Site: SHARP FACTORIES (Tenri & Nara)

Date Visited: October 9, 1991

Report Author: J. Covert

ATTENDEES

JTEC:

Covert
Hoffman
Slusarczuk
Tannas

HOSTS:

SHARP (Tenri):

Minoru Fukuoka

Group General Manager, LCD Group

Yasunori Nishimura

Deputy Head, A1153 Project Team, LCD Group

Hiroshi Take

General Manager, 2nd Product & Planning Dept, CDL Group

Jyunichi Matsuda

General Manager, Foreign Trade Relations Dept.,
International Business Group

Yasukazu Mori

Manager, Foreign Trade Relations Dept, International Business Group

SHARP (Nara):

Hisaaki Nakajima

Plant General Manager of Nara, LCD Group

Etsuo Mizukami

Manager, Products Planning Dept, LCD Group

Hideo Isozaki

Manager, Products Planning Dept, LCD Group

Hiromu Watanabe

Department General Manager, Engineering Dept 1,
Duty Panel Development Center, LCD Group

Toshio Wakatsuki

Department General Manager, Production Engineering Dept, LCD Group

Yukihiro Inoue

Department General Manager, Engineering Dept 2,
Duty Panel Development Center, LCD Group

Sharp was founded in 1912. They marked their 79th year of business in 1991, with 31 manufacturing plants in 23 countries. Their first product was the Ever-Sharp Pencil (invented by founder Tokuji Hayakawa), from which the company name and trademark are derived. Products include business equipment (computers, etc.), personal equipment (calculators, organizers, etc.), optoelectronics (TFT LCD, LED, EL display, etc.), commodity distribution systems, measuring and control equipment, manufacturing systems, TV/video, audio, air conditioning, kitchen appliances, home products, integrated circuits, and general components (thin-film sensors, tuners, etc.).

TENRI VISIT

The Sharp Tenri factory produces only TFT LCDs using a-Si. We were the first foreign group ever to see this new facility. The building, completed in February 1991, has a total floor area of 36,000 sq meters (83 m W x 87 m L x 31 m H), with production planned in one half of this area. The building took two years to design, and six months was planned for installation of equipment. At the time of the JTEC panel's visit, clean room equipment was being installed, and initial production was planned for January 1992. Their existing plant at this site is a 6,000 sq meter facility; 65% of this area is in production, currently making 70,000 TFT LCDs per month. By March 1992, when the new facility became operational, this facility was capable of producing 120,000 displays per month. About 1,000 people, of whom 60% are production workers, are working TFT LCD production at this facility.

Sharp's a-Si TFT 3-śm design rule process was described as using five masks for TFT fabrication and two for passivation and terminals. Test of a single transistor per panel on sample (not all) panels is used to characterize panel yield. Sharp buys its color filters from outside vendor(s). Tenri fabricates the TFT substrate and assembles the glass panel using automatic alignment techniques. The glass panel is tested before drivers are attached. Other subcontractor facilities assemble the electronics. Anisotropic adhesive is used to attach drivers.

Tenri builds the following TFT LCD products as shown in Table Sharp.1. Detailed specifications are attached for some products. They supply various modules to outside customers.

The new clean room facility we toured was undergoing final assembly of clean room equipment. The building floor-to-floor height was 8 meters, with 3.5-meter room height. A 40,000 panel/month production rate is planned similar to the 50,000 per month indicated earlier in our discussions. It appears that two more facilities like this can be repeated on two more floors.

Table Sharp.1
Tenri's TFT LCD Products

The CVD and sputtering equipment were all located in a class 100 area, and all appeared to be Anelva equipment we saw later in our visit to Anelva on October 11, 1991. The CVD equipment appeared to be with substrates starting at the Clean/Etch end of the line and being returned from the far end of the CVD line by robots on tracks to reenter the Clean/Etch area through a load-lock. Sputtering equipment was being installed. A 20- (or more) panel cassette was seen in the Clean/Etch area, but it was difficult to judge the substrate size. Our notes from this visit indicated that the next size, 500 mm x 550 mm, would take 4-5 years.

We also visited the basement area, where gasses were handled. We saw three recovery tanks. The area is huge by Japanese standards and can accommodate trucks for handling of materials. Nitrogen is made on site, but other materials are delivered. Thermal shock ESPEC, reliability, environmental, temperature, and vibration testing capability is in the basement. The facility is currently operating around the clock.

Concerning plans for substrate size, it was indicated that internally Sharp does not agree on its approach. Our hosts suggested that two approaches should be used: For small products, smaller mother glass should be used; and for workstation sizes, mother glass larger than 10 inches should be used.

It was indicated that the useful life of the machinery was 3-5 years. Equipment is depreciated over 5-6 years, consistent with equipment life. Plans are to build 3-inch displays in the old plant and all larger displays in the new plant. Sharp plans to build new lines every 2 years and insert new technology. Their experience in building plants is that it takes 1 year to complete the plant, 6 months to install equipment, and 1 year to achieve design yield, with the learning curve depending on the size of display.

Updating/engineering of equipment for a new plant is mostly done by Sharp, not by the equipment manufacturer. They mature a new process in their old facility before establishing a new facility. Production workers come straight from high school, and learn the job in the plant through team training. There is no formal training to learn this business. Our hosts indicated that the morale in the LCD Group is the highest in all of Sharp. They have dorms at Tenri for the production facility.

Sharp is planning for passive matrix LCD assembly in the United States, and, if this works out, it may be possible to build display manufacturing capability in the United States.

Sharp is doing p-Si research and is interested in both high- and low-temperature processes. They do not believe p-Si will be important for large displays. They believe that the cost of external drivers is lower than that of integrated p-Si drivers. They indicate that p-Si drivers cannot match Sharp's driver performance. They currently use 20 mHz analog drivers. When asked to compare TFT and MIM LCDs, they indicated that several requirements of MIM cell thickness, resistivity, and capacitance would limit application of MlMs.

When asked about their plans for producing cockpit displays, they indicated that on a commercial basis they need more than 5,000 displays per month for 2 years of a single design to produce such a display. They currently build auto displays with specification of -30 to +80 degrees centigrade for storage and -10 to +70 degrees centigrade for operation, and are planning to develop modules with -30 to +80 degrees centigrade operational capability. At temperatures around -55 degrees centigrade, LC crystallization is a problem.

NARA VISIT

Four LCD production lines are operational at Nara (only two are open to outsiders), producing TN and STN displays. Sharp's long history in displays was presented in detail. Their monthly production capacity of TN LCDs for calculators and measuring instruments etc. is 10 million, and that of STN LCDs for graphics displays is 350 thousand and that of EL displays, mostly for machine control, is 12 thousand pieces. A total of 300 people do LCD production in four groups and three shifts around the clock. LCD 1 A & B lines are on the first floor. Line 1 B was added this summer with emphasis on speed and economy of process. The LCD 2 and LCD 3 lines are on the second floor. A line is chosen for production on the basis of customer needs.

The LCD2 line we saw was 10 years old, and was updated last year. It is a TN display line that produces 10 million small displays per month for calculators. It uses 300 x 320 mm substrates. Humidity control is the primary control for static, and control techniques are used at each stage. Hot press is done for 10 minutes to set the seal and to control cell gap. Scribe and break are fully automated. There was some automatic inspection after exposure, but they indicated it was used only to tune the run and not needed later. Batch cassette-to-cassette with manual machine- to-machine transfer was used. Our hosts indicated that this was the same for the other lines as well. Equipment is cleaned every day, with fixed-time maintenance every 2 months.

Yield is top secret because it determines profit and cost of product, but it is about 80-90% overall. Yields are about the same for EL. The STN lines at Nara usually don't do repair. Resist was put on with rollers, and uniformity was the main issue for yield. It has taken about 6 months to 1 year to reach design yield for STN.

Experience has shown that many problems must be solved for each step in the process. Total Productive Maintenance (TPM), a management method, is used by all Nara people to improve yield. Fifty teams have been in place for 3 years, with one more year planned in this cycle. Engineering and management meet with the TPM teams once per month to understand their activities. Sharp works with its own precision machine company as well as other manufacturers to design new machines for display production. New machines are debugged on an existing line before a new line is built.

We received detailed specifications for Sharp Nara's new STN 640 x 480 display. Sharp started in the EL business in 1983, and growth was flat until 2 years ago when it jumped to 50%/year with expansion of the equipment controller market. They believe EL can continue at the 50% growth rate. They believe EL and plasma will coexist, and both will experience a 50% growth rate. "Full-color EL is a dream for Sharp." Use of a color filter is a possible solution. Yellow source with red or green filter could enable red and green operation. Blue emission efficiency is a problem. Sharp is working on blue phosphor at its Central Research Laboratory. They indicated that a 1024 x 1280 dots EL display was demonstrated at the Japan Electronics Show in 1991.


Published: June 1992; WTEC Hyper- Librarian