Template-Assisted Fabrication of Patchy Particles with Uniform Patches

Letter pubs.acs.org/Langmuir

Template-Assisted Fabrication of Patchy Particles with Uniform Patches Zhenping He and Ilona Kretzschmar* Department of Chemical Engineering, The City College of New York, 140th Street & Convent Avenue, New York, New York 10031, United States S Supporting Information *

ABSTRACT: Patchy particles with uniform patches of specific shape and size have been predicted to have a rich potential in fabricating new structures; however, an effective method to control the patch shape and size is still missing. In the method presented here, a template is used to assist the fabrication of patchy particles with patches of uniform shape and controlled size by use of the glancing angle deposition method (GLAD). Uniform shadowing effects are caused by the wall of the grooves carved into the surface of a silicon wafer. The ratio of template dimension to particle diameter and the angle of incidence of the metal vapor rays determine the patch shape and size. Mathematical calculations are applied to predict the patch shape and size. Scanning electron microscopy is used to demonstrate the efficiency of the method. Scaling analysis shows that the template-assisted GLAD method leads to a 3100-fold increase in patchy particle fabrication volumes compared to the template-free GLAD method.

■

INTRODUCTION Patchy particles with an anisotropic surface chemistry have attracted much attention due to their unique properties and a wide range of potential applications.1−6 The unique control over directional interactions with other particles (through specific patches) and physical fields7−9 opens several new possibilities for the fabrication of materials with more complex structures. Many simulations have been performed to reveal the assembly potential of patchy particles.10−16 One of the highlights, for example, is Glotzer’s work11 that demonstrates that the right patch shape and placement leads to the soughtafter diamond structure. Formation of polymer chains,13 assembly dynamics, and aggregation behavior of patchy particles have also been investigated.12,14−17 In the meanwhile, many methods of patchy particle fabrication have been reported.18−31 A careful review of the simulation results shows that many of the proposed structures are based on the essential premise that the patchy particles used have uniform patches with a very specific patch shape and area. Despite the large number of synthetic routes proposed, an effective method with accurate control over both patch shape and size for low-symmetry patches is still missing. Embedment into polymers22,23,28−30 and chemical etching31 have been used to assist in patch size and shape control but are limited to highly symmetric, cap-shaped patches. Silver-assisted colloidal synthesis has also provided a new route of controlling the patch size but lacks in the control of patch size and number.20 Velegol’s particle lithography technique27 leads to highly symmetric, unfunctionalized patches, while the rest of the particle surface is functionalized. In previous work,24−26 we © 2012 American Chemical Society

introduced a convenient, template-free way to create patches on particles with controlled patch surface area. However, the method suffers from a variation of the shape of the patches as a function of monolayer domain orientation, especially at small incident angles of θ < 25°.24 Here, we report a template-assisted patchy particle fabrication method that combines a template that controls the particle position with the glancing angle deposition (GLAD) method and allows for the repeated use of the template. In our method, a silicon wafer template is used to help control patch shape and size by tilting it based on the angular requirement. We show that with the help of the silicon wafer template the patches created on the particle surface are uniform in shape and size. Further, we study the effect of incident angle and dimensions of the template on the patch geometry and support our observations with mathematical calculations that verify the patch shape and size. For validation of our methodology, we also perform a scaling analysis, which shows that the patchy particle production is increased by 3100 times in the template-assisted GLAD method compared to the more general GLAD method.

■

EXPERIMENTAL DETAILS

A silicon (Si) wafer with either V- or square-shaped grooves, hereafter called template, is prepared using an anisotropic wet etching technique (see Supporting Information),32 and is used to assist with the selfReceived: April 29, 2012 Revised: June 8, 2012 Published: June 18, 2012 9915

dx.doi.org/10.1021/la3017563 | Langmuir 2012, 28, 9915−9919

Langmuir

■

assembly of sulfate-terminated polystyrene (PS) beads into linear close-packed aggregates by controlling the ratio of groove dimension to particle diameter.33 The grooves enable controlled partial exposure of the particles to the metal vapor leading to patch formation. The coated area percentage, A%, (i.e., area of patch divided by the entire surface area of the PS particle) is predicted for varying θ using a simple geometric model. The model includes parameters such as groove depth and width, which affect the coated area percentage.

Letter

RESULTS AND DISCUSSION The advantages of the method described are the achievement of uniform patch shape and size, the reusability of the template, and the increase in particle production volume. In the following, the dependence of patch size on parameters such as the width and depth of the groove, the incident angle, the size of the particles, and also the type of the template are discussed. Generally, templates will be reused, since they are fabricated on a Si wafer and can endure long-term, multiple usage. Thus, for one template we can simply consider the influence of the incident angle and size of particle while all other templaterelated parameters are kept constant. Equations 2a and 2b show the relation of patch height, h, with particle radius, r, and incident angle, θ: ⎛ ⎞ ⎛r⎞ h = r − (L2 + r 2)1/2 sin⎜54.74° − a tan⎜ ⎟ − θ ⎟ ⎝L⎠ ⎝ ⎠

in which L is defined as shown in eq 2b: w L= − r tan 54.74° 2 cos(54.74°)

(2b)

Figure 2 shows the coated area percentage, A%, calculated as shown in eq 1 as a function of incident angle, θ, for varying particle sizes in V-shaped (A) and square-shaped grooves (B). In Figure 2A, the width of the V-shaped groove is 4.0 μm. When the incident angle is equal or larger than the angle of the sidewall (54.74°) Janus particles with 50% gold coating are obtained. As this is true for any particle size, all lines in Figure 2A join at the same point (54.74°, 50%). As the radius of the particles increases, the effective range of incident angles creating patches on the particles consequently increases as well. The linear appearance enables a straightforward choice of the correct incident angle required for a specific coated-area percentage. Note the relationship between A% and θ is not straight forwardly linear, but rather complex as shown by combination of eqs 1, 2a, and 2b leading to a noticeable deviation from linearity at θ values near the 1/4 and 3/4 point of the entire range, which is more easily observed for larger particle radii. Figure 2B shows the same relationship for square-shaped grooves with dimensions of 4.0 μm width and 3.2 μm depth assuming that the particle is located at the center of the groove. The changing radius of the particle is denoted by different colors. The patch height, h, is calculated using eq 3 and converted to coated area percentage, A%, using eq 1:

Figure 1. Schematic of template-assisted patchy particle fabrication using glancing angle deposition method and parameters affecting patch shape and area, where w and d are the width and depth of the groove, respectively. θ is the incident angle, r is the radius of the particle, h is the height of patch cap, and v is the off-center distance if the particle is not located at the center of the groove. Particle-template geometries for (A) particle in V-shaped groove and (B) particle in square-shaped groove leading to patches of A% = 13.4% and 20.8%, respectively. Figure 1 represents a schematic view of the template-assisted patchy particle fabrication using the GLAD method for both (A) V- and (B) square-shaped grooves. The angle θ corresponds to the incident angle of the Au vapor flow measured from the substrate side. w indicates the width of the groove, d is the depth of the square grooves (the depth of the V-shaped grooves changes with their width), and r is the radius of the PS particles. For V-shaped grooves, the particles are always located at the center of the groove (Figure 1A), whereas an off-center distance v (defined as the distance from the center of the particle to the center of the groove) needs to be considered for the square-shaped groove (Figure 1B). The schematic in the inset of both parts of the figure shows the particle-template geometry and incident angle for two specific cases. Note that the shape of the resulting patch is always a half-spherical cap due to the perpendicular evaporation geometry. The height of the spherical cap is denoted as h. The coated area percentage, A%, is calculated according to eq 1 πrh h A% = = 4r 4πr 2

(2a)

h=r−

w ⎡ 2 sin⎣⎢a tan

(

⎞⎤ ⎛ w tan⎜ ⎟⎥ ⎝ 2(d − r ) ⎠⎦

(1)

in which h is related to θ, w, and the slope of the sidewall (54.74°, a constant value here) for a V-shaped groove and θ, w, d, and r for a square-shaped groove. The dashed circle in Figure 1B shows the case in which a particle is not located at the center of the square groove. For this case, the coated area percentage is calculated by moving the particle along the direction of the metal vapor flow until its center reaches the center of the groove, which results in replacing d with (d + ν tan θ) while all other parameters in the calculation remain unchanged.

w 2(d − r )

⎡ sin⎢90° − θ − a ⎤ ⎣ ⎦⎥

)

(3)

Note that, in contrast to Figure 2A, there is no θ limitation. The point at which 50% coating, i.e., Janus particle, is reached depends on the ratio of groove dimension to particle radius. If the width and depth of the groove are equal to the diameter of the particle, the values of incident angles creating patchy particles ranges from θ = 0° to 90°. Interestingly, all lines in Figure 2B join in the point Z at A% = 38.3% and θ = 58.0°. The point Z is of interest as it enables evaporation of patches with 9916

dx.doi.org/10.1021/la3017563 | Langmuir 2012, 28, 9915−9919

Langmuir

Letter

Figure 3. 1.500 ± 0.026 μm patchy polystyrene particles in V-shaped grooves of w = 3.7 μm after gold evaporation. (A) Large-scale image of wafer with particle-filled V-shaped grooves. Solid and dashed arrows indicate two types of packing defects observed (see text for details). Enlarged images of particles with patches produced at incident angles of (B) θ = 26°, (C) θ = 22°, and (D) θ = 19°. (D) Bright areas are gold patches, 2l indicates the distance between the two corner points on the flat side of the patch, and scale bars are 1.0 μm.

highlighted in Figure 3A by the solid and dashed arrows, respectively, which results in about 6 ± 2% of the particles having defective patches. The few stray particles that are outside the grooves can be removed by stamping with a precured PDMS film. In our previous work,26 we reported average domain sizes of 3000 particles within which all patches have a uniform shape and size. However, the particles at the edge of such a domain within a close-packed monolayer do not benefit from the same shadowing effect as particles within the domain leading to an additional patch size and shape variation of approximately 7 ± 1%. The template-assisted GLAD method introduced here has the additional benefit that only the template size limits the number of patchy particles that can be produced in one fabrication cycle. For example, a template on a 1 square inch Si wafer with 1-inch-long grooves at a density of 3845 (groove width = 3.7 μm, groove pitch = 2.8 μm) that is filled to 15% with particles (1.500 ± 0.026 μm diameter) leads to 9 300 000 patchy particles with uniform patch size and shape compared to the ca. 3000 particles in one ordered domain reported for the template-free GLAD technique.26 Note that functionalization of particles for a 1 mL particle latex solution with 4% solids would require 4600 experiments using the 1 × 1 in.2 template described above. This number can be dramatically decreased to 28 times if an 18-in. Si wafer (currently used by Intel) would be used for template preparation. Figure 3B−D shows enlarged images of particles and their uniform 35 nm gold patches obtained at θ = 26°, 22°, and 19°. The SEM images confirm that the patch shape (bright area on particle) is a half-spherical cap, which is caused by the shadowing of the metal vapor flow by the wall of the groove located closer to the source. Besides having the same shape and size, all patches are symmetric around the axis parallel to the metal vapor flow and normal to the groove surface. As θ increases from 19° to 26°, the patch size increases from A%19° = 3.9% to A%26° = 9.9% in good agreement with the prediction

Figure 2. Coated-area percentage, A%, plotted as a function of incident angle, θ, for (A) particles with diameters ranging from 0.4 to 1.0 μm in a V-shaped groove of 4 μm width and (B) particles with diameters ranging from 0.7 to 1.6 μm located at the center of a square groove of 4 μm width and 3.2 μm depth. The inset in (A) shows the particle-template geometry for the data graphed in (A). The inset in (B) shows the particle-template geometry for point Z at which a fixed A% value of 38.3% is achieved for a 4-μm-wide and 3.2-μm-deep square-shaped groove at a θB of 58° independent of the particle size.

identical surface areas onto particles of varying radii using the same template and incident angle θ. Note that the specific A% and θ values for point Z depend strongly on the dimensions of the groove. The inset in Figure 2B schematically shows the particle-template geometry for point Z where θ = a tan (2 d/w), which leads to h = r + r cos(a tan(2 d/w)) and consequently a particle-size independent coated-area percentage as shown in eq 4: A% = [1 + cos(a tan(2d /w))]/4

(4)

Equation 4 can be corrected for particles not located at the center of the groove by setting d = d + v tan θ. Figure 3 shows scanning electron microscopy images for 1.500 ± 0.026 μm polystyrene particles with patches created with a V-shaped template. The V-shaped template is chosen, as it enables easy centering of particle in the groove due to its Vshaped geometry. Figure 3A shows a large area view of the distribution of particles in a template with V-shaped grooves (w = 3.7 ± 0.1 μm, groove pitch = 2.8 ± 0.1 μm). Most of the particles are found to form linear close-packed aggregates within the grooves. Averaging over several regions on the template yields a maximum particle packing average of 14 ± 5%. There are some defects such as particles not staying in grooves or not forming a perfect linear arrangement as 9917

dx.doi.org/10.1021/la3017563 | Langmuir 2012, 28, 9915−9919

Langmuir

Letter

shown in Figure 2B. To verify the coated area percentage at the three incident angles, the distance between the two corner points on the flat side of the patches (denoted as 2 l in Figure 3B) is measured for each patch. The patch height, h, is then calculated using eqs 5a and 5b (note, eq 5a is used when the coated area is less than one-fourth of the entire particle surface, whereas eq 5b is used when it is larger than one-fourth): h = r − (r 2 − l 2)1/2 Acap


1 A sphere 4

(5b)

The 2l for the patches shown in Figure 3B,C,D is measured to be 2l26° = 1.18 ± 0.05, 2l22° = 1.04 ± 0.04, and 2l19° = 0.79 ± 0.09 μm, respectively. These 2l values result in h values of h26° = 0.29 ± 0.05, h22° = 0.21 ± 0.03, and h19° = 0.11 ± 0.05 μm. Using the h-values, coated percent areas of A%26° = 9.9 ± 0.8, A %22° = 7.1 ± 0.5, and A%19° = 3.9 ± 1.0% are calculated and are in good agreement with the computationally predicted values of 10.5%, 7.1%, and 3.6%, respectively. The increased variation of patch size at the smallest incidence angle (Figure 3D) is a result of non-uniform etching of the template and/or the variation in particle size. Both sources of patch variation can be addressed by further optimization of the template or choosing of particles with a narrower particle size distribution. From eqs 1, 2a, and 2b, we see that for V-shaped grooves besides the incident angle, θ, two other parameters, r and w, play an important role with respect to the patch size and shape. The r and w values can be combined into one dimensionless parameter, w/r, which generalizes the graph (Figure 2A) for Vshaped grooves and particles with w/r ratios ranging from 4 to 10. The dimensionless w/r ratio allows us to choose or draw a plotted line for the purpose of finding the right angle of incident for a specific coated area percentage to be obtained. Similarly, for the square grooves, the three parameters r, w, and d can be combined into two dimensionless parameters w/r and d/r. Figure 2B is based on the practical consideration that particles of different sizes can be loaded into the same template. To understand the influences of the groove dimensions, a more systematic study is necessary. In addition, it may prove difficult to locate particles at the center of the square groove, as capillary forces are likely to pull particles toward one or the other groove wall.33 This tendency may be suppressed by reshaping the flat bottom of the groove into a concave-shaped bottom warranting further study of the evaporation geometry with the squareshaped groove. Figure 4 illustrates the trend of coated-area percentage as a function of incident angle with the change of the two dimensionless parameters w/r and d/r. The diameter of the particle is set to 2.4 μm, whereas the width of the square groove is varied from 2.4 to 4.8 μm and the depth of the square groove is adjusted from 2.4 to 4.8 μm. When the width of the groove, w, is fixed, i.e., constant w/r ratio, the incident angles at which patchy particle fabrication (smallest patch) begins and where Janus particle fabrication (50% patch) occurs increase with increasing groove depth, while the effective range of incident angles resulting in patchy particles narrows. When the groove depth, d, is fixed, i.e., constant w/r ratio, the beginning and ending points of the θ-range in which patchy particles are fabricated decreases with increasing groove width with a similar narrowing of the range. From a practical point of view, a groove with width and depth similar to the diameter of the particle is

Figure 4. Coated-area percentage, A%, plotted as a function of incident angle, θ, for a 2.4 μm particle located at the center of a square groove with varying width, w, and depth, d. For a specific groove width, w, the groove depth, d, is varied from 2.4 μm (solid circle, leftmost line) to 4.8 μm (solid pentagram, right-most line). The inset shows the particle-template geometry; see text for details.

preferred, since it provides the largest effective range of incident angles and a more precise control of the patch size with respect to θ. Another advantage is that the patches will have identical shapes if they cover the same area independent of the template used, because the patch is created by the shadow of the groove walls. The slightly curved nature of the lines discussed in the context of Figure 2 can also be observed in Figure 4. In summary, we are able to use a Si wafer template to assist with the fabrication of patchy particles with uniform patches through the glancing angle deposition technique. If the vapor flux is kept perpendicular to the groove direction, the patch shape produced is always symmetric and takes the shape of one-half of a spherical cap. The patch shape and size is controlled by the incident angle, θ, particle size, r, and the dimensions of the template. Note that a change of the vapor flux direction with respect to the groove direction enables access to a large variety of asymmetric patch shapes that are the topic of future research. The ability of fabricating uniform and controlled patches, the ease of use in carrying out the experiments, the reusability of the template, and the 3100fold increase in particle volume make this template-assisted method interesting, versatile, and scalable. In addition, the use of industrially employed substrates and lithography techniques may allow for easier implementation in an industrial-scale process. Simple geometric models illustrate the relationship between coated-area percentage and the incident angle of vapor flux, the template dimensions, and the particle size.

■

ASSOCIATED CONTENT

S Supporting Information *

Experiment details on template preparation. This material is available free of charge via the Internet at http://pubs.acs.org. 9918

dx.doi.org/10.1021/la3017563 | Langmuir 2012, 28, 9915−9919

Langmuir

■

Letter

(22) McConnell, M. D.; Kraeutler, M. J.; Yang, S.; Composto, R. J. Patchy and Multiregion Janus Particles with Tunable Optical Properties. Nano Lett. 2010, 10, 603−609. (23) Lin, C.-C.; Liao, C.-W.; Chao, Y.-C.; Kuo, C. Fabrication and Characterization of Asymmetric Janus and Ternary Particles. ACS Appl. Mater. Interfaces 2010, 2 (11), 3185−3191. (24) Pawar, A. B.; Kretzschmar, I. Fabrication, Assembly, and Application of Patchy Particles. Macromol. Rapid Commun. 2010, 31 (2), 150−168. (25) Pawar, A. B.; Kretzschmar, I. Multifunctional Patchy Particles by Glancing Angle Deposition. Langmuir 2009, 25 (16), 9057−9063. (26) Pawar, A. B.; Kretzschmar, I. Patchy Particles by Glancing Angle Deposition. Langmuir 2007, 24 (2), 355−358. (27) Yake, A. M.; Snyder, C. E.; Velegol, D. Site-Specific Functionalization on Individual Colloids: Size Control, Stability, and Multilayers. Langmuir 2007, 23 (17), 9069−9075. (28) Cui, J.-Q.; Kretzschmar, I. Surface-Anisotropic Polystyrene Spheres by Electroless Deposition. Langmuir 2006, 22 (20), 8281− 8284. (29) Charnay, C.; Lee, A.; Man, S.-Q.; Moran, C. E.; Radloff, C.; Bradley, R. K.; Halas, N. J. Reduced Symmetry Metallodielectric Nanoparticles: Chemical Synthesis and Plasmonic Properties. J. Phys. Chem. B 2003, 107 (30), 7327−7333. (30) Skirtach, A. G.; Kurth, D. G.; Mohwald, H. Laser-embossing nanoparticles into a polymeric film. Appl. Phys. Lett. 2009, 94, (9). (31) Chen, Q.; Diesel, E.; Whitmer, J. K.; Bae, S. C.; Luijten, E.; Granick, S. Triblock Colloids for Directed Self-Assembly. J. Am. Chem. Soc. 2011, 133 (20), 7725−7727. (32) Knotter, D. M. Etching Mechanism of Vitreous Silicon Dioxide in HF-Based Solutions. J. Am. Chem. Soc. 2000, 122 (18), 4345−4351. (33) Rycenga, M.; Camargo, P. H. C.; Xia, Y. Template-assisted selfassembly: a versatile approach to complex micro- and nanostructures. Soft Matter 2009, 5, 1129−1136.

AUTHOR INFORMATION

Corresponding Author

*E-mail [email protected]. Notes

The authors declare no competing financial interest.

■

REFERENCES

(1) Manna, L.; Scher, E. C.; Alivisatos, A. P. Synthesis of Soluble and Processable Rod-, Arrow-, Teardrop-, and Tetrapod-Shaped CdSe Nanocrystals. J. Am. Chem. Soc. 2000, 122 (51), 12700−12706. (2) Sajanlal, P. R.; Sreeprasad, T. S.; Samal, A. K.; Pradeep, T. Anisotropic nanomaterials: structure, growth, assembly, and functions. Nano Rev. 2011, 2, 5883. (3) Halder, A.; Kundu, P.; Viswanath, B.; Ravishankar, N. Symmetry and shape issues in nanostructure growth. J. Mater. Chem. 2010, 20 (23), 4763−4772. (4) Romano, F.; Sciortino, F. Colloidal self-assembly: Patchy from the bottom up. Nat. Mater. 2011, 10 (3), 171−173. (5) Chastek, T. T.; Hudson, S. D.; Hackley, V. A. Preparation and Characterization of Patchy Particles. Langmuir 2008, 24, 13897− 13903. (6) Glotzer, S. C.; Solomon, M. J. Anisotropy of building blocks and their assembly into complex structures. Nat. Mater. 2007, 6 (8), 557− 562. (7) Zhang; Glotzer, S. C. Self-Assembly of Patchy Particles. Nano Lett. 2004, 4 (8), 1407−1413. (8) Noya, E. G.; Vega, C.; Doye, J. P. K.; Louis, A. A. The stability of a crystal with diamond structure for patchy particles with tetrahedral symmetry. J. Chem. Phys. 2010, 132, 234511. (9) Kretzschmar, I.; Song, J. H. K. Surface-anisotropic spherical colloids in geometric and field confinement. Curr. Opin. Colloid Interface Sci. 2011, 16, 84−95. (10) Zhang; Keys, A. S.; Chen, T.; Glotzer, S. C. Self-Assembly of Patchy Particles into Diamond Structures through Molecular Mimicry. Langmuir 2005, 21 (25), 11547−11551. (11) Chen, T.; Zhang, Z.; Glotzer, S. C. Simulation Studies of the Self-Assembly of Cone-Shaped Particles. Langmuir 2007, 23 (12), 6598−6605. (12) De Michele, C.; Gabrielli, S.; Tartaglia, P.; Sciortino, F. Dynamics in the presence of attractive patchy interactions. J. Phys. Chem. B 2006, 110 (15), 8064−8079. (13) De Michele, C.; Bellini, T.; Sciortino, F. Self-Assembly of Bifunctional Patchy Particles with Anisotropic Shape into Polymers Chains: Theory, Simulations, and Experiments. Macromolecules 2011, 45 (2), 1090−1106. (14) Sciortino, F.; Zaccarelli, E. Reversible gels of patchy particles. Curr. Opin. Solid State Mater. Sci. 2011, 15 (6), 246−253. (15) Romano, F.; Sanz, E.; Sciortino, F. Crystallization of tetrahedral patchy particles in silico. J. Chem. Phys. 2011, 134, 17. (16) Sciortino, F.; Giacometti, A.; Pastore, G. A numerical study of one-patch colloidal particles: from square-well to Janus. Phys. Chem. Chem. Phys. 2010, 12 (38), 11869−11877. (17) Goyal, A.; Hall, C. K.; Velev, O. D. Phase diagram for stimulusresponsive materials containing dipolar colloidal particles. Phys. Rev. E 2008, 77 (3), 031401. (18) Zhang, G.; Wang, D.; Möhwald, H. Nanoembossment of Au Patterns on Microspheres. Chem. Mater. 2006, 18, 3985−3992. (19) Zhang, G.; Wang, D.; Möhwald, H. Patterning Microsphere Surfaces by Templating Colloidal Crystals. Nano Lett. 2004, 5 (1), 143−146. (20) Bao, H.; Butz, B.; Zhou, Z.; Spiecker, E.; Hartmann, M.; Klupp Taylor, R. N. Silver-Assisted Colloidal Synthesis of Stable, Plasmon Resonant Gold Patches on Silica Nanospheres. Langmuir 2012. (21) Kraft, D. J.; Hilhorst, J.; Heinen, M. A. P.; Hoogenraad, M. J.; Luigjes, B.; Kegel, W. K. Patchy Polymer Colloids with Tunable Anisotropy Dimensions. J. Phys. Chem. B 2010, 115 (22), 7175−7181. 9919

dx.doi.org/10.1021/la3017563 | Langmuir 2012, 28, 9915−9919