Click Here! Click Here!
Cloud Elimination In DVD Production

Solved in CD manufacturing, cloud is again an issue.

By Franky K. L. Fan, Dr. Sam M. S. Yeung and William M. C. Ng

Sticking stain, or cloud mark, thought to be a solved cosmetic problem for compact disc, is appearing on the DVD discs of some well-known manufacturers. This is not surprising, as the process window of DVD is smaller than that of CD. Even though such cosmetic defects do not usually cause playback problems, they make the disc look unprofessional. In Hong Kong, consumers have learned to identify pirated discs by the cloud on the surface, as discs with cloud are not normally accepted by the big movie and music companies - thus, elimination of cloud on optical discs has become very important in the Asian market.

In this article, the methodology of identifying and eliminating the different types of cloud mark from optical discs is discussed based on experimentation conducted by Franky K. L. Fan, of Harmonic Hall Ltd., Hong Kong, with the cooperation of Meiki Co., Ltd. and GE Plastics.

As more and more CD manufacturers enter the DVD business, we believe one of the problems they will encounter is cloud mark. To solve the cloud problem in optical disc, close cooperation between the mastering and replication departments is vital. In the past, solutions for cloud have generally been proposed by the injection molding machine makers and plastic suppliers. A perfect stamper will sometimes produce discs with both a terrible appearance and signal, causing many process engineers to look for solutions in the replication process. However, in more than 90 percent of the cases where cloud is a problem, it can be solved by a mastering process adjustment.

High density optical disc mastering is different from the traditional CD mastering process. The difference is not only in the precision of the optical path design, but also in the process steps. Many problems which are not significant in traditional CD production become critical for DVD. For example, pit distortion and replication loss, not necessarily important in the CD replication process, become critical to the signal amplitude and symmetry of DVD. The pit slope of DVD has to be higher than CD in order to achieve good jitter and block error rate (BLER), due to the high spatial frequency of the information pits. In order to overcome symmetry loss and achieve good birefringence, engineers prefer to use high molding temperatures for the DVD family of products. In DVD-RAM replication, for example, a mold temperature as high as 135 degrees C is necessary to precisely replicate the pits and grooves, making it even more likely for cloud to occur. Moreover, the high mold temperature used in the replication process also replicates the minor imperfections on the stamper surface, such as photoresist and chemical residue. The shallower the grooves or the pits, the easier it is for the human eye to observe such marks.

What Is Cloud?

Cloud refers to cosmetic defects on an optical disc which are caused by imperfections on the stamper, or by problems encountered during the replication process. Pit damage and differences on a disc's surface cause light to be diffracted and reflected in a way that makes a disc look "dirty." This type of cloud becomes visible to the naked eye when the pit damage is uneven on different parts of the disc.#

Depending on the nature of the cloud, high BLER or jitter problems may develop. Cloud usually does not cause playback difficulties in CD, but serious cloud becomes critical in DVD, due to its more stringent specifications. Experimentation and research being conducted at Harmonic Hall has shown that there are several types of cloud, and that different solutions should be applied for different cases. The following are some of the reasons for cloud occurrence in a normal process setting:

  • • High pit slope
  • • Observable defect on stamper
  • • Residue photoresist or chemical on the stamper surface
  • • Improper back polishing
  • • Shrinkage in replication
  • • Non-parallel mold release

Cloud Due To High Pit Slope

Pit or groove slope is affected by many factors, such as wavelength of the recording laser, alignment conditions, optical path design, photoresist characteristics, exposure power and developing conditions. Assuming all factors are kept constant, the pit slope increases as the developing time increases. When the pit slope exceeds the maximum no-cloud pit angle of the system, cloud occurs. The steeper the slope, the more serious the cloud. The maximum no-cloud pit angle of the stamper is defined by the following formula:

MPA= CA - MRA-SC-SSC

  1. MPA: Maximum No-Cloud Pit Angle
  2. CA: Critical Angle for mold release
  3. MRA: Substrate Mold Release Angle
  4. SC: System Constant
  5. SSC: Stamper Surface Condition

The CA is a function of pit geometry and substrate cooling. It varies with pit height, pit width and cooling time. The narrower and shallower the pit, the higher the CA. The SC is related to the injection molding machine and mold design. The poorer the injection molding machine and mold design, the higher the SC. The value of SC is determined by experience. The higher the stamper surface and backside roughness, the higher the SSC. Different types of injection molding machines will show different characteristics. Generally speaking, a mold with both mechanical and vacuum clamping for holding the stamper is the best scenario.

#When every parameter is kept fixed, the solution to cloud is to reduce the pit slope. This can be done by reducing the development time or the stop value in the automatic developing control unit. However, mastering engineers tend to develop the master a little more in order to obtain good jitter and HF signal. Underdeveloped pits tend to produce bad jitter, low HF signal and high symmetry value even though the pit slope is low. This makes the disc difficult to read in the high speed mode of most ROM drives. Besides, the BLER is also much higher than normal. To preserve the signal quality while reducing the pit slope, it is necessary to change the input beam width of the objective lens, change the focus offset value, change the photoresist baking condition and change the signal duty.

If problems only happen to the symmetry when everything else is at the optimum point, change the duty cycle of the EFM signal or adjust the bias voltage of the modulator. The change in duty can be observed with a spectrum analyzer as changes in the higher order harmonics. Actually, signal duty change has been a standard feature of the new formatters and laser beam recorders. In our experiment, the Panasonic Dual Beam Laser Beam Recorder (LBR) was used. The advantage of this LBR is that we can handle all popular formats - including CD, DVD, CD-RW and CD-R - with a single laser. And, we found that duty control is an extremely good feature for DVD and CD mastering - it makes cloud control easy and effective. We used a UV laser for both DVD and CD mastering, and have not encountered any difficulties in cloud control thus far. Apart from mastering, different replication systems have different pit angles for cloud mark. Based on our experience, some injection molding machines can mold stampers with a maximum pit slope as high as 70 degrees without cloud, while the others can only accept a maximum angle of about 45 degrees. We believe the difference comes from the substrate release sequence, the injection molding process design and the parallelism of the mold during the mold opening process. In a high temperature molding system, the mold release process is very different from that of low temperature molding replication.

Even though cloud can be reduced by making changes to the injection molding process, we do not recommend solving the cloud problem during replication, as this will affect birefringence, dishing, diffraction efficiency and even the dynamic balance of the substrate, which can further affect the signal quality of the disc. Moreover, CD-ROMs with poor dynamic balance can damage the spindle motor of the CD-ROM drive due to disc vibration during high-speed rotation. For a disc spinning at 36X speed, the centripetal force is 1296 times higher than those spinning at 1X speed. Hence, the injection molding process for CD-ROM is more critical than for CD-Audio. In our view, replication engineers should try their best to maintain good substrate quality and leave the cloud problems to mastering, as long as the machine setting is normal. They should not change more parameters than the mold release sequence if the machine is in a normal production setting.

Sometimes, process engineers may find that as the pit slope increases, the cloud mark becomes invisible. This is because every part of the disc has cloud. As there is no contrast difference, the cloud cannot be observed just by viewing it with the naked eye. However, if the substrate is analyzed by AFM or SEM, the pit damage can be clearly seen. By using order measurement equipment, cloud can also be seen as a big variation and increase in first diffraction order of the pits. It can also be observed as a reduction in zero order reflectivity. Generally speaking, the lower the pit slope and the narrower the pit width, the less the cloud. The smaller the mold non-parallelism and the smaller the substrate release angle, the less the cloud. The smoother the stamper surface, the less the cloud.

Cloud Due To Observable Defects On Stampers

Cosmetic defects on a stamper usually come from the photoresist coating process, or the metalizing process if the chemical method is used. Such stains can easily be observed on the stamper by blowing some water vapor on the cold stamper surface. Cosmetic defects, or surface stains, will appear as areas with contrast differences. Photoresist will sometimes peel off from the glass master or the stamper surface after galvanizing due to improper coating, also resulting in a visible stain. To avoid the photoresist coating stain, take the following steps:

  • • Use absolutely clean and defect-free glass;
  • • Thoroughly flush the chemical piping; use proper coating conditions; and
  • • Change the nozzle positions.

Cloud Due To Residue Photoresist Or Chemical On The Stamper Surface

During the standard mastering process, stampers are exposed to UV and washed with sodium hydroxide to remove the photoresist. It is very difficult to remove the trace amounts of photoresist or chemical residue sticking on the stamper surface using a chemical method. The residue is invisible but can cause cloud during high temperature molding replication. The problem is particularly serious in nickel sputtering systems, where photoresist and nickel crosslink becomes strong. In chemical metalizing systems, the problem is less serious. Such residue photoresist makes the PC or PMMA substrate difficult to release from the stamper, resulting in cloud mark. Sometimes, process engineers will use chemicals such as hydrogen peroxide or acetone to clean the stamper surface to reduce the quantity of residue photoresist and chemicals. However, these chemical methods will not completely solve the problem, which becomes increasingly serious when shallow grooves are replicated in a high- temperature molding condition. There are several solutions. The first is the traditional plasma ashing method, where oxygen plasma reacts with the photoresist, or residue chemicals, in a high-vacuum environment. This method is quite popular in Japan, but precise control is necessary.

The second method is to further reduce the slope of the pits or grooves. From our experience in CD mastering, the maximum no-cloud pit angle (MPA) for stampers which use sputtering for the first metal layer deposition is about 5 degrees lower than those using the wet nickel method, if the photoresist is only to be removed by wet process. At Harmonic Hall, we use the post sputtering method which is a proprietary process developed for DVD and CD-RW stampers. In this process, a second thin film is coated on the finished stamper surface by sputtering. The purpose of this thin film is to cover any residue photoresist or chemicals and to make the surface homogeneous. By using a different sputtering rate, substrate temperature and thin film type, a different stamper surface smoothness can be achieved, increasing the MPA of the stampers.

#From our production experience in DVD replication, we find this method effective in preserving the pit profile while removing the clouds. By applying different thin film layers, we believe that it is possible to further lower the adhesion between the substrate and the stamper and, hence, effectively reduce the cloud while increasing the MPA. This is very critical for high-density optical disc production, as the optimum pit slope for high-density optical discs is higher than traditional CDs. In our experiment, we found that the MPA for stampers with post sputtering is higher than those with just plasma cleaning. We believe that by applying a different kind of second thin film after plasma ashing by post-sputtering, it is possible to reduce adhesion between the PC and the stamper to achieve easier mold release. This is critical for the future production of super density discs.

Cloud Due To Improper Back Polishing

Stamper back polishing is related to cloud. If the back roughness is high after polishing, the heat transfer efficiency of the stamper will be lowered as the contact surface between the stamper and the mold is decreased. This causes cloud due to inadequate cooling. To eliminate such cloud, the maximum back roughness of less than 0.5um is necessary for DVD stampers. However, loss in cooling efficiency can also be caused by the water precipitation inside the mold. We recommend the use of pure water for cooling purposes.

Cloud Due To Shrinkage In Replication

When serious shrinkage happens in replication, cloud occurs. The pattern of such cloud varies for different mold design, but the pattern is generally circular. Usually, it appears as dark circular bands. If the transparent substrate is observed under polarized light, circular shrinkage bands can be observed as dark rings from the inner to outer radius. Cloud caused by shrinkage cannot be solved in mastering. It must be solved during replication. If a standard machine is used, cloud can be eliminated by the following measures:

  • • Higher mold temperature
  • • Higher injection stroke
  • • Higher injection speed/pressure
  • • Higher hold pressure
  • • Lower initial clamping force
  • • Larger nozzle

Future Direction For Cloud Control

As optical disc manufacturing moves forward, we believe the optimum pit slope for good signal should be even higher as the pits get smaller. This makes mold release difficult and cloud more likely. We will continue to research second thin film coating on stampers in order to solve the problem. By coating some thin film on the stamper surface, we believe it is possible to improve the stamper surface smoothness and reduce the adhesion between the substrate and stamper. This will make mold release easier. On the other hand, the hardness of the stamper has to be increased to reduce stamper deform during mold release. As the physics of cloud have become more clear to researchers, standard solutions for cloud elimination have been developed. The time has come for mastering equipment manufacturers to integrate new processes in the LBR for cloud elimination.

On the replication side, mold design, which ensures the parallel release of substrate from the stamper, is important. To ensure the maximum stability of the mold release sequence, it is suggested that small air compress buffers should be installed in the air blow circuitry of the injection molding machines. As machines with higher precision are necessary to reduce the system constant (SC), we feel the development of electric servo type of injection molding machines will be a direction for future development.

For more information about Cloud Elimination, contact, Harmonic Hall Optical Disc Ltd., Tel: 852-241-21388; Fax: 852-241-42333 or refer to the references sited below.

References:

  1. Franky, K.L. Fan, "Towards Shorter Replication Cycle Time," TapeDisc Business, February, 1997.
  2. K. Hibi, Lecture Presentation, Meiki Co., Ltd, July, 1995.
  3. Clouds, G.E. Plastics