Effects of humidity and temperature on laser-assisted dip-pen nanolithography array using molecular dynamics simulations

(1)

Effects of humidity and temperature on laser-assisted dip-pen

nanolithography array using molecular dynamics simulations

Cheng-Da Wu

a

_{, Te-Hua Fang}

a,⇑

_{, Tsung-Tse Wu}

b

a

Department of Mechanical Engineering, National Kaohsiung University of Applied Sciences, Kaohsiung 807, Taiwan

b

Institute of Mechanical and Electromechanical Engineering, National Formosa University, Yunlin 632, Taiwan

a r t i c l e

i n f o

Article history:

Received 27 September 2011 Accepted 20 January 2012 Available online 28 January 2012 Keywords: Dip-pen nanolithography Array Humidity Molecular dynamics Temperature

a b s t r a c t

Two-dimensional dip-pen nanolithography (DPN) combined with laser-assisted heating is studied using molecular dynamics (MDs) simulations. The effects of humidity, deposition temperature, heating rate (laser-assisted patterning), and cooling rate on ink molecules are evaluated in terms of molecular trans-ference, alkanethiol meniscus characteristics, surface binding energy, number of transferred chains, pat-tern characteristics, and the diffusion coefficient of ink molecules. The simulation results clearly show that the number of molecules transferred significantly increases with increasing humidity, which leads to increases in meniscus size and pattern size. The surface binding energy decreases and the diffusion coefficient of ink molecules increases with increasing humidity and deposition temperature. The dwell stage has the largest number of molecules transferred and the largest diffusion distance of ink molecules. The number of vaporous water molecules increases when the temperature is above 300 K, which limits meniscus growth and leads to unstable deposition. The DPN transfer efficiency can be significantly enhanced by increasing the laser pulse energy/heating rate. The transfer efficiency improves as the sys-tem humidity increases to saturation (374 water molecules).

1. Introduction

Dip-pen nanolithography (DPN)[1]has become a versatile and widely used technique for the direct patterning of versatile organic or inorganic materials on solid substrates. DPN platforms are capa-ble of generating features as large as 10

l

m and as small as 50 nm. In DPN, an atomic force microscope (AFM) tip is used to deliver the desired molecules, such as self-assembled monolayers (SAMs)[1], proteins[2,3], polymers[4], biomolecules[5], and inorganic mate-rials[6,7], onto a variety of substrates[8–12]via a solvent menis-cus, which naturally forms in the ambient atmosphere. Several molecular inks[1–7] can be deposited or aligned on a substrate by the tip because AFM can precisely control the position of depo-sition. DPN with a single probe tip can arbitrarily create patterns with high resolution and registration, but its throughput is usually limited[13,14]. Recently, two-dimensional DPN (a tip array) has emerged as a viable and robust large-area patterning method that can produce patterns with nanoscale feature sizes[15,16].

In the DPN process, the meniscus is the channel via which ink molecules are transported to the substrate. Previous experiments and simulations found that humidity [1,17] and temperature

[16–18]directly affect meniscus size. Hence, an understanding of these effects on the transport and adsorption of ink molecules in

the DPN process is essential to developing and fully controlling DPN. Giam et al.[19]experimentally conﬁrmed a model in which ink deposition rates are governed by ink surface coverage. How-ever, most fundamental properties (e.g., transfer mechanism, mechanics, molecular spread on the substrate, etc.) are extremely difﬁcult to obtain from experiments because molecules are quickly transported to the substrate before feature growth. Therefore, the development of a molecular-level model that considers the com-plex physical and chemical interactions between molecules is required. Molecular dynamics (MDs) simulations are a powerful atomic modeling technique for studying material behavior and atomic/molecular interactions at the nanometer scale. We have recently studied the formation and physical mechanisms in the DPN process using an MD simulation[20]and found that molecu-lar transfer ability strongly depends on temperature. Ahn et al.[21]

and Heo et al.[22,23] studied the effect of molecule–substrate binding energy in the DPN process. They found that increasing the molecule–substrate binding energy increases the molecular deposition rate and makes the monolayer well-ordered. For laser–material interaction, Shibahara and Kotake[24,25] studied the interaction between metallic atoms and a laser beam; they found that the structural change of metallic atoms is due to laser beam absorption. Lai et al. [26] studied microscopic spallation mechanisms at the solid-state interface induced by a pulse laser and found that the occurrence of structural spallation is mainly characterized by extraordinary expansion dynamics and tensile

⇑Corresponding author.

E-mail address:[email protected](T.-H. Fang).

Journal of Colloid and Interface Science 372 (2012) 170–175

Contents lists available atSciVerse ScienceDirect

Journal of Colloid and Interface Science

(2)

stress. Huang and Lai[27]studied the pressure-induced solid-state lattice mending of nanopores in single-crystal copper by femtosec-ond laser annealing processes and found typical lattice mending phases.

In the present study, terminal linear alkanethiols with the gen-eric formula (CH3(CH2)15SH, hexadecane thiol (HDT)), a

prototyp-ical molecule in DPN, are deposited on a substrate. The effects of humidity, deposition temperature, and laser-assisted patterning (ink molecules heated by a pulsed laser) on the DPN array are sys-tematically investigated. The deposition results are evaluated in terms of molecular trajectories, surface binding energy, number of transferred chains, diffusion coefﬁcient, and pattern formation. 2. Methodology

2.1. Physical model

Fig. 1shows a schematic model of DPN array deposition of an MD simulation. The physical model consists of three silicon (Si) probe tips, a gold (Au) substrate, HDT molecules adsorbed onto the tips, and 748 water molecules adsorbed onto the substrate. To simplify the DPN array analysis, the tips are assumed to be a ri-gid body with a radius of 1.5 nm. The Au substrate consists of a perfect face-centered cubic (FCC) single crystal with a length, width, and height of 17.1, 5.3, and 2.3 nm, respectively. Two ﬁxed layers of Au atoms are imposed beneath the substrate to constrain the whole system in the vertical direction. The 297 HDT molecules adsorbed onto each tip underwent an MD equilibrium run of 300 ps in order to achieve energy relaxation. In the simulation,

the tips have a common constant unit displacement of

2 105_{nm per time-step moving along the Z-axis to deposit a}

SAM pattern and pull off from the substrate. The time-step unit was 1015_{s. A periodic boundary condition was applied to the}

X-and Y-axes of the Au substrate to simulate a large system by mod-eling a small part that is far away from the edge. The distance be-tween tips was set to 8 nm. Before the DPN simulation started, the distance between the tips (including the HDT molecules) and the substrate was set to 1.2 nm for all tested cases. The tips were then indented to a given depth. The whole system was simulated at a temperature of 300 K.

2.2. Potential functions

Morse potential[20] is used here to describe the interaction among Au atoms. The potential form is UðrijÞ ¼ Dfexp½2

a

ðrij

r0Þ 2 exp½

a

ðrij r0Þg. The D,

a

, and r0 parameters in the

simulation were 0.4826 eV, 1.6166 Å1_{, and 3.004 Å, respectively,}

where D represents the bonding energy between two atoms,

a

rep-resents the material constant, and r0 represents the equilibrium

distance between two atoms. The alkylthiol chain description pre-sented by Hautman and Klein[28]is used in this study. The CH2

and CH3groups were treated as a single spherical molecule to

sim-plify the model to 17 united molecules per chain. Because the bonds undergo distortion and the chain length decreases under high compression, the Hautman and Klein model was modiﬁed to correct the potential functions to allow for bond stretching and the 12-3 chain-surface, as mentioned in Ref.[29]. The bond-bending terms were modeled using a harmonic potential. The tor-sional terms were assumed to have a Ryckaert–Bellemans dihedral potential form, which is a power series expansion of the dihedral angle[28]. The intermolecular interaction and intramolecular non-bonding interaction for atoms separated by more than three bonds along a chain were represented by the Lennard–Jones potential. The Lennard–Jones potential function was also employed to de-scribe the physisorption interaction between the tip and HDT mol-ecules[30]. To describe molecular transfer, Luedtke and Landman

[31]modeled the HDT-Au interaction as a corrected Morse poten-tial function (S atoms and Au substrate have chemisorption inter-action) and ﬁtted the parameters to experimentally obtained binding energies. The potential parameters of water–water are adopted from Ref.[32].

2.3. Laser absorption mechanism

An increase in temperature can effectively increase molecular diffusion. DPN combined with laser-assisted heating is studied to observe the transfer mechanism of ink molecules. To simplify the complicated light-to-heat conversion mechanism with a nonlinear energy characteristics, the laser energy is assumed as uniform absorption by the ink molecules adsorbed on tips and converted into kinetic energy by scaling the velocities [26,27,33] of these molecules with an approximate factor (the factor is a function of laser incident energy); namely, all ink molecules have the same in-crease in kinetic energy with time for a given incident energy. Total incident energy densities of 0.99–7.92 J/m2_{and duration of 100 ps}

are performed sequentially during simulation. 3. Results and discussion

3.1. Effects of humidity

To study the effect of humidity on the molecular transfer pro-cess, a layer of water ﬁlm consisting of 187, 374, and 561 mole-cules was physisorbed onto the substrate surface before indentation to represent surfaces with various levels of humidity; those correspond the masses of water molecules per unit volume of 10.3, 20.6, and 30.9 kg/m3, respectively. In the simulation, the indentation was carried out in the ﬁrst 100 ps, followed by a dwell period of 20 ps. The tips were then gradually pulled off. For DPN, the transfer process of ink molecules begins once the meniscus forms at the interface between the tips and the substrate. The meniscus naturally forms when the tips are about to make contact with the substrate surface. The meniscus gradually grows as the tips are pulled away from the substrate. The formation mechanism of the meniscus is mainly controlled by the strong chemisorption force at the Au substrate-SAM interface and Van der Waals (VDW) interactions between ink molecules/ink molecules and ink molecules/tips. At the pull-off stage, the meniscus is continuously pulled by the tips until it breaks, at which point the molecular transfer process ends.Fig. 2a–d shows snapshots of the DPN pro-cess at the pull-off stage for systems with 0, 187, 374, and 561

Fig. 1. Schematic illustration of the array deposition of the DPN process under a humid environment. The humid environment was simulated by adding a layer of water ﬁlm on the substrate surface before indentation.

(3)

the SAM on the substrate/ink molecules in the meniscus) being greater than the interactive forces between ink molecules in the meniscus and ink molecules adsorbed on the tips, which leads to more ink molecules being transferred onto the substrate. When the cooling rate is increased, the temperature of the ink molecules quickly cools down to room temperature, and the number of trans-ferred molecules decreases due to the lower kinetic energy of mol-ecules.Fig. 10a and b shows the numbers of transferred chains for the two heating and cooling conditions. The numbers of trans-ferred chains are 176, 191, 203, and 226 for heating rates of 1, 2, 6, and 8 K/ps and a cooling rate of 1 K/ps, and 202, 197, 173, and 159 for heating rates and cooling rates of 1, 2, 6, and 8 K/ps, respec-tively. The results indicate that the number of transferred chains increases with increasing heating rate and decreasing cooling rate.

Fig. 11a and b shows the variation of the diffusion coefficient with time for the two heating and cooling conditions. The variation of the curves of the diffusion coefficient with time has a trend similar to that of the curves of binding energy inFig. 8. This indicates that the diffusion coefficient of the ink molecules is sensitive to the changes of heating rate and cooling rate.

4. Conclusion

MD simulations were utilized to investigate the laser-assisted dip-pen nanolithography array process. The effects of humidity,

deposition temperature, heating rate, and cooling rate on molecular transfer were analyzed. The following conclusions were obtained:

(1) The number of molecules transferred increases signiﬁcantly with increasing humidity, which leads to increases in menis-cus size and pattern size.

(2) The surface binding energy decreases and the diffusion coef-ﬁcient of ink molecules increases with increasing humidity and deposition temperature.

(3) The number of transferred ink molecules and their diffusion ability are highest at the dwell stage.

(4) The transfer efﬁciency of the DPN process improves as the system humidity increases to saturation (374 water mole-cules in the system).

(5) Vapor water molecules limit meniscus growth for tempera-tures above 300 K.

(6) Laser heating of ink molecules can effectively increase the efﬁciency of molecular transfer by increasing the heating rate.

Acknowledgment

This work was supported by the National Science Council of Tai-wan under Grants 2628-E-151-003-MY3 and NSC-100-2221-E-151-018-MY3.

References

[1] R.D. Piner, J. Zhu, F. Xu, S. Hong, C.A. Mirkin, Science 203 (1999) 661. [2] K.B. Lee, S.J. Park, C.A. Mirkin, J.C. Smith, M. Mrksich, Science 295 (2002) 1702. [3] B. Elena, M.D. Rocio, R.M. Daniel, L. Anabel, M. Daniel, Adv. Mater. 22 (2010)

352.

[4] A. Noy, A.E. Miller, J.E. Klare, B.L. Weeks, B.W. Woods, J.J. DeYoreo, Nano Lett. 2 (2002) 109.

[5] C.C. Wu, D.N. Reinhoudt, C. Otto, V. Subramaniam, A.H. Velders, Small 7 (8) (2011) 989.

[6] M. Su, X. Liu, S.Y. Li, V.P. Dravid, C.A. Mirkin, J. Am. Chem. Soc. 124 (2002) 1560. [7] X. Liu, L. Fu, S. Hong, V.P. Dravid, C.A. Mirkin, Adv. Mater. 14 (2002) 231. [8] C.A. Mirkin, ACS Nano 1 (2007) 79.

[9] T.H. Fang, W.J. Chang, C.D. Wu, Microelectron. Eng. 85 (2008) 223.

[10] T.H. Fang, W.Y. Chang, S.J. Lin, C.N. Fang, J. Colloid Interface Sci. 345 (2010) 19. [11] O.A. Nafday, M.W. Vaughn, B.L. Weeks, J. Chem. Phys. 125 (2006) 144703. [12] Y. Cho, A. Ivanisevic, Langmuir 22 (2006) 8670.

[13] S. Kramer, R.R. Fuierer, C.B. Gorman, Chem. Rev. 103 (2003) 4367. [14] R. Maoz, S.R. Cohen, J. Sagiv, Adv. Mater. 11 (1999) 55.

[15] J. Haaheim, V. Val, J. Bussan, S. Rozhok, J.W. Jang, J. Fragala, M. Nelson, Scanning 32 (2010) 49.

[16] D. Bullen, S.W. Chung, X. Wang, J. Zou, C.A. Mirkin, C. Liu, Appl. Phys. Lett. 84 (5) (2004) 789.

[17] S. Rozhok, R. Piner, C.A. Mirkin, J. Phys. Chem. B 107 (2003) 751.

[18] R.G. Sanedrin, N.A. Amro, J. Rendlen, M. Nelson, Nanotechology 21 (2010) 115302.

[19] L.R. Giam, Y. Wang, C.A. Mirkin, J. Phys. Chem. A 113 (2009) 3779. [20] C.D. Wu, T.H. Fang, J.F. Lin, Langmuir 26 (2010) 3237.

[21] Y. Ahn, S. Hong, J. Jang, J. Phys. Chem. B 110 (2006) 4270.

[22] D.M. Heo, M. Yang, S. Hwang, J. Jang, J. Phys. Chem. C 112 (2008) 8791. [23] D.M. Heo, M. Yang, S. Hwang, H. Kim, L.C. Saha, J. Jang, J. Phys. Chem. C 113

(2009) 13813.

[24] M. Shibahara, S. Kotake, Int. J. Heat Mass Transfer 40 (1997) 3209. [25] M. Shibahara, S. Kotake, Int. J. Heat Mass Transfer 41 (1998) 839. [26] H.Y. Lai, P.H. Huang, T.H. Fang, Appl. Phys. A 86 (2007) 497. [27] P.H. Huang, H.Y. Lai, Nanotechnology 19 (2008) 255701. [28] J. Hautman, M.L. Klein, J. Chem. Phys. 91 (1989) 15. [29] K.J. Tupper, D.W. Brenner, Langmuir 10 (1994) 2335. [30] I.H. Sung, D.E. Kim, Appl. Phys. A 81 (2005) 109.

[31] W.D. Luedtke, U. Landman, J. Phys. Chem. B 102 (1998) 6566.

[32] Q. Zhang, J. Zheng, A. Shevade, L. Zhang, S.H. Gehrke, G.S. Heffelﬁnger, S. Jiang, J. Chem. Phys. 117 (2002) 808.

[33] X. Wang, J. Phys. D: Appl. Phys. 38 (2005) 1805. [34] P.V. Schwartz, Langmuir 18 (2002) 4041.

[35] P.E. Sheehan, L.J. Whitman, Phys. Rev. Lett. 88 (15) (2002) 156104. [36] Q. Xia, C. Keimel, H. Ge, Z. Yu, W. Wu, S.Y. Chou, Appl. Phys. Lett. 83 (2003)

4417. Fig. 11. Variation of the diffusion coefﬁcient with time for (a) heating rates of 1–

8 K/ps and a cooling rate of 1 K/ps and (b) equal heating and cooling rates (1–8 K/ ps).