Novolac Polymer Research Articles

Mass spectroscopic and degassing characteristics of polymeric materials for 157 nm photolithography

We report on the mass spectroscopic and the degassing characteristics of an epoxy novolac-based chemically amplified photoresist, with triphenylsulfonium hexafluoroantimonate as photoacid generator, using an F2 pulsed discharge molecular laser at 157 nm. This photoresist has been used previously for optical (248 nm), e-beam, and X-ray microlithography both in wet and dry development processes. For laser energy of 1 mJ per pulse focused on the sample and 5 MW/cm2 intensity, photofragments with m/e less than 40 amu have a higher probability of formation. These fragments are mainly attributed to the breaking of the substituents of the aromatic rings. The specific fragments observed are: H, H2, C, CH, CH2, CH3, O, OH, H2O, HF, C2, C2H, C2H2, C2H3, CO, C2H4, C2H5, C2H6, H2CO, OCH3, O2, C3H3, C3H4, C3H5, C3H6, C3H7, C2H3O, C3H8, C2H5O, C3O, C3HO, C3H2O, C3H3O, C3H4O, C3H5O, C2HO2, C2H2O2, C2H3O2, C2H4O2, C2H5O2 and C2H6O2. Comparison of these results with the ones obtained from samples with pure epoxy novolac polymer show that the direct fragmentation of the side polymer bonds is predominant over bond breaking arising from the presence of the triphenylsulfonium salt sensitizer.

Melt flow of the silylated areas during the desire process

The DESIRE@ process has been proposed as an attractive solution to lithographic problems, combining the performance of multilayer systems to the simplicity of monolayer processes. Despite the large number of studies devoted to this type of process, the various mechanisms involved during the silylation and dry development steps are not yet totally understood. The effects of substrate heating, occuring during the dry development process, on the viscous strain of the silylated areas have been recently reported (1). The main results are listed hereafter : * The etch rate selectivity between silylated and non-silylated areas decreases from 15 to15 when the film temperature increases from 20 to 100 “C. * RBS measurements confirm the thermally activated diffusion of silylated chains in the novolac matrix. * Silylation induces a net decrease (from 70 to 5°C) of the glass transition temperature of the novolac polymer. Therefore, the so-called polymer fluid state can be reached during etching, and consequently allows the flow of polymer chains one with respect to each other. Thus, it was shown that the top silylated material can spread and coat the sidewalls of the pattern during the dry development step.

Etching of polymers by oxygen plasmas: Influence of viscoelastic properties

A study of novolac polymer etching in an oxygen microwave multipolar plasma with independent rf wafer biasing is reported. A step-like etch rate variation with temperature is observed for these polymers. Experiments conducted on chemically identical novolacs with different molecular weights allow this phenomenon to be correlated with their glass transition temperatures. Etch rate variations are caused by the thermal effect of ion bombardment, emphasizing the role of viscoelastic properties in polymer plasma etching.

New Products

To meet the new demands placed on solder resist technology, Coates have developed a new two‐pack epoxy solder resist, with a mixed pot life of over 36 hours, which they have named XZ2000. Based on high performance Epoxy Novolac Polymers and a completely novel hardener, XZ2000 is a screen printable solder resist system having improved application properties and a high technical performance with low toxicity.

Oxygen reactive ion etching mechanisms of organic and organosilicon polymers

The oxygen reactive ion etching characteristics of a silyl novolac polymer and an organic novolac polymer were investigated as a function of pressure, self-bias voltage, power density, radio frequency (rf) frequency, and other system variables. For each etching condition, sheath thickness measurements were used as a diagnostic to estimate the flux and average bombardment energy of ions and the energetic neutral products of charge transfer collisions. The estimated flux allows the observed etching rate to be converted into a yield per bombarding particle. The trends for the etching rate as a function of pressure and other system variables were found to reflect a single trend for the yield as a function of the average bombardment energy. The yield of the organic novolac is proportional to the bombardment energy and is equal to 12 carbon atoms per 100 eV. The silicon atom yield of the organosilicon novolac is equal to the sputtering yield of SiO2 as predicted by the steady-state model of organosilicon polymer etching. The selectivity in bilayer lithographic schemes (the ratio of the organic to organosilicon etching rates) increases as the bombardment energy decreases because the SiO2 sputtering yield is negligible for bombardment energies less than 50 eV.

Factors controlling the etching rate and etching profile in the O2 reactive ion etching pattern transfer step in multilevel lithography

AbstractPhysical bombardment plays a dominant role in the O2 reactive ion etching (RIE) pattern transfer step in multilevel lithography. Etching rates are determined by the flux and energy distribution of the bombarding ions and energetic neutrals (charge transfer products), while anisotropy is determined by their directionality. Measurements of the sheath thickness and voltage drop may be used to estimate flux, energy distribution, average energy, and angular distribution of ions and the energetic neutral products of charge transfer collisions. The estimated flux of bombarding particles allows measured etching rates to be converted into yields. The trends for the etching rate as a function of pressure, bias voltage, and other system variables reflect a single fundamental trend for the yield as a function of bombardment energy. Etching rates of an organic novolac polymer are proportional to the energy flux from bombarding particles while the yield per bombarding particle is proportional to its energy. These kinetics are combined with angular distribution and interface evolution models to predict etching profiles in multilevel lithography.

Mechanistic studies of oxygen plasma etching

Oxygen dissociation was measured in reactive ion etching (RIE) plasmas using a quadrupole mass spectrometer which sampled the particle flux through a pinhole in the lower electrode of a parallel-plate reactor. The oxygen atom flux increased with increasing power and generally decreased with increasing pressure. Ion currents and bombardment energies also decreased with increasing pressure, while the etch rate of novolac resists increased. In the low-pressure (20–75 mTorr) regime of oxygen RIE plasmas, etch rates of novolac polymers correlated most directly with the supply of oxygen molecules to the surface.

Experimental tests of the steady-state model for oxygen reactive ion etching of silicon-containing polymers

A steady-state model based on a silicon material balance has been proposed to predict the oxygen reactive-ion-etching resistance of organosilicon polymers. This model assumes that the rate determining step is sputtering of SiO2 film that forms on the surface of the organosilicon polymer. It predicts the etching rate of organosilicon polymers relative to the sputtering rate of SiO2 , based on the mass density of silicon in the polymer. The steady-state etching rate of a silyl novolac polymer is accurately predicted by the model over a wide range of etching conditions. Silyl methacrylates etch at the predicted rate under high-bombardment-energy conditions typical of trilevel processing, but exceed the predicted rate under low-bombardment-energy conditions. Surface analysis shows that the SiO2 film thickness continues to increase with time under these conditions, invalidating the steady-state approximation. Low silicon content (4.1 wt. %) polymers do not etch according to the model but form highly porous oxides that continuously accumulate on the surface of the polymer.

Molecular weight distributions in novolactype phenol-formaldehyde polymerization

To model the reversible novolac polymerization, five reactive species A to E have been defined. Molecules having bound CH 2OH (Q n ) are distinguished from those without it (P n ) and it is assumed that molecules of Q n do not have more than one bound CH 2OH group. A kinetic model has been written and, based upon it, balance equations for molecules of novolac polymer in batch reactors have been derived. Based upon our earlier studies, the phenomenon of molecular shielding has been neglected. As a result, the reactivities of the ortho and para positions of phenol which are available in the literature could be used. The kinetic model for the molecular weight distribution ( MWD) of reversible novolac polymer formation thus involves only one parameter. The study of the MWD of novolac polymer reveals two very important design variables: the phenol-formaldehyde ratio, [P] 0 [F] 0 , in the feed and the vacuum applied on the reactor. As the [P] 0 [F] 0 ratio is increased, the breadth of the distribution is found to increase and it undergoes a maximum at [P] 0 [F] 0 ⋍ 1.4 for the set of rate constants chosen. At this ratio, the chain length average molecular weight is also found to be the largest. Industrially, the [P] 0 [F] 0 ratio used in producing novolac polymer is 1.67 and it is usually desired that the polymer be linear with minimal branching. On application of vacuum, for a given time of polymerization, the chain length molecular weight is found to increase when the results are compared with those of batch reactors. The breadth of the distribution is also found to reduce thus giving a lower polydispersity index of the polymer formed.

Calculation of the isomeric composition of polycondensation polymers by the monte carlo method

The Monte Carlo method was used fo calculating the isomeric composition of phenolformaldehyde novolac polymers. Good agreement between calculated and experimental results was obtained.

Isomeric composition of phenolformaldehyde novolac polymers

Isomeric composition of phenolformaldehyde novolac polymers