Should I believe that Spiber can produce their fibers in an economical way competitive with natural fibers when GTman tells me this is impossible? As I said, I am uncertain about their costs but believe that the situation may be more nuanced than GTman suggests, partly because of the Randy Lewis paper arguing that sub-$100/kilo costs are achievable and partly because the venture capitalists funding Spiber are surely sensitive to production costs but still invested hundreds of millions into Spiber.
Well it's pretty obvious the paper above wasn't read. Either that or some are willfully disregarding the glaring technological gaps which the paper admits to assuming.
Also, the author, Alan Edlund (not Randy Lewis), is a grad student who wrote this paper for his masters thesis. And get this, his masters is in mechanical engineering, not microbiology, and it shows. Randy Lewis wasn't even his PI. Details matter.
That being said, Mr. Edlund still recognizes that the only way to ever come close to a sub 100/kilo fiber is to assume technological advances that have never been achieved in protein expression using fermentation. He is not shy of detailing these assumptions either. Here is one of many quotes supporting my point:
Assuming that both high protein expression and cell concentrations are possible, then reaching a protein content of 25 g l-1 would reduce the cost of production by 96% below the baseline, Case 1, to $32 kg-1 . Since the inception of the work, RSSPs levels have constantly been increasing. In comparison to other recombinant proteins, expression levels for spider silk have remained relatively low. This challenge is largely thought to be due to the size of the protein which is expected to limit the final yield compared to what has been achieved with other proteins. Economic modeling shows this to be the most valuable area where advancements can be made.
The current state of the art that Spiber, Bolt Threads, and Amsilk were using would yield about 1g/L of protein (and much lower for their larger sized proteins). In this paper, in order to achieve sub $100/kilo, the yields would need to be 25 times more efficient!. Are you serious? not a 5%, or 10%, or even 20% better, but 2500% better efficiency. The science for protein fermentation is very developed. they've been optimizing it since the 1980's. Tweaks of maybe a few percent in yield are possible, but 2500% !!! ? are you kidding me? And that would be for a fiber that has a lower tensile strength than regular silk.
Also, the author notes something that I've repeated many times in the past. There is an inverse relationship with the protein yields and the protein size. Simply put, the higher the yield, the smaller the protein. This has been proven many times over. As recently as the Nature paper posted which celebrated the 8x yield improvement without recognizing the dismal protein size and mechanical properties. Reread this quote:
This challenge is largely thought to be due to the size of the protein which is expected to limit the final yield compared to what has been achieved with other proteins
I'm starting to see a pattern where articles are posted without ever reading said article.
The Edlund paper also has a myriad of other pitfalls for the fermentation companies but the list is too long to really get into. But if anyone wants to read why the fermentation method isn't even "sustainable" like Spiber claims, read the last section. Even at it's most wishfully optimized situation where it can magically become 2500% more effecient then the state of the art, it produces almost10 times more CO2 then polyester (plastic). At it's current state, it produces 100 times more CO2 than polyester. Compare that to sericulture which actually sequesters CO2. Just because Spibers end product doesn't contain petroleum, doesn't mean it doesn't use a ton of it in the process.
It's strange that the last two papers posted on here and was told to read, completely validated how awful the fermentation method is, and confirms that they are nowhere near competitors of Kraig Labs.
I know people don't have the time to read a 60 page paper, and maybe my post is even too long and dense, but below are some passages taken straight from the text of the paper. Feel free to read as much as you want, or don't:
Spider silk represents a new technology where little is understood about the large-scale flow rates and exact sizing specifications. The lack of process definition has an effect on capital cost estimation causing designs to be generic, and hard to optimize. For modeling purposes, pioneer plants are considered to be first of a kind (FOAK) facilities. The technologies are hard to troubleshoot this early on, so the first plant is expected to have high capital costs from lack of proper optimization that could be realized as greater process 24 definition occurs. As additional plants are built, a technology becomes better understood and optimized, resulting in lower capital costs due to reduced engineering, and lower contingency fees [65]. Considering this consequence, the benefit of being a more mature plant was included as a part of the optimized scenarios. The mature plant scenarios were considered to be the 17th plant constructed of their type under the assumption of a 0.06 learning rate as defined by NETL (2013).
4.2 Sensitivity Results A sensitivity analysis was leveraged to identify statistically significant inputs that can be improved to support the identification of a commercial viable pathway. Based on a 95% 30 confidence interval, inputs with a t-ratio less than 2.12 are considered to be insignificant. Only five input parameters were identified as substantial. Protein yield has the single greatest impact on sale price, Figure 5. Case 1 represents a low, but demonstrated protein yields of 1.0 g l-1 (Unpublished Data). Laboratory protein expression has reached higher levels but has not been consistently demonstrated. Increasing the yield by 3X corresponds to a 67% reduction in sale price, bringing the cost to $254 kg-1 . Protein yield’s dramatic influence on price is largely due to the batch nature of the first three sub-processes (fermentation, harvesting, and protein purification), which also make up 99% of the operational expenses. Increasing protein yield has a direct impact on increasing the amount of available product with minimal change in process volumes resulting in effectively no additional cost for downstream processing. Fermentation time is an important model input with a reduction representing an improvement in economic viability. IPTG pricing, capital costs, and NTA beads are also hefty financial burdens. Other individual material pricing estimates and model inputs had lower t-ratios and have relatively low impact on sale price. Results from the sensitivity analysis were used to identify the alternative scenarios explored in this modeling work.
4.3 Optimization The commercialization of synthetic spider silk through E. coli fermentation will require the integration of alternative technologies to those of the pioneer plant as well as improved protein yield. Alternative technologies are explored that represent advancements in strategic areas. Modeling improvements come from three sources: 1) process optimization, 2) experienced design for subsequent plants, and 3) increased protein expression (10 g l-1 ). 31 Individually, each of these methods are considered for their effect on pioneer plant economics. All of these advancements combined are considered for their mutual economic benefit in an optimistic scenario, Case 2.
IPTG induction is the process of triggering the E. coli to produce RSSPs and is a costly component of fermentation process. Since the price of IPTG is high, the implementation of heat induction (Case 1c) results in a significant benefit. A 28% sale price reduction is observed. The application of this method would have significant benefits towards achieving economic viability. While heat induction has not been demonstrated with the laboratory strain, it has been demonstrated on E. coli [43]. Some possible concerns with the implementation of heat induction are reduced protein expression or protein degradation but this has not been explored. The combined impact of integrating all the fermentation improvements (1a, 1b, and 1c) result in a 33% decrease in the sale price over Case 1.
Post fermentation, the E. coli are harvested via a centrifuge, and are then mixed with a salt solution. By skipping this first centrifugation step, and adding the salt components directly to the media (Case 1d), a 1.7% cost reduction is observed. Additional optimization during harvesting could come from reducing the pressure of the HPH process. Based on a literature review, successful lysis occurs between 1000- 1500 bar. Since 1500 bar was used in the model, using only 1000 bar would cut $1.75 kg-1 in production cost resulting in a 0.23% reduction in total cost (not shown in Figure 6). The integration of these two processes have minimal impact on the economics of the system.
Initially, flocculation in the laboratory could maintain high purity only when using volumes of a few milliliters, but this has since been rectified. This process has now been verified in volumes of up to several liters (Unpublished Data). The implementation of this technique will undoubtedly save costs for large-scale production, but issues with purity at this scale may need to be addressed.
The techniques described have the ability to be implemented together without interference, and the collective effect is significant. Their combined effect (Case 1g) results in a 55% decrease in the levelized cost bringing the minimum product sale price to $344 kg-1 . Most of these methods are feasible for large application, with some additional testing necessary. These identified methods of process optimization have substantial impact on price, but further optimization of protein yield will be required for the delivery of an economically viable product.
Combining all of the advantages discussed above, with the addition of increased protein expression, represents the optimistic scenario, Case 2. This scenario represents an optimized production facility, using the more economical processing options (Case 1g), and is the 17th production facility of its kind with a protein yield of 10 g l-1 . All of these inputs combined (Case 2) result in a production capacity of 3,550 tons per year and a sale price of $23 kg-1 . This is a 97% reduction in sale price from Case 1. As expected, the product price is dramatically impacted by the protein production level. Process optimization also has an important impact, but not at the same level as protein expression.
Applications using small amounts silk are better suited for economic viability. Due to its hypoallergenic properties and high strength, a likely candidate is the medical industry where only a few grams of spider silk might be used to make a replacement ligament, tendon, or other implant. The diverse properties of synthetic spider silk make it a promising material for a variety of applications.
Increasing protein expression is not only the most beneficial method of optimization, but is expected to advance through research and development. E. coli are commercially used to produce more than 150 recombinant proteins [68]. The techniques used for production and purification have been highly developed. The limit on protein production has been explored in E. coli. Increased yield can come through increased dry cell weight (DCW- grams of E. coli per liter of media) or genetic optimization for greater expression. Laboratory fermentations typically result in a DCW of 45 g l-1 (Unpublished Data).
Assuming that both high protein expression and cell concentrations are possible, then reaching a protein content of 25 g l-1 would reduce the cost of production by 96% below the baseline, Case 1, to $32 kg-1 . Since the inception of the work, RSSPs levels have constantly been increasing. In comparison to other recombinant proteins, expression levels for spider silk have remained relatively low. This challenge is largely thought to be due to the size of the protein which is expected to limit 39 the final yield compared to what has been achieved with other proteins. Economic modeling shows this to be the most valuable area where advancements can be made.
Another method of decreasing the emissions per unit mass is to increase the production through improved protein expression. The net effect of going from Case 1, at 1.0 g l-1 of protein expression, to Case 2 at 10 g l-1 of protein expression results in a 90% reduction of emissions to 55 kg CO2 eq. kg-1 . In spite of the optimization and increased yield, the reduced emissions of 55 kg CO2 eq. kg-1 are still higher than that of many other high strength materials. Linear low-density polyethylene and poly propylene have emissions of 1.62 kg CO2 eq. kg-1 and 1.59 kg CO2 eq. kg-1 [75]. Average steel has an emissions of 1.36 kg CO2 eq. kg-1 steel [39]. When compared with carbon fiber the emissions for silk are not quite as 0 50 100 150 200 250 Fermentation Harvesting Purification Drying Fiber Spinning kg CO2 eq. kg-1 product Materials Natural Gas Electricity 45 substantial. Emissions for the PAN carbon fiber are 31 kg CO2 eq. kg-1 [76]. There are additional benefits that an optimized plant would experience, such as reduced energy demand, which cannot be fully quantified at this stage of analysis.
Optimization through processing techniques and increased protein expression will be required before any of the three protein production methods can become feasible for large scale production. Additionally, as global warming awareness increases an eventual national carbon tax is conceivable. This would further affect the 47 economics as emissions are high. Further work to advance spider silk products should focus on the key model inputs identified though the sensitivity analysis, namely increased protein production. Increasing protein output sets off the high cost and emissions weight that are tied with low yields. Higher yields will not only increase the economic viability, but will hopefully increase social acceptance. Under this assumption it is likely that spider silk based products will start to enter the market in at least some industries. Additional market penetration is probably upon achieving higher levels of process optimization and protein expression
