Systems Equipment Division At Ferrofluidics Part A

Introduction In 1986, the Systems Equipment Division (SED) at Ferrofluidics1, was in big trouble.2 Sales, which had peaked in 1985 at $9 million, were less than $3 million. They did not receive a new order during the last half of the 1986. By the end of the year it was apparent that the main product, pullers, had severe quality problems. There were no backorders, and no prospective customers. The future looked bleak. General Corporate Background Ferrofluidics was founded in October 1968 by Dr. Ronald Moskowitz and Dr. James Rosensweig to pioneer the technical and market development of magnetic fluid technology. These men invented and patented this specialized materials technology while working as re searchers on NASA sponsored projects in the early 1960s. NASA investigated magnetic fluid technology because of its potential as a sealed bearing which isolates hazardous environments from ambient normal conditions along a rotating shaft. These fluids, called ferrofluids, can be magnetized by suspending very fine magnetic particles in a liquid. The results are an extremely stable colloidal magnetic fluid. When a magnetic field is applied, the ferrofluid acquires a magnetic moment and can be precisely positioned and controlled. Ferrofluids have superior properties as lubricants, sealing agents, bearings, and dampening agents. These materials have applications in many areas including the manufacture of contact lenses, hard disks, stereo speakers and semiconductors. For example, ferrofluids are used as a frictionless sealed bearing which allows a hard disk to spin at incredible rates. Another advantage of ferrofluids is that the seal prevents foreign particles from damaging the disks. The company developed numerous products based on ferrofluid technology. Ferrofluids were sold as raw materials to some manufacturers, however, most sales and product offerings were small component parts, which were based on ferrofluid technology. In the later 1970s Ferrofluidics developed an innovative ferrofluidic sealed bearing for equipment used in the production of silicon. These huge machines, called silicon crystal pulling furnaces, or pullers, were much more efficient with the addition of the ferrofluid seal (See Table 5 for Physical Dimensions of puller equipment). The product was clearly superior to the sealed bearing technology, which had been employed,
1

The Systems Equipment Division was originally called the Crystal Systems Division.

This case study would not have been possible without the permission granted by Dr. Kedar Gupta, Vice President and General Manager, Ferrofluidics Systems Equipment Division, and without the help of Jonathan Talbott, Operations Manager, Ferrofluidics Systems Equipment Division, Michael Osbourne, Manufacturing Manager, Kay Dorr, Executive Secretary, Rick Meier, Electrical Engineering Manager, Carl Chartier, Director of Process Technology, Dr. Jurek Koziol, R&D Manager, and many others.

and became very popular. Sales of these retrofit vacuum sealers reached $600,000 in the first year, and were the first real commercial success for Ferrofluidics. These seals became the flagship technology and were instrumental in making Ferrofluidics a profitable organization. (See figure 1 for an example of a puller.) In the early 1980s, Ferrofluidics realized that ferrofluid technology had many applications. Corporate management also reasoned they might be able to increase profits by forward integrating into industries where ferrofluid technology was important. With this strategy in mind, the company searched for possible acquisition candidates. When it became known that Varian wanted to divest itself of a silicon crystal puller division, management thought they had a perfect match. They were familiar with silicon crystal pullers, and they thought the semiconductor silicon industry had great potential. During 1981, Ferrofluidics purchased the puller division from Varian for approximately $1.5 million. Pullers contained millions of parts, many major subsystems and sold for upwards of $500,000. Overnight, Ferrofluidics became a capital equipment supplier. SED Division History Management at Ferrofluidics was attracted to SED for a number of reasons. They projected that the silicon puller market would grow from average sales of 70 pullers per year to sales of 100 to 150 pullers per year. They wanted to be part of this burgeoning market. In addition, they perceived the production of silicon pullers as a relatively low technology industry, where the application of their technological expertise would give them a huge advantage. Ferrofluidics had built their reputation by utilizing their strong research and development capabilities to deliver innovative products. The management at Ferrofluidics was convinced they could apply their materials based technological capabilities to the fabrication and assembly of silicon pullers. During 1982 Ferrofluidics marketed the puller design, which they had inherited from Varian, and met limited success. During the year, Ferrofluidics began development of a new puller, the Six-Four-Two puller. This puller derived its name from the fact that it could produce silicon ingots with a six inch diameter, as well ingots with diameters of four and two inches. After a six-month design phase, the new puller was introduces to the market. The general manager of SED, Walter Hegaland, believed that the Six-FourTwo puller would be the technology leader in the industry. It was the most automated and most technologically advanced puller available. As this quote from the 1981 annual report shows, Ferrofluidics anticipated great things from the new product: The successful integration of ferrofluid technology to subsystems and systems is exemplified in the development of our innovative Six-Four-Two computer controlled, silicon crystal growing furnace. The system was designed to meet the total processing requirements for converting polycrystalline silicon into semiconductor grade, single crystal ingots at high productivity and yield. This new system incorporates a number of evolutionary advances including a sophisticated process control computer, complete vacuum integrity with Ferrofluidic rotary sealing, and a proprietary new simplified materials handling system, which in

aggregate results in a revolutionary machine that meets the needs of the industry in the 1980s.3 The market for pullers is tied to the demand for silicon. The demand for silicon, in turn, is dependent on the demand for microelectronic devices. During the early 1980s, the silicon industry was experiencing rapid growth due to the demand for microelectronic devices. Silicon producers such as SHE, Monsanto-MEMC, and Wacker purchased pullers to produce the silicon ingots, which were converted into wafers. These wafers were the substrate on which most microelectronic devices were produced. To better market the newly designed puller, Ferrofluidics attempted to increase their international presence in the market. Ferrofluidics purchased Sloan Technology Gmbh of Germany to market the pullers in Europe and established Nippon Ferrofluidics to sell the puller in Japan. As they improved their international presence, they also continued to improve the Six-Four-Two puller. Based on preliminary marketing survey information, the customers overwhelmingly preferred the new automated system to the more manual puller inherited from Varian. During the years 1982 and 1983, Ferrofluidics spent $2.4 million developing the new puller. This included a total writedown ($1.3 million) of the Varian technology and assets, which were considered obsolete. Despite attempting to market both the new and the old machines, 1982 was a dismal year. They did not sell a puller of either type. The $366,000 in sales was limited to replacement parts. Although 1982 was a horrible year, Ferrofluidics was still enthusiastic about the future of SED. In the 1982 letter to stockholders, President Moskowitz stated that Ferrofluidics would be delivering the third generation of the Six-Four-Two puller in early 1983. The letter also stated SED was expanding their customer base by taking the first steps to develop pullers for manufacturers of other crystal based materials, such as gallium arsenide. The first completely automated silicon production system, the fourth generation of the Six-Four-Two puller, was sold to Osaka Titanium Corporation (OTC) in 1983. This culminated over two years of development and innovation. In addition, Ferrofluidics entered into a program to jointly develop a gallium arsenide puller with Harris Semiconductor. Revenue for the year increased to $929,000 but the operating loss of $884,000 was larger than anticipated. In 1983, Richard Regan4 was interviewed in the stock market newsletter Over The Counter Storyline. He discussed Ferrofluidics chances in the crystal puller industry. Mr. Regan stated that the puller industry was: ..tight little industry where the market is already well served by current crystal puller suppliers. They (SED) could make it bigor flop on their fannies. .What is likely to happen is that a crystal producer will buy a couple of pullers, take his

3 4

Ferrodfluidics Annual report Fiscal 1981 Winthrop, M.M., (1983), Is Ferrofluidics New Furnace Hot?, OTC Review, January 1983.

time and evaluate them, and then take some more time before deciding to order more.5

By the end of 1984 demand for semiconductor grade silicon had outstripped supply. Silicon producers needed to add production capacity. Because of the increased demand for pullers, the traditional puller manufacturers had a large number of backorders with long lead times. These conditions provided the impetus for major silicon manufacturers to invest in the innovative but unproven Six-Four-Two puller from SED. Early in the year, Ferrofluidics received their second order for the new, fourth generation Six-Four-Two silicon crystal puller. Shortly thereafter they received the first multiple order for the puller. Texas Instruments had performed extensive testing of the SED crystal puller and placed an order for two pullers. Later they received the first multiple order from a major Japanese producer of silicon. By the end of the year SED has generated sales of about $3,548,000 and had a backlog of $1.8 million. During 1985, the market was still strong and puller sales at SED tripled to over $9 million. The division contributed over 29% of total corporate revenues for the year. SED was finally becoming a profitable division. In addition, a fifth generation SixFour-Two puller was in design and development. The new puller featured enhanced computer control, and the ability to continuously feed raw materials into the puller. At this point it seemed that the future for SED looked bright. In 1985 the purchasers of the fourth generation Six-Four-Two puller began to report quality problems. The consistent complaint was that the machines were not reliable and were off line too much of the time. Interest in the Six-Four-Two puller began to wane. During 1985, corporate management realized that SED had troubles. Management had expected SED to become a world leader and was disappointed that the division was not performing up to expectations. They were particularly dismayed that the fourth generation Six-Four-Two puller was experiencing reliability and quality problems. Upper management at Ferrofluidics decided the solution was to replace the general manager at SED. A new general manager would bring fresh ideas and creative ways to solve the problems experienced at SED. The man selected for the job was Dr. Kedar Gupta, a manager from the semiconductor materials division at Cincinnati Milacron. The president of Ferrofluidics, Dr. Ronald Moskowitz, had hired Gupta with the charge develop the type of equipment company that you (Dr. Kedar Gupta) would buy from. At the time, management at Ferrofluidics thought a minor overhaul would get the division back on track. At the beginning of 1986 SED was busy filling the backorders from 1985, but new orders did not arrive. The word had gotten out. The market new knew that the SixFour-Two puller had severe quality problems. Customers were not purchasing the highly automated pullers from SED, but were investing in the more proven pullers from competitors. It quickly became obvious that 1986 was going to be a disastrous year.
5

Ibid.

When Gupta accepted the position he knew that the division was in need of leadership but he was unaware that the division was in total decline. SED slumped badly by the end of 1986. Revenues declined to about $3 million, and most of these sales were on backlog at the end of 1985. They did not make a new sale during the last half of the year. The operating loss was $1,396.000. In addition, they were continuing to fund the development of a gallium arsenide puller but were having considerable difficulty. It appeared they would not have a viable gallium arsenide puller ready for quite a while. (See table 2 for a review of the financial information of the division.) At the time Gupta was hired, management at Ferrofluidics had been concerned about SED, but had been hopeful about its future. By the end of 1986, management had gone from the point of being concerned to the point of being disgusted. They felt that the diversification strategy was an unmitigated failure and were considering selling the division. The mode of upper level management at Ferrofluidics was summed up in this quote from Dr. Moskowitz: If the systems equipment division cannot recover and it fails, it will not make much of an impact on the company. Ferrofluidics loses approximately $500,000 in annual sales. If the systems equipment division grows much faster than anticipated, and becomes a $3 million division, it will still not have much of an impact on the corporation. In 1986 Ferrofluidics was taking a hard look at SED. The division was in trouble and the future did not look particularly bright. At best, Ferrofluidics was apathetic about the future of SED. It was not a surprise to Gupta that Ferrofluidics mandated that SED reduce the staff. The staff was to be cut to a total of six individuals, including Gupta.
Year No. of Customers Silicon Production N. America Silicon Production Pacific Rim Silicon Production Europe Other % Market Top 5 Producers (rough est.) 50% 60%

1981 1986

53 44

44% 37%

33% 43%

19% 16%

4% 4%

Table 1. Customer Profile by Location Over the course of his first year at SED, Gupta realized that the division was in severe trouble. When Gupta analyzed the market for silicon pullers, he recognized trends which would have adverse affects on future sales. The number of silicon producers was decreasing and the market was becoming more concentrated. In addition, silicon production was shifting from North America to the Pacific Rim. (See Table 1 for a summary of market trends.) Gupta figured he had six major competitors. Most of his competitors had strong financial backing and all had a larger percent of the world market. In addition, many of the competitors had a geographic advantage. Most silicon producers wanted to purchase

from local puller suppliers. Gupta suspected that the production of silicon in North America would steadily decline and silicon production in the Pacific Rim would continue to increase. (See Table 3 and Table 4 for a review of the competition.) Gupta knew that Ferrofluidics was not going to be overly supportive and realized that the technological core capabilities necessary for the production of ferrofluid products were not suited for the fabrication and assembly of pullers. Gupta was not sure that staying at SED was in his best interests. He placed his situation in perspective with a single thought, a thought instilled in him at Cincinnati Milarcon: Do I have an acorn that a good management team, with care and watering, can develop into an oak tree?
Ferrofluidics Including SED (000) 1980 Revenue Operating Income Net Income Earnings per Share Working Capital Total Assets Long Term Debt Net Worth SED Revenue Operating Profit Identifiable Assets Capital Expenditures 366 (2,414) 877 75 929 (884) 1,311 2,400 3,548 (618) 2,156 46 9,135 883 3,737 562 3,106 (1,396) 4,822 232 3,153 199 196 0.03 655 2,516 582 1,2242 1981 5,056 437 279 0.04 685 5,099 470 2,052 1982 7,020 (2,418) (2,176) (0.24) 1,670 7,207 2,038 3,178 1983 10,786 366 127 0.01 1,624 10,103 2,089 3,326 1984 21,131 1,739 1,176 0.12 10,492 26,210 8,971 9,831 1985 31,289 2,077 1,207 0.12 10,517 30,740 11,368 12,575 1986 28,184 (261) (2,512) (0.13) 27,449 54,216 20,564 25,025

Table 2. Financial Information (1980-1986)

Company Type of products Specific industry Captive supplier for semiconductor silicon Vacuum systems for high technology users Financial size Type of puller manufacturer Captive

SHE (Japan) Lybold Hearus (Germany)

Substrate materials High technology fabrication & assembly

Over $1 billion

Over $1 billion

Merchant

Hamco (USA) Fugi Koshi (Japan) Siltec (USA) Ferrofluidics SED (USA) Komatsu (Japan)

High technology fabrication & assembly High technology fabrication & assembly High technology fabrication /assembly and materials Specialty materials

Pullers Semiconductor processing equipment Semiconductor substrates & silicon processing equipment Ferrofluids Captive supplier for semiconductor silicon

Over $15 million

Merchant

Over $100 million

Merchant

Over $20 million

Merchant/captive

Over $11 million

Merchant

Substrate materials

Over $100 million

Captive

Table 3. Competitor Profile

COMPANY SEH Lybold Hearus Hamco Fugi Koshi Siltec Kormatsu Ferrofluidics SED All others

1981 30% 13% 30% 10% 5% 7% 1% 4%

1986 30% 15% 28% 11% 4% 7% 1% 4%

Table 4. Market Share by Firm

Height Floor Area Total Weight Furnace Power supply Console

Closed: 6,121.4 mm (241 inches) Full Open : 6,883.4 mm (271 inches) 4,108 mm (162) x 1,610 mm (62) Approximately 8,000 kgs. (17,640 lbs.) Approximately 6,000 kgs. (13,230 lbs.) Approximately 1,500 kgs. (3,307.5 lbs.) Approximately 500 kgs (1,102.5 lbs.)

Table 5. Physical properties of puller