A chemical vapor deposition (CVD) polymer technology has been developed to use low cost monomers and provide highly crystalline and thermally stable polymer films for the sub-65μm ICs and OLED display applications. This thin film is expected to be a critical enabler for making flexible touch screens, TFTs, solar cells and displays, and very thin fuel cells. The applications include sheet computers, e-books, e-paper, actuators, sensors, memories, RFID tags, solar panel and batteries.
New reactor chemistries have been developed , so low-cost monomers can be used to make various parylenes that have a general composition of (I) (-CX2-Ar-CX2-)n, wherein, Ar is an aromatic moiety, X can be H or F. CVD equipment  that uses the reactor chemistries for cracking monomers is commercially available for deposition of various parylenes, including PPX-N and PPX-F (Ar =C6H4, X=H & F). PPX-F has a higher thermal stability and mechanical strength than any existing parylenes, including PPX-N. It has been qualified for commercial applications as a low-k intermetal dielectric (IMD) [3, 4] for sub-65nm ICs, and as encapsulation of OLED displays [5-7]. It is now undergoing evaluations for manufacturing of flexible and transparent touch screen and thin-film transistors (TFT), solar panels, and fuel cells.
Over the past 50 years, various parylenes  or poly-para-xylylenes have been made using the known Gorham method  that involves thermal cracking of dimers (II) (-CX2-Ar-CX2-)2 in vacuum. Due to the presence of the bucked benzene rings, dimers have high ring strain energy (31 Kcal/Mol for Ar = C6H4, X=H in (II)) , thus during synthesis of dimers, formation of parylene is favored over that of dimers . To gain high yield, low concentration of monomers (III) (Y-CX2-Ar-CX2-Y, Y=Br or I) in large equipment is employed to prevent multi-molecular collisions of reactive diradical intermediate (IV) (*CX2-Ar- X2C*, *= unpaired electron) and formation of parylenes. The presence of bulkier F groups prevents the F-dimers (Ar= C6H4, X=F in (II)) from becoming commercially available in large scale [12, 13].
In general, commercial dimers are very expensive (~$400-500/Kg for N-dimer) compared to monomers. Therefore, over the last five decades, the Gorham method found only limited applications for industrial-coating applications.
New monomer chemistry and implications
Before the invention of the Gorham method, many attempts  were made to thermally crack low cost monomers in vacuum. However, even the temperature of the tubular reactor, or “cracker” was as high as 1000°C. Due to poor thermal conductance and the short residence time (a few seconds) of the monomers in a vacuum of a few mTorrs, these efforts resulted in a large amount of carbon deposit inside the reactor and many C-Y bonds at polymer chain-ends. The resulting parylenes have low molecular weight and poor thermal stability (<150–200°C) and are useless for industrial applications.
New reactor chemistry
Since 2002, new reactor chemistries  have been used to convert monomers into diradical (IV) (*CX2-Ar- X2C*) with more than 99.9% yield. This was accomplished even though the reactor temperature was only about 660°C, similar to processing conditions for cracking of commercial dimers.
A newly designed reactor was constructed from a chosen transition metal and was irradiation heated to the inside space of the reactor, in contrast to the conductively heated stainless tube used for cracking dimers. New regeneration processes have been developed specifically for a given transition metal to remove carbon deposits inside the reactor. The regeneration chemistries comprise an oxidation step to remove the carbon deposit by oxygen and a hydrogen reduction, and a second step to recover the pure metallic inner surfaces from their metal oxide state.
To maintain continuous deposition at a constant rate, a dual-reactor set is used per CVD equipment, so when one virgin reactor is under operation, the other deactivated reactor will be under regeneration. To qualify the CVD equipment for commercial operation, the regeneration chemistries have been fully tested by continuous depositions of >500 wafers that resulted in a +/- 2.5% wafer-to-wafer-film-thickness variation at 1σ.
Features of the CVD polymer process
In the CVD process flow , a reactive diradical is formed in a reactor located outside the deposition chamber. The diradical is transported from the reactor by diffusion to the deposition chamber. Inside the chamber, the diradical is adsorbed onto a low-temperature substrate and polymerizes very rapidly—even at a substrate temperature as low as –50°C—to form a polymer thin film.
This process is capable of making very thin coatings, which is important for coating applications relevant to nanotechnology. For the deposition process using a monomer that has a high vapor pressure at temperatures ranging from 60–80°C, a vapor flow controller (VFC) kept at 120°C can be used to control not only the monomers’ feed rate, but also the deposition end point. The deposition rate of this CVD process depends only on the feed rate of the monomers and the substrate temperature, or the adsorption rate of the diradical on the substrate. When the feed rate of monomers is kept at 3sccm by VFC, the deposition rate of PPX-F on Si or glass substrate increases from 100 to 3000Å/min., as the substrate temperature decreases from 10 to –40°C. Due to good end-point control using VFC, a consistent thickness of PPX-F film at 20Å has been demonstrated. Therefore, this CVD process is now available for making the organic buffer layer  for ITO and as the protective layer  for the organic emitter compound.
In contrast, since all conventional dimers are solid and have low vapor pressure, they need to be heated to temperatures above 120–140°C to obtain sufficient feed rates. But at 140°C, condensation of dimers still occurred inside the VFC, thus rendering the VFC useless for flow control. When VFC is above the 140–160°C range, the electronic components inside the VFC failed quickly. In general, VFC is not used in conventional equipment for controlling feed rate of dimers. Instead, dimer temperature in a heated crucible regulates the dimers’ feed rate, while a mechanical valve controls the deposition endpoint. Existing commercial equipment is inherently incapable of accurately controlling low feed rates from 0.1–1sccm due to temperature fluctuation of dimers in vacuum, or for deposition of very thin films that have only few Angstroms of thickness, due to the absence of clear end-point deposition control.
Deposition can occur at a substrate temperature lower than 25°C, in contrast to conventional CVD, which is important for coating applications in the coming age of organic-electronics, and is critical when the substrate is a heat-sensitive organic material. Furthermore, because there is no plasma present in the deposition chamber, plasma sensitive devices such as OLED displays can be coated using the CVD deposition system.
Additionally, in contrast to all plasma CVD systems, the thermal CVD chamber never needs to be cleaned because it is kept above the maximum temperature for the adsorption of diradicals, or the ceiling temperature, Tc. For a diradical derived from the F-dimer, the Tc is ~10ºC, thus at room temperature, none of the diradical will form a thin film on the inner wall of the deposition chamber.
Figure 1. Gap-fill of 0.25μm Al standouts showing 95%
Lastly, this CVD deposition is a molecular layer deposition (MLD) process in that the thin film is formed one molecular layer at a time on low-temperature surfaces. The MLD deposition process can fill small gaps with high aspect ratios (Fig. 1) and results in conformal coating for products with very complex topology.
These new reactor chemistries enable the manufacture of newly designed parylenes that would not be available from the Gorham process, due to the presence of ring strain energy and unavailable dimers. Table 1 compares calculations based on five new parylenes with two known parylenes.
In Table 1, k is dielectric constant, Tg is the glass transition temperature, E and G are Young’s modulus and the shear modulus of these parylenes, respectively. All of these parylenes have much better thermal, mechanical, and electrical properties than PPX-N and PPX-F. For instance, the FF-PPX (-CF2-C6F4-CF2-)n would have a dielectric constant of 2.06, a thermal stability up to 450°C, and therefore is an ideal low-k IMD for IC applications at <34nm.
Verified applications. The PPX-F prepared from the patented monomer process  resulted in thin films consisting of more dimensionally stable β crystals with high crystallinity (Fig. 2) and have a Young’s modulus up to 14GPa. In contrast, the dimer process [18-21] resulted in unstable ά crystal (Fig. 3) with low crystallinity and a low Young’s Modulus of 3.5GPa. Although the F-dimer process failed, the monomer process  rendered the success integration of PPX-F into sub-65nm ICs [3, 4].
Figure 2. Comparative differential scanning calorimetries for PPX-F prepared from both monomer and dimer processes.
By coating only 0.1μm of PPX-F on a cathode, one can avoid using liquid descant and expensive etched glass for packaging of the resulting OLED displays and achieve a cost savings >50% for OLED display packaging .
Potential applications. PPX-F exhibits excellent chemical resistant, electrical insulation and mechanical properties (Table 2) among all existing polymers. PPX-F also has better light transparency  and is less expensive than clear polyimides (CPI) . The following summarizes the current studies being undertaken for applications of PPX-F or PPX-N in flexible electronics and displays.
Figure 3. Comparative wide-angle X-ray diffractions for PPX-F for the a) dimer process at 40°C and b) dimer process at 200°C, and c) monomer process at 25°C.
PPX-F can replace both SiN, the topcoat for TFT, and the polyimides (PIM), the alignment layer (AL) for LCD in active matrix LCD (AMLCD) displays. Compared to SiN, PPX-F has a lower dielectric constant and a lower leakage current [3, 4]. PPX-F also has a lower (0%) water adsorption than PI (2-3%) and SiN (0.5–1%). In addition, since both the topcoat and the alignment layer can be deposited at one step in a CVD equipment to avoid the high cost spin-on coating process for PIM , the cost of ownership for manufacturing LCD can be lowered when PPX-F is used.
PPX-N can be used to bind nanotubes and replace the brittle ITO in the commercial touch screen for a more flexible and transparent touch screen (FTTS) . To make the FTTS, the carbon nanotubes (CNTs) are first spray-coated over a plastic substrate, such as PET or siloxane-hard-coated polycarbonate (HCPC) . After drying the surfactant and solvents in coating, a thin layer of PPX-N is deposited into pinholes in the CNT layer. The PPX-N coating will bind all the CNTs together, and also bind the CNT layer to the plastic substrate. For this application, the coating cost, including the material and equipment is only ~5 cents/m2. Note that a (PPX-F+CNT) film is a patternable  organic conducting layer that can also be used to fabricate a transparent TFT or make solar cells .
Currently, to make flexible TFTs , SiO2 is deposited as a thermal barrier layer to protect a lower cost plastic substrate, such as PET. The SiO2 barrier layer is needed when PECVD is used to deposit the ά H-Si at 400°C and when an excimer laser is used to crystallize the ά H-Si into polycrystalline silicon, or LTPS . However, since the SiO2 is brittle, an expensive clear polyimide (CPI)  and a PES  have been studied for replacing the SiO2/PET substrate. To replace the brittle SiO2 on PET, the CPI and PES substrate, PPX-F thin film is used to overcoat the lower cost PET or a silicones-hard coated-PC (HC PC) for the above applications . Potential advantages of PPX-F over CPI and PES include:
PPX-F is a better dielectric material. It has lower water adsorption (0% vs.2%) and the lowest leakage current [3, 4] due to the absence of polarizable imide-groups in CPI.
PPX-F has lower CTE, so can provide smaller increases in threshold voltage (Vth) that result from thermal annealing-induced interface stress .
PPX-F has a higher Young’s modulus, thus it can result in lower intrinsic stress during fabrication  of LTPS.
Figure 4. PPX-F as hydrogen/proton separator in fuel cells.
A PPX-F thin film can be coated over the cathode and anode as the backings (Fig. 4) . PPX-F anode backings, after annealing at 400°C, would have a crystallinity of ~70% and their polymer chain-to-chain distance would be at ~1.45Å. The annealed PPX-F thin film is an ideal hydrogen/proton separator to prevent diffusion of hydrogen through the proton exchange membrane (PEM) and enter the cathode, and also to prevent diffusion of oxygen through the PEM to the anode. In principle, the PPXX-F can also be chemically modified by sulfonation so that the –SiO3H groups will be incorporated onto the aromatic moieties of the PPX-F. The sulfonated PPX-F can be used to replace the Nafions that are currently used as the PEM inside the membrane electrolyte assembly (MEA) shown in Fig.5.
The new reactor chemistries described here offer not only a lower cost process to make existing parylenes, such as PPX-N from monomers, but also provide parylenes such as PPX-F, that have more attractive material attributes. Additionally, the new chemistries offer an opportunity to make non-existing parylenes, such as FF-PPX, that will have a lower dielectric constant, but higher mechanical strength and higher thermal stability than PPX-F.
Flexible electronics that include organic TFTs  and transparent and flexible touch screens have become hot research subjects the last few years [20-28]. Suh  and Lee  have mentioned many advantages for ultra-light and thin, flexible displays and electronics and their potential applications for e-books, e-paper, actuators, sensors, memories, RFID tags, and solar cells. Also, ultra-thin fuel cells having much greater efficiencies are also needed for the coming age of mobile electronics and displays. The CVD polymer technology described in this article is expected to be a critical enabler for making these flexible electronics and displays.
Nafions is a trademark of DuPont.
- C. J. Lee, US Patents 6,881,447 (2004), 6,825,303B2 (2004).
- C. J. Lee, et al., US Patent 7,309.395 B2 (2007).
- C. J. Lee, A. Kumar, “Integrating Nonporous low-k (2.2) Film), Solid State Technology, pp. 69-74, July 2004.
- N. Kaji, et al., “Non-porous F-Arylene Film of k=2.2 with High Young’s Modulus and Adhesion for Cu/low-k Interconnect,” ADMETA Conference, Tokyo, 2003.
- C. Chen, A. Kumar, C. J. Lee, “Low Temperature Thermal CVD Process for Thin Film Encapsulation of OLED,” Symp. Proc., SID, Sect 60.1/C, 2007.
- C.H. Hsu et al., “High Work-function Hole Injector Material for Enhanced OLED Performance,” Symp. Proc., SID 5.4, 2006.
- X.L. Zhu et al., “Very Bright and Efficient TE-OLED with Yb Electron Injector,” SID 29.3, 2003.
- R.W. Beach et al., Encyclopedia of Polymer Sci. & Eng. Vol.17, 2nd ed., pp. 990-1025, 1989.
- W. F. Gorham, US Patent, 3,288,728 (1966).
- R.H. Boyd, “The Heat Combustion and the Strain Energy of (2,2) Para-cyclophane,” Tetrahedron Vol. 22, p. 119, 1966.
- S.W. Chow et al., US Patent 3,274,267 (1966); 5,302,767 (1994).
- W. R. Dolbier et al., US Patent 5,210,341 (1993).
- B.J. Jorgensen et al., “Synthesis and Properties of Fluorinated Parylenes,” Proc. JOWOG Meeting 28-A-Organic Coating, Oct. 6-8, 1987.
- Lee et al., US Patent Application Series Nos. 10/854,776 (2004), 10/897,797 (2004), 10/900,878 (2004), 11/155,209 (2005), and 11/642,383 (2006).
- C. J. Lee et al., US Patents 6,703,462 (2004), 6,881,447 (2007), 7,238,626 (2007).
- M. Morgan “Comparison of Crystalline Phase Transition in Fluorinated and Non-fluorinated Parylenes Thin Film,” Macromolecules Vol. 32, No 22 p. 7555, 1999.
- A.S. Harrus et al., “Parylene AF-4 low-k Material Candidate for ULSI Multilevel Interconnect Applications,” MRS Symp. Proc. Vol. 443, p. 21, 1997.
- M.A. Plano et al., “The Effects of Deposition Conditions on the Properties of Vapor Deposited Parylene AF-4,” MRS Symp. Proc., Vol. 476, pp. 213, 1997.
- J.B. Easton, “AF-4 as the low-k Dielectric for Gap Fill or IMD”, DUMIC Proc., 7.C, Feb 8-9, 1999.
- K.Y. Ho et al., “Temperature Instability of LT Deposited a-Si:H TFTF
Fabrication on Plastic Substrate,” SID, 16/6, p. 683, 2008.
- C. J. Lee, US Patents in application.
- A. Schindler et al., “Suspension-deposited CNT for Flexible AM Displays,” J. Soc. Inf. Display 16/6, pp. 651, 2008.
- J.Y. Kwon et al., “2.2 inch qqVGA AM OLED Derived from LTPS TFT,” SID 34/2, 2006.
- Y.D. Son et al., “AM OLED using CW Laser Crystallized Poly Si TFT,” SID, 28/1, 2006.
- Y. Fusjisaki, et al., “LCD Panel Driven by Organic TFT,” SID 10/4, 2006.
- K. McGrath, “Nanotechnology in Fuel Cells and Batteries for Portable Electronics,” Nanoelectronics, Nov. 9, 2006 (www.qsinano.com).
- N. Yoneya, et al., “All Organic TFT-Driven AM Polymer Dispersed LCD,” SID 10/5, 2006.
- M.C. Suh, et al., “Organic TFT-based AMOLED Displays,” SID 10/3, 2006.
- W. Lee, “Transmissive Plastic a-Si TFT-LCD Development,” SID 6/7, 2006.
Chung J. Lee received his PhD at RPI in 1975 and is founder and president of Dielectric Systems Inc. at 45500 Northport Loop West, Fremont, CA 94538; ph: 510-386-1209; email: email@example.com.