Patent Grant
7800818
The invention relates to the field of super continuum light generation in microstructure fibres. In particular, this invention relates to a super continuum source configured to produce light within a selectable bandwidth blue shifted relative to the pump wavelength in microstructure fibres by trains of Tera Hertz repetition rate pulses.
Super continuum (SC) generation is a nonlinear phenomenon characterised by dramatic spectral broadening of intense light pulses passing through a nonlinear material. SC generation occurs in various media and finds use in numerous applications ranging from spectroscopy to ultra-short-pulse generation. Inter alea, optical radar and ranging (LIDAR), spectroscopy, optical computing, and reaction rate studies. Spectral slicing of the generated SC is the main mean to design such multi wavelength optical sources. The presently available SC sources hold a spectral density below 0.1 mW/nm (−10 dBm/nm).
The newly developed micro structured fibres (MF) possess unique optical properties which allow generating SC with a broader bandwidth than what is possible in bulk silica or in standard optical fibres. Since the first report on SC generation in a MF in 1999 by Ranka et al. (Optics Letters, Vol. 25, no. 1, (2000), pp. 25-27) comprehensive efforts have been made to understand the physical mechanisms leading to the generation of light with a broad spectrum in this type of fibres, and an extensive literature has been published on the subject. The bulk part of these studies utilises femto-second pulses (10−15 s) to generate the SC. The physical mechanism responsible for the SC generation is believed to be the creation and fission of higher order solitons according to J. Herrmann et al., Phys. Rev. Letters, Vol. 88, No 17, 2002. It has also been shown that it is possible to create SC by use of pico- and nanosecond pulses, and the mechanism responsible for these SCs are believed to be a combination of four wave mixing and stimulated Raman scattering (Coen et al., Optics Letters, Vol. 26, (2001), pp. 1356-1358, and Town et al. Applied Physics B (Lasers and Optics), vol. B77, no. 2-3, September 2003, pp. 235-238). The possibility of tailoring the properties of MFs for improving the efficiency of SC light generation using pico- or nanosecond pulses has, however, been little explored. The use of longer pulses is, however, attractive as it does not require a complex and expensive femto second laser. This has so far been the main obstacle to the creation of commercially viable SC light sources.
The spectral slicing of a SC only utilizes a small part of the launched energy. This energy is symmetrically distributed around the pump and primarily generated through a four wave mixing process or alternatively red shifted relative to the pump when stimulated Raman scattering dominates the generation process in the case of inefficient phase matching of the four wave mixing process. The hereby generated blue light will be limited to the half pump wavelength due to energy conservation of the four wave mixing process. Here the blue shifted light (idler) is generated through the action of two pump photons and a red shifted (signal) photon. The idler light wavelength generated through the four wave mixing process is determined through the conservation of energy equation: h νidler=2 h νpump−h νsignal<=>1/λidler=2/λpump−1/λsignal, where ν and λ denote frequency and wavelength, respectively, and h is Planck's constant. For the hypothetic situation of the infrared part of the SC extending to infinity the idler wavelength minimum is to be found at the half pump wavelength.
The four wave mixing or Stimulated Raman Scattering will hereby either require considerable pump energy or unattractive short pump wavelength when generating light in the near infrared (760-1300 nm), visible (400 nm-760 nm) and/or at UV wavelengths (<400 nm) and cannot generate light below a wavelength of λpump/2. In prior art SCs shown in
In prior art SC from Price et al. (Optics Express Vol. 10., No. 8, Mar. 20, 2002) shown in
Thus, there is a need for a light source providing a spectrum extending below λpump/2 with a spectral density exceeding −10 dBm/nm.
In prior art SC generated by ps or ns pulses the power intensity of the red shifted part measured in mW/nm is equal to or larger than the intensity of the blue shifted part. Examples of such generated spectra are shown in
In the following a new method for generating light is presented. This method can advantageously be applied to yield a new SC source that can be managed to have a substantial part of its output in the UV, visible or near infrared when a pump wavelength in the wavelength range of for example 900-1100 nm is used.
The invention teaches that a pulsed pump source with substantially constant pulse peak amplitude propagating through a dispersive medium, will undergo amplitude modulation, with the modulation depth growing exponentially with time, provided that the pump wavelength lies in a region of anomalous dispersion of the transmission medium. This amplitude modulation means that part of the energy in the pump has been shifted into sidebands. The generated sidebands are the result of modulation instability gain that exists near the pump frequency.
Upon launch of a high peak power pulse into a fibre the Kerr effect will contract the pulse which in the frequency domain corresponds to broadening of the spectrum. Once the spectrum is broad enough to cover the maximum modulation instability gain this takes over and breaks the pulse into a pulse train of short pulses of Tera hertz repetition rate (1012 Hz). A positively chirped pulse will hereby help the contraction to meet the maximum gain at the earliest possible point and hereby initiate the modulation instability gain takeover at the earliest possible point.
The modulation instability gain takes over the process and sidebands are generated at the angular frequencies Ωmax determined by equation 1:
where γ, Ppeak and β2 are explained below. The generated sidebands will, if the pump power is sufficiently strong, generate their own sidebands through a modulation instability gain given that the sidebands are generated at a frequency with anomalous dispersion. The repetition rate of the generated pulse train is determined by Ωmax/2π. Now, having generated a train of short pulses the formation of solitons through the interplay between self phase modulation (SPM, the Kerr effect) and anomalous dispersion of the fibre can take place. I.e. the dispersion length of the fibre LD=T02/β2 has become of comparable size to the nonlinear length LNL=1/(γPPeak) of the fibre due to the formation of the train of short pulses. The self phase modulation generates a frequency chirp such that the leading edge of the soliton is red-shifted while its tailing edge is blue shifted from the central frequency. The anomalous dispersion contracts the pulse as the pulse is positively chirped by the SPM. These two effects will for certain pulse durations and peak power levels balance each other out and a soliton is created. The soliton order number N is determined by equation 2:
Here PPeak is the launched peak power, T0 is the soliton duration (which is equal to TFWHM/1.763, where TFWHM is the Full Width Half Maximum duration for a Gaussian pulse), and the second order dispersion parameter β2=−λ2/(2πc)D, where D=d/dλ(1/vg) is the group velocity dispersion and vg is the group velocity, and c is the speed of light in vacuum, and γ is the nonlinear parameter given by equation 3:
Here n2 is the nonlinear refractive index of the fibre material and Aeff is the effective mode area of the fibre.
Now, if the modulation instability gain of the pump or one of the generated sidebands of the pump holds an overlap with a zero dispersion point, soliton formation will primarily take place in the vicinity of the zero dispersion point of the fibre due to the very low threshold for soliton formation. The fundamental soliton (N=1) can be created with very low peak power as the second order dispersion parameter is zero at the dispersion zero point. Here the higher order terms of the dispersion will take over the process and determine the threshold for soliton formation (β2(ω)=β2(ω0)+(⅓) β3(ω0)(ω−ω0)+( 1/12)β4(ω0)(ω−ω0)2. Solitons will also form from the sidebands (the maximum gain) or directly under the pump pulse (at the pump wavelength)—however, the generation process is less efficient as the creation power is considerable higher as compared with the zero dispersion point. Despite this fact it is the solitons created at the pump pulse wavelength that will dominate the SC generation process.
The light generation by solitons can be understood through the following teaching. The launched quasi-CW pump pulse can be regarded as a light particle (“light bullet”) travelling through the glass fibre with a given speed (group velocity). If the quasi_CW pump pulse group velocity exceeds the linear phase velocity of a given wavelength, a dispersive wave can be created. Explained in another way the electrons in the glass are displaced and polarized by the passing quasi-CW pulse photons. Some of these glass molecule electrons are excited to states above their ground state (by multiple photon excitation). Photons are re-emitted as the electrons in the glass restore to equilibrium during and after the quasi-CW pulse has passed. Under normal circumstances the re-emitted photons destructively interfere with each other and no radiation is detected.
The modulation instability gain has however transformed part of the energy in the quasi-CW pulse into a Tera hertz pulse train of solitons. As these solitons gain energy they contract in time and start to impose Raman amplification to themselves which leads to red shifting of the soliton central wavelength.
Where the red shifted soliton phase match with the randomly generated energy of the quasi-CW pulse energy it will be re-emitted in the blue part of the spectrum. This takes place where the quasi-CW pulse travels faster than the re-emitted photons, the re-emitted photons constructively interfere and intensify the radiation given that the soliton pulse and the re-emitted photons match in phase. The result is genuine Cherenkov radiation and the emitted light does not have to belong to the soliton spectrum as the soliton only seeds the re-emission process rather than to take part of it directly with its energy.
The light bullets will generate Cherenkov radiation if phase match between a dissipative mode in the UV or blue part of the spectrum exists for the given microstructure fibre medium. This phase match AK is determined by equation 4:
Δκ=β(ωd)−(β(ωbullet)+γP0) (4)
Where β(ωd) and β(ωbullet) are the phase of the dissipative wave and the soliton with the peak power P0, respectively, and γ is the nonlinear parameter defined in equation (3).
The intensity of the generated Cherenkov radiation will be determined by the soliton self frequency shift of the light bullet. If the soliton self frequency shift of the light bullet is small, the central wavelength of the light bullet will remain for a longer period near the zero dispersion point and the generated Cherenkov radiation will increase in intensity as a function of propagation distance. The Cherenkov radiation is fed with energy from the soliton which again is fed with energy by the pump through the modulation instability gain. If the self frequency shift is large for the generated light bullet no significant level of Cherenkov radiation will be generated. The latter situation is what can be observed in prior art experiments with pico- or nanosecond pulse generation of SCs as a weak pilot beam blue shifted to the generated SC. No attention has, however, been paid to this pilot beam in the pico- or nanosecond pulse generation. In the Price et al. (Optics Express Vol. 10., No. 8, Mar. 20, 2002) the blue extended spectrum is obtained as a result of such a phase match but no explanation was given for the observed spectrum and no route is given to describe how a higher spectral density is to be obtained.
The soliton self frequency shift Ω(z) at a distance z is given for the fundamental soliton by equation 5:
Here TR is the Raman parameter—with a relaxation time in the 2-4 fs range.
Particularly intense radiation can be achieved in situations where the soliton central frequency is stabilized by the action of a negative dispersion slope as known from Skryabin et al. (Science Vol. 30, pp 1705-1708, 19 Sep. 2003). Here intense radiation with wavelength in the vicinity of the stabilized solitons near the second (red shifted) zero dispersion wavelength is expected.
In the present invention it is, however, not energy from the soliton that leads to generation of the blue part of the spectrum but merely the small perturbation of the electrons in the material caused by the soliton that seeds a phase match for energy exhange between the quasi-CW pulse and the Cherenkov radiation. It is observed that solitons stabilize at the negative dispersion slope from the infrared absorption edge, for a silica material this is observed at a wavelength near 2220 nm (4500 cm−1), here termed the ‘infra red absorption band edge stabilization point’. As long as the stabilized soliton holds an overlap with the quasi-CW pump pulse exchange of energy takes place.
An objective of the invention is to provide a scheme for extending the spectrum beyond (i.e. below) the half pump wavelength border (λpump/2) normally defined by the four wave mixing process. It will hereby simultaneously be possible to use a pump wavelength in the 900 nm-1100 nm range and generate light at UV or visible wavelengths with spectral density above −10 dBm/nm (>0.1 mW/nm).
Another objective of the present invention is to provide a light source with an increased power intensity of the blue shifted light relative to the pump wavelength compared with the red shifted light relative to the pump wavelength with spectral density above −10 dBm/nm (>0.1 mW/nm).
The present invention relates to a blue extended super continuum light source comprising: a pump laser which operates at a wavelength λpump and produces pulses of a duration longer than 0.2 picoseconds with a repetition rate higher than 1 kHz, and a micro structured optical transmission medium adapted to transmit radiation at λpump the medium having at least one wavelength region of anomalous dispersion, and the pump wavelength λpump is chosen to lie within the region of anomalous dispersion of the transmission medium thereby breaking said pump pulses into trains of Tera Hertz repetition rate pulses. Hereby efficient generation of radiation is achieved such that the power in the blue shifted part of the spectrum is larger than the power in the red shifted part of the spectrum, both relative to the pump wavelength:
where the multiplication factor M≧1, and λpump−λ1 defines the low wavelength border of the generated light as: I(λpump−λ1)/I(λpump+λ1)≦0.01, and λpump−λ2 defines the high wavelength border of the generated light, where 0<λ2<λ1λpump, wherein a low wavelength border λpump−λ1 of the generated blue shifted light is below the half pump wavelength, such that 2(λpump−λ1)<λpump<=>λpump<2λ1 and the intensity of at least part of the generated radiation I(λ)>−25 dBm/nm, where λ1<λ<λ2.
It has surprisingly turned out that spectra generated with very different excitation conditions holds similar spectral distribution of the energy while holding the peak power of the pump pulses constant.
In an aspect of the invention, a blue extended super continuum light source is provided, the light source comprising:
Some of the characteristics of a super continuum spectrum I(λ) according to the invention (including the above mentioned wavelengths λ1, λ2 and λpump) are schematically indicated in
An advantage of the present invention is the possibility to use longer pump pulses, which is attractive as it does not require a complex and expensive femto second laser.
This has so far been the main obstacle to the creation of commercially viable SC light sources.
The term ‘the duration of pulses’ is in the present context taken to mean the Full Width Half Maximum duration for Gaussian pulses of essentially constant peak power.
The term ‘a micro structured optical transmission medium’ is in the present context taken to mean a medium comprising a photonic crystal fibre, also known as microstructured fibres or holey fibres, cf. e.g. T. A. Birks, J. C. Knight, and P. St. J. Russell, “Endlessly single-mode photonic crystal fibre,” Optics Letters, vol. 22, pp. 961-963 (1997). Various aspects of the design, properties, manufacturing and applications of microstructured optical fibres are e.g. discussed in Bjarklev, Broeng, and Bjarklev in “Photonic crystal fibres”, Kluwer Academic Press, 2003 ([Bjarklev et al.]).
The term ‘anomalous dispersion’ is in the present context taken to indicate a dispersion coefficient larger than 0 (such as larger than +2 ps/nm/km, e.g. +5 ps/nm/km).
In a particular embodiment, the power in the blue shifted part of the spectrum is larger than the power in the red shifted part of the spectrum, both relative to the pump wavelength:
where the multiplication factor M≧1, and:
I(λpump−λ1)≦−15 dBm, where 0<λ2<λ1<λpump.
In an embodiment, the low wavelength border λpump−λ1 of the generated blue shifted light is below the half pump wavelength, such that 2(λpump−λ1)<λpump<=>λpump<2 λ1.
In an embodiment, I(λpump−λ2)/I(λpump+λ2)≦0.1, such as <0.05, such as <0.01.
By placing the pump wavelength in the anomalous dispersion region of the microstructure fibre (cf.
Here Ω=(ω−ωpump), where ωpump=2πc/λpump and ΩC2=4γPPeak/|β2|.
The creation of blue shifted light relative to the pump wavelength by excitation of Cherenkov radiation is an attractive alternative to the before mentioned processes as it generates light from the launched pump energy primarily in the bands where the light is wanted when the fibre is dispersion managed relative to the pump wavelength and pulse peak power.
To obtain a substantial increase in the blue part of the spectrum, the quasi-CW pulse duration is to be tailored to the available fibre. If a high intensity Cherenkov radiation line is wanted the Raman solitons are to be shifted to the infrared absorption edge (for silica-based fibres, near 2220 nm) as fast as possible by the soliton self frequency shift. This will lead to the smallest amount of energy shed to other wavelengths than the in this situation wanted narrow band of wavelengths near the phase matching between the Raman soliton at the infrared absorption edge and the Cherenkov radiation.
The fast soliton self frequency shifting can be achieved by choosing a fibre with a high dispersion (β2) (e.g. lager than 10 ps/(nm·km)) at all wavelengths longer than or equal to the pump wavelength and/or by increasing the launched peak power to achieve a shorter Raman soliton duration (T0) according to equation (5).
Attention has to be paid to that the group velocity vg(q-CWp) of the quasi-CW pulse needs to exceed the group velocity vg(Cherenkow) of the wavelengths where the Cherenkov radiation is wanted.
A wider band with high spectral density Cherenkov radiation can be achieved when the soliton self frequency shifting is held at a moderate level (e.g. β2 in the range from 5 to 10 ps/(nm·km)) such that the soliton rests for a longer stretch of fibre at a given central wavelength. Here again attention has to be paid to the fact that the group velocity of the quasi-CW pulse needs to exceed the group velocity of the wavelengths where the Cherenkov radiation is wanted (vg(q−CWp)>vg(Cherenkow)).
To reach attractive spectral density levels substantial pump pulse energy is required. This is done by increasing the pulse energy of individually launched pulses and/or their repetition rate, while maintaining the peak power level of the launched pulses.
To find the condition given by equation 4 it is required to determine the group velocity dispersion of the fibre of interest to determine the dispersion as well as the dispersion slope (i.e. β2 and β3=dβ2/dω) and preferentially even higher order terms. A white light interferometer measuring method for this purpose is disclosed in ECOC 2002 paper 3.4.2 by Andersen et al. “A photonic crystal fiber with zero dispersion at 1064 nm”.
Further, dependent on the actual wanted spectrum, the fibre parameters have to be chosen accordingly. One example of this is shown in connection with example 1.
The spectrum of generated radiation may go beyond the four wave mixing limit. This is determined by the fibre parameters that have to be chosen such that phase match between the dissipative wave (Cherenkov radiation) and the generated solitons yield radiation at such wavelengths beyond the four wave mixing limit.
In some applications it may be useful that the wavelength range of the generated radiation extends up to the pump wavelength. In these cases the wavelength parameter λ2 is zero.
The duration of the initially launched pulse should hold a length which makes it viable to initiate the modulation instability gain. If the pulse duration is too short this instability gain may not come into play as the dispersion of the fibre will counteract the initial pulse broadening of the spectrum set by the self phase modulation of the pulse. The initially launched pulse should therefore be kept above a minimum duration of 0.1-0.2 picoseconds, e.g. above 0.15 ps such as above 0.2 ps. Going below this duration will lead to a less efficient generation of a super continuum by the modulation instability. As the cost price of a laser system among other things is dependent on the pulse duration it is preferred that the duration of the initial launched pulses is larger than 0.25 ps, such as larger than 0.5 ps, such as larger than 1.0 ps, such as larger than 5 ps, such as larger than 10 ps, such as larger than 50 ps, such as larger than 1 ns, such as larger than 2 ns, such as larger than 10 ns.
It is an advantage that as much as possible of the generated light goes into the blue part of the spectrum as the SC upon creation is subject to a spectral slicing.
It is hereby advantageous that the multiplication factor M is larger than 1.2, such as larger than such as larger than 1.5, such as larger than 2, such as larger than 2.5, such as larger than 3. This can be achieved by increasing the launched peak power of the pump. The level of needed pump peak power is strongly dependent on the actual fibre dispersion and nonlinear coefficients.
Dependent on the wavelength range that is wanted for the specific application the pump laser wavelength is to be chosen longer than 600 nm. This is mainly due to the fact that current technology only allows for microstructure fibres with a zero dispersion wavelength above 550 nm-580 nm. It might, however, be lower than 600 nm, if an appropriate transmission medium were available. Given that the pump wavelength has to be placed in a region with anomalous dispersion this sets the lower limit of the applicable laser wavelength to about 600 nm. It is however, advantageous to choose a longer wavelength such as in the range 900 nm to 1300 nm, so that a powerful and still economic laser source becomes available. Lasers with wavelengths in the range of 1000 nm to 1100 nm gives at present the best combination of economy and laser pump power.
The transmission medium may comprise any appropriate optical wave-guiding medium exhibiting an anomalous dispersion and for which a phase match can be produced between the solitons and the dissipative wave. Preferably, the transmission medium comprises a silica based optical fibre, e.g. a micro structure optical fibre. Many different transversal arrangements of micro structure features of the optical fibre of the transmission medium may be used. A very efficient medium for achieving the high nonlinear response and simultaneously to manage the dispersion is found in a microstructure optical silica fibre with a waveguide structure having a longitudinal direction, said microstructure optical fibre comprising: a solid core region extending along the longitudinal direction, and a cladding region extending along the longitudinal direction, said cladding region comprising a triangular hole pattern separated by a pitch A and with a hole diameter d relative to the pitch d/Λ≧0.2. [Bjarklev et al.] describe e.g. dispersion properties (cf. e.g. chapter 5.3.4, pp. 148-151) and the fabrication (cf. e.g. chapter 4, pp. 115-130) of microstructure optical fibres (including triangularly structured fibres).
This microstructure fibre should advantageously hold at least one first zero dispersion wavelength below 1300 nm, such as below 1100 nm, such as below 1064 nm, such as below 1000 nm, such as below 900 nm. The tailoring of the dispersion properties of microstructure fibres is e.g. discussed in WO-02/12931, WO-02/39159, WO-02/088801, WO-02/084350 and WO-03/079074. The choice of zero dispersion wavelength for the fibre depends among other things on the choice of laser source.
Further it is advantageous if the fibre is arranged to support propagation of the wavelength λpump in a single transverse mode (cf. e.g. T. A. Birks, J. C. Knight, and P. St. J. Russell, “Endlessly single-mode photonic crystal fibre,” Optics Letters, vol. 22, pp. 961-963 (1997). For some applications it is advantageous if the fibre is arranged to support propagation of all generated wavelengths in the spectrum of wavelengths in a single transverse mode. This will guarantee that the generated light only is delivered in one single transverse mode.
For those applications that need polarized light it will be an advantage that the microstructure fibre is arranged to be polarization maintaining. (e.g. by introducing an anisotropy in the core region (by form, stress etc., cf. e.g. WO 00/49436 or WO 2005/059612)). This is due to the fact that the generated radiation obtained when using a homogeneous microstructure fibre (with equal longitudinal propagation of the two degenerate states of the ground mode) is not polarized. For a non-polarization maintaining fibre, polarized light can be obtained by placing a polarizer in front of the fibre, however at the expense of half the generated radiation power.
As the output power of the super continuum light source is directly a measure of the needed pump power level, some applications might only need a moderate output power level which is to be achieved with a more economic pump laser. In these cases the generated radiation us advantageously larger than −10 dBm/nm for at least part of said generated radiation. However, as the sensitivity of many applications will gain with a more powerful light source it might be advantageous to increase the output power level of the generated radiation. This to a level such as larger than −5 dBm/nm, such as larger than 0 dBm/nm, such as larger than 3 dBm/nm, such as larger than 5 dBm/nm, such as larger than 10 dBm/nm, such as larger than 12 dBm/nm, such as larger than 15 dBm/nm, all within the wavelength range λ1<λ<λ2 where the intensity level is for at least part of said generated radiation. In one embodiment, ‘at least part of said generated radiation’ is taken to mean λ1−λ2 larger than 0.01 nm, such as larger than 0.05 nm, such as larger than 0.1 nm, such as larger than 0.5 nm, such as larger than 1 nm, such as larger than 5 nm, such as larger than 10 nm, such as larger than 50 nm, such as larger than 100 nm, such as larger than 300 nm, such as larger than 500 nm.
An article comprising a super continuum light source as described above, in the detailed description and in the claims is furthermore provided. The article may e.g. be optimized for use in applications such as spectroscopy, ultra-short-pulse generation, optical radar and ranging (LIDAR), optical computing, reaction rate studies, etc.
In a further aspect of the invention a method of manufacturing a blue extended super continuum light source is provided, the blue part of the generated spectrum substantially extending from a low wavelength border λpump−λ1 to a high wavelength border λpump−λ2, the method comprising:
In a particular embodiment, the power in the blue shifted part of the spectrum is larger than the power in the red shifted part of the spectrum, both relative to the pump wavelength:
where the multiplication factor M≧1, and:
I(λpump−λ1)≦−15 dBm, where 0<λ2<λ1<λpump.
The method provides the same advantages as the corresponding product. The features of the product as described above and in the detailed description below and in the claims are intended to be used in combination with the method, where appropriate.
Further objects of the invention are achieved by the embodiments defined in the dependent claims and in the detailed description of the invention.
It should be emphasized that the term “comprises/comprising” when used in this specification is taken to specify the presence of stated features, integers, steps or components but does not preclude the presence or addition of one or more other stated features, integers, steps, components or groups thereof.
In the following, the invention will be explained in more details with reference to the New blue extended super continuum light source according to the invention shown in the drawing, in that
In the following, the dimensioning and application of the new blue extended super continuum light source will be discussed in connection with a number of examples.
The microstructure fibre of example 1 consists of a 3.5 μm diameter silica core surrounded by an array of 1.1 μm diameter air holes in a hexagonal close packaged arrangement (also known as a triangular pattern) with a hole centre to centre distance of about 2.44 μm. The fibre is produced by Crystal fibre A/S (of Birkerøod, Denmark) and has the product name NL-3.5-975. Experimentally the fibre appears to be endlessly single mode and insensitive to bend loss up to at least 1700 nm. The fibre holds a zero dispersion wavelength at 974 nm and anomalous dispersion for all wavelengths above this wavelength. The pump laser source applied in the present example is produced by Fianium (of Southampton, UK) and has the product name FEMTOPOWER 1060, operation at 1064 nm and delivering 10 W of average power. The 6 ps pulses with a repetition rate of 80 MHz and a peak power of 16 kW are launched into the microstructure fibre. This generates by the action of modulation instability a pulse train with 32.2 THz repetition rate in the current fibre. This pulse train is responsible for the generated spectrum at full excitation power shown in
The calculated Cherenkov radiation wavelength as function of “light bullet” soliton central wavelength is shown in
Comparing the calculated wavelengths with the generated spectrum shown in
The optimum fibre length has not been determined, however increasing the length of the fibre (from the current 10 m) will be at the expense of increased fibre absorption loss due to the background loss of the fibre of about 0.4 dB/m.
To apply this spectrum as a light source for example in connection with spectroscopy or confocal microscopy, spectral slicing of the spectrum by use of either glass filters or mirrors is an easy way to produce light with the wanted bandwidth. Adjusting the peak power level of the excitation source is one way to adjust the power level of the generated light. Another possible way is through use of neutral density filtering of the spectra.
In this example the fibre of example 1 is subjected to excitation pump pulses with considerably longer pulse duration and launched quasi-cw pulse energy. The generated spectra of a 6 ps launched pulse width of 80 MHz repetition rate and a 2 ns launched pulse width of 28 kHz repetition rate both with a peak power level of 2.5 kW are shown in
Surprisingly, it is to be observed that the two spectra are similar despite the very different excitation conditions. As the two spectra both are generated with equal peak power it is, however, according to the teaching of the present invention, equal modulation instability conditions that is set by the fibre. It will therefore be pulse trains containing individual pulses of equal duration and peak power that is generated by the two different launched pulses. As these individual pulses are substantially equal, the generated Cherenkov radiation will be at substantially the same wavelengths.
It is to be observed that both spectra in
It is worth to notice that the 2 ns excitation pulse will lead to a catastrophic breakdown of the fibre at the fibre entrance when a too high pulse peak power level is launched into the fibre (above 2.7 kW). This is due to the much higher energy per pulse that is available in the 2 ns excitation pulse as compared with the 6 ps excitation pulse. This will lead to that the initial coupling of the pulse to the fibre becomes critical in the nanosecond pulse regime. Typically the pulse energy is in the 1-10 μJ/pulse range for the nanosecond pulses as compared with 1-10 nJ/pulse for the picoseconds pulses.
It is further worth to notice that the output power per nm ratio between the two spectra corresponds well with the launched energy per time period. This such that the 6 ps system holds 1.2 J launched energy per second compared with 0.14 J launched energy per second for the 2 ns system. Comparing the spectral density of the two systems it is noted that the output in the visible is 26 μW/nm for the 2 ns system compared with 360 μW/nm for the 6 ps leading to a 1:14 ratio compared with the launched energy per time ratio of 1:9. The ratio of the launched energy per time frame for the two spectra is substantially identical with the ratio of the output power per nm. The lacking energy of the 2 ns system is believed to be due to the local heating by the long pulse.
To obtain a wavelength range different from the one in example 1 and 2, the fibre dispersion profile has to be managed. One way to perform this is by calculating the expected Cherenkov radiation wavelengths for various dispersion profiles and hereby to determine the wanted dispersion zero point, and dispersion slope and higher order derivates of the dispersion both at the pump wavelength of choice as well as at the infrared absorption edge stabilization point.
Intense blue lines will be found where phase match with the 2220 nm absorption edge position for solitons trapped here. This requires, however that there is a temporal overlap between the generated soliton and the quasi-CW pump pulse once the soliton has been self frequency shifted to the 2220 nm position. Optimal fibre design and quasi-CW pulse parameters (peak power and pulse duration) may preferably be determined by an optimization process, e.g. a trial and error-process.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2005 01010 | Jul 2005 | DK | national |
| Filing Document | Filing Date | Country | Kind | 371c Date |
|---|---|---|---|---|
| PCT/DK2006/050027 | 6/28/2006 | WO | 00 | 1/7/2008 |
| Publishing Document | Publishing Date | Country | Kind |
|---|---|---|---|
| WO2007/006316 | 1/18/2007 | WO | A |
| Number | Name | Date | Kind |
|---|---|---|---|
| 20050117841 | Braun et al. | Jun 2005 | A1 |
| 20070216989 | Nerin et al. | Sep 2007 | A1 |
| 20080226242 | Buchter et al. | Sep 2008 | A1 |
| Number | Date | Country |
|---|---|---|
| WO 0049436 | Aug 2000 | WO |
| WO 0212931 | Feb 2002 | WO |
| WO 0239159 | May 2002 | WO |
| WO 02084350 | Oct 2002 | WO |
| WO 02088801 | Nov 2002 | WO |
| WO 03079074 | Sep 2003 | WO |
| WO 03096490 | Nov 2003 | WO |
| WO 2005059612 | Jun 2005 | WO |
| WO 2005062113 | Jul 2005 | WO |
| Number | Date | Country | |
|---|---|---|---|
| 20090262764 A1 | Oct 2009 | US |
| Number | Date | Country | |
|---|---|---|---|
| 60697389 | Jul 2005 | US |
