay lifetime of the phosphor in a standard MCP-phosphor imaging detector. This approach has considerable appeal, in that precise arrival times may be recorded for an essentially unlimited number of ions. However, the dynamic range of the measurement is determined by the phosphor lifetime. Extremely high precision (on the nanosecond time scale or better) may be achieved in the timing measurement using a fast phosphor with a decay lifetime of ∼100 ns, but when using a much slower phosphor matched to the arrival time distribution of ions of multiple masses, the timing resolution is likely to be insufficient to resolve images for individual masses. Recent advances in CCD and complementary metal−oxide− semiconductor (CMOS) technology have opened up the possibility of developing integrated cameras that effectively incorporate a delay line into each pixel. This removes the requirement for low ion count rates and allows multimass imaging to be carried out using popular nanosecond laser systems operating at repetition rates of tens of Hertz. We have shown previously12 that a framing CCD camera can be used to perform multimass VMI. However, while such cameras offer excellent time resolution, down to around 10 ns, the number of images that may be acquired is limited by the number of memory registers on the CCD chip, and the ion arrival times must be known in advance in order to program the exposure sequence during data acquisition. An alternative detector that shows significant promise for such applications is the TimePix chip,13 originally developed at CERN for particle physics applications. TimePix has been demonstrated with great success for electron imaging in chemical dynamics studies,14 for ion imaging in combination with a microchannel plate detector (also as part of a chemical dynamics study)15 and, within the past year or so, for imaging mass spectrometry.16−18 TimePix currently has one drawback for multimass imaging studies, in that each pixel may only be activated once per time-of-flight cycle, with the result that ions arriving at a given pixel later in an acquisition cycle are shadowed by lighter ions that arrived earlier, and are therefore not detected. Other novel pixel detectors originally developed for the particle physics community are likely to make an impact as detectors for velocity-map imaging in the future; for example, the Gigatracker hybrid silicon pixel detector19 is currently under development at CERN as part of the NA62 experiment and will have subnanosecond timing resolution.
camera. The frame rate of commercial CCD cameras is typically around 50 Hz and is therefore only sufficient to acquire a single image on each ∼100 μs time-of-flight cycle. In order to image a single mass, the potential applied either to the MCPs or to an image intensifier located in front of the camera is time-gated such that the signal is only detected when the chosen mass arrives at the detector. The signal for the selected mass is usually integrated over a number of time-of-flight cycles until an image with sufficiently high signal-to-noise is obtained. In the case of multiple fragmentation pathways, while it is certainly possible to acquire images for each mass of interest in sequence, for a large number of fragments, this quickly becomes timeconsuming, and the image quality for different fragments is susceptible to drift in the experimental conditions over time. Two previous approaches to multimass velocity-map imaging both rely on incorporating an additional electric or magnetic field into the flight tube of a velocity-map imaging apparatus in order to achieve spatial separation of the various masses on the detector.5,6 In measurements of this type, the mass range that can be probed in a single experiment is limited by the finite size of the imaging detector. The alternative is to use a standard VMI instrument and to employ an ultrafast detector capable of recording separate images for each mass as the ions arrive at the detector. Such a detector requires a time resolution on the nanosecond time scale to match the typical widths of ion timeof-flight peaks, and a dynamic range of a few tens to a few hundreds of microseconds to match the arrival time distribution of the ions. Delay line detectors allow the simultaneous recording of arrival time and position for individual particles using a combination of one or more microchannel plates and a delayline anode.7−10 The anode consists of two or more wires or striplines arranged to form a grid pattern. The arrival time is recorded at the MCP, and the positional information is determined from a measurement of the propagation time of the electron current pulse incident from the back face of the MCP to each end of the wires making up the delay line anode. Though subnanosecond time resolution is achievable in the arrival time measurement, this arrangement typically allows one ion to be recorded within a given 50−100 ns time window and therefore limits the data acquisition rate to one ion per mass per laser shot in a velocity-map imaging experiment of the type described above. Delay line detectors are widely used in coincidence measurements employing ultrafast lasers operating at repetition rates in the kilohertz range7 but are not well-suited to measurements employing nanosecond laser systems operating at tens of Hertz, in which hundreds of ions may be produced per mass per laser shot. 10898
dx.doi.org/10.1021/jp309860t | J. Phys. Chem. A 2012, 116, 10897−10903
The Journal of Physical Chemistry A
Article
provided by aSpect reads out data from the chip as five 72 × 72 pixel image frames, the first four frames containing the arrival times stored in each timestamp memory and the fifth containing the analogue image (i.e., the total signal acquired on each pixel during the data acquisition period). LabVIEW software has been developed in-house to calibrate the camera in order to correct for variation in sensitivity for different pixels and to process the incoming data. Following processing, the data is stored as a data point in the format (x, y, t, n, m), where x and y denote the pixel coordinates, t is the arrival time, n the timestamp register, and m the acquisition cycle. The cost of a complete PImMS camera system is similar to that of an intensified CCD or electron multiplying CCD camera of the type already used widely within the velocity-map imaging and fluorescence microscopy communities. 2.2. Velocity-Map Imaging Measurements. Two different velocity-map imaging instruments were employed for the reported measurements on the Br2 and N,N-dimethylformamide (DMF), respectively, both sharing a similar internal layout (see Figure 2). The first employed a pulsed Jordan valve to
Here, we report the first multimass imaging experiments employing a new event-triggered intelligent pixel CMOS sensor capable of acquiring images for every mass channel on each time-of-flight cycle in a velocity-map imaging, imaging mass spectrometry, or other time-of-flight-based particle imaging experiment.20,21
2. EXPERIMENTAL METHODS 2.1. Pixel Imaging Mass Spectrometry (PImMS) Sensor. The monolithic active pixel sensor is designed in a 180 nm CMOS technology,22 with each intelligent pixel containing over 600 transistors, of both NMOS and PMOS types. The sensor is currently configured as a visible light image sensor, but future versions will also be capable of direct charged particle detection. The first prototype, reported here, consists of a 72 × 72 pixel array (a larger 324 × 324 pixel sensor is currently in the manufacturing stages), in which each pixel can independently detect events and record their time of arrival via a process known as timestamping. Rather than acquiring complete image frames, the sensor effectively records an (x,y,t) data point for each detected particle, vastly improving data handling efficiency relative to framing cameras through a dramatic reduction in the volume of recorded information. In addition, since each ion is logged independently, no prior information is needed on the ion arrival times. The signal conditioning circuitry within each pixel is shown schematically in Figure 1a. At the start of a data acquisition (or time-of-flight) cycle, the sensor is reset and may subsequently be triggered by an external signal. On irradiation of a pixel, charge is collected on four sensing diodes connected in parallel. The signal is then amplified and shaped by low-noise analogue electronics. The time measurement is carried out by a programmable threshold comparator. Once an event has been detected, it is stored as a timecode in one of four 12-bit memories, and the pixel then resets itself prior to detection of further events, with a dead time of 500 ns. The choice of four memories is discussed later. Each pixel also includes an analogue output (used primarily when focusing the camera), logic to control writing and reading of the four memories, as well as a 4-bit digital-to-analogue converter, which allows adjustment of the comparator offset at the individual pixel level. This allows the sensor to be calibrated in order to reduce pixelto-pixel variability. The resolution of the timing measurement can be set via the camera’s computer interface, with a best time resolution of 25 ns. With this time resolution, the 12-bit depth of the on-pixel memories means that data may be acquired for up to 102.4 μs on each data acquisition cycle, with correspondingly longer acquisition times possible at lower time resolutions. The time codes are fully programmable from off-chip, allowing irregular patterns to be generated if desired. The sensor chip also includes digital and analogue logic circuitry to control configuration and readout of the sensor. During the readout phase after each acquisition cycle, the sensor is not active, and following readout, the sensor requires around 250 μs to reset before the next acquisition cycle commences. The sensor is integrated into a camera designed in collaboration with aSpect Systems (http://www.aspect-sys. com). Because of the on-chip data reduction, a USB2 interface is sufficient to allow a repetition rate as high as 500 Hz for the prototype sensor and 50 Hz for the larger 324 × 324 pixel sensor, with expected frame rates of up to 350 Hz for the latter via a CameraLink interface. The base software and firmware
Figure 2. Internal layout of the velocity-map imaging instruments used in the Br2 and N,N-dimethylformamide fragmentation studies.
create the molecular beam (
