A.

Spectral characterization

In order to understand the usability spectral range of Bayfol® HX material, we recorded the transmission spectrum before and after the light exposure to incoherent white light with a fluence of ~ 50 J/cm² (Fig. 2). As can be seen in the figure the behavior in the NIR spectral region is dominated by the polymer substrate. Indeed, the material is transparent up to 2.2 µm, where the infrared overtones of the polyamide start to absorb strongly. It seems that no evident absorptions come from the photopolymer system. In the UV region, the photopolymer shows a steep transition opaque to transparent at 312 nm. Moreover, in the unexposed film, we clearly identify the absorption peaks of the sensitizers that make the material panchromatic. After the bleaching step, some weak residual absorption bands are present in the visible. In general, higher residual absorption occurs increasing the film thickness and this is especially true approaching the UV cut-off. Therefore, a particular attention must be paid to VPHGs working in this region, where the absorption could reduce the final VPHG’s diffraction efficiency.

Fig. 2.

UV-vis-NIR spectra of the Bayfol® TP 16 photopolymeric material before (solid line) and after photo-bleaching (dashed line).

B.

Tunability: Δn as function of power density and beam ratio intensity

As previously anticipated, the most wanted property of a holographic material for volume diffraction grating is the ability to tune the refractive index modulation, in order to assess the spectral requirements that have been defined in the design phase.

The Bayfol® HX photopolymeric material generates the index difference through a reaction-diffusion mechanism, which involves a competing process between the migration of the monomers inside the film and the polymerization of the material itself [10]. The photopolymerization occurs in the exposed regions of the pattern (constructive interference), while due to a concentration gradient, the monomers migrate from the dark to the exposed regions being polymerized themselves.

The photopolymerization of this system depends upon the writing power density [11]. Therefore, increasing the writing power, the monomer conversion occurs with a higher velocity competing with the diffusion mechanism.

For this reason, the maximum achievable index modulation starts to drop.

This behavior finds confirmation in the data presented in Figure 3, which reports the values of Δn as function of the writing power density ranging from 0.143 to 21.5 mW/cm² for samples with increasing line density (600, 1000, 1400 1/mm). The total illumination dose was kept constant at 42 mJ/cm², therefore the exposure times for each sample changed accordingly.

Fig. 3.

Δn as function of the mean value of the intensity of the two writing arms. The total dose was kept constant at 42 mJ/cm² (referred to double beam writing process).

Another peculiar effect that is visible in figure, is the dependence of Δn on the line density: the curves of samples with higher line densities are shifted at larger Δn values. This effect is again easily understandable if we think at the reaction-diffusion mechanism of the photopolymer. An Increased Λ value means a shorter path to travel for the monomer, thus a facilitated process of Δn formation. This is true in the range of line density here considered, for much higher Λ, a Δn cut-off could occur because of the nonlocal effects. It is important to notice that Bayfol® HX photopolymers showed a much smaller nonlocal response parameter, which turns into a larger resolution in comparison with standard AA/PVA photopolymers [12].

We also notice a dependence of the Δn with the wavelength. This could be ascribed to the dispersion of the refractive index with the light frequency; by increasing the frequency, the refractive index increases (normal dispersion). Such increment changes in the exposed and unexposed regions because of the different polarizability. In particular, we could expect a more pronounced rise of refractive index for the exposed regions, with a consequent growth of Δn for shorter wavelengths, as experimentally verified. Moreover, the amount of change is compatible with the chemical composition of the photopolymer system.

Another possible strategy for the Δn tuning is related to the intensity ratio of the two holographic writing beams. It is well known that in order to obtain the largest contrast in the interference pattern, the intensity of the two laser beams must be as equal as possible. This is for sure the best achievable condition for maximizing the refractive index modulation while, if lower Δn values have to be achieved, unbalanced writing intensities can be used.

We have also evaluated this behavior performing writings at different ratios, while maintaining constant the total irradiation dose at 42.8 mJ/cm². The outcome was a strong dependence of the Δn with the intensity ratio occurs, confirming that the best condition for the recoding of transmission VPHGs is the balanced (50:50) beam ratio, but lower values can be addressed precisely when different ratios are set. This strategy is especially useful when very low Δn values are required and high values of laser power cannot be provided to constrain the Δn with the balanced beams.

From these results, it is evident how with this material we can achieve precise Δn values in accordance with the design goals, in a range strictly related to the lines per millimeter of the target VPHG. For example, if we need a diffraction grating with 1400 1/mm and we can deal with powers from 0.143 to 21.5 mW/cm², the achievable Δn can be from 0.026 to 0.039.

A.

ALFOSC devices

The three elements for ALFOSC are characterized by a resolving power of approximately 1000 and they cover the 350-1000 nm spectral range. In Table 1, the specifications of the devices that were named #18, #19 and #20 according to the telescope’s internal numeration are reported.

Table 1.

NOT’s delivered GRISMs main specifications summary: (1) Working central wavelength of the grating; (2) pixel dispersion of the grating; (3) Wavelength range; (4) Resolution, R = λ/Δλ; (5) pitch of the grating; (6) final refractive index modulation of the grating; (7) prism apex angle (the prisms were in BK7).

In the case of the devices #19 and #20, a coloured filter has been embedded to cut the second order spectrum, which had a non-negligible efficiency. For the device #18, devoid of order sorting filter, the coupling fluid OCF-452 has been used instead of cedar oil in order to shift the UV cut-off at shorter wavelengths, granting a high transmittance (60%/mm) even at 330 nm. To completion of the description, we report in Table 2 the diffraction efficiencies of the three transmission GRISMs that we have manufactured for ALFOSC. The measurements were completed in laboratory before the commitment at the telescope and they were performed at different wavelengths. As has been reported, each device possesses a different spectral range, therefore is not possible to present the efficiency data all at the same wavelength. It is evident from the table how each device, for wavelengths inside its own spectral range, possess very high overall efficiencies, that in most cases surpass the 80%.

Table 2.

ALFOSC GRISMs transmission efficiencies at different wavelengths. The values correspond to raw measurements, without reflection losses corrections or materials absorption corrections.

After the integration into the telescope, the GRISMs are compared with similar gratings, in terms of dispersion, spectral range and resolution, which were already available in the spectrograph. In Figure 4, for example, can be seen the spectra of photometric-standard star SP0305+261 taken with one of our device in comparison with the already mounted one. It is evident that the throughput of the VPHGs surpasses the ruled dispersive element in the spectral region of interest. In some cases, (e.g. GRISM #18 vs. #16, Figure 1) a gain of 85% is reached, meaning that the exposition time to have the same S/N (signal to noise ratio) is nearly halved.

Fig. 4.

Comparison of photometric-standard stars’ spectra SP0305+261. Total system throughput of the NOT telescope with Grism #16 and Grism #18. Slitwidth used in the observations is 10.0 arcsec.

The GRISM #20 has been successfully used to observe the source SDSS J004054.65-0915268 which was a candidate for being a high redshift (z = 5.00) bright BLLAC object3 (i = 18.2). With the integration of this new GRISM with extended range, we had the opportunity to confirm the featureless nature of this object from 9000 to 10000 Å, targeting the C IV strong emission line in its near-IR spectrum. In that case this source could be the farthest known BLLACs at that redshift. In any case if its BLLAC nature will not be confirmed it would be anyway an interesting source for spectroscopic studies (weak emission lines QSOs5) [13].

B.

AFOSC devices

In the past decade, the AFOSC instrument already took advantage of the Volume Phase Holographic Grating technology thanks to six newly designed DCG based dispersive elements [15]. Those innovative GRISMs well matched the most important spectral features observable with Asiago’s spectrograph with an improved resolution up to R = 5000. Following the procedure of the NOT’s devices described in the previous section, we designed and produced three new photopolymer based VPHG GRISMs for the AFOSC spectrograph. The astronomical specifications of these devices are reported in Table 3.

Table 3.

Asiago’s delivered GRISMs main specifications summary: (1) Working central wavelength of the grating; (2) pixel dispersion of the grating; (3) Wavelength range; (4) Resolution, R = λ/Δλ; (5) pitch of the grating; (6) final refractive index modulation of the grating; (7) prism apex angle. *) single non-right triangle.

Two of the elements show a low dispersion (ca. 280 l/mm with R=500–700), while the third (VPHG8) has 600 l/mm and a resolution of 1300. As we discussed before, VPHGs at very low dispersion (line density) require low Δn and thick films. Achieving precise low Δn is not straightforward with the risk to send too much light into unwanted diffraction orders. In Table 4 we report the diffraction efficiencies of the three GRISM that we have manufactured for AFOSC. The measurements were completed in laboratory before the commitment at the telescope and they were performed at different wavelengths.

Table 4.

AFOSC GRISMs transmission efficiencies at different wavelengths. The values correspond to raw measurements, without reflection losses corrections or materials absorption corrections.

The VPHG6 was designed for observations of supernovae, while the VPHG8 was aimed for the inspection of the Hα region with a very high efficiency. Details of these VPHGs were already been published by the authors [16].

The most interesting device among these GRISMs is the VPHG7, which shows a novel architecture based on slanted fringes VPHG and only one BK7 prism (Figure 7), with advantages such as less production costs and number of optical interfaces, reducing the reflection losses and the risk of ghosts [17]. This novel design was chosen because the target astronomical application required the highest possible near-UV efficiency and the minimisation of the number of layers and interfaces allowed us to produce a dispersive element with a good efficiency down to 3500 Å.

Figure 6 reports the comparison between the throughput of the single prism GRISM VPHG7 with the corresponding ruled GRISM commonly used to observe the bluest part of the objects’ spectra. It is apparent the shift in the efficiency curve at shorter wavelengths. There is not a large increase in the peak efficiency and this is due to the low line density of the grating (280 l/mm). Indeed, for low line density gratings, also the ruled gratings showed a good efficiency. On the other hand, by a suitable choice of low UV absorption materials like substrates and coupling fluids a huge increase in the efficiency is achieved up to 4300 Å.

Fig. 5.

Scheme of the VPHG7 single GRISM with slanted fringes, ϕ is the slanting angle, A and B the prism apex angles, D1 the illuminating beam diameter.

Fig. 6.

Comparison of photometric-standard stars’ spectra. Total system throughput of the Asiago’s telescope with GRISM VPHG7 and grating GR4. Slitwidth used in the observations is 4.22 arcsec.