Open Access
Add to BibSonomy
Abstract
We report on the action of exposure time and peak Intensity on the growth of long-scale focal volumes in multiphoton polymerization. Using modified engineered Bessel beams, we explore the effects that rise during the voxel growth, while we present a counterintuitive action of the expected expansion of the polymerized volumes that is revealed for a specific range of peak Intensities. We show that there is a regime where the higher exposure time (number of pulses) creates shorter polymerized strings in comparison to a lower number of pulses, and we determine the polymerization thresholds.
© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Corrections
10 July 2019: A typographical correction was made to the author listing.
1. Introduction
Multiphoton Lithography [1,2] is an exceptional technique for the fabrication of complex 3D objects with nanoscale features. When the beam of an ultrafast laser is tightly focused into the volume of a transparent, photosensitive material, the polymerization process can be initiated by non-linear absorption within the focal volume pixel (voxel) [3–6]. By point-by-point scanning the voxel in space, following a path derived from a computer designed 3D model, one can thus create arbitrary 3D structures. This powerful technique has found numerous applications in fields such as micro-optics [7,8], metamaterials [9], photonic crystals [10,11], microfluidics [12], biomedical implants [13] and 3D scaffolds for cell culture and tissue engineering [14].
Recently there is an increasing interest for rapid fabrication, parallel processing and 3D printing technology involving the fabrication of high aspect ratio (AR) structures. Several techniques including multi-beam focus [15], holographic lithography [16] or the use of shaped laser beams like Bessel, vortex or ring-Airy beams [17–20] are employed for this purpose. For Direct Laser Writing (DLW) to advance in this fabrication regime, a better understanding on the single voxel properties and their correlation to the initial conditions is essential. However, along this direction, the majority of the studies have been performed only for low aspect ratio (AR<10), micron scaled, voxels [21–25].
In what follows, we report on the growth of long, tens of microns scaled, high aspect ratio (AR>40) multiphoton polymerized voxels as a function of exposure time and pulse peak intensity. Using a Bessel beam, with engineered asymmetric on-axis distribution, for the multiphoton polymerization (MPP), we reveal the scaling laws that relate the initial parameters to the polymerized voxel physical dimensions. We show that for a specific range of peak intensities in contrast to the expected behavior, the voxel size decreases, instead of increasing, as the exposure time (number of pulses) is increased.
2. Experimental section
2.1 Generation of the engineered Bessel beams
Bessel beams are nondiffracting beams that invariantly maintain their transverse intensity profile along the propagation direction [26]. Bessel beams have a conical wavefront and can be easily generated by shaping the phase of a Gaussian beam with the use of an axicon [27] or a Spatial Light Modulator (SLM) [17]. The required phase to generate a Bessel beam is given by [28] $\varphi (r )= - (2\; \pi /\lambda )\tan (\gamma )r$, where r is the radius, λ is the wavelength, and γ is the Bessel cone angle. We note that Bessel beams have already been employed in DLW by MPP [17,27]. Here, we are not interested in the invariant propagation of the Bessel beam itself, but on the ability to engineer the dependence of the FWHM longitudinal length of the focal intensity distribution to the peak intensity or total power of the beam. Since the physical dimensions of the polymerized voxel, are related to this distribution we can achieve a smooth size transition as the peak intensity or the exposure time is varied. To achieve that, we properly engineer the Bessel distribution along the propagation axis creating an asymmetry at its axial extent as shown in Fig. 1(iii). To achieve this asymmetric intensity distribution, an extra quadratic phase, equivalent to the action of a spherical lens, is added on the conical phase mask applied to the Spatial Light Modulator (SLM), thus the required phase modulation now becomes:$\; \varphi (r )= - (2\; \pi /\lambda )(\tan (\gamma )r + {r^2}/R)$, where R is the radius of curvature we add on the conical phase distribution. Taking into account that MPP is actually a threshold process, polymerization can take place only at areas where the intensity I is high enough so that Iκ > C, where κ is the order of the multi-photon polymerization process and C is a constant. Thus, control of the intensity enables progressive tuning of the polymerization threshold and scaling of the asymmetric Bessel focal volume range dimensions (see schematic illustration of the square of Intensity distribution I2 in Fig. 1(iii)).
Fig. 1. Experimental setup. f (300 mm); Obj. (Microscope objective, 20x NA = 0.4); (i) Typical SLM conical phase mask used. (ii) Experimental Intensity I2 distribution of the engineered Bessel beam along propagation as captured by a CCD camera in air after the objective lens. Bessel zone length 130µm (FWHM); (iii) Schematic representation of the engineered Bessel beam normalized intensity I2 distribution inside the volume of a photoresist droplet on a cover glass substrate. (iv) Intensity distribution I2 (x-y plane) of the Bessel beam at the focus. Focal Spot size (FWHM) 1.8µm.
We experimentally generated asymmetrically engineered Bessel beams using the experimental setup described in Fig. 1. The phase of a Gaussian beam was spatially modulated using an appropriate, as described above, conical phase mask applied on a Hamamatsu LCOS-X10468-2 phase-only reflecting spatial light modulator (SLM), generating the designed engineered Bessel distribution. A Ti:Sapphire femtosecond laser (800 nm, 35 fs, 1kHz) was used as a light source. In order to adapt to the MPP application, the generated Bessel beam was further reduced in size, by M = 32 times in the x-y plane, using a 4f optical system composed by a 300 mm lens and a 20x (NA = 0.4) microscope objective. The experimentally generated Bessel beam propagation was studied using a compact microscope imaging system coupled with a linear 14-bit charge-coupled device (CCD) camera. Figure 1(i) shows a typical two-dimensional phase mask used for the phase modulation of the Gaussian beam. The square of the intensity distribution I2 along the asymmetric Bessel beam propagation axis (z) can be seen in Fig. 1(ii), as synthesized by multiple cross-sectional CCD images (taken at different z positions in air). The values of I2 represent better, compared to the intensity I, the conditions under which the two photon polymerization process takes place inside the material. From these images, the longitudinal focal range at the FWHM of the Bessel distribution along the propagation was measured to be 130 µm. A schematic representation of the engineered Bessel beam normalized intensity distribution I2 inside the volume of a photoresist droplet on a cover glass substrate is shown in Fig. 1 (iii) while the intensity distribution of the Bessel beam cross-section focus in the x-y plane is shown in Fig. 1 (iv), with the measured focal spot size at FWHM to be 1.8µm. For the multiphoton polymerization, the modified Bessel was let to interact with the photosensitive material (SZ8020). Samples of the photosensitive materials were prepared by drop-casting onto 100 µm thick salinized glass substrates. After polymerization, the samples were immersed in an appropriate development solvent and the non-polymerized material was washed away to leave the free-standing voxels. A 4-methyl-2-pentanone mixed with iso-propanol in 1:1 volume ratio solution was used as a developer. The polymerized voxels were studied by Scanning Electron Microscope (SEM) after being coated with a thin gold layer through sputtering. Taking into account the transmissivity of our optical system, the amount of energy corresponding to the central Bessel lobe at the focus was determined in order to estimate the nominal pulse peak intensity ${I_{peak}}$. Also, using a shutter (Uniblitz VS25, min. exposure time 6 msec), the number of pulses during exposure was controlled, while the beam was attenuated using a rotating λ/2 plate followed by a GLAN prism placed before the SLM.
2.2 Long-scale multiphoton polymerization voxel growth
Figure 2 shows Scanning Electron Microscope (SEM) images of polymerized voxels fabricated using the engineered Bessel beams of various exposure times and peak intensities. Furthermore, for each case of peak intensity value, voxels created by different number of pulses, from 6 to 1000 (6 msec to 1sec exposure), are presented too. Voxels fabricated using peak intensities Ipeak varying from 8 to 42 TW/cm2 (∼ 0.17 µJ to 0.90 µJ pulse energy measured after the SLM) are presented in Fig. 2 (i,j). The values of voxel width and length were varied between 1.8–24.6µm and 45 - 167µm respectively, depending on the intensity and exposure time during polymerization. Figure 2(a)-(d) presents selected examples of voxels fabricated using Ipeak = 20 TW/cm2 and (e)–(h) Ipeak = 31 TW/cm2, for exposure with 1000, 600, 60 and 8 pulses. SEM images of the whole sample at 45° view and vertical side view can be seen in Fig. 2 (i) and (j) respectively. The directions of the intensity increase, pulse decrease and light propagation direction are shown with the respective arrows. From Fig. 2 (j) we can clearly observe that the length of the polymerized voxels becomes longer with the increase of peak intensity. In addition, as will be discussed later, for high intensities, an increase of the number of pulses leads to increase of the polymerized voxel length and width dimensions. For the case of Fig. 2(a)-(d) though, we observe a counterintuitive action, the length of polymerized voxels decreases with the increase of exposure time, while the diameter follows the expected monotonically increasing behavior as the number of pulses is increased. It is characteristic that comparing case 2(d) and 2(c), the photopolymerization grows radially starting from the end of the string. For a higher number of pulses, this broadening appears earlier and nearer to the glass substrate, while at 1000 pulses it is extended over the whole length. The dashed arrows in Fig. 2 (a)-(c) show the direction and extension range of the voxel diameter broadening effect in respect to the number of pulses, while the scale bars are showing the measured lengths along the voxel’s longitudinal structure .
Fig. 2. SEM images of the polymerized voxels using engineered Bessel beams. (a) - (d) Voxels fabricated using Ipeak=20TW/cm2 (dashed arrows show the voxel diameter broadening direction and range of extension in respect to number of pulses), and (e)–(h) Ipeak =31 TW/cm2, for 1000, 600, 60 and 8 number of pulses. Scale bars are showing the measured lengths along the voxel’s longitudinal structure. (i)-(j) SEM images of the whole sample at 45° view and vertical side view. The directions of the Intensity increase, pulses number decrease and light propagation direction are shown by the respective arrows.
In Fig. 3 the polymerized voxel lengths as a function of the number of pulses are presented for various Ipeak values. Successive pulses with numbers ranging from 6 to 1000 were used for the fabrication of the voxels. Figure 3(b) shows a detailed view of the dashed area (<300 pulses) of Fig. 3 (a). We observe that for more than 100 pulses (> 100 msec exposure time), the polymerized voxel length remains almost invariant. On the other hand, for the first 100 pulses, the length gradually increases before reaching a saturation value. This behavior is expected since this time window corresponds to the required time for the completion of the polymerization reactions, stabilization of the radical’s distribution and monomer to polymer transformation before saturation [23,25].
Fig. 3. (a) Voxel Length as a function of the Number of Pulses (6 to 1000 pulses). (b) Inset of the dashed area of figure (a). For a number of pulses higher than 100 (more than 100msec) saturation of the polymerized volume length is observed.
Likewise, Fig. 4 presents the polymerized voxel widths of various Ipeak values as a function of the number of pulses. For low intensities, and for more than 100 pulses (exposure >100 msec) the voxel width practically saturates. On the other hand, for Ipeak higher than ∼24 TW/cm2, the voxel width monotonically increases by successive pulses. Moreover, for the first 100 pulses, the behavior is not stable. A rapid increase of the voxel width, followed by a decrease and then a saturation to a lower value as the number of pulses increases is observed. We believe that this behavior is related to the time window parameter range which is relevant to the monomer to polymer transformation times, similarly to what we observed for the voxel length case for the same time range in Fig. 3.
Fig. 4. (a) Voxel width as a function of the Number of Pulses (6 to 1000 pulses). (b) Inset of the dashed area of figure (a).
The increase of the physical dimensions of the voxel with exposure time (or pulses), can be explained as follows: Polymerization is initiated first in regions of intense illumination, so the lateral dimension of the voxel is smaller than that of the optical beam. Light scattering and the diffusion of free radicals beyond the illumination region contribute to the further increase of the voxel diameter. At longer exposure times, newly produced radicals will diffuse either away from the focal spot or towards the existing voxels depending on their locations. As a result, photopolymerization is initiated at the boundaries of the existing polymerized structure, broadening thus the voxel dimensions.
The voxel width and length as a function of Ipeak values are presented in Fig. 5, for different number of pulses exposure, from 1000 to 10 pulses. In the case of Fig. 5 (a) where the voxel width is presented as a function of the intensity, we observe, as expected, that the voxel diameter increases with increasing intensity for all exposure times. Likewise, in Fig. 5 (b), where the voxel length is presented as a function of the intensity, we observe that for Ipeak above ∼31TW/cm2, increasing number of pulses results to longer voxels, while for lower powers (shown as region (ii) in Fig. 5(b)) the increase of exposure time creates shorter voxel lengths. Thus, it appears that the peak intensity of Ipeak ∼ 31TW/cm2 corresponds to a switching point of this complex scaling behavior.
Fig. 5. Polymerized Voxel Width (a) and Voxel Length (b) as a function of the pulse peak Intensity for 1sec to 10msec exposure time.
In order to understand the origin of this peculiar behavior observed for the intensity zone (ii) in Fig. 5(b), we must examine the intrinsic physical events that take place during polymerization and influence the growth of such long volumes. As the intensity increases the index of refraction of the medium [29] also increases through the nonlinear index leading to self-focusing. The nonlinear induced lens-like index profile will self-focus primarily the central part of the beam which will lead to waveguiding effects and finally to the extension of the voxel length beyond the limits imposed by the linear intensity profile through the material. In the case of successive pulses, the refractive index changes at the center quickly saturate, preventing the longitudinal extension of the voxel caused by waveguiding. This reduces the strength of the self-focusing and the beam power is now distributed over the entire cross-section of the effective waveguide, creating a shorter and wider polymerized voxel. As the tails of the optical field guided by this effective waveguide extend beyond its core, the photopolymerization is initiated at the boundaries which broadens the polymerized voxel diameter. After the switching point at ∼ 31 TW/cm2 (Fig. 5(b), zone (iii)), the intensity is high enough to rapidly create, i.e. in 10 pulses, quite long voxels which will monotonically extent, both in the lateral and longitudinal direction in response to successive pulses. It is worth noting that the observed switching intensity corresponds to the one of intensity clamping observed in the filamentation of ultrashort laser pulses in dense media [30]. When this intensity is reached, the intensity in the filamentary propagation cannot attain higher values, because of dynamic balance with nonlinear defocusing from electrons plasma. Thus, as the input power is further increased the filamentary zone, and the polymerized one as well, is increased rather than obtaining higher intensities.
The polymerized voxel length and width as a function of intensities for the lowest time of exposure, T = 6 msec (6 pulses), employed for fabrication, is presented in Fig. 6(a). From Fig. 6 we can identify the polymerization threshold for Tth = 6 msec at Ipeak=14 TW/cm2. For this threshold, the polymerized voxel was measured to be 76 µm in length and 1.8 µm in width, resulting in aspect ratio AR = 42. Figure 6(b) shows the voxel length and width as a function of the number of pulses for the previous threshold intensity. However, if we take into account the width and length instabilities which are revealed in Fig. 4 and 5 for the first 100 msec, which is the time required for polymerization reactions completion, we can set the time threshold to be TTh = 100 msec at Ipeak = 8 TW/cm2, where L.:56 µm, W: 2.2µm, AR∼ 25 (Fig. 3 & 4). Moreover, after the 100 msec exposure, the voxel length and width is found to saturate or gradually increase. Based on the above we can safely assume that the stabilization of the polymerization system sets a lower boundary to the range of exposure times (or number of pulses) to use for reliable results in the polymerization process.
Fig. 6. (a) Polymerized voxel length (solid black line) and width (dashed blue line) as a function of Ipeak for the lowest exposure time threshold TTh =6msec. (b) Polymerized Voxel length (solid black line) and width (dashed blue line) as a function of the number of pulses for IPeak = 14 TW/cm2 (•).
3. Conclusions
In summary, we have demonstrated the growth scaling laws of high aspect ratio polymerized voxels with respect to initial conditions of intensity and exposure time. Asymmetrically engineered Bessel beams were experimentally generated for the multiphoton polymerization of scalable focal volumes in a photosensitive material. We observed that for a number of pulses higher than 100 (exposure > 100msec) both the length and width of the polymerized volume remain almost invariant or monotonically increase, while for the first 100msec exposure time, instabilities for both cases were present due to the requirements for polymerization reactions completion. In addition, a counterintuitive action of successive number of pulses to the longitudinal extension of the polymerized volumes was revealed for a specific range of Peak Intensities and the regime where the higher exposure time can create shorter polymerized strings, in comparison to a smaller number of pulses, is shown. Finally, the polymerization thresholds were determined.
Funding
Qatar National Research Fund (QNRF) (NPRP9-383-1-083); Horizon 2020 Framework Programme (H2020) (Laserlab Europe (EC-GA 654148)); State Scholarships Foundation (IKY) (program: 5003404, grant No. 2017-050-0504-10113).
References
1. J. Fischer and M. Wegener, “Three-dimensional optical laser lithography beyond the diffraction limit,” Laser Photonics Rev. 7(1), 22–44 (2013). [CrossRef]
2. M. Malinauskas, M. Farsari, A. Piskarskas, and S. Juodkazis, “Ultrafast laser nanostructuring of photopolymers: A decade of advances,” Phys. Rep. 533(1), 1–31 (2013). [CrossRef]
3. H.-B. Sun and S. Kawata, “Two-Photon Photopolymerization and 3D Lithographic Microfabrication,” in NMR 3D Analysis Photopolymerization (Springer, 2006), pp. 169–273.
4. M. Farsari and B. N. Chichkov, “Two-photon fabrication,” Nat. Photonics 3(8), 450–452 (2009). [CrossRef]
5. S. Juodkazis, V. Mizeikis, and H. Misawa, “Three-dimensional microfabrication of materials by femtosecond lasers for photonics applications,” J. Appl. Phys. 106(5), 051101 (2009). [CrossRef]
6. C. N. LaFratta, J. T. Fourkas, T. Baldacchini, and R. A. Farrer, “Multiphoton Fabrication,” Angew. Chem., Int. Ed. 46(33), 6238–6258 (2007). [CrossRef]
7. E. Brasselet, M. Malinauskas, A. Žukauskas, and S. Juodkazis, “Photopolymerized microscopic vortex beam generators: Precise delivery of optical orbital angular momentum,” Appl. Phys. Lett. 97(21), 211108 (2010). [CrossRef]
8. M. Malinauskas, A. Žukauskas, V. Purlys, K. Belazaras, A. Momot, D. Paipulas, R. Gadonas, A. Piskarskas, H. Gilbergs, A. Gaidukevičiūtė, I. Sakellari, M. Farsari, and S. Juodkazis, “Femtosecond laser polymerization of hybrid/integrated micro-optical elements and their characterization,” J. Opt. 12(12), 124010 (2010). [CrossRef]
9. G. Kenanakis, A. Xomalis, A. Selimis, M. Vamvakaki, M. Farsari, M. Kafesaki, C. M. Soukoulis, and E. N. Economou, “Three-Dimensional Infrared Metamaterial with Asymmetric Transmission,” ACS Photonics 2(2), 287–294 (2015). [CrossRef]
10. I. Sakellari, E. Kabouraki, D. Gray, V. Purlys, C. Fotakis, A. Pikulin, N. Bityurin, M. Vamvakaki, and M. Farsari, “Diffusion-assisted high-resolution direct femtosecond laser writing,” ACS Nano 6(3), 2302–2311 (2012). [CrossRef]
11. G. Bickauskaite, M. Manousidaki, K. Terzaki, E. Kambouraki, I. Sakellari, N. Vasilantonakis, D. Gray, C. M. Soukoulis, C. Fotakis, M. Vamvakaki, M. Kafesaki, M. Farsari, A. Pikulin, and N. Bityurin, “3D Photonic Nanostructures via Diffusion-Assisted Direct fs Laser Writing,” Adv. OptoElectron. 2012, 1–6 (2012). [CrossRef]
12. L. Amato, Y. Gu, N. Bellini, S. M. Eaton, G. Cerullo, and R. Osellame, “Integrated three-dimensional filter separates nanoscale from microscale elements in a microfluidic chip,” Lab Chip 12(6), 1135 (2012). [CrossRef]
13. C. Schizas, V. Melissinaki, A. Gaidukeviciute, C. Reinhardt, C. Ohrt, V. Dedoussis, B. N. Chichkov, C. Fotakis, M. Farsari, and D. Karalekas, “On the design and fabrication by two-photon polymerization of a readily assembled micro-valve,” Int. J. Adv. Des. Manuf. Technol. 48(5-8), 435–441 (2010). [CrossRef]
14. M. T. Raimondi, S. M. Eaton, M. M. Nava, M. Laganà, G. Cerullo, and R. Osellame, “Two-photon laser polymerization: from fundamentals to biomedical application in tissue engineering and regenerative medicine,” J. Appl. Biomater. Funct. Mater. 10(1), 56–66 (2012). [CrossRef]
15. L. Yang, A. El-Tamer, U. Hinze, J. Li, Y. Hu, W. Huang, J. Chu, and B. N. Chichkov, “Parallel direct laser writing of micro-optical and photonic structures using spatial light modulator,” Opt. Lasers Eng. 70, 26–32 (2015). [CrossRef]
16. G. Vizsnyiczai, L. Kelemen, and P. Ormos, “Holographic multi-focus 3D two-photon polymerization with real-time calculated holograms,” Opt. Express 22(20), 24217 (2014). [CrossRef]
17. L. Yang, A. El-Tamer, U. Hinze, J. Li, Y. Hu, W. Huang, J. Chu, and B. N. Chichkov, “Two-photon polymerization of cylinder microstructures by femtosecond Bessel beams,” Appl. Phys. Lett. 105(4), 041110 (2014). [CrossRef]
18. J. Ježek, P. Zemánek, T. Čižmár, and V. Nedĕla, “Formation of long and thin polymer fiber using nondiffracting beam,” Opt. Express 14(19), 8506–8515 (2006). [CrossRef]
19. M. Manousidaki, D. G. Papazoglou, M. Farsari, and S. Tzortzakis, “Abruptly autofocusing beams enable advanced multiscale photo-polymerization,” Optica 3(5), 525 (2016). [CrossRef]
20. E. Stankevicius, T. Gertus, M. Rutkauskas, M. Gedvilas, G. Raciukaitis, R. Gadonas, V. Smilgevicius, and M. Malinauskas, “Fabrication of micro-tube arrays in photopolymer SZ2080 by using three different methods of a direct laser polymerization technique,” J. Micromech. Microeng. 22(6), 065022 (2012). [CrossRef]
21. H.-B. Sun, T. Tanaka, and S. Kawata, “Three-dimensional focal spots related to two-photon excitation,” Appl. Phys. Lett. 80(20), 3673–3675 (2002). [CrossRef]
22. H.-B. Sun, M. Maeda, K. Takada, J. W. M. Chon, M. Gu, and S. Kawata, “Experimental investigation of single voxels for laser nanofabrication via two-photon photopolymerization,” Appl. Phys. Lett. 83(5), 819–821 (2003). [CrossRef]
23. H.-B. Sun, K. Takada, M.-S. Kim, K.-S. Lee, and S. Kawata, “Scaling laws of voxels in two-photon photopolymerization nanofabrication,” Appl. Phys. Lett. 83(6), 1104–1106 (2003). [CrossRef]
24. J. B. Mueller, J. Fischer, F. Mayer, M. Kadic, and M. Wegener, “Polymerization kinetics in three-dimensional direct laser writing,” Adv. Mater. 26(38), 6566–6571 (2014). [CrossRef]
25. A. Pikulin, N. Bityurin, and V. I. Sokolov, “Model of diffusion-assisted direct laser writing by means of nanopolymerization in the presence of radical quencher,” AIP Adv. 5(12), 127215 (2015). [CrossRef]
26. J. Durnin, “Exact solutions for nondiffracting beams. I. The scalar theory,” J. Opt. Soc. Am. A 4(4), 651 (1987). [CrossRef]
27. B. Bhuian, R. J. Winfield, S. O’Brien, and G. M. Crean, “Pattern generation using axicon lens beam shaping in two-photon polymerisation,” Appl. Surf. Sci. 254(4), 841–844 (2007). [CrossRef]
28. S. Suntsov, D. Abdollahpour, D. G. Papazoglou, P. Panagiotopoulos, A. Couairon, and S. Tzortzakis, “Tailoring femtosecond laser pulse filamentation using plasma photonic lattices,” Appl. Phys. Lett. 103(2), 021106 (2013). [CrossRef]
29. A. Žukauskas, I. Matulaitienė, D. Paipulas, G. Niaura, M. Malinauskas, and R. Gadonas, “Tuning the refractive index in 3D direct laser writing lithography: towards GRIN microoptics,” Laser Photonics Rev. 9(6), 706–712 (2015). [CrossRef]
30. P. Panagiotopoulos, D. G. Papazoglou, A. Couairon, and S. Tzortzakis, “Sharply autofocused ring-Airy beams transforming into non-linear intense light bullets,” Nat. Commun. 4(1), 2622 (2013). [CrossRef]
