In a move that could affect multi-gigawatts of residential, commercial, and utility-scale installations in 2022, in March, the US Commerce Department announced that Auxin had presented sufficient evidence for an investigation into transshipments from China through Southeast Asia. The order is region-wide and not manufacturer-specific; for example, Hanwha Q-Cells, a South Korean manufacturer with cell and module manufacturing in Malaysia, is included in the investigation, as is Taiwan-based UREC (Malaysia, Vietnam, Thailand). Using supporting information from a Bloomberg Report and NREL, Auxin claims that cell processing is a minor part of the total module cost. Based on the Bloomberg Report, Auxin claims that wafers, silver and aluminum paste, and silane represent the majority of cell and module processing costs. Auxin, a module assembler based in San Jose, California, stated that the cell manufacturing and module assembly processes are relatively trivial, further claiming that little IP is involved in cell processing. Tariffs, if imposed, would be retroactive. 

The petition does not cover thin-film manufacturers in Southeast Asia.

Once the investigation was announced, developers reported that manufacturers began canceling outstanding orders, even those for 2022. Modules delayed because of supply chain problems turned into canceled orders, and manufacturers outside of Southeast Asia raised prices accordingly. As many developers were waiting for delayed orders for multimegawatt installations already in progress, module mismatch is an unintended consequence of the investigation. Having their orders canceled, developers must look to other countries to provide modules with the appropriate specifications for their projects. As module assemblers in Europe, India, and elsewhere get their cells from Southeast Asia and China, developers risk new and potentially retroactive tariffs on new orders.

The DOC has 150 days to investigate and another 150 days to decide and make recommendations. At this point, it is unknown whether or not the DOC will impose tariffs, whether or not they will be retroactive if imposed, and whether or not the bifacial exemption will apply.

Analysis:

In its petition, Auxin avers that cell manufacturing is a relatively minor part of the solar technology chain and that the investment required to produce cells from wafers and then assemble modules is minimal compared to that of investment required for polysilicon/ingot/wafer production. Auxin relied on a report published in February 2021 by Bloomberg for much of its assessment. Quoting from the report, Auxin asserted that “the technical hurdles required for cell manufacturing are lower than for polysilicon through-wafer manufacturing.” This assertion is repeated nine times without listing any specific technical hurdles.

To offer proof of the difference in investment, Auxin offers a range of $643 million to $2.1 billion for polysilicon capacity versus $7.7-million to a maximum of $160- million to add solar cell capacity. Yes, polysilicon production requires significant capital and a lot of time. Comparing the investment required for polysilicon production to the investment required for cell manufacturing requires an explanation as to why the comparison is relevant. The comparison is not relevant because the two processes cannot be compared. Making the comparison via a range without any detail shows a lack of critical thinking, unacceptable even in a high school essay.

Auxin states that R&D is primarily conducted in China and that all materials necessary to produce cells and modules are manufactured in China and shipped to subsidiaries in Southeast Asia. Further, Auxin, referring to the report, stated that the cost of silicon wafers, silver paste, solar glass, aluminum frames, junction boxes, EVA, and backsheets supposedly supplied by China to companies in Southeast Asia make up the majority of module costs.

Auxin is both correct and incorrect. The investment required at each point in the chain (polysilicon, ingots and wafers, cells, and modules) is different – with the investment required for polysilicon production the most significant. Making a judgment based on combining the investment required for polysilicon, ingots, and wafers compared with the cost of producing cells and modules will obviously result in the conclusion that poly/ingot/wafer production requires more money than cell manufacturing and module assembly—a conclusion that splits hairs to meaninglessness. 

As to Auxin’s point that the photovoltaic value chain is primarily located in China, it is accurate that China has chosen to build up and support its domestic PV manufacturing base, including expansions into Southeast Asia. It is equally valid that the US, Europe, Japan, and for many years India accepted low prices for China’s cells and modules instead of supporting domestic manufacturing at home. At this point, the PV manufacturing scales are so tilted in China’s favor that it will take years to resolve the imbalance.

Assuming that cell manufacturing is somehow less technical and has fewer hurdles is a leap not well explained by either Auxin or Bloomberg. Likewise, assuming that R&D is a function performed only in China is, well, a leap into science fiction. China’s manufacturers have transferred R&D knowledge to subsidiaries – but all PV knowledge is transferred from universities, national labs, and other manufacturers. The US has deep PV R&D in its universities, national labs, and startups and, other than First Solar, it just hasn’t transferred to domestic production. Australia’s universities are still the gestation point for much of today’s commercial cell technology, for example, PERC – and Australia has not chosen to support domestic manufacturing despite being the world leader in PV wafer and cell research.

During the 201 hearings in 2021, Auxin testified that it was ready to begin cell manufacturing, which indicates that it understands the technical aspects of cell processing. Suppose Auxin is prepared to manufacture cells at its San Jose, California facility. In that case, it understands that cell processing is highly complex, highly technical, and high risk – this indicates that using the assertion that cell processing has fewer technical hurdles than wafer and polysilicon processing is disingenuous. 

  • Polysilicon production is a messy, dirty, high-energy process. Currently, there is more investment in the Siemens process than the fluidized bed process. Polysilicon manufacturing requires considerable money and at least two years from investment to production. 
  • The simplest way to describe the wafer process is: the polysilicon is melted into a crucible to form the ingot, and the ingot is sliced into wafers (there is also kerfless wafering). The aforementioned is not meant to stand in as a definitive description of wafer manufacturing. 
  • The basic steps in cell processing are: precheck and pretreat, texturing, acid cleaning, diffusion, etching and edge isolation, post-etch washing, anti-reflective coating deposition, contract printing and drying, testing and sorting. 

Comment: The US has launched an investigation to protect domestic manufacturing that, other than First Solar’s CdTe, it does not have, nor will it have significant domestic manufacturing any time soon.

Should the investigation and decision take the allotted 300 days, developers and installers in the US will continue facing higher prices and fewer choices.

What if the US encouraged China to locate cell manufacturing in the US? Would it then investigate wafer manufacturers in China claiming that cell manufacturing alone was of little value? Would Auxin withdraw its petition if a Chinese wafer/cell manufacturer invested in Auxin for expansion into cell manufacturing?

Lesson: Currently, the US has strong climate change goals – though the country is one election away from those goals going up in a puff of carbon dioxide – it will not realize its goals without domestic cell manufacturing or relaxation on tariffs.