POLYSILICON MANUFACTURERS

OCI Company: Chemical know-how for high polysilicon purity

DC Chemical had a good nose for the market of the future. In 2000 the South Korean chemicals group began to develop its own technology to produce polysilicon based on the Siemens process in cooperation with the Korea Research Institute of Chemical Technology (KRICT).

That was a smart strategic decision, given the fact that the polysilicon market did not look good then. After the downturn of the semiconductor sector in 1998 and the burst of the dotcom bubble in 2000 (‘double dip’), the polysilicon industry was plagued with overcapacity, and prices tanked (see our polysilicon market analysis).

But DC Chemical was right in betting on the rapidly growing polysilicon demand of the solar industry. The company was even a bit late when it started construction of its first polysilicon plant, which had an annual production capacity of 5,000 metric tons (MT), in August 2006. By that time, the polysilicon price on the spot market had already soared to almost US$300 per kilogram, up from US$28/kg in 2003. When the factory commenced commercial production in March 2008, it was at the last moment before the spot price began to decline again (polysilicon price trend).

Renamed OCI Company in April 2009, the South Korean manufacturer became one of the most successful new entrants in the polysilicon industry because it did several things right:

• Thanks to its vast chemical experience and careful preparation, the company smoothly managed to ramp up production. It immediately achieved a high polysilicon purity of 99.9999999% (9N), which it improved to 10N in 2009.
• OCI also went quickly for large capacity additions to reach economies of scale. By the end of 2011, the production capacity of its polysilicon factory, consisting of three single plants in Gunsan, North Jeolla province, totaled 42,000 MT.
• The company took advantage of the time window arising from the polysilicon shortage to create a broad base of established solar customers with long-term contracts, covering approx. 80% of its production capacity.

In its 2010 Annual Report, OCI announced another polysilicon plant (“P4”) in Gunsan with a capacity of 20,000 MT and envisaged a bright future: “With P4, we will achieve a total manufacturing capacity of 62,000 metric tons and become the largest polysilicon supplier in the world.”

This dream, however, did not come true. Oversupply on the polysilicon market forced OCI to stop the project in 2012. Instead, the company increased its production capacity by 10,000 MT through low-cost debottlenecking in 2015, thus reaching a total of 52,000 MT.

New low-cost production base in Malaysia

In 2017 OCI purchased the new Malaysian polysilicon plant from Japanese manufacturer Tokuyama. The factory is fed by low-cost electricity from hydropower; however, Tokuyama was unable to fully ramp it up to its nameplate capacity of 20,000 MT and make it cost-competitive. OCI has revamped the plant to a capacity of 30,000 MT within three years.

This secured OCI the rank as the world’s third-largest polysilicon manufacturer through 2019. In February 2020, however, the company succumbed to the price pressure from low-cost Chinese plants and shut down its Korean factory in Gunsan – except for the P1 plant with a capacity of 6,500 MT, which is now producing electronic-grade polysilicon for the semiconductor industry.

GCL-Poly Energy: Bold capacity expansion, late adaption to high purity

The way GCL-Poly Energy Holdings dealt with the oversupply phase in 2011/2012 tells much about Chinese mentality. Against all odds, the company boosted the capacity of its China-based subsidiary Jiangsu Zhongneng Polysilicon Technology Development Company from 25,000 MT to a whopping 65,000 MT in 2011.

The consequence was a very low utilization rate in 2012. But supported by the market recovery in 2013 and the in-house polysilicon consumption of GCL’s wafer subsidiary, Jiangsu Zhongneng got out of the doldrums and became the world’s largest polysilicon manufacturer in 2013. Nevertheless, GCL-Poly’s debt load is still high.

The beginnings of the company were rather modest in comparison. Jiangsu Zhongneng commenced the construction of a small 1,500 MT production facility in Xuzhou in the eastern Chinese province of Jiangsu in July 2006 – almost at the same time as OCI did. After starting up production in September 2007, Zhongneng raised the capacity step by step to 25,000 MT before it made the big leap to the world’s top in 2011.

Despite the economies of scale, the company’s production costs suffered from the high electricity rates in eastern China. Therefore, Zhongneng commissioned a captive power plant in 2015 to generate electricity at half of the cost that it paid for power supply from the grid before.

Second polysilicon plant in Xinjiang with very low electricity costs

For many years, the purity of Zhongneng’s polysilicon product was not overly high, but good enough to supply GCL’s wafer subsidiary and make it the world’s largest producer of multicrystalline solar wafers. This model, however, has come under pressure since the global market share of monocrystalline solar wafers, which require polysilicon of higher purity, began to rise in 2015.

GCL has responded to this challenge by building a new polysilicon plant with a capacity of 48,000 MT in the Xinjiang Uyghur autonomous region in northwestern China and reducing the capacity of Zhongneng in Xuzhou to 42,000 MT. The new plant in Xinjiang, which started production in October 2018, tackles the challenge in three ways:

• GCL has established a strategic alliance with Zhonghuan Semiconductor, the world’s second-largest producer of monocrystalline solar wafers; Zhonghuan has acquired 27% of the shares in Xinjiang GCL New Energy Materials Technology Co., Ltd., the holding company of the polysilicon plant. Both partners have also founded Inner Mongolia Zhonghuan GCL, a joint venture for monocrystalline silicon ingots, which are sawn into wafers.
• The new polysilicon plant has been designed to produce high-purity material. 100% of the capacity is foreseen for monocrystalline wafers. (However, only 50% of the actual output is said to be mono grade.)
• By obtaining electricity from a nearby coal-fired power plant at a very low tariff, Xinjiang GCL is able to produce at very low costs.

After a major explosion at its distillation unit in July 2020 delayed the planned capacity increase to 60,000 MT, Xinjiang GCL completed the expansion in April 2021.

New focus on fluidized bed reactor technology for granular polysilicon

In Xuzhou, GCL-Poly’s focus has shifted to fluidized bed reactor (FBR) technology to produce granular polysilicon. Building on FBR assets acquired from bankrupt SunEdison in 2017, Jiangsu Zhongneng began to gradually convert its conventional polysilicon factory to a monosilane-based FBR plant in 2020. The company is aiming to reach an FBR capacity of 54,000 MT by the first quarter of 2022. In October 2020 GCL-Poly celebrated the groundbreaking for a second FBR plant with a capacity of 60,000 MT in Leshan, Sichuan province.

The “Seven Sisters”: Polysilicon supply oligopoly until 2005

GCL and OCI were the two most successful among a few dozens of new polysilicon manufacturers that entered the industry beginning in 2005. Until then, an oligopoly of only seven companies, known as the “Seven Sisters”, had ruled the market:

• Hemlock Semiconductor (USA)
• Wacker Chemie (Germany)
• Advanced Silicon Materials (ASiMI, today REC Silicon/USA)
• Tokuyama Corporation (Japan)
• MEMC Electronic Materials (USA and Italy)
• Mitsubishi Materials (Japan and USA) and
• Osaka Titanium Technologies (formerly Sumitomo Titanium/Japan).

The Seven Sisters ran ten polysilicon plants in the United States, Germany, Italy and Japan. The only China-based polysilicon manufacturer at that time, Emei Semiconductor, played a negligible role as it merely had a tiny annual production capacity of 100 MT (see chart).

Polysilicon plants and production capacities worldwide in 2004

Wacker, Mitsubishi, Sumitomo Titanium and Hemlock already started to produce polysilicon on an industrial scale between 1959 and 1961; the rest followed suit between the mid-1970s and mid-1980s. For decades, all of them produced exclusively electronic-grade polysilicon for the semiconductor sector; only Tokuyama and Mitsubishi are still focused on electronic grade today. Osaka Titanium shut down production at the end of 2018, MEMC in the U.S. in 2015 and in Italy in 2011.

The photovoltaic (PV) branch solely lived on scrap silicon from the semiconductor sector until the late 1990s. When the polysilicon demand from the PV industry strongly increased, Renewable Energy Corporation (REC) from Norway first formed a joint venture with ASiMI in 2002, then acquired the company in 2005 to use part of its production capacity for solar-grade instead of electronic-grade polysilicon. But it was mainly Hemlock and Wacker that invested massively in new production capacities for solar-grade polysilicon, starting in 2006. Both closed long-term supply contracts with customers who financed the expansion by making high prepayments.

Wacker and Hemlock: Top polysilicon manufacturers with different fates

These contracts became a problem when the polysilicon spot price began to undercut the average contract price in May 2011. Wacker first tried to resolve the situation by offering its customers additional shipments without further payment, thus diluting the contract price on the one hand, but securing the off-take of its production volumes on the other. With growing polysilicon oversupply, however, this strategy did not pan out; the company had to renegotiate its long-term contracts.

In the various trade disputes the USA and the European Union (EU) have had with China, Wacker has been on the lucky side. After the EU closed a minimum price agreement with Beijing on the import of solar panels from China in 2013, the Chinese Ministry of Commerce did the same with Wacker on the import of polysilicon from its two German plants. It made those shipments exempt from any punitive duties beginning in 2014.

This has enabled Wacker to maintain its customer base in China. After its new 20,000 MT polysilicon plant in the U.S. state of Tennessee opened in 2016, the company even became the world’s largest polysilicon manufacturer for four years through 2019.

The Tennessee plant initially shipped solar-grade product to customers outside China in order to avoid the high Chinese duties on polysilicon imports from the United States. After these markets have almost vanished (solar value chain), the plant mainly produces electronic-grade polysilicon for the semiconductor industry.

The curse of Chinese import duties on Hemlock Semiconductor

The fortunes of Hemlock Semiconductor were not as good. The former polysilicon market leader lost its place in the sun in 2012 when new management began to insist on sticking to the high polysilicon prices fixed in the long-term contracts while the spot price collapsed. Several customers refused to take shipments at the high contract price, which prompted the manufacturer to haul them to court.

What ultimately cost Hemlock its number one position were the prohibitive duties that China introduced on polysilicon imports from the United States in 2013. Thus, the company has practically lost the largest market for polysilicon. Consequently, Hemlock’s new 12,000 MT plant in Tennessee has never come on stream and ended as the biggest stranded investment in the polysilicon industry in 2014. The company has slid to number nine in our ranking of the world’s top ten polysilicon manufacturers.

Xinte, Daqo, Tongwei and East Hope: Chinese polysilicon industry achievers

China’s introduction of duties on polysilicon imports in 2013 was the start signal for its domestic polysilicon industry to expand. Since then, four manufacturers have moved to the foreground:

• Xinte Energy (a subsidiary of TBEA)
• Daqo New Energy
• Tongwei (parent of Sichuan Yongxiang) and
• East Hope New Energy.

Between 2013 and 2017, Xinte roughly doubled its production capacity from 17,000 MT to 36,000 MT. Although Daqo and Sichuan Yongxiang were still relatively small manufacturers in 2013 (with capacities of 6,150 MT and 4,000 MT, respectively), they strongly grew to a size of 20,000 MT by 2017. Newcomer East Hope started right away with 20,000 MT in 2017.

But that is not all. What Xinte, Daqo and Tongwei have in common is a big expansion leap to a capacity of roughly 80,000 MT in 2018/2019. All their new capacities are dedicated to high-purity polysilicon at low-cost locations, and all three have closed supply contracts with China-based Longi Green Energy Technology, the world’s largest manufacturer of monocrystalline solar wafers. Longi is planning to increase its wafer production capacity eightfold from 15 GW at the end of 2017 to 120 GW by the end of 2022.

That explains the massive capacity expansion at Xinte, Daqo and Tongwei. East Hope is not known for aspiring to produce high-purity material, but is also on an ambitious expansion course. Originally, the company was expected to double its capacity to 80,000 MT in 2020; instead, the capacity was increased to 70,000 MT in 2021.

With its even more far-reaching plans, the new world market leader Tongwei is heading for a capacity of 300,000 MT in 2023.

The Big Six: Emergence of a new polysilicon super league

Three of the four mentioned companies have now gained promotion to the 80,000 MT league, which Wacker and GCL-Poly have already been playing in. Thus, a top group of the “Big Six” will form a new polysilicon super league in the industry (see chart).

Polysilicon capacities of the Big Six in 2020

This group alone reached a production capacity of 470,000 MT in 2020 – almost as much as all polysilicon manufacturers had in total worldwide at the end of 2015. It makes obvious what tremendous pressure the capacity of the Big Six puts on other manufacturers of solar-grade polysilicon; only a few of them will survive.

But even among the Big Six, competition will be intense. With their extremely low production costs, the five Chinese top manufacturers are putting Wacker under pressure anyway. OCI has already hoisted the white flag of surrender with the shutdown of its Korean factory. The climate is getting tougher in the polysilicon industry.