4th Year Project (2008)

P. Molyneux

PLUME (Picosatellite for Leicester University’s Micrometeoroid Experiment) is a student satellite being developed by undergraduates at the University of Leicester. Its primary payload is a highly sensitive nanometeoroid detector based on the concept of thin film penetration by dust. The PLUME detector will make use of a microchannel plate (MCP) to support the film, allowing a thinner film layer to be used which results in increased sensitivity to low mass particles.

1. 
   Introduction......................................................................................................1
   
2. 
   Dust in Low Earth Orbit.................................................................................2
   
   1. 
      Sources of Cosmic Dust...................................................................2
      
   2. 
      Why Study Cosmic Dust?.................................................................3
      
   3. 
      Dust Flux in Low Earth Orbit.........................................................5
      
3. 
   Typical Dust Detector Designs.......................................................................6
   
   1. 
      Remote Sensing...............................................................................6
      
   2. 
      Impact Sensing................................................................................6
      
      1. 
         Solid Target Detectors.............................................6
         
      2. 
         Thin Film Detectors.................................................7
         
      3. 
         Impact Ionisation Detectors.....................................7
         
   3. 
      Lobster ISS samples: An Accidental Dust Detector........................7
      
4. 
   Design Options for a CubeSat Dust Detector...............................................8
   
   1. 
      The MCP Base.................................................................................8
      
   2. 
      Impact Light Flash Detector............................................................9
      
   3. 
      Electron Cascade Design................................................................9
      
   4. 
      Scintillator Design.........................................................................10
      
   5. 
      Chosen Design...............................................................................10
      
5. 
   Feasibility Study for an Impact Light Flash Detector................................10
   
   1. 
      Light Flash Theory.........................................................................10
      
      1. 
         Light Flash Studies.................................................11
         
      2. 
         Signal Timing of Flash...........................................11
         
      3. 
         Temperature of Flash.............................................11
         
      4. 
         Impact Flash Intensity and Energy........................11
         
      5. 
         Intensity as a function of Target Melting Point......12
         
   2. 
      Photon Detection System................................................................12
      
   3. 
      Photomultiplier Theory..................................................................12
      
   4. 
      Modelling the expected response of the PMT................................13
      
      1. 
         Choice of PMT model.............................................13
         
      2. 
         Detector Spectral Response....................................14
         
      3. 
         Discussion of parameters used in model................15
         
6. 
   Conclusions.....................................................................................................17
   
   1. 
      Further Work Required..................................................................18
      
7. 
   References........................................................................................................19
   

PLUME (Picosatellite for Leicester University’s Micrometeoroid Experiment) is a student satellite being developed by undergraduates at the University of Leicester, conforming to the CubeSat standard developed by California Polytechnic State University and Stanford University’s Space Systems Development Lab. It is expected to launch in mid 2009 and its primary payload will be a high-sensitivity nanometeoroid detector, the design of which will be discussed in this report.

The CubeSat programme was created primarily to allow undergraduate students to plan and execute a real space mission, gaining valuable skills and experience in the space industry. All CubeSats must conform to a standard size and shape- a 10 cm sidelength cube of total mass not exceeding 1 kg, with a centre of mass within 2 cm of the geometric centre of the satellite. This clearly puts severe restrictions on the types of payload that can be carried by CubeSats: developers must be creative in order to design scientifically interesting experiments that fit within such strict mass and volume limits.

The dust detector being developed to fly on PLUME is based around a microchannel plate (MCP- a thin glass plate with many parallel channels running through it). The MCP will be used to support an extremely thin foil, and below this structure will be a system that detects the effects of dust collisions with the foil. This payload is ideal for flight on a CubeSat, as MCPs are thin and lightweight and the exposed surface of the detector may be scaled up or down to fit as much area as is available on the satellite

surface after solar panels have been attached. Foils previously flown with MCP bases have been found to contain impact holes apparently caused by dust particles smaller than any previously detected [1], but none have been flown specifically in order to detect dust. This means that the PLUME detector should be able to make a significant contribution to models of dust populations in the near Earth environment. More importantly, the PLUME mission will act as a proof of concept for the detector design: if successful, similar dust detectors may well be flown in the future on missions to low Earth orbit (LEO) and beyond.

The novel design of the detector has attracted interest from other CubeSat teams looking for payload ideas. The PLUME team has already been in contact with groups from the University of Florida in the USA, the University of Rome in Italy and the University of Patras in Greece, all of whom wish to fly copies of the detector on their satellites. This is clearly beneficial in terms of outreach between Leicester and the collaborating universities, and the more data is collected the more reliable any conclusions drawn from the experiment will be. It will also be interesting to compare the data obtained by detectors flying on satellites in different orbits.

The phrase ‘cosmic dust’ is used to describe dust particles in interstellar or interplanetary space, but does not distinguish between particles of different origins. The dust in near-Earth space, for example, may come from one of a number of sources. Some of the dust will have been ejected from celestial bodies elsewhere in the solar system or galaxy and travelled towards the Earth, while other particles may be due to electrostatic blow-off from the moon (see figure 1). Some particles near the Earth will be artificial: Al₂O₃ grains from solid rocket motors, for example, or fragments ejected when artificial satellites experience explosions or collisions.

Some of the particles in low Earth orbit may be associated with dust grains ejected when interplanetary material collides with the moon. The percentage of ejected matter that ends up in the Earth’s magnetosphere depends on the phase angle of the moon, but at certain times in the lunar cycle may be up to 80% [2]. The dust particle experiment on the 1959 Eta (Vanguard III) satellite detected a particularly large number of events over a period which coincided with the Leonid meteor shower [3]. It is believed that this was due to dust being ejected from the lunar surface into the near-Earth environment after impacts by Leonid meteors. It has been suggested that these secondary ejecta make the moon the primary source for all particulate matter in the magnetosphere with radii less than 0.5μm [3].

The authors of [2] have a different explanation for the increased flux of dust during the Leonid shower. They investigated the possible effects of the Earth’s magnetosphere on the motion of submicron-sized dust grains associated with the Leonid meteor stream. The parent comet of the Leonids is 55P/Tempel-Tuttle, which has an eccentric retrograde orbit. This causes the Leonid dust grains moving towards Earth to have large approach velocities (~70km s^-1). Generally, since dust particles carry a charge, the effects of the Earth’s magnetic field and currents in the magnetosphere cause the dust grains to be ejected from the magnetosphere or to crash into the Earth a short time after entering the region. (For a detailed description of the effects of the magnetosphere on dust, see [4].) However, it has been found that magnetospheric perturbations are not strong enough to change the orbits of dust with velocities as high as those reached by the Leonid dust [2]. This is true for dust particles produced by the parent comets of most meteor showers.

surface of the moon.

2) Al₂O₃ grains etc. from solid rocket

motors

2) Disruption of particles by electrostatic force/rotation

of interplanetary dust.

3) Fragments of artificial satellites from

explosions/collisions etc.

Sinks 1) Impact with Earth/satellite.

2) Gradual erosion due to sputtering (bombardment causing atoms to

be ejected into gas phase) by energetic ions.

Fig.1: Sources and Sinks of dust in the Earth’s magnetosphere, adapted from fig. 1 in [5]

2.2 Why study cosmic dust?

Once a dust particle has been ejected from an object in space, it moves through interplanetary or interstellar space, carrying information about the composition and structure of its parent body. The dust particle is then affected by the interplanetary material in a number of ways. Its composition will be changed by photolysis and erosion (sputtering) by the solar wind, volatile molecular species will be evaporated from it, and it will also be affected by temperature changes.

Dust particles (whether interstellar, interplanetary, cometary or meteoritic) are known to contain a type of molecule called polycyclic aromatic hydrocarbons (PAHs) [6]. The authors of [7] recreated the effects that PAHs experience as they travel through space by bombarding frozen anthracene (a PAH) containing water with low energy hydrogen ions. By doing so, they were able to synthesize quinones, a class of molecule that is present in most living organisms. Hence it is hypothesised that the action of the solar wind on dust particles could lead to the production of biologically active molecules like quinones, which may later undergo physical or chemical changes that could lead to the evolution of cells and eventually more complex organisms. Since dust grains are so small, they are decelerated gently enough while entering the Earth’s atmosphere for some of the material carried by them to survive and fall to Earth. Because of this, research into dust particles and their constituent molecules is central to the study of theories of the origin of life on Earth, particularly panspermia.

The study of dust also has a practical application to the field of X-ray astronomy. X-rays cannot be focussed in the same way as visible light. While the index of refraction for visible light varies depending on what medium is considered, the index of refraction for X-rays is approximately one for all materials. This means that the path of the X-rays is not changed significantly when they move through one material into another, and so the focal length for an X-ray lens would be impractically large. Parabolic mirrors (which are often used in optical telescopes) cannot be used either, since when X-rays hit a material at near-normal incidence they are generally absorbed rather than reflected. However, when X-rays hit a mirror at grazing incidence, they do get reflected, and so X-ray telescopes use mirrors that lie nearly parallel to the telescope's line-of-sight (Wolter type 1 mirrors [8]), with a Charge Coupled Device (CCD) detector at the focal point of the system. This configuration leaves the CCD more exposed to the space environment than the detectors in conventional telescopes (see figure 2), and so it is vulnerable to dust impacts. A CCD detector on the European Space Agency’s (ESA) cornerstone x-ray observatory XMM Newton has already been damaged by a micrometeoroid impact (figure 3) and a knowledge of the dust population in space is crucial for spacecraft risk assessments in future X-ray astronomy missions.

The Grün interplanetary flux model gives the average flux of micrometeoroids above a certain mass, m, per m² per year incident on a randomly tumbling plate at 1 AU from the Sun, ignoring the presence of the Earth. Mathematically, it is described by the following equations [11]:

where

The shape of the flux model is shown in figure 4.

While this model has been found to fit closely with the fluxes measured by previously flown dust detectors, the most sensitive detector ever flown- the dust detector on HEOS 2- was sensitive only to particles of mass greater than 2x10^-16g [1], so the model has not been tested at the less massive end of the micrometeoroid spectrum. The other factor that affects the reliability of the model is the presence of the Earth: Earth shielding will reduce the flux seen by a spacecraft (this effect decreases as the altitude of the spacecraft increases), while at the same time the Earth’s gravitational field will attract dust particles, increasing the flux. For a discussion of how these effects may be compensated for, see [11].

In-situ methods of detecting cosmic dust can be broadly separated into two categories: remote- or pre-impact- sensing where the dust is detected before it hits the spacecraft, and impact-sensing, in which the physical effects of the dust impacting the detector are studied.

It is possible to observe dust particles using an imaging telescope in space, because individual sun-illuminated particles passing the spacecraft will appear in images as streaks of light. In some cases, it might be preferable to provide an artificial light-curtain to illuminate the particles rather than relying on the sun [12]. Another remote detection technique involves the use of a thin wire or other conductor. When a dust particle comes close to the conductor, it induces a charge on it, since the particle itself is naturally charged. It may be possible to deduce from the induced charge the velocity, trajectory and orbit of the particle [12].

Collisions in interplanetary space tend to occur at hypervelocity speeds: speeds which are greater than the speed of sound in the colliding objects (around 5 km s^-1 or faster [13]). When such collisions take place between a dust particle and a solid target, shock waves propagate outwards from the collision point through each of the colliding bodies, compressing and heating them. If the shock waves reach the back edge of the dust particle (the edge that did not collide with the target), rarefaction waves start to propagate back through the material, and this compression and rarefaction causes the particle to break up, melt, and evaporate. The vapour produced emits light in the form of a blackbody spectrum. Other products of the collision include an impact crater in the target, and ejecta which may include solid and liquid particles, neutral and ionised molecules and electrons [12].

The impacting particle and ejecta all transfer momentum to the target and generate a pulse of high pressure (up to 5 TPa) and temperature (up to 0.5 MK) [12]. The ejecta expand outwards from the collision point and strike adjacent surfaces where they may produce more craters, light flashes and ejecta. Any of these effects may be measured in a dust detection system based around particle impact onto a solid target.

3.2.2 Thin film detectors

When a fast-moving dust particle hits a thin foil, shock waves propagate through both particle and foil in the same way as in collisions between dust and solid targets. In this case, however, shock waves may reach the back surface of the foil as well as the edge of the particle, causing the back of the foil to buckle. If the foil is thin compared to the diameter of the particle (particle diameter / foil thickness ~ 0.33 [12]) this buckling, coupled with the crater caused by the impact, will result in an impact hole through the foil.

The physical effects experienced by the particle and target in a solid target detector are all present in thin film detection systems. A dust particle impact with a foil will result in similar ejecta, a crater or hole, a light flash and momentum transfer. Secondary ejecta and craters may also be produced. Ejected material may emerge from both front and back surfaces of the foil.

An example of a dust detector design based on thin foil perforation is the pressurised cell detector, in which a drop in the cell pressure- caused by a dust particle perforating the cell wall and allowing gas to escape- triggers the system to register a dust impact. This design was first flown in 1962 on the Explorer 16 satellite [12]. Similar detectors were later flown on the Pioneer 10 and 11 spacecraft [12].

High-speed collisions between dust grains and both solid and thin film targets result in the production of ejecta including ions and electrons. In impact ionisation detectors, these are separated by an electric field into positive ions and negative ions and electrons, which can be detected, and in the most sensitive systems even counted individually [12]. The addition of a mass spectrometer allows different ion species to be separated, and their relative concentrations measured. The total number of electrons detected is related to the mass and velocity of the impacting dust particle. An advantage of impact ionisation detectors is that it is easy to distinguish between the signal created by a particle impact and noise: if the signal is due to a particle, electron and ion signals will be seen simultaneously.

Between 2002 and 2004, microchannel plates coated with a thin (60 nm) film of aluminium were exposed to the space environment outside the International Space Station [1]. Similar MCPs were intended to form the optics of the Lobster X-ray telescope, and the purpose of the ISS exposure was to test their suitability for this application. The walls of the channels in MCPs focus X-rays in a similar way to the mirrors shown in figure 2: X-rays entering the telescopes at angles almost perpendicular to the plates are reflected down the channel walls towards a detector.

The aluminium film on the Lobster samples was intended to act as a thermal control layer for the system. When the MCPs were returned to Earth, they were examined at the University of Leicester to determine the effects of the exposure on the absorptivity and emissivity of the film. The aluminium layers were found to have been perforated by a large number of tiny holes from impacting particles. The foils were studied using a Field Electron Gun Scanning Electron Microscope (FEG-SEM), and impact holes as small as 100 nm in diameter were found [14] (see figure 5). The Carey-McDonnell-Dixon (CMD) formula [15], which relates the diameter of a particle to the diameter of hole resulting from its impact with a thin foil, predicts that a 100 nm hole through a 60 nm foil may be created by a 33 nm particle impacting at 5 km s^-1, or by a 13 nm particle at 20 km s^-1 [14]. Dust detectors measuring the flux of particles penetrating a thin foil have been flown before, but since none had a supporting base like an MCP, their ability to detect the smallest, lightest particles was limited by the fact that they had to be thick enough to support themselves. A detector making use of a supporting MCP base may therefore increase our knowledge of the distribution of the smallest dust particles in space, allowing models such as the Grün flux model to be tested in the low mass particle regime.

The impact features found on the Lobster sample plates are believed to have been produced by nanometeoroids with masses as small as ~10^-18 g [1]. This makes the filmed MCPs sensitive to particles approximately 200 times smaller than those detected by the dust detector HEOS 2, the most sensitive detector flown to date.

The PLUME nanometeoroid detector will be an active version of the passive dust detection system that was flown on the ISS in the form of the Lobster samples, allowing information about dust impacts to be transmitted to Earth without the need for retrieval of the detector. The development of an active detector will also allow future use of the system on unmanned missions anywhere in the solar system.

The two main designs considered for the PLUME dust detector are built around a microchannel plate structure. The MCP consists of a thin (~1 mm) glass plate, through which run a large number of parallel channels with diameters of 12.5µm. The channels are arranged hexagonally, as shown in figure 6, with a distance between the centre of one channel and the centres of the channels surrounding it of 15µm. This gives an open area fraction for the MCP of 63%.

The first design considered works by detecting the light flash produced when a hypervelocity dust grain hits a thin foil and creates a clear hole by vaporising part of the target foil. At least a fraction of the photons emitted in the flash moves through the perforated foil and down the MCP channel towards a detector. This detector is either a CCD (charge coupled device) or CMOS (complementary metal-oxide semiconductor) imaging chip or a photomultiplier tube (PMT).

As an alternative to the light flash detector, a detector that measures the charge released during particle impact may be flown. If a voltage is applied across an MCP, the plate can be used as a charge multiplier. Incoming radiation or particles hit the wall of a channel and excite electrons out of the glass. These electrons are accelerated down the channel by the potential difference. They hit another wall, exciting more electrons which in turn are accelerated down the channel, exciting yet more electrons. This electron cascade results in a large charge which is detected at the base of the MCP. Generally, MCPs are run in saturation (high voltage) mode, which means that the overall charge in the cloud of electrons that reaches the detector is always large enough to cancel out the electric field gradient down the channel. This means that any event causes the same size signal at the detector. If used in low gain (lower voltage) mode, information about the charge that initially caused the electron cascade may be deduced from the magnitude of the charge that hits the detector.

In the electron cascade dust detector, a voltage is applied across the MCP and it is the charged particles in the plasma produced during impact that cause the electron cascade. The charge is collected on a metal plate at the base of the MCP. The voltage across the MCP should be chosen such that dust impacts will cause saturation but impacts by cosmic rays or photons will not, allowing genuine signals to be distinguished from noise.

4.4 Scintillator Design

It has been suggested that instead of an MCP-based dust detector, PLUME could fly a detection system consisting of a scintillator to convert the energy of the particle into visible light, and an imaging chip to detect this light. Scintillators are normally used to convert high energy radiation such as gamma rays into lower energy radiation like visible light. It is not well known how scintillators respond to particle impact, or how sensitive such a detector would be.

The design chosen by members of the PLUME payload team for further development was the light flash design. It was felt that this was the most elegant design, and the fact that the imaging system could be bought off the shelf, making development faster and simpler, was seen as a major advantage. Another advantage it has over the electron cascade design is that the cascade system relies on a high voltage which requires a lot of power. Power available on a CubeSat is limited to the amount of power that the solar panels can generate to charge batteries, and this must be shared between the communications, on-board data handling and attitude determination subsystems, as well as the dust detector and camera.

The following section is a feasibility study for a dust detection system in which the intensity of the light flash produced by the vaporisation of a dust particle impacting foil is measured by a PMT.

As stated in the section dealing with typical dust detector designs, hypervelocity collisions between dust particles and targets cause shock waves to travel through both objects, compressing and decompressing the materials in turn and resulting in the vaporisation of the particle. The flash seen when this happens has been studied using dust accelerators. It should be noted that all of the research summarised in the rest of this section was gathered from experiments involving dust impacting solid targets: it has not yet been experimentally determined whether or not the flash looks the same for thin foil targets.

Studies of the light flashes emitted during the collision of high velocity dust particles with solid targets of various different materials have been carried out using dust accelerators [16]. In these accelerators, dust particles are first electrically charged and then accelerated by a 2MV Van de Graaff generator towards a target in a vacuum chamber. Using this set-up, the dust particles are accelerated to a velocity given by

where V is the accelerating voltage, m is the mass of the particle and q is the charge on it. The light emitted in the flash is usually detected using photomultiplier tubes.

The light flash comprises of two parts: the main impact flash and a second signal due to secondary impacts of ejecta from the projectile and target [17]. The first part of the flash has a rise time shorter than 100 ns and a length of 10-20 µs. The length of the second part is velocity-dependent, and can be 200 µs at low velocities. The relationship between the amplitudes of the two peaks is independent of particle mass but increases with impact velocity, until at ~8 or 9 km s^-1 the second peak is covered by the first peak as it falls [18].

In some studies, a number of photomultiplier tubes are used to detect the light flash, each with a different wavelength filter. The light energy signals from the tubes are plotted as a function of wavelength. This distribution is integrated over all wavelengths to give the total energy in the flash. If a blackbody curve is fitted to the plot, the temperature of the light flash can be estimated using Wien’s law:

This method gives a plasma temperature estimate in the range of 2500 K to 5000 K for typical light flashes [19].

The maximum intensity of the flash, I, and the total energy in the flash, E, depend on the velocity, v, and mass , m, of the impacting particle as I = c₁m^α1v^β1 and E = c₂m^α2v^β2. The parameters c₁, β1, c₂, and β2 all depend on the projectile-target combination used. β1 has a value between 3.8 and 4.6, with an average of 4.1, while β2 varies from 3.0 to 3.6 with an average of 3.2. α1 and α2 are both equal to 1[19]. Hence:

Depending on the projectile and target materials, the light energy in the flash varies from 2 x 10^-6 to 10^-2 in units of the projectile energy [19].

It has been shown that the intensity of the first peak in the light flash is higher for targets made from metals with high melting points (e.g. molybdenum, rhenium, tantalum and tungsten) than for other targets [19]. These metals also only show one peak, instead of the normal two. A suggested explanation for this is that the temperature generated in the flash (assumed to be 2500 K) is not hot enough to liquefy these metals significantly and produce enough ejecta to have a second peak [19]. The energy that would be used for melting other targets is instead used for heating and evaporating the projectile, generating more hot gas and plasma than usual.

When the impact flash detector was initially proposed, it was intended to use a CCD or CMOS imaging chip under the filmed MCP to detect each flash. CCDs are often used in digital cameras, but they are quite difficult to use in space as they must be cooled and are highly susceptible to radiation damage. CMOS chips are much better at coping with exposure to radiation, and if radiation damage to a pixel does occur it is easy to isolate and ignore the pixel. This is not possible for CCDs. However, CMOS images are typically much noisier than CCD images, so the ability of the system to detect the lowest energy particles would be compromised if a CMOS chip rather than a CCD was used.

Since neither CCD nor CMOS chips are particularly well-suited to application as the imaging system for an impact flash detector, it was suggested that photomultiplier tubes (PMTs) should be considered instead. Photomultiplier tubes have already been successfully used to study dust impact flashes with solid targets on Earth. They require high-voltage power supplies, which is not ideal for CubeSat applications, but are available with built-in power supply units that require a 5V supply. These PMT units would be well-suited to use in the PLUME detector.

5.3 Photomultiplier Theory



Fig. 9: Schematic diagram of a photomultiplier tube [20]

A Photomultiplier tube converts light into electrical signals via the photoelectric effect (see figure 9 for an illustration of PMT operation). Light enters the tube through a glass window and hits the photocathode, causing photoelectrons to be emitted into the vacuum. The photoelectrons are accelerated and focussed onto the first dynode, depositing energy which causes secondary electrons to be released from the dynode surface. These electrons go on to excite more electrons from the next dynode, and so on down the tube. The electrons emitted from the last dynode are collected at the anode.

The output of a photomultiplier tube is a current pulse, the charge in which is proportional to the energy absorbed at the photocathode and the gain of the PMT (e.g. the number of secondary electrons that are excited by each photoelectron). The pulse may be small, so a preamplifier is used to convert it into a voltage pulse large enough to be analysed properly. The voltage pulse falls away exponentially in time, but if the time constant of decay is large and the time between signals (dust impacts) is small, the pulses may pile up, making it difficult to analyse each one individually. For this reason a shaping amp is used to shorten the pulse. It also smoothes out the peak in order to make detection of the maximum value possible. Without the shaping amplifier the maximum would be too sharp to be read out properly. The shaped pulse is then converted into a digital signal.



Fig. 10: Signal processing for a PMT system

At this point the signal is ready to be analysed to determine the energy that was absorbed at the photocathode: the energy in the impact light flash. In the case of the PLUME dust detector, data about the shaped pulse will be written to memory stack by direct memory access (DMA). If it is determined that the pulse is a valid signal rather than noise (e.g. if the energy in the pulse is similar to that expected to be contained in the light flash), an interrupt signal will be sent to the satellites on-board data handling (OBDH) system and the data will be written to non-volatile memory.

The material from which the photocathode is produced determines the spectral response of a PMT- its sensitivity to different wavelengths of light. PMTs are available which respond best to UV, visible or IR radiation. In order to choose a suitable PMT unit to use in the dust detector, the spectrum of the light flash must be considered. The temperature of the flash is known to be between 2500 and 5000 K. The peak wavelength of the flash can be determined from Wien’s Law,

and is found to be between 579 nm at T=5000 K (towards the yellow part of the spectrum) and 1160 nm at T=2500K (in the near infrared).

Hamamatsu produce a wide range of photomultiplier tubes with cathodes of various materials. The main photocathode types are: Cs-Te (caesium telluride) or Cs-I (caesium iodide), which are sensitive to UV rays only; bialkali materials, which are typically sensitive to UV and visible radiation; multialkali cathodes, which respond well in the visible and near IR regions of the spectrum; and GaAs (gallium (II) arsenide) cathodes, which cover the UV, visible and near IR regions. For the dust detector, a multialkali model was chosen, as these are readily available and cover a large proportion of the spectral range that the light flash’s peak wavelength is likely to fall into. For the purposes of modelling the dust detector’s response to particle impacts, a specific PMT model was chosen: the Hamamatsu R7400U-02. This unit has a multialkali cathode and responds to light with wavelengths between 300 and 880 nm [21].

5.4.2 Detector spectral response

The typical spectral responses of multialkali PMTs in the Hamamatsu R7400U series are shown in figure 11.

The spectral response of the detectors is given in terms of the cathode radiant sensitivity. This can be converted to quantum efficiency using the following equation:

where h is Planck’s constant, c is the speed of light in a vacuum, e is the electron charge, λ is the wavelength of incident light in nanometres and S_λ is the radiant sensitivity at that wavelength.

In order to determine whether a PMT would be a suitable photon detection system for the dust detector, a spreadsheet model was produced in which the spectral response curve of the Hamamatsu R7400U-02 PMT was convolved with the expected spectrum of a typical hypervelocity light flash. This was achieved by treating the light flash as a blackbody. The temperature of a typical light flash is between 2500 K and 5000 K, so two blackbody curves were plotted based on these two extreme temperatures. This allowed the minimum and maximum expected PMT response to be compared.

A blackbody can be described by Planck’s law:

, (7)

where I(ν)Δν is the energy per unit surface area, per unit time, per unit solid angle emitted in the frequency range ν to ν + Δν, h is Planck’s constant, c is the speed of light in a vacuum, k is Boltzmann’s constant and T is the temperature of the blackbody.

The purpose of the spreadsheet was to calculate the number of photons from the flash that would be detected by the PMT. The calculation was split up so that the number of photons in each of a large number of frequency bins was counted, giving an idea of the range of wavelengths over which the strongest signals would be expected from the detector.

The first step in creating the spreadsheet was to take readings from the spectral response curve (figure 11) at regular wavelength intervals and convert them into quantum efficiencies using equation (6). It was only possible to take readings every 50 nm along the curve because of the small size of the graph, so the radiant sensitivity was extrapolated between these readings to provide estimates for the quantum efficiency every 10 nm. The wavelengths were then converted to frequencies to allow easier use of Planck’s law.

Next, the blackbody equation was used to calculate I(ν) at the two extreme temperatures for each of the frequencies considered. The value given was then multiplied by largest expected time signal of the flash (200 µs) and by a solid angle of 4π. The energy value also had to be multiplied by the surface area of the flash. The heat flow of the vapour released in a hypervelocity dust impact has been modelled before [17]. It was found that the temperature of the vapour is around 3000 K at a distance of 3r from the impact site, where r is the radius of the dust particle. Hence, the surface area of the flash was approximated to a sphere with this radius. The radius of the impacting particle was calculated from a typical dust density, ρ, and mass, m:

(8)

The particle mass used was 10^-16 g, since no system previously flown has detected particles this light, and the main aim of the detector is to expand on current knowledge of the low-mass dust population in low Earth orbit by bettering previous detector sensitivities.

To get I(ν)Δν, the value of I(ν)calculated above was multiplied by the frequency difference between the row of the spreadsheet being considered and the next row down; so for the first row of the spreadsheet- with a wavelength of 280 nm- I(ν) was multiplied by the difference in frequency between light with wavelengths 280 and 290 nm.

The energy of a photon in each row was calculated using E=hν, where the symbols have their usual meanings. I(ν)Δν was divided by this to give an estimate of the number of photons in each frequency bin, over all solid angles. Assuming that the photons in the flash would be evenly distributed over 4π steradians, half of them could be ignored as they would be moving away from the detector rather than towards it. However, not all of the photons in the remaining 2π steradians would reach the detector. Glass is not a very reflective material, so a large proportion of the photons would be absorbed by it if they hit the MCP channel wall. For the purposes of the spreadsheet, the worst case scenario was considered; that in which all photons incident on the channel walls would be absorbed. In this case, only photons in a narrow solid angle would reach the detector. The magnitude of this solid angle was calculated using

where θ is the angle shown in figure 12. When calculating θ, the microchannel diameter and length were taken to be 12.5 μm and 1.0 mm respectively.

The number of photons of a particular wavelength in this solid angle was calculated and multiplied by the quantum efficiency of the detector to give the total number of photons detected by the PMT in the corresponding frequency bin. This was the required output of the spreadsheet.

The finished spreadsheet is shown in full in appendix 1.

Initially, an error in the spreadsheet meant that the number of photons hitting the PMT was multiplied by the radiant sensitivity of the detector, rather than the quantum efficiency. This resulted in the number of photons detected being underestimated by a factor of ~100 and so, according to the model, the PMT would detect at most 1 or 2 photons in each frequency bin (see appendix 2). It was suggested that the MCP channel walls might be coated in a reflective material to increase the number of photons reaching the PMT. However, it is not known whether the film depositing technology is compatible with MCPs that have had their channels coated in this way.

Because of the apparent inefficiency of the PMT, it was decided that an electron cascade detector (figure 8) might be a more suitable design choice for the dust detection system. The electronics chain required to process a signal from an electron cascade detector is the same as that used by an PMT light flash detector (figure 10), although the detector itself would require a high-voltage power supply, unlike the PMT detector which would have one built into it. A preamplifier and shaping amplifier have been set up in the MCP lab at the University of Leicester’s Space Research Centre. A test pulse generator has been connected to the system, as has an oscilloscope which can be used to view the shape of the pulse at various stages in the signal processing chain. The components will now be tested to find the optimum settings to produce clear voltage pulses (see appendix 4). Once this has been completed, the detector front and high-voltage power supply may be connected and the complete setup tested at a dust accelerator.

The error in the spreadsheet has now been noted and corrected, and the results can be seen in appendices 1 and 3. At T = 2500 K (which corresponds to low energy particles), the PMT is still not expected to be able to detect photons in any frequency bin. However, higher energy impacts, producing flash temperatures of 5000 K, are expected to result in tens of photons being detected in most of the frequency bins. This is enough to result in a charge pulse, and so the light flash detector should be suitable for the detection of high-energy dust impacts. However, since the energy in the flash is related to the mass of the impactor, a detector that is only sensitive to high-energy collisions may limit the success of the PLUME mission in measuring the flux of low-mass particles in the near Earth environment.

The electron cascade detector is potentially much more sensitive to dust impacts than the PMT design, because the voltage across the MCP can be tweaked to give the best response to dust collisions. For this reason, the electron cascade design will be developed further, although since the electronics required for both designs are the same it may be possible to test both without much extra work.

One of the main advantages of using the light flash detector is that it would make it easy to distinguish between valid dust impact signals and noise. If the electron cascade design is used this process would be considerably harder. An MCP electron multiplier responds to electromagnetic radiation such as X-rays, and also charged particles- whether due to dust impacts, solar wind or cosmic rays. It should be possible to select the voltage across the MCP so that dust impacts produce a saturated signal while lower energy protons and electrons do not. However, X-rays may have similar energies to dust particles, making them hard to exclude. An important step in the development of the electron cascade detector is therefore to research the charged particle (proton, neutron, alpha and beta particles) and radiation (X-ray and γ-ray) flux expected in low Earth orbit, and also to investigate further any possible overlaps between the energy of the dust particles and the energy of background particles and radiation.

Another area that needs further work is the development of the signal processing system for the detector. An electronics chain for the electron cascade detector has been set up and will be finished once the settings have been adjusted to give optimum signals. Once this has been done a filmed MCP and high-voltage power supply can be attached. A vacuum chamber has been sourced for the detector to sit in while being tested, and the PLUME team have been promised time at the dust accelerator at the Max-Planck-Institut für Kernphysik in Heidelberg by summer 2008. It is hoped that it will be possible to test both of the detector designs that are currently being considered.

7. References

[1] J. D. Carpenter, T. S. Stevenson, G. W. Fraser, J. C. Bridges, A. T. Kearsley, R. J. Chater and S. V. Hainsworth (2007): Nanometer Hypervelocity Dust Impacts In Low Earth Orbit, J. Geophys. Res., 112, Issue E8, December 2004

[2] A. Juhász, M. Horányi: Magnetospheric Screening of Cosmic Dust, Journal of Geophysical Research, Volume 104, Issue A6, June 1999, p. 12577-12584

[3] W.M. Alexander, J.D. Corbin: Interaction of Lunar Ejecta and the Magnetosphere of the Earth, Solid particles in the solar system; Proceedings of the Symposium, Ottawa, Canada, August 27-30, 1979. Dordrecht, D. Reidel Publishing Co., 1980, p. 425-428.

[4] M. Horányi, H.L.F. Houpis, D.A. Mendis: Charged Dust in the Earth’s Magnetosphere- 1. Physical and Dynamical Processes, Astrophysics and Space Science, vol. 144, no. 1-2, May 1988, p. 215-229

[5] D.A. Mendis: Entry of Dust Particles into Planetary Magnetospheres, Advances in Space Research, vol. 4, no. 9, 1984, p. 111-120

[6] L. J. Allamandola, A. G. G. Tielens, and J. R. Barker: Interstellar Polycyclic Aromatic Hydrocarbons: the Infrared emission bands, the excitation/emission mechanism, and the astrophysical implications, Astrophysical Journal Supplement Series 71, 733 (1989).

[7] M. Tuleta, L. Gabła, and J. Madej: Bioastrophysical Aspects of Low Energy Ion Irradiation of Frozen Anthracene Containing Water, Physical Review Letters, vol. 87, no. 7, 2001.

http://www.spenvis.oma.be/spenvis/help/background/metdeb/metdeb.html#METFLUX