- Research Article
- Open Access
Novel method for state selective determination of electron-impact-excitation cross sections from 0° to 180°
© Weyland et al.; licensee Springer. 2014
- Received: 3 April 2014
- Accepted: 5 June 2014
- Published: 7 August 2014
We use an improved target recoil momentum spectroscopy setup to determine differential cross sections for excited metastable state production in atoms and molecules by electron impact and show its capabilities for an atomic helium target. A crossed beam setup with a supersonic helium jet and a pulsed electron beam at energies close to the excitation threshold of 19.82 eV was used. Measuring the recoil momentum vector of the target instead of the momentum of the scattered electron removes common restrictions to the accessible scattering angles while the microchannel plate detector ensures a high counting efficiency. Using a photoemission electron source we reach an energy resolution of about 200 meV at 1 µA peak current. Results are compared with simulations using theoretical convergent-close-coupling (CCC), R-matrix with pseudo-states (RMPS) and B-spline R-matrix (BSR) calculations and show good agreement.
- Inelastic scattering
- Electron impact excitation
In traditional electron impact experiments, the scattered electron is measured using a movable detector. The scattering angles are scanned by changing the position of the detector . This technique allows measurement within a limited angular range, as the electron detector would not be able to measure the scattered electrons in the backward and forward directions due to the interference of the spectrometer with the incoming or outgoing projectile beam. Detecting the recoil momentum of the excited target instead of the scattered electron has the advantage that the projectile beam has no influence on the measurement. This principle has been shown by Murray and Hammond , for electron impact excitation at intermediate projectile energies. Our improved setup allows for state selective measurements close to the threshold energy by using a time- and position-sensitive microchannel plate detector, where events from all electron scattering angles can be registered simultaneously and by using a photo-emission electron source.
For a given projectile momentum and a particular excited state energy Eexc, the absolute value of the final electron momentum is fixed. The target momenta therefore lie on a sphere with origin and radius for all possible scattering angles. In Figure 1, momentum transfer is shown only in the x-y plane for simplicity: therefore, all events that correspond to the excitation of the same state have target momenta on the dotted dark gray circle while the momentum of the outgoing electron lies on the full circle. The radius of this circle is proportional to the amount of electron momentum after the collision, putting energetically lower lying states - which correspond to a higher excess energy - on larger radii.
where d is the distance between the interaction region and the detector plane. Unfortunately, our measurements showed that the initial longitudinal momentum spread in the supersonic jet due to its finite temperature is about 0.75 a.u. and therefore of the same order of magnitude as the transferred momentum, making the momentum transfer in z-direction inaccessible for all practical purposes. The rotational symmetry of the electron scattering cross section around the y-axis, however, allows for reconstruction of the complete differential cross section, as will be described in Section ‘Data analysis’.
Helium has long been a benchmark target for the investigation of electron-atom collisions due to its simple structure. Many experiments have been conducted to measure electron impact excitation cross sections ,– and theories were developed to explain the experiments –. Total cross sections are known very precisely from experiment  and agree very well with theoretical calculations , but measurement of differential scattering cross sections has for a long time been restricted to an intermediate range of projectile scattering angles, usually ranging from 10° to 130° ,.
One way of accessing electron scattering angles around 180° is the use of a magnetic angle changer , in which a strong localized magnetic field in the interaction region changes the ejection direction of the electrons, depending on their energy. The few experiments that have been conducted to measure inelastic electron scattering on helium using this technique – show discrepancies between experiment and theory especially at high scattering angles. Detection of the scattered helium atoms, as performed with the instrument described here, is therefore a sensible addition to the existing methods.
In earlier experiments, the momentum transfer to the scattered helium was used to determine electron scattering angles by Zajonc et al. . They only measured the time-of-flight of metastable helium atoms, which varied due to the momentum transfer in the flight direction of the helium. This method gave access to all scattering angles, but was unable to distinguish between different excited states and was therefore only applicable in the energy range from 19.8 eV to 20.6 eV, where just the 23S state can be excited. Murray and Hammond  used a rotatable detector setup to measure the deflection angle of helium atoms after electron impact excitation. In their arrangement the incoming electron beam crosses a gas jet perpendicularly and electron scattering within the plane determined by both beams is studied. Thus, scattered electron momentum determination relies on the measurement of the atomic recoil momentum along the incoming electron beam and along the gas jet. Since the resolution along the latter direction, due to the thermal velocity spread, is limited to about 0.76 a.u., a larger projectile excess energy was chosen such that different electron scattering angles result in more strongly varying recoil momenta. Also, excitation to the different accessible states was not resolved but summed differential cross sections were obtained.
Using a microchannel-plate detector with time of flight resolution and x-y-position resolution for detection of the recoiling metastable excited atoms, we obtain 4π-acceptance for electron scattering. In the plane perpendicular to the gas jet we obtain strongly improved momentum resolution, discriminating excitation of different states at low energies above the excitation threshold.
Our new setup is not limited to a helium target, but rather it can be used with all light targets, for which the deflection from their original direction after the collision is large enough to be resolved.
The recorded momenta associated with the excitation of metastable states provide diverse information. On the one hand, the excited atom yield and, thus, the total metastable state excitation cross section as a function of impact energy are obtained. On the other hand, excited state resolved cross sections can be extracted as, e.g., excitation functions and cross sections differential in the projectile scattering angle.
There is no significant difference between CCC, RMPS and BSR calculations in the differential cross section for 23S state excitation, and the experiments are in good agreement with all theories. For the 21S and 23P state excitation, however, the CCC calculations show lower yield in the broad backward scattering peak and a more pronounced peak at about 110°, showing a reminiscent feature of the 23S state excitation (cf. Figure 4). However, all theories are in good agreement with the experimental results within our measurement uncertainty.
Our setup allows for the measurement of electron impact excitation to metastable states in light targets, resolving all scattering angles as well as different excited states. We demonstrated the setup’s ability by using a helium target. Differential cross sections for the excitation of the 21S and 23P state in helium close to the excitation threshold energy are shown for the first time at all scattering angles, including the scattering angles in the vicinity of 0° and 180°. This instrument provides an additional tool to check theoretical predictions of electron impact excitation cross sections.
We have created a supersonic jet of the target gas by expansion through a 30 μ m nozzle at a backing pressure of 5 bar. The jet is collimated by two skimmers 250 μ m and 400 μ m in diameter. Each skimmer is set in a differentially pumped stage, maintaining pressures of 2×10−3 mbar in the first jet stage, 5×10−6 mbar in the second jet stage, and 5×10−8 mbar in the main chamber during operation. The longitudinal jet temperature is 3 K. With this setup a beam diameter of about 2 mm and a target number density of 1.5×1011 cm −3 can be obtained in the region of interaction with the electron beam. To increase resolution, the beam is additionally collimated by inserting a 500 μ m aperture a few millimeters above the interaction point, which is held at the same potential as the surrounding region at all times. The beam is dumped directly into the chamber without an additional dump stage.
The electron beam is produced in a photoemission electron gun . A GaAs cathode is coated with a monolayer of cesium and oxygen to obtain negative electron affinity (NEA) conditions. In a NEA semiconductor, the conduction band minimum is above vacuum energy. Illuminating the surface with a laser excites electrons to the conduction band and can induce electron emission by tunneling through the thin surface barrier. The electron gun has been described in detail in the work by Schröter et al. . To keep the surface stable, ultra-high vacuum is needed. The electron gun is therefore placed in a separately pumped chamber, which can be sealed off completely from the measurement vacuum chamber by a gate valve. This so-called gun chamber was heated to 220°C for one week to reach a base pressure below 1×10−11 mbar at room temperature afterwards. During the measurements, the two chambers are connected by an aperture 1 mm in diameter, which allows a pressure of 1×10−10 mbar to be maintained in the electron gun chamber.
The pulsed electron beam is produced by a laser pulse, illuminating the cathode. Previous studies have shown that multi-mode lasers can induce a high energy spread in the electron beam, as do high current densities . Kolac et al.  explain this large energy spread from multi-mode lasers with high temporal intensity fluctuations which lead to an inhomogeneous space charge distribution in the emission current.
In our setup, the cathode is illuminated by a single-longitudinal-mode laser with a constant temporal intensity profile. The laser has a wavelength of 671 nm corresponding to a photon energy of 1.84 eV, which is high enough to populate the conduction band in the used crystal that has a band gap of 1.42 eV . When operated in continuous-wave mode at a laser intensity of 25 mW, currents of several μ A are reached. The pulsed electron beam is produced by guiding the incident laser through an acousto-optic modulator and coupling the light of the first diffraction maximum into an optic fiber, where it is guided to the focussing optics outside the vacuum chamber. Using the acousto-optic modulator allows switching the laser within 50 ns and obtaining extinction ratios better than 10−3. The laser is focused to a diameter of 1 mm at the cathode, which restrains the electron source to an area of 1.4 mm × 1.0 mm, due to the laser incidence angle of 45°. Additional attenuation filters and an aperture are used to regulate the laser beam intensity such as to compensate for a decreasing quantum efficiency of the cathode over time.
The electrons are focused by four electrostatic lenses and two pairs of deflection plates onto the 1 mm aperture, which separates the electron gun chamber from the detection chamber. Additional steering to the interaction point is supplied by a pair of Helmholtz coils, creating a magnetic field of 1.2 mT pointing from the center of the aperture to the center of the interaction region, thereby confining the electron beam motion perpendicular to the direction of the magnetic field. Fine control of the beam position in x- and z-direction is possible with two additional pairs of coils that create a low magnetic field up to 0.1 mT perpendicular to the direction of the electrons.
The excitation process takes place in an electric-field-free region that is provided by a number of electrically connected metal rings. The distance between two adjacent metal rings is 10 mm and the symmetry axis of the rings is coincident with the electron beam direction. The rings are set to a variable voltage to change the electron impact energy. The electric-field-free region between the two central rings is accessible to the electron beam and the target jet. Rings blocking the flight path of the scattered target atoms are cut on the lower end to clear the path to the detector. The electron gun, interaction region boundary rings and the magnetic field are aligned precisely to keep the projectile beam on the axis of the apparatus when the interaction potential is changed. Any change in overlap of the electron beam and the jet leads to changing count rates. In the aligned setup, this effect was negligible for scanning amplitudes up to 2 V. The scanning voltage is read by an analog-to-digital converter (Hytec ADC 521).
After the inelastic electron-atom scattering, the excited atoms impinge on an MCP detector 80 mm in diameter with a position-sensitive delay line anode, located about 110 mm below the interaction region. Here they are efficiently detected since their internal energy gives rise to electron release at the detector surface. It is therefore necessary to shield the MCP from other electrons. An additional grid in front of the detector, set to -200 V, and the presence of the magnetic field eliminate background due to charged particles. The MCP is mounted at 32 mm offset of the target jet in positive y-direction to account for momentum transfer to the atoms and to detect excited atoms for all projectile scattering angles at excess energies up to 30 eV. Time-of-flight information and delay line signals are read by a time-to-digital converter (LeCroy 3377 TDC) operated in common stop mode and gated to accept only events in the expected time-of-flight range from 55 μ s to 87 μ s. Data from the TDC and ADC are read into a personal computer by a CAMAC system and are analyzed in offline mode. To correct for varying detection efficiency at different positions of the delay line anodes, the detector was homogeneously illuminated using an 241Am α-source. The efficiency map obtained this way was used to weigh counts in the electron-scattering measurements.
Experiments on helium
The cathode potential of the electron gun is fixed to -22.9 V. Different impact energies are created by changing the potential of the interaction region while keeping all electron gun settings. For impact energy scans, a triangular scan with a 1 Hz repetition rate is carried out on the voltage applied to the interaction region boundary, usually providing the energy scan range from 19 eV to 23 eV. Energy scans are used to measure the electron-impact-energy dependence of the total cross section for metastable state excitation that is used to calibrate the impact energy scale based on RMPS calculations. From the measured energy dependence of the cross section the width of the electron energy distribution is determined by comparison with theoretical cross-sectional data  that are convoluted with Gaussian distributions of different standard deviation.
Differential cross sections were measured with constant impact energy and count rates of around 300 counts per second were achieved. The counts were accumulated over an acquisition period of 10 to 24 hours. The discussed exemplary cross sections were obtained for an impact energy of 22.0 eV. At this energy, theoretical calculations predict only minor changes in differential scattering cross sections within several 100 meV, making the measurement less sensitive to the finite energy spread of the electron beam.
Flight times in this setup are about 70 μ s, making the measurement insensitive to fast decaying states like the 21P state in helium, which has a radiative lifetime of 0.6 ns  and decays to the undetected ground state. The three lowest excited states 23S, 21S and 23P are metastable because radiative dipole transitions are forbidden. They have sufficiently long lifetimes of about 8000 s (23S) and 20 ms (21S) and are detected before a significant amount of the excited metastable atoms decays ,. The 23P state radiatively decays with a lifetime of about 100 ns , to the 23S state, which is then detected.
The 21P state produces photons of 21.2 eV when it decays. These photons can be detected by the MCP, but are not recognized, because they arrive before the start of the TDC gate and are therefore ignored.
In the experiment, the momentum transfer in the detector plane is measured very accurately. Error propagation analysis for Eq. (6) with typical values results in an uncertainty of Δ qx,y = 0.068 a.u. that originates mainly from the transversal size of the gas jet. Analogously, the uncertainty of the momentum in the jet direction given by Eq. (7) is an order of magnitude higher: Δ q z = 0.8 a.u., which is almost exclusively due to the initial velocity distribution in the gas jet. Because the highest transferred momentum is 0.4 a.u. at 2.2 eV excess energy, no useful information about the momentum transfer in z-direction is available.
Therefore, in the analysis, projections of the momenta on the x-y plane are used. Two methods of data analysis are applied. First, restricting analysis to events near the peak of the time-of-flight distribution is used to suppress large q z -values. Second, the fact that the scattering has rotational symmetry around the y-axis allows the use of an inverse Abel transform , to reconstruct the full 3D-scattering cross section from the x-y projection, based on the axial symmetry of the scattering. Both sets of results are compared to the predictions from simulations.
The first approach offered the advantage that only little post-processing was necessary. Scattering out of the detector plane was partly suppressed by using only events with velocities in a 50 ms −1 window around the maximum of the velocity distribution (cf. Figure 8). In this case, only out-of-plane scattering events with a vertical momentum transfer up to |q z |=0.17 a.u. contribute to the momentum transfer maps. On the other hand, atoms from the wings of the velocity distribution can contribute if they experience a vertical momentum transfer such that their resulting time-of-flight is within the acceptance interval. This contribution is smaller due to the lower number of atoms in the wings of the velocity distribution. Owing to this condition, the obtained momentum transfer maps show pronounced in-plane scattering features.
The second way to analyze the data was to reconstruct the momentum transfer in the detector plane by exploiting the rotational symmetry around the electron impact direction. Using all events we obtained the projection of all momenta on the detector plane. The projection of a cylindrically symmetrical object on a plane that contains the symmetry axis is an Abel transformation . In the experiment, the symmetry axis was along the electron impact direction in y-direction and the projection plane was the detection plane, i.e. the x-y plane. The original distribution was reconstructed using the inverse Abel transform . Here, the results are independent of the velocity distribution in the jet and show better momentum resolution in x- and y-direction. These results are compared to a simulation where scattering was confined to the detector plane, thus eliminating the jet velocity distribution from the calculations.
To allow for a comparison of the experimental results with theoretical cross sections, the expected momentum transfer maps were simulated by a Monte Carlo-type program. This program takes into account the differential cross sections provided by R-matrix with pseudo-states (RMPS), B-spline R-matrix (BSR), and convergent-close-coupling (CCC) calculations , as well as the following experimental factors: target initial velocity, collision position, and collision energy for every event. Finally, the momentum of the excited target atom is mapped as a projection on the x-y plane. The time-of-flight is calculated from the initial velocity of the atom and the momentum change during the collision. Events produced in this program are filtered by their time-of-flight using the same velocity condition as was used in the experiment. Thus, out-of-plane scattering is taken into account in the same way as in the experiments and the measured momentum distributions are reproduced. The finite size of the jet has been modeled as an additional momentum spread which is specific to the setup geometry. All random variables are assumed to have Gaussian or logistic distributions and parameters of the distributions are taken from measured properties where possible.
The method of using only events with a certain velocity strongly improves the visibility of details in the momentum distributions. At the same time, the result depends on experimental parameters like electron beam resolution and all jet properties. A comparison with theoretical cross sections is difficult, as all parameters enter in the simulation as well.
The momentum distributions that were reconstructed using the inverse Abel transformation were compared to a similar simulation. Scattering angles were restricted to the x-y plane and no velocity condition was applied to reconstruct the expected momentum transfer maps for this case.
a atomic unit of momentum, 1 a.u. =1.993×10−24 kg m s−1.
This work was supported, in part, by the United States National Science Foundation under grants No. PHY-1068140, PHY-1212450, and PHY-14230245, as well as the XSEDE allocation PHY-090031 (KB and OZ). Support of the Australian Research Council and iVEC via the The Pawsey Centre’s advanced computing resources is gratefully acknowledged as well (DVF and IB).
- Hall RI, Joyez G, Mazeau J, Reinhardt J, Schermann C: Electron impact differential and integral cross sections for excitation of the n = 2 states of helium at 29.2 eV, 39.2 eV and 48.2 eV. J Phys Paris 1973, 34: 827–843. 10.1051/jphys:019730034010082700View ArticleGoogle Scholar
- Murray AJ, Hammond P: Laser probing of metastable atoms and molecules deflected by electron impact. Phys Rev Lett 1999, 82: 4799–4802. 10.1103/PhysRevLett.82.4799ADSView ArticleGoogle Scholar
- Murray AJ, Hammond P: Studies of electron-Excited targets using recoil momentum spectroscopy with laser probing of the excited state. Adv Atom Mol Opt Phy 2001, 47: 163–204. 10.1016/S1049-250X(01)80057-4ADSView ArticleGoogle Scholar
- Schulz GJ, Fox RE: Excitation of metastable levels in helium near threshold. Phys Rev 1957, 106: 1179–1181. 10.1103/PhysRev.106.1179ADSView ArticleGoogle Scholar
- Holt HK, Krotkov R: Excitation of n = 2 states in helium by electron bombardment. Phys Rev 1966,144(1):82–93. 10.1103/PhysRev.144.82ADSView ArticleGoogle Scholar
- Dugan JLG, Richards HL, Muschlitz EEJr: Excitation of the metastable states of helium by electron impact. J Chem Phys 1967,46(1):346–351. 10.1063/1.1840391ADSView ArticleGoogle Scholar
- Pichou F, Huetz A, Joyez G, Landau M, Mazeau J: Electron impact excitation of helium: absolute differential cross sections of the n = 2 and 3 3 S states from threshold to 3–6 eV above . J Phys B 1976,9(6):933–944. 10.1088/0022-3700/9/6/017ADSView ArticleGoogle Scholar
- Brunt JNH, King GC, Read FH: A study of resonance structure in helium using metastable excitation by electron impact with high energy resolution. J Phys B 1977,10(3):433–448. 10.1088/0022-3700/10/3/012ADSView ArticleGoogle Scholar
- Zajonc AG, Pearl JC, Zorn JC: Differential cross section for electron impact excitation of metastable helium. Phys Rev A 1978,1(4):1408–1414. 10.1103/PhysRevA.18.1408ADSView ArticleGoogle Scholar
- Hoshino M, Kato H, Tanaka H, Bray I, Fursa DV, Buckman SJ, Ingólfsson O, Brunger MJ: Benchmark differential cross sections for electron impact excitation of the n = 2 states in helium at near-ionization-threshold energies. J Phys B 2009, 42: 145202.Google Scholar
- Fursa D, Bray I: Calculation of electron-helium scattering. Phys Rev A 1995,52(2):1279–1297. 10.1103/PhysRevA.52.1279ADSView ArticleGoogle Scholar
- Bartschat K: Electron-impact excitation of helium from the 1 1 S and 2 3 S states . J Phys B: At Mol Opt Phys 1998, 31: 469–476. 10.1088/0953-4075/31/10/005ADSView ArticleGoogle Scholar
- Lange M, Matsumoto J, Lower J, Buckman S, Zatsarinny O, Bartschat K, Bray I, Fursa D: Benchmark experiment and theory for near-threshold excitation of helium by electron impact. J Phys B 2006,39(20):4179–4190. 10.1088/0953-4075/39/20/016ADSView ArticleGoogle Scholar
- Read FH, Channing JM: Production and optical properties of an unscreened but localized magnetic field. Rev Sci Instrum 1996,67(6):2372–2377. 10.1063/1.1147004ADSView ArticleGoogle Scholar
- Cubric D, Mercer DJL, Channing JM, King GC, Read FH: A study of inelastic electron scattering in He covering the complete angular range from 0 to 180°. J Phys B 1999, 32: 45–50. 10.1088/0953-4075/32/3/027ADSView ArticleGoogle Scholar
- Allan M: Excitation of the 2 3 S state of helium by electron impact from threshold to 24 eV: measurements with the ‘magnetic angle changer’ . J Phys B 2000, 33: L215-L220. 10.1088/0953-4075/33/6/104ADSView ArticleGoogle Scholar
- Ward R, Cubric D, Bowring N, King GC, Read FH, Fursa DV, Bray I, Zatsarinny O, Bartschat K: Differential cross sections for electron impact excitation of the n = 2 states of helium at intermediate energies (80, 100 and 120 eV) measured across the complete angular scattering range (0 to 180°). J Phys B 2011, 44(4):045209.Google Scholar
- Murray AJ, Hammond P: A novel spectrometer combining laser and electron excitation and deflection of atoms and molecules. Rev Sci Instrum 1999, 70: 1939–1950. 10.1063/1.1149692ADSView ArticleGoogle Scholar
- Feigerle CS, Pierce DT, Seiler A, Celotta RJ: Intense source of monochromatic electrons: photoemission from GaAs. Appl Phys Lett 1984,44(9):866–868. 10.1063/1.94960ADSView ArticleGoogle Scholar
- Schröter CD, Rudenko A, Dorn A, Moshammer R, Ullrich J: Status of the pulsed photoelectron source for atomic and molecular collision experiments. Nucl Instrum Meth A 2005,536(3):312–318. 10.1016/j.nima.2004.08.090ADSView ArticleGoogle Scholar
- Groves T, Hammond DL, Kuo H: Electron-beam broadening effects caused by discreteness of space charge. J Vac Sci Technol 1979,16(6):1680–1685. 10.1116/1.570272ADSView ArticleGoogle Scholar
- Kolac U, Donath M, Ertl K, Liebl H, Dose V: High-performance GaAs polarized electron source for use in inverse photoemission spectroscopy. Rev Sci Instrum 1988, 59(9):1933. Kolac U, Donath M, Ertl K, Liebl H, Dose V: High-performance GaAs polarized electron source for use in inverse photoemission spectroscopy. Rev Sci Instrum1988, 59(9):1933.Google Scholar
- Blakemore JS: Semiconducting and other major properties of gallium arsenide. J Appl Phys 1982,53(10):R123-R181. 10.1063/1.331665ADSView ArticleGoogle Scholar
- Toennies JP, Winkelmann K: Theoretical studies of highly expanded free jets: influence of quantum effects and a realistic intermolecular potential. J Chem Phys 1977, 66(9):3965. Toennies JP, Winkelmann K: Theoretical studies of highly expanded free jets: influence of quantum effects and a realistic intermolecular potential. J Chem Phys1977, 66(9):3965.Google Scholar
- Haberland H, Buck U, Tolle M: Velocity distribution of supersonic nozzle beams. Rev Sci Instrum 1985,56(9):1712–1716. 10.1063/1.1138129ADSView ArticleGoogle Scholar
- Gabriel AH, Heddle DWO: Excitation processes in helium. P Roy Soc A 1960,258(1292):124–145. 10.1098/rspa.1960.0178ADSView ArticleGoogle Scholar
- Moos HW, Woofworth JR: Observation of the forbidden 2 3 S 1 → 1 1 S 0 spontaneous emission line from helium and measurement of the transition rate . Phys Rev Lett 1973,30(17):775–778. 10.1103/PhysRevLett.30.775ADSView ArticleGoogle Scholar
- Van Dyck RS, Johnson CEJr, Shugart HA: Radiative lifetime of the metastable 2 1 S 0 state of helium . Phys Rev Lett 1970,25(20):1403–1405. 10.1103/PhysRevLett.25.1403ADSView ArticleGoogle Scholar
- Pretzler G: A new method for numerical Abel-inversion. Z Naturforsch 1991, 46a: 639–641.ADSMathSciNetGoogle Scholar
- Killer C: Abel inversion algorithm. Published 2013-09-26, Updated 2014-02-18., [http://www.mathworks.com/matlabcentral/fileexchange/43639-abel-inversion-algorithm]
- Bracewell RN: The Fourier Transform and Its Applications. McGraw-Hill Higher Education, Singapore; 2000.Google Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License(http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.