Commercial supercontinuum light sources are becoming more widely used in instrumentation and measurement applications. They use intense pico- or femto-second laser pulses at high pulse repetition rate (i.e., ~1-100 MHz) coupled into a photonic-crystal fiber to produce a broadband optical continuum through highly nonlinear frequency conversion processes. The photonic-crystal fibers confine the light to small cross-sectional areas over large distances and have engineered dispersion to optimize the non-linear processes required for broadband generation [1
1. J. M. Dudley and G. Genty, “Supercontinuum light,” Phys. Today 66(7), 29 (2013). [CrossRef]
2. R. R. Alfano, ed., The Supercontinuum Laser Source, 2nd Ed. (Springer-Verlag, 2006).
]. The output spectra can cover most of the solar spectrum with a spectral range of 400 through 2400 nm with some sources only extending down to ~500 nm. However one should note that the shorter and longer wavelengths of the source’s spectral range typically have fairly low spectral power density. The optical output can have multi-watt average power and is emitted from a single-mode fiber with a circular and usually non-polarized beam pattern although some sources may use a birefringent photonic crystal fiber for polarized output [3
3. C. Xiong and W. J. Wadsworth, “Polarized supercontinuum in birefringent photonic crystal fibre pumped at 1064 nm and application to tuneable visible/UV generation,” Opt. Express 16(4), 2438–2445 (2008). [CrossRef] [PubMed]
]. It can be collimated, yielding a laser-like beam if chromatic aberration can be reduced or avoided. Typical applications either use the full bandwidth of the beam or use switchable or tunable spectral filters to sequentially select specific wavelengths [4
4. P. S. Johnston and K. K. Lehmann, “Cavity enhanced absorption spectroscopy using a broadband prism cavity and a supercontinuum source,” Opt. Express 16(19), 15013–15023 (2008). [CrossRef] [PubMed]
6. N. Sharma, I. J. Arnold, H. Moosmüller, W. P. Arnott, and C. Mazzoleni, “Photoacoustic and nephelometric spectroscopy of aerosol optical properties with a supercontinuum light source,” Atmos. Meas. Tech. 6(12), 3501–3513 (2013). [CrossRef]
]. The two common collimation techniques used for the broad band sources are an achromat doublet that is designed to have the same focal point for two wavelengths or an off-axis parabolic mirror that is free from chromatic aberration [7
7. W. J. Smith, Modern Optical Engineering, 4th Ed. (McGraw-Hill, 2007).
This paper investigates the collimation of commercial supercontinuum light sources using a Fianium SC-400-2 light source with a proprietary, permanently integrated, achromatic lens-based collimator and a Fianium SC-400-6-06 light source with an unterminated optical output sealed behind a plate of flint glass that has been collimated with a custom collimator using an off-axis parabolic (OAP) mirror. Beam collimation and quality is investigated by determining spectrally resolved beam profiles at multiple distances from the fiber output.
Knowledge of beam collimation and quality are of the essence for acquiring a supercontinuum light source because the decision between an integrated lens-based collimator and an unterminated output is semi-permanent; the laser must be returned to the manufacturer to remove or add the lens collimator. In this paper, we describe measurements taken and compare the beam properties of a supercontinuum light source utilizing a factory lens-based collimator with that of a light source utilizing an OAP mirror-based collimator.
One of the supercontinuum sources used was a Fianium SC-400-2 that was supplied with a permanently integrated collimator. The collimator was an achromatic doublet lens with 6.25-mm diameter and a focal length of 10 mm. This source emits short pulses (< 10 ps) at a repetition rate of 20 MHz with a total output power of ~2 W. The second supercontinuum source was a Fianium SC-400-06 with an unterminated output that is sealed behind a flint-glass plate to prevent damage to the fiber. This source also emits short pulses (< 10 ps) but at a higher repetition rate of 60 MHz with a total output power of ~6 W. It was collimated using a 90-degree silver-coated OAP mirror with a diameter of 12.7 mm and an effective focal length of 15 mm. This is the OAP mirror used in the Thorlabs RC04 collimator housing. The OAP mirror was placed at 90° relative to the output of the fiber and iteratively adjusted until the beam was visually well collimated along a 2-m length beam path with minimal coma. Both sources were designed to have a spectral range from ~400 nm to greater than 2200 nm.
The supercontinuum source’s one-dimensional (i.e., horizontal) beam profile was characterized by scanning a 75-μm diameter pinhole across the beam and measuring the spectral power density transmitted through this pinhole. This was done by using a sphere assembly consisting of an integrating sphere with a pinhole as the entrance aperture and a detector fiber-coupled to the exit aperture. Integrating spheres are ideal for power measurements as they homogenize spatial, directional, and polarization distribution of the light entering through the entrance aperture and in addition reduce the power incident on the detector [8
8. A. Carrasco-Sanz, S. Martín-López, P. Corredera, M. González-Herráez, and M. L. Hernanz, “High-power and high-accuracy integrating sphere radiometer: design, characterization, and calibration,” Appl. Opt. 45(3), 511–518 (2006). [CrossRef] [PubMed]
]. The sphere assembly was scanned across the collimated beam by attaching it to a translation stage with 0.01-mm spatial resolution. An ASD FieldSpec® 3 Max spectroradiometer was used to obtain a measurement of the relative spectral power density exiting the pinhole by attaching its fiber input to the exit aperture of the integrating sphere. The spectroradiometer had a wavelength range from 350 nm to 2500 nm with a spectral resolution of 3 nm at the wavelength of 700 nm and 10 nm at wavelengths of 1400 nm and 2100 nm. The experimental setup is shown in Fig. 1
Fig. 1 Experimental setup including supercontinuum source, spectroradiometer, collimator assembly (A); integrating sphere with 75-μm diameter pinhole attached (B), and linear translation stage (C). The linear translation stage is used to scan the pinhole across the beam, thereby acquiring power spectral density beam profiles.
. The sphere assembly was moved perpendicular across the beam in 0.1-mm increments, with a measurement of power spectral density, for wavelengths λ
from 400 nm to 2100 nm, being performed for each increment. For wavelengths shorter than 400 nm and longer than 2100 nm the power density was too low to make a reasonable measurement. These scans were repeated at multiple distances from the collimator in order to characterize beam collimation at 50-nm steps across the wavelength range. A line drawing of the experimental setup is shown in Fig. 1
For each full scan the radiance measured was plotted against pinhole position for each wavelength (Figs. 2
Fig. 2 Horizontal radiance profile of collimated beam from lens-based collimator through 75-μm diameter pin hole at a wavelength of 1250 nm normalized and centerd at zero. (A) Profile measured at a distance of ~53 mm from the collimator. (B) Profile measured at a distance of ~307 mm from the collimator (note the slight deveation from a Gaussian profile). (C) Profile measured at a distance of ~561 mm from the collimator (note the two spikes in the profile). (D) Profile measured at a distance of ~815 mm also displaying two spikes.
Fig. 3 Horizontal radiance profile of collimated beam from off-axis parabolic mirror through 75-μm diameter pin hole at a wavelength of 1250 nm normalized and centerd at zero. (A) Profile measured at a distance of ~110 mm from the center of the parabolic mirror. (B) Profile measured at a distance of ~364 mm from the center of the parabolic mirror. (C) Profile measured at a distance of ~491 mm from the center of the parabolic mirror. (D) Profile measured at a distance of ~745 mm from the center of the parabolic mirror.
). To retrieve the beam waists at the distances from the collimators shown in Tables 1
Table 1. Gaussian Beam Waist for Lens-based Collimator
| | |
Table 2. Gaussian Beam Waist for OAP Mirror-based Collimator
| | |
, these experimental beam profiles were fitted with a Gaussian function using the Levenberg–Marquardt method as
) is the beam intensity, x
is the position of the translation stage perpendicular to the laser beam propagation axis, A
is the area under the curve, w
is the beam waist radius at 1/e2
of the fitted peak intensity, and x0
is the peak position for the Gaussian curve. The uncertainty shown in the tables is the standard error (σ) of the parameter estimates reported from the fit. Representative beam profiles are shown for the lens-based collimator and the OAP mirror-based collimator in Figs. 2
, respectively. Figure 2
shows the horizontal profile of the beam from the lens-based collimator at the wavelength of 1250 nm at multiple distances from the collimator. It can be seen that for distances less than 0.3 m from the collimator the horizontal profile is Gaussian. However, at a distance of ~0.3 m, the profile begins to show signs of deviation from Gaussian. At a distance of ~0.5 m, the profile shows very distinctive, symmetric spikes. For the beam from the lens-based collimator, these features were prevalent at all wavelengths and became more distinct and larger in magnitude with increasing distance from the collimator, that is in the far field. The sharpness of the spikes and the fact that they appear in the far field indicate that they were not side lobes due to beam truncation by the lens but that the collimation lens imaged irregularities at the fiber termination into the far field. As we have characterized only a single supercontinuum light source with lens-based collimator, we do not know if these spikes are due to an imperfection of the fiber termination of the specific light source used in our study, or if they are systemic to supercontinuum light sources with lens-based collimators.
The horizontal radiance profile of the beam from the OAP mirror-based collimator is shown in Fig. 3
for a wavelength of 1250 nm and multiple distances from the collimator. Unlike the beam from the lens-based collimator, the beam from the OAP mirror-based collimator did not exhibit the spikes and remained Gaussian in shape at all wavelengths and distances.
To evaluate the collimation of the beams from the lens-based and OAP mirror-based collimator, it was assumed that the fiber is single mode at all wavelengths and thus the beam quality factor M2
is approximately equal to 1. With this assumption, the measured beam waist (see Tables 1
) was plotted as function of distance from the collimator and fitted with Eq. (2)
weighing (Figs. 4(a)
Fig. 4 (A) Gaussian beam waist radius plotted as a function of distance from the collimator for the lens-based collimator. (B) Gaussian beam waist plotted as a function of distance from the collimator for the OAP mirror-based collimator.
and 4 (b)
is the distance from the collimator, and w0
is the spot radius located at z0
. The results for the fitting parameters are shown in Tables 3
Table 3. Propagation Parameters for Lens-based Collimator
| | |
Table 4. Propagation Parameters for OAP Mirror-based Collimator
| | |
. Also evaluated was the far field divergence angle calculated from the spot size using Eq. (3)
is the half-angle spread of the 1/e2
amplitude of the peak intensity and ZR
is the Rayleigh length. From Table 2
and Fig. 5
Fig. 5 The far-field divergence angles of the beams as function of wavelength. Shown as black dots is the far-field divergence angle for the lens-based collimator; notice the two minima around 550 nm and 2100 nm. The red squares show the far-field divergence angle for the OAP mirror-based collimator. The OAP mirror-based collimator divergence angles are much smaller than those of the lens-based collimator for most wavelengths.
it can be seen that the beam from the OAP mirror-based collimator is fairly well collimated at all wavelengths with a beam divergence ranging from 0.14 to 0.7 mrad. However the beam from the lens-based collimator shows apparent chromatic aberration, with reasonable collimation in the visible at wavelengths of 500 and 750 nm, maximum beam divergence around 1400 nm, and becoming reasonably collimated again around 2100 nm. We infer from theseobservations that the chromatic focal shift for the achromat doublet is near zero around 550 nm and 2100 nm. The apparent chromatic aberration for the OAP mirror-based collimator shown in Fig. 4(b)
and in Table 3
is likely caused by the imperfections in mirror alignment.
We have characterized beam quality and collimation of commercially available supercontinuum sources with a commercial lens-based collimator and an off-axis parabolic mirror-based collimator. Our results show that the lens-based collimator produces a beam with significant chromatic aberration especially in the near-infrared, further away from the reasonably well-collimated visible wavelengths. Symmetric spikes surrounding the Gaussian peak from the lens-based collimator were observed in the far field. We speculate that these spikes may be due to an imperfection of the fiber termination (imaged into the far field) of the specific laser used in our study. The OAP mirror-based collimator produced a Gaussian beam at all distances and wavelengths characterized with good and fairly uniform collimation at all wavelengths, largely devoid of chromatic aberration.
In conclusion, we recommend supercontinuum light sources with simple lens-based collimators if the user application requires a reasonably well-collimated beam in the visible spectrum. However, if the application requires good collimation and uniform Gaussian beams over the whole spectrum, the use of achromatic OAP mirror-based collimators is of the essence.