Levis Green Technology

NEW APPROACHES IN
COTTON CROSSLINKING
GREEN TEAM FINAL REPORT TO LEVI STRAUSS & CO.
JOE CHARBONNET
JENNIFER LAWRENCE
LEAH RUBIN
SARA TEPFER
GREENER SOLUTIONS 2013

UNIVERSITY OF CALIFORNIA, BERKELEY
i
TABLE OF CONTENTS
Green Team Final Report
GOAL 1
CONTEXT + APPROACH 1
The Current Problem 1

Inspiration from Nature 2
Translation to Textile Crosslinking 3
STRATEGIES 4
Process Considerations 4
Linkages to Cellulose: Non-Covalent and Structural Bonds 6
Linkages to Cellulose: Covalent Bonds 9
Crosslinking Strategies: Covalent Bonds 10
FRAMEWORKS + EVALUATIONS 13
Boundary Conditions 13
Technical Framework 13
Health Framework 16
Health Evaluations 19
RECOMMENDATIONS 23
BIBLIOGRAPHY 24
APPENDICES 28
Appendix A: 12 Examples of Crosslinking Within Nature* 28

Appendix B: Interaction of Bonding Themes in One Organism 61
Appendix C: Why Not Metals? 62
Appendix D: Technical Evaluation Framework – Detailed Description 63
Appendix E: Detailed Technical Evaluation Results 65
*These descriptions were prepared by the Biomimicry 3.8 Institute

ii
FIGURES AND TABLES

Figure 1a and b. Existing crosslinking technologies 1

Figure 2a-c. One-step crosslinking compared to potential two-step processes 3
Figure 3. Textile manufacturing process 5
Figure 4. Polyethylene terephthalate 6
Figure 5. Poly(vinyl alcohol) 7
Figure 6. Thiolated poly(vinyl alcohol) 8
Figure 7a-d. Potential compounds for non-covalent crosslinking applications 8
Figure 8. Common naturally occurring poly(carboxylic acids) 9
Figure 9. Cyclic anhydride intermediate 9
Figure 10. In situ thiolation of poly(carboxylic acid) 10
Figure 11. Disulfide linkages 11
Figure 12. Imine bond formation 12
Figure 13. Reductive amination 12
Figure 14. Technical evaluation of each strategy 15
Figure 15. Technical evaluation of combined strategies 15
Figure 16. Health evaluation process 18
Figure 17. Comparison of health evaluation framework to GreenScreen 19
Figure 18. Health evaluation weighting scheme 19
Figure 19. Baseline health data 20
Figure 20. Disulfide bond health data 21
Figure 21. Imine/amine bond health data 21
Figure 22. Poly(carboxylic acid) health data 22
Figure 23. Polymer thiolation health data 23
Figure 24. Polymer coating health data 23
Table 1. Examples of crosslinking within each of the four bonding themes 2
Table 2. Eight technical evaluation factors 14
Table 3. GreenScreen Benchmark 1 criteria 17

FINAL REPORT
Joe Charbonnet, Jennifer Lawrence, Leah Rubin, Sara Tepfer
GOAL
The goal of our collaboration with Levi Strauss & Company [LS&Co.] is to identify biologically
inspired opportunities to modify crosslinking technologies currently used in their wrinkle-resistant
and water repellent finishes. The new crosslinking technologies must comply with current
manufacturing processes, produce cost-competitive products that meet consumer expectations, and
reduce occupational and environmental health hazards.
CONTEXT + APPROACH
The Current Problem

The garment and textile industries have made great strides in meeting consumer demands for high-
performance and low-maintenance 100% cotton apparel. Two of these performance features,
wrinkle-resistance and water-repellency, are conferred using crosslinking chemicals. Unfortunately,
there are significant health hazards surrounding the application of these features. Formaldehyde,
released from crosslinking chemicals that are used to impart wrinkle resistance, is a known
carcinogen and can cause nasopharyngeal irritation and sensitization (Figure 1a).1 While LS&Co.
regulates maximum formaldehyde concentrations in its final distributed products, the use of
formaldehyde-based resins does pose significant occupational hazard prior to product distribution
and consumer use. Diisocyanates, used to bind durable water repellent (DWR) coatings, pose
hazards of similar concern. These chemicals can cause respiratory tract irritation, occupational
asthma, and possibly cancer (Figure 1b).2 There
O
O
is an identified
D M Dneed for
HEU [FO R M ALS&Co.,
L D E H Y D E ] and the textile
HO N N OH
industry as a whole, to develop an alternative approach to crosslinking
OH HO
HO
O N
OH
N O
chemistry.
NASOPHARYNGEAL IRRITATION
NASOPHARYNGEAL SENSITIZATION
DMDHEU HO OH KNOWN CARCINOGEN
O
O DMDHEU [FORMALDEHYDE]
HO N N OH
OH HO O N N O NASOPHARYNGEAL IRRITATION
HO OH
NASOPHARYNGEAL SENSITIZATION
DMDHEU HO OH KNOWN CARCINOGEN
DIISOCYANATES
N DWR R DWR RESPIRATORY TRACT IRRITATION

C OCCUPATIONAL ASTHMA
O
POSSIBLE CARCINOGEN
DIISOCYANATES
Figure 1a. DMDHEU, the most N

C
common formaldehyde-based
DWR R resin
used RESPIRATORY
DWR for cottonTRACT
crosslinking,
IRRITATION contributes
OCCUPATIONAL ASTHMA
to free formaldehyde in fabric and Oposes significant occupational health hazard.
POSSIBLE CARCINOGEN
Figure 1b. Diisocyanates, which bind water repellant coatings to cotton, pose potential respiratory hazard.
2
Inspiration from Nature

Nature has been evolving for over three billion years. The animals, plants, and microbes found in
nature have used this time to engineer solutions to problems within their environment in a manner
that is appropriate and sustainable to life. Biomimicry is a design discipline that draws upon nature’s
abilities and seeks sustainable solutions to humans’ problems through the emulation of its patterns
and strategies.3 This discipline has procured innovative solutions to many challenges across a wide
variety of industries, and we have used it to identify sustainable crosslinking technologies for the
textile industry.
In partnership with the Biomimicry 3.8 Institute, we identified 12 examples of crosslinking within
nature (Appendix A). Upon analysis, we noticed that nature employs a wide variety of bond types to
achieve crosslinking. We distilled these bond types into four themes: covalent, non-covalent, metal-
containing, and structural (Table 1).
Table 1. Examples of crosslinking within each of the four themes

Bond Type Strategy in Nature Place of Use
Covalent Bonds Disulfide bonds - Chinese soft shelled turtle
- Human joint
- Malaysian tree frog
- Pearl Oyster
Imine bonds - Slug
Peroxide-mediated bonds - Flax stem
Non-Covalent Bonds Hydrogen bonds - Chinese soft shelled turtle
- Chiton tooth
- Deep sea sponge
- Flax stem
- Woody plants
Metal-Containing Bonds Coordination complexes - Blue sea mussel
Direct metal crosslinking - Slug
Metal ion interfaces - Chiton tooth
- Human joint
Structural Bonds Anti-parallel sheets - Chinese soft shelled turtle
Checkerboard reinforcement bonds - Deep sea sponge
Flexible coils - Blue sea mussel
- Vineyard snail
Large, branched structures - Flax stem
Micelles - Chinese soft shelled turtle

Greener Solutions 2013
3
Additionally, we noticed that none of these aforementioned bonds work in isolation. Instead,
several different bonding themes come together in a single organism to impart a stable and robust
crosslink (Appendix B). We decided to emulate these two ideas in our biologically inspired
crosslinking technologies.
Translation to Textile Crosslinking

Biological inspiration led us to consider textile crosslinking in a new way. Traditional textile
crosslinking relies on a covalent bond, but we consider other bonding types. While all four bonding
strategies demonstrate potential, metal-containing bonds are outside the scope of this report
(Appendix C). Our suggestions focus on covalent bonds, non-covalent bonds, and structural bonds.
Traditional textile crosslinking also relies on one bonding step alone to impart the desired crosslink:
in one step, the cellulose and crosslinking chemicals are joined (Figure 2a). We look at crosslinking
as a two-step process. First, a bond is employed that attaches the crosslinking chemical to cellulose
(Figure 2b). Second, a bond is employed that imparts the desired crosslinking property. In the case
of wrinkle resistance, the crosslinking chemical will employ a second bond and attach to itself,
thereby connecting the strands of cellulose (Figure 2c). In the case of water repellency, the
crosslinking chemical will employ a second bond and attach to a durable water-repellent, thereby
connecting the strand of cellulose to the durable water-repellent (Figure 2d).
Figure 2a. Traditional Crosslinking imparts desired attributes in one step (Top, wrinkle resistance. Bottom, water
repellency).
+ +
!"#
+ + !"#
Figure 2b,c. Biologically Inspired Crosslinking imparts desired attributes in two steps (Top, wrinkle resistance. Bottom,
water repellency).

4
STRATEGIES
Process Considerations
Addressing cellulose bonds and crosslinking bonds separately provides us freedom to perform the
crosslinking step any time after the initial bond with cellulose has been formed. We propose
cellulose bond treatments that could be applied in the yarn, fabric, or garment form (Figure 3,
following page). The crosslinking bond treatment could then occur during the fabric finishing stage
or in garment form.
The large span of steps between the treatment’s initial and secondary applications forces us to
consider potential interference from intermediate processes on the crosslinking chemicals. The most
unavoidable chemical conditions include the presence of water and air, which may trigger premature
hydrolysis or oxidation of chemicals bound to the cellulose. Depending on the point of application,
these chemicals must also withstand high temperatures, strong bases, and strongly oxidizing bleaches.
The anticipated performance of these chemicals during the manufacture process is assessed in the
Technical Evaluation.

5
TEXTILE PROCESSING FOR DUMMIES
STEP SUBPROCESS OVERVIEW CHEMICAL INPUTS

FORMATION
RAW COTTON
YARN
WAX REMOVAL_ hydrophobic coating is removed

FIBER PREPARATION to allow dyes to penetrate into the cellulose
SPINNING

Lengthwise yarns are held in tension and weft
FORMATION
WARPING yarns are later woven through in the weaving step.

FABRIC
polyvinyl alcohol
Formulations added to warp yarns to reduce
SIZING friction and fraying; adds at least 20% by weight.
polyacrylic acid KNITTING
carboxymethyl cellulose
For Levi’s_ indigo dye applied

WEAVING prior to weaving
Removal of starches and other water- amylase
DESIZING soluble compounds in sizing formulations other enzymes
Remove fats, waxes, tannins, pectins, boiling conc. NaOH
may include: desizing, scouring, bleaching [for SCOURING proteins, and any dirt or husk residues. 1 hr
PREPARATION pale fabrics], and/or mercerizing
PROCESSING
For Levi’s_ indigo dye applied BLEACHING Remove naturally occurring pigments 100+ºC H 2O 2
DYEING/PRINTING
WET
prior to weaving
25-30% caustic soda

DMDHEU; Diisocyanates; MERCERIZING them longitudinally, making them more
May include DWR or permanent press application; receptive to dyes
FINISHING traditionally spray or dip process, though more
-
recently CVD. LS&Co. uses a dip process.
based DWRs

CUTTING
FINISHING
For Levi’s_ DWR properties

SEWING imparted here [garment form]
CURING
Lifecycle performance_ 10-30

FINISHED GOODS washes, depending on line
Figure 3. Textile manufacturing process, with potential application points of crosslinking treatment highlighted.

6
Linkages to Cellulose: Non-Covalent and Structural Bonds

A major obstacle to crosslinking cellulose is the low reactivity of hydroxyls, cellulose’s most available
functional group. By inserting functional groups that are more reactive than hydroxyls onto the
cellulose, better “handles” can be made for crosslinking. A non-covalently bound or blended
polymer that is more easily cross-linked could become these “handles.” Polymers could be blended
into the threads of the fabric or applied to the textile at a variety of points its manufacture. Such an
addition is the most radical alternative that our group proposes in this report; these treatments are
upstream of LS&Co.’s typical crosslinking work space.
Polymer Blending
Polyester polyethylene terephthalate (PET) is frequently incorporated into LS&Co. fabric (Figure
4).4 Certain styles of Dockers incorporate PET blends in proportions upwards of 40%, while Levi’s
WasteLess jeans include polyester in proportions up to 29%.5 Within LS&Co., PET is sourced from
recycled drink bottles and is an excellent source of sustainability for the company. Additionally,
PET has phthalate ester handles, an attractive target for crosslinking. The expanded use of PET
within LS&Co. is very attractive; it could potentially achieve wrinkle repellency and water resistance.
Figure 4: PET, the most common polyester in textiles.6
PET is almost exclusively used for the weft fibers.4 It is unclear how much crosslinking to only the
weft fibers of a fabric could solve either the permanent press or DWR crosslinking challenge.
Crosslinking between synthetic and cotton fibers could work well in permanent press applications. It
is possible that the weft fibers alone of jeans may provide enough surface area for the EcoRepel
DWR treatment to adhere, but there may be complications as most weft fibers are located on the
inside of the garment. If LS&Co. investigates this alternative, they should consider treatments on a
variety of blends of fabric to understand its full potential.
Polymer Application
Applying a polymer to the cellulose rather than blending it is an attractive alternative, as this
technique could be integrated into existing manufacturing processes. However, several obstacles
must be overcome for this strategy to become viable. First, the polymer must provide more reactive
functional groups than are available on cotton for crosslinking. Secondly, the polymer must be
rendered sufficiently insoluble in water for it to hold fast to both the cotton and a cross-linker under
the conditions of manufacture and use. This is primarily a concern for alternatives in a sizing-type
coating. It is essential that the polymer coating not be removed in the desizing process, as this would

7
eliminate the crosslinking functionality. This leads to a third challenge: the treated garments must
have a pleasing feel to the hand. Some polymers may interfere with the softness of 100% cotton
clothing. North Korea manufactures a high-performance clothing called vinylon entirely out of
poly(vinyl alcohol) (PVA) that is described as feeling like wallpaper.7 This level of discomfort would
outweigh any benefit in functionality. Finally, the added polymer must not have interactions harmful
either to itself or the fabric in subsequent steps in the manufacture. Research reveals numerous
possibilities for overcoming these obstacles (described below), and their expected performance is
codified in the Technical Performance Evaluation.
Most unconventional would be the application of polymers for crosslinking in the fiber phase. The
concept of coating fibers with polymers, however, is not foreign to textile manufacture. Analogous
procedures take place during the sizing process, where chemicals including poly(vinyl alcohol) (PVA)
(Figure 5), carboxymethyl cellulose (CMC), acrylates, starch, and wax coat warp fibers to impart
tensile strength and facilitate weaving. We investigated alternatives using fiber procedures similar to
the sizing already used by LS&Co. that could impart easier crosslinking functionality. While these
coatings would be held onto the cellulose by numerous hydrogen bonds – a crosslinking strategy
common in nature – the systems of physical application are derived directly from procedures already
employed by the industry. EcoRepel DWR is able to coat individual fibers, thus imparting water
resistance without losing breathability. This strategy complements the DWR crosslinking application.
Figure 5. PVA, a water-soluble polymer frequently used in the textile industry.
Of the chemicals used in sizing, PVA and starch only offer more hydroxyls to crosslink, while CMC,
acrylates, and wax might offer more reactive carboxylic acid, ester, and vinyl functional handles. But
a variety of functional groups can be added to PVA, including thiols, esters, and urethanes.8,9 These
functional groups complement methods we propose for crosslinking applications. For instance,
thiolated PVA (Figure 6) would lend itself to forming disulfide bonds (discussed below). While the
variety of functional groups that can be added to polymers is promising, LS&Co. should take a
cautious approach while considering modifying polymers. Functionalizing treatments often occur
under intense physical or chemical conditions, many of which do not advance the goal of reducing
health hazard. For instance, PVA is typically made insoluble by treatment with heat and
formaldehyde, which is unacceptable from a health hazard standpoint.

8
Figure 6. Thiolated PVA could be used with disulfide bond crosslinking.
Rather than limiting itself to coatings derived from currently used sizes, LS&Co. should also
consider polymers that will coat fibers with better crosslinking potential than sizes. Poly(vinyl
acetate), poly(lactic acid), poly(methacrylic acid) and polyimides (Figures 7a-d) all offer strong
possibilities for noncovalent crosslinking applications and should coat fibers well. These polymers
are also more resistant to desizing and subsequent processes than PVA is. Nonetheless, some of
these polymers do raise concerns. PLA, for instance, may hydrolyze if exposed to water before being
crosslinked, and the chemical precursors of polyimides have many of the same health concerns as
the di-isocyanates used in crosslinking DWR finishes. While limited research has been done on the
applications of these polymers to a cellulose substrate, other polymer-coated textiles show good
performance in hand and feel.10 Teams have succeeded in coating natural and synthetic fibers with
polymers to achieve enhanced functionality with similar physical characteristics and handling as the
virgin material.11,12 Others have modified PLA to decrease its brittleness and increase its tensile
strength.13 This modification could even allow PLA to be used as a sizing agent that, rather than
being desized, is retained for functionality later in the manufacture process. PET fibers are also sized,
so this alternative could also be applied to cotton-poly blends.
a. b. c. d.
Figure 7: a. poly(lactic acid), b. poly(vinyl acetate), c. poly(methacrylic acid), d.polyimide.
Polymers could also be applied in either the textile or the garment phase, which are much more
conventional points in the process to impart crosslink functionality. Polymer coatings are often
applied in the fabric form. This is currently the case for DWR finishes like EcoRepel and the shape-
holding features of Levi’s Revel line.4 The polymers could be screen printed onto the garment, as in
the Revel line, or be polymerized on the surface of the fabric in situ as is done with electrically
conductive fabrics often used in technology like “smart braces”.14 These polymers, however, would
serve to enable later crosslinking functionality. This could be a viable pathway for coating textiles, as
the chemical intermediates of the polymerization can penetrate into the fibers themselves, making

9
the polymer an integral part of the fabric.11 Such polymers have demonstrated physical and chemical
stability under many relevant conditions (though they may require the fabric to be dry-clean only)
and can resist shearing away from the fabric at stress upwards of 6MPa.11
In conclusion, to augment the reactivity of cellulose in cotton clothing and enable easier crosslinking
in either the permanent press or DWR crosslinking application, polymeric handles could be non-
covalently bound to the cellulose. These handles could be coated on in fiber form like the sizes
currently used in textile manufacture, woven into the fabric in cotton-poly blends, or formed in situ
on the fabric. These alternatives offer radically different approaches than those presently used by the
textile industry but this new perspective may yield functional and nontoxic results.
Linkages to Cellulose: Covalent Bonds

Polycarboxylic acids have been studied for nearly 50 years as possible formaldehyde-free wrinkle-
release finishes for cellulose. They are also biologically-inspired, since many of them are found in
nature (Figure 8), and recent review stated that they are “the most promising non-formaldehyde
durable press finishes”.15 Despite this, they have not proven to be a fully effective substitute for
DMDHEU (dimethylol dihydroxyethyleneurea, the current technology) due to problems with loss
of tensile strength,15 the catalyst,16 cost,17 and fabric yellowing.18
Figure 8. Some polycarboxylic acids commonly found in nature.
Polycarboxylic acids work as cotton crosslinkers by covalently bonding to cellulose fibers. When
heated, they form cyclic anhydride intermediates that react with the hydroxyl groups on cellulose to
form ester linkages (Figure 9). Thus, the acids that have been studied typically have at least three
carboxylic acid groups, as two are required to form each cyclic anhydride. Examples include citric
acid, 1,2,3,4-butanetetracarboxylic acid (BTCA), and 1,2,3,4-cylopentanetetracarboxylic acid.16–23 Of
these, BTCA has proven the most effective but is cost-prohibitive.24 The cost of citric acid is
comparable to that of DMDHEU, but it causes fabric yellowing.16–18 Unsaturated dicarboxylic acids
have also been studied, such as maleic and itaconic acids: the double bonds found in these
compounds enable further crosslinking activity without cyclic anhydride formation.16,20,25–27
Figure 9. Formation of a cyclic anhydride enables polycarboxylic acids to react with cellulose.

10
The loss of tensile strength seen with polycarboxylic acid crosslinking is attributed to the stiffness of
the crosslinking itself, as well as the uneven heating in the curing step.22 To overcome this, longer-
chain acids are proposed as a way of making the crosslinker more flexible.15 One example of in situ
oligomerization of maleic and itaconic acids resulted in improved tensile strength over DMDHEU,
with lower cost than BTCA and no fabric yellowing.25 Microwave heating has also been proposed as
a way of curing the fabric more evenly.22 Finally, pre-cationization of the fabric with an ammonium
reagent, followed by ionic crosslinking, has also been shown to improve tensile strength.28
The most common catalyst used to perform crosslinking with polycarboxylic acids is sodium
hypophosphite (SHP). Unfortunately, it causes changes in dye shade and can degrade cellulose
fibers.16,24 It is also an environmental hazard, leading to eutrophication.16 Other phosphorus-based
catalysts18 and non-phosphorus-based catalysts16,24 have been studied, however, SHP remains the
standard.23 The yellowing that occurs when α-hydroxy acids such as citric acid are used has been
attributed to the formation of unsaturated acids at high temperatures, which can be reduced by
addition of boric acid16 or triethanolamine.18 Esterification of citric acid with a polyol can also reduce
unsaturated acid formation.23
In our case, we propose to use a difunctional carboxylic acid, such as succinic acid, which can form a
single linkage with cellulose. The acid would be applied and linked to the fabric in the finishing step.
At this point, an extra carboxylic acid group would be available for further reactivity. We then
propose using common bioconjugation reagents to react with this carboxylic acid in situ to attach a
thiol group (Figure 10).29 Alternatively, we propose using a diamine to crosslink with the carboxylic
acid groups, either through reductive amination or amidation.29,30 As this two-pronged approach has
not yet been tried with polycarboxylic acid crosslinking, we think it may open up new avenues of
research that could help mitigate some of the current problems.
Figure 10. A polycarboxylic acid bound to the cellulose could react with N-hydroxysuccinimide (NHS) and DCC (N,N’-
dicyclohexylcarbodiimide) to form an NHS-ester that can then react with an amine to attach a thiol group.
Crosslinking Strategies: Covalent Bonds

Disulfide Bonds
Disulfide bonds are ubiquitous in nature, and are therefore an obvious choice for a biologically-
inspired crosslinking strategy. In biological systems, disulfide bonds are formed by the oxidative
reaction of two cysteine residues (Figure 11) and stabilize secondary and tertiary protein structure.
Formation of the disulfide linkage requires an oxidizing agent (electron acceptor), which is usually
oxygen, a redox-active cofactor, or another thiol.31

11
Figure 11. The thiol (-SH) groups on cysteine can react to form disulfide linkages.
Disulfide bonds have also been extensively explored for synthetic applications,32 and a wide variety
of oxidants and catalysts can be used. (The term “oxidant” will be used here to denote the reagent
that accepts electrons from the thiol groups, meaning that it must be used in amounts equivalent to
the amount of thiol groups. “Catalyst” denotes an additional, optional chemical that is unchanged by
the reaction and is likely used in small amounts.) The simplest option is to simply allow the thiol
groups to oxidize in air, although this can take days,33 and more often stronger oxidants, catalysts, or
heat are used to accelerate the reaction process. Hydrogen peroxide is a common oxidant and can be
made even more effective by acidic pH or elevated temperature.34–36 Ellman’s reagent, 5,5'-
dithiobis(2-nitrobenzoic acid), is also often used, usually at alkaline pH.37 Sulfoxides are effective
oxidants at elevated temperatures, or in either acidic or alkaline media, and are often used with an
additional catalyst.32,38 Other common oxidants used include sodium periodate,36 tetranitromethane,36
and potassium permanganate.39
Recently some “greener” options for disulfide bond formation have been reported. Some use a pre-
oxidized catalyst such as graphite oxide, although in that case the reagents used to oxidize the
graphite are not especially “green”.40 Others use air as an oxidant with an additional catalyst, such as
potassium fluoride on alumina,41 potassium phosphate,42 metal salts on silica,32 iron metal-organic
frameworks,43 or nickel44 or gold nanoparticles.45 One additional advantage of some of these systems
is that they work in solvent-free conditions, which could be an advantage for setting a permanent
crease.
Although disulfide bonds have not, to our knowledge, been used in textile applications before, there
are examples of disulfide bond crosslinking to form biodegradable polymers,34 hydrogels,35 or
bioconjugated systems.46 For our application, we propose that thiol groups be attached to cellulose
either through a thiolated polymer coating46–49,8 or by reaction with a poly(carboxylic acid). This
application would likely occur in the finishing step, at which point the crosslinking could also occur
(for example, attaching a DWR finish). Alternatively, the crosslinking step could be performed in the
garment form, as would be required to set a permanent crease. Either way, the crosslinking step
would occur upon addition of the necessary oxidant or catalyst.

12
Imine and Amine Bonds
Imine bonds are formed reversibly in the slug, Arion subfuscus, to harden the mucus secreted by the
animal and form a protective layer (Figure 12). Iron and copper ions catalyze carbonyl formation on
protein side chains; the carbonyls subsequently react with the amine group on lysine to form the
imine bonds.50 Imine bond formation is a dehydration reaction, and imine bonds are easily
hydrolyzed back to the starting amine and carbonyl groups. While this reversibility is useful for the
slug, we were concerned about an easily hydrolyzable crosslinking bond.
Figure 12. Imine bond formation, resulting in a release of water.
To impart added stability to our proposed crosslinking bond, we suggest taking the imine bond
formation one step further and performing a reductive amination (Figure 13). The starting materials are
still an amine and a carbonyl, which react to form an imine intermediate. Upon addition of a catalyst
and a reducing agent (electron donator), this intermediate is then reduced to an amine, which is
more stable and resistant to hydrolysis than the imine.51
Figure 13. Amine bonds can be formed by the reduction of an imine.
Reductive amination is most commonly performed using catalytic hydrogenation or borohydride

reagents. Catalysts for hydrogenation are usually nickel, platinum, or palladium.52 Sodium
cyanoborohydride is the most extensively studied borohydride reagent, but it is extremely toxic, in
part because it can form hydrogen cyanide during the reaction workup. 52–54 More recent results
report that sodium borohydride and sodium triacetoxyborohydride are also effective reducing agents
for reductive amination, with or without catalysts. Catalysts for sodium borohydride reduction
include titanium (IV) isopropoxide53, iron triflate,54 and acetic acid.55 Although reductive amination is
usually performed using aldehydes or ketones, carboxylic acid-amine reactions may work,30 and
analogous alcohol-amine reactions have also been reported.56
Reductive amination has been explored for crosslinking in other applications, with varying success.
In one case, polyamines were successfully used to crosslink polysaccharides using sodium
borohydride.57 In another, a poly(caprolactone)-based polymer was reacted with a diamine and
sodium cyanoborohydride, which resulted in polymer chain cleavage instead of crosslinking.58 This
suggests that success depends greatly on both the nature of the polymer and the reductant. For
textile applications, we propose to use a diamine crosslinker with a carbonyl-containing cellulose
treatment. The cellulose treatment could be polyethyelene terephthalate woven with the cotton, a

13
polymeric fabric coating such as poly(lactic acid), or a poly(carboxylic acid) that can covalently bind
to cellulose. The diamine would be applied during the finishing step, and crosslinking would be
effected upon addition of the reducing agent and catalyst.
FRAMEWORKS + EVALUATIONS
Boundary Conditions
We limited our proposed crosslinking and cellulose binding strategies to ones that are minimally
disruptive to the overall garment manufacturing process. Radical overhauls are not feasible
solutions, as there are currently no regulatory drivers forcing significant changes to crosslinking
technologies. While we did develop and implement an approach to evaluating the technical
performance of our proposed strategies beyond manufacturing feasibility, we did not evaluate these
strategies on the basis of either cost or measured performance. These are beyond our purview as we
have neither the data nor the laboratory equipment and expertise to evaluate cost and performance
parameters. Some of the considerations in the Technical Evaluation, nonetheless, are excellent initial
approximations for cost and measured performance considerations.
We also limited our proposals to those that showed promising health evaluations. Reducing human
health and environmental impacts were top priorities as we researched numerous approaches to
both crosslinking and cellulose binding; therefore, we neither pursued in depth, nor included in this
report, those strategies that required input chemicals whose impacts offered little or no
improvement over current technologies.
Technical Framework
For our technical evaluation, we chose eight factors that we felt we could qualitatively evaluate using
the time and resources available to us. The first four factors deal with innovation and disruption to
the current system, and are in some ways good proxies for cost as they represent the possible
research investment required before a strategy could be fully implemented. The remaining four
factors deal with specific problems associated with binding to cellulose, crosslinking, and other
effects on the fabric, and are thus reasonable proxies for performance.
In all eight categories, we gave each strategy a score from “red” to “green” (Table 2). “Red” scores
indicate that a technology is a pretty big change or that we foresee numerous technical and
performance hurdles. “Green” scores are given for technologies that fit well into the existing fabric
treatment framework or do not have a lot of anticipated hurdles. “Yellow” scores fall somewhere in
the middle.

14
INNOVATION ADDITIONAL MAJOR HURDLES MINOR HURDLES OPTIMIZATION

RESEARCH NEEDED ANTICIPATED ANTICIPATED ONLY
SPECIAL LIMITED WIDE

CHEMICAL SUPPLY
MANUFACTURE AVAILABILITY AVAILABILITY
DISRUPTION OF FABRIC MODIFY EXISTING USES EXISTING

INFRASTRUCTURE NEW PROCESS
APPLICATION PROCESS PROCESS
NEW CHEMICALS, HEAT OR AIR

CROSSLINKING STEP NEW EQUIPMENT
SOLVENTS CURED
CONTROLLABLE TOO REACTIVE OR SPECIAL CONDITIONS ADD CATALYST,

CROSSLINKING UNREACTIVE OR EXTRA CHEMICALS REAGENT, HEAT
ROBUSTNESS RESILIENCE DURING LIKELY PROBLEMS POSSIBLE PROBLEMS

NO FORESEEABLE
MANUFACTURING PROBLEMS
RESILIENCE DURING NO FORESEEABLE

LIKELY PROBLEMS POSSIBLE PROBLEMS
CONSUMER USE PROBLEMS
REQUIRES PROBLEM POSSIBLE NEED FOR NO FORESEEABLE

SIDE EFFECTS EFFECTS ON FABRIC
CHEMICALS PROBLEM CHEMICALS PROBLEMS
Table 2. Eight technical evaluation factors, with scoring requirements from red to green.
It is important to emphasize that a “red” rating does not necessarily make a strategy bad, any more
than a “green” rating makes it good. We want to be clear that “red” ratings merely reflect technology
disruptiveness and our ability to predict potential problems. “Green” ratings, meanwhile, may
disguise problems that we are unable to predict, or simply indicate that a technology is well-
researched. However, at this point, no rating corresponds to the eventual success or failure of
strategy.
Detailed descriptions of the eight technical evaluation factors are given in Appendix D.
Combined Strategies Yielding Solutions
In order to get a technical evaluation of a complete cellulose binding-and-crosslinking solution, the

technical results for one cellulose binding strategy and one crosslinking strategy can simply be
averaged. In other words, if the final proposed solution was to use a diamine crosslinker with the
PET weave, the results for PET weave would be averaged with the results for imine & amine
crosslinking. Since our health framework depends on specific chemical evaluations, we chose some
specific strategy combinations to evaluate using both our technical framework and our health
framework (Figures 14 and 15).
Based on our combinations of strategies, we can see that imine bond crosslinking with poly(ethylene
terephthalate) weaving is the most promising solution according to our current information,
although any of the combinations may be ultimately viable. More detailed descriptions of how these
classifications were determined are provided in Appendix E.

15
INNOVATION DISRUPTION ROBUSTNESS SIDE EFFECTS

ADD’L CHEMICAL FABRIC CROSSLINK CONTROLLED PROCESS USE PHASE PROCESS
RESEARCH
INNOVATIONSUPPLY APPLICATION
DISRUPTION STEP ROBUSTNESSDURABILITY SIDE
CROSSLINK DURABILITY EFFECTS
EFFECTS
NEEDED DISRUPTION DISRUPTIONFABRIC
DISRUPTION
ADD’L CHEMICAL CROSSLINK CONTROLLED PROCESS USE PHASE PROCESS
POLYMER WEAVE RESEARCH SUPPLY APPLICATION STEP CROSSLINK DURABILITY DURABILITY EFFECTS
NEEDED DISRUPTION DISRUPTION DISRUPTION
FIBER COATING
POLYMER WEAVE
FABRIC COATING
FIBER COATING
IN SITU POLYMERIZATION
FABRIC COATING
POLY[CARBOXYLIC ACIDS]
DISULFIDE BONDSACIDS]
POLY[CARBOXYLIC
IMINE/AMINE BONDSBONDS
DISULFIDE
IMINE/AMINE BONDS
Figure 14: Technical evaluation of each strategy.
INNOVATION DISRUPTION ROBUSTNESS SIDE EFFECTS
ADD’L
INNOVATIONCHEMICAL FABRIC
DISRUPTIONCROSSLINK ROBUSTNESSUSE PHASE
CONTROLLED PROCESS SIDE EFFECTS
PROCESS
RESEARCH SUPPLY APPLICATION STEP CROSSLINK DURABILITY DURABILITY EFFECTS
ADD’L CHEMICAL FABRIC CROSSLINK CONTROLLED PROCESS USE PHASE PROCESS
NEEDED DISRUPTION DISRUPTION DISRUPTION
APPLICATION
RESEARCH SUPPLY STEP CROSSLINK DURABILITY DURABILITY EFFECTS
POLY[CARBOXYLIC ACIDS] NEEDED DISRUPTION DISRUPTION DISRUPTION
+ POLY[CARBOXYLIC
DISULFIDE BONDSACIDS]
+ DISULFIDE BONDS
THIOL. POLY. [FABRIC
THIOL. BONDS
COAT] + DISULFIDE POLY. [FABRIC
COAT] + DISULFIDE BONDS
CARBONYL POLYMER +
CARBONYL
IMINE/AMINE POLYMER +
BONDS
IMINE/AMINE BONDS
POLYMER WEAVING +
POLYMER WEAVING +
IMINE/AMINE BONDS
IMINE/AMINE BONDS
+ IMINE/AMINE BONDS
+ IMINE/AMINE BONDS
Figure 15: Technical evaluation of combined strategies.

16
Health Framework
GreenScreen Precedent
We used a hazard-based approach to evaluate the health impacts of the chemicals used in
our proposed crosslinking alternatives. While there are numerous hazard- and exposure-
based methods for assessing health risk, LS & Co., and other major stakeholders in the
textile industry, have embraced a hazard-based approach. To align as closely as possible with
this current paradigm, our health evaluation framework was derived from the GreenScreen
Chemical Hazard Assessment Procedure, which LS&Co. already implements.62 The
GreenScreen framework is a comparative hazard assessment tool that builds upon nationally
and internationally recognized hazard evaluation frameworks to assess and classify hazards
and to apply benchmarks to chemicals.63 Each chemical evaluation includes a hazard
assessment of the parent chemical as well as potential environmental transformation
products. Individual chemical evaluations draw upon measured data gathered from the
literature, modeled data, data for suitable chemical analogs, and hazard lists published by
governmental agencies, academic researchers, and non-governmental agencies.64 Chemicals
are evaluated for 18 different endpoints sorted into five hazard groups. Each endpoint is
assigned a hazard classification [vH, H, M, L, vL] and confidence level. These hazard
classifications are aggregated into a benchmark score based on the GreenScreen’s weighting
system.65
Framework scope and goals

The context of our research posed two key challenges that required us to develop and
implement an abbreviated version of the GreenScreen framework: First, limited time and
scope do not allow a comprehensive evaluation of the scientific literature; second, many of
the proposed chemicals are not well studied and therefore have very limited toxicological
and epidemiological data. As a result, our proposed framework will not result in
GreenScreen benchmarks for individual chemicals. Rather, our framework aims to do the
following:
1. Identify [and subsequently eliminate] chemicals meeting any Benchmark 1 criteria [Table
3].
2. Compare, on a relative basis, chemicals for which Benchmark 1 designation does not
apply using a weighting system derived from GreenScreen12.
3. Compare hazard of individual chemicals and the lifecycles in which they operate,
including the inputs and conditions for their synthesis, subsequent crosslinking syntheses
in which they participate, and use phase and end of life hazards.
4. Compare hazard of each of the four crosslinking strategies on a system level.

17
Table 3: GreenScreen Benchmark 1 Criteria

Hazard Classification Criteria
PBT High P + High B + [very high T (ecotoxicity or Group II

Human) or high T (Group I or II* Human)]
vPvB Very high P or very high B
vPT Very high P + [very high T (ecotoxicity or Group II Human) or

high T (Group I or II* Human)]
vBT Very high B + [very high T (ecotoxicity or Group II Human) or

high T (Group I or II* Human)]
High T Group I Human
See GreenScreen Guidance documents10 for definitions of the five GreenScreen endpoint
groupings: Group I Human, Group II Human, Group II* Human, Ecotoxicity and Fate,
and Physical Hazards.
Evaluation
We researched five strategies, including two crosslinking and three cellulose-binding
strategies. Lifecycle hazard evaluation for each of these strategies included:
1. Chemical inputs and conditions required to make crosslinking or binding chemical

2. Chemical inputs and conditions required to crosslink or bind
3. Probable environmental transformation products occurring in use or end of life phases
This scope of evaluation was expanded from that of the GreenScreen to include of the
parent crosslinking chemical both to be more complete and because these chemicals can
serve as surrogates for some proposed chemicals whose health hazards are not well
understood.
For each chemical, we first gathered data from authoritative lists designated by
GreenScreen11. Unlike GreenScreen, we did not weight some lists more heavily than others.
Where list data is unavailable, inconclusive, or conflicting, we turned to the primary literature
to gather supplementary hazard information. Where toxicological and epidemiological data
were limiting or inconclusive, we will use modeled data or data from suitable analogs.
Modeled data was used almost exclusively for estimations of persistence and
bioaccumulativity.
We completed the GreenScreen hazard table to the extent that was possible given time and
data constraints. Where possible, we identified and eliminated Benchmark 1 chemicals. Our
framework assessed remaining chemicals relative to other chemicals serving comparable
roles. Our assessment then prioritized the evaluation of those endpoints weighted most

18
heavily by the GreenScreen weighting system. We performed this data collection and relative
ranking process for each chemical within each of the proposed crosslinking and binding
strategies. The process cumulatively enableed comparison both within and among the four
strategies [Figure 16], though due to data and time limitations, we believe that the more
useful comparison is among chemicals performing similar roles within single strategies.
Results from these health evaluations can be used to guide further research, either through
prioritization or elimination of a candidate chemical for the strategy of interest.
SINGLE CHEMICAL
EVALUATION
GROUP I HUMAN GROUP II + II* HUMAN E TOX FATE PHYS

ST N
C M R D E AT sgl rep sgl rep SnS SnR IrS IrE AA CA P B Rx F
GROUP I HUMAN GROUP II + II* HUMAN E TOX FATE PHYS

ST N
C M R D E AT sgl rep sgl rep SnS SnR IrS IrE AA CA P B Rx F
CHEMICAL HAZARD
COMPARISON
SYSTEMS HAZARD COMPARISON
Figure 16: Health evaluation process.
Alignment with GreenScreen

While our evaluation framework is an abbreviated version of the full GreenScreen, the
underlying structure of the assessment is the same, allowing the data collected for this initial
screening to feed directly into a full GreenScreen, in the future, should one be performed.
Figure 17 shows the compatibility of our adapted evaluation with the GreenScreen
framework.

19
Figure 17: A comparison of scope, data collection and prioritization, and outcomes of the GreenScreen with
our proposed adapted evaluation framework.
Health Evaluations
The following health evaluations rely on graphics that have been created from data gathered
from authoritative lists and the primary literature using the approach described above. The
raw data from which these graphics have been developed, as well as the corresponding
references, can be found in the accompanying Excel File: HealthEvaluationData.xlsx. Figure
18, below, describes the color coding system that was used to translate from the raw data to
the tables shown in this section.
MEETS GREENSCREEN BENCHMARK 1 CRITERIA FOR AT LEAST ONE

BM 1 ENDPOINT; MUST BE ELIMINATED.
PROBABLE HIGH HAZARD FOR GROUP 1 HUMAN AND ECOTOXICITY
ENDPOINTS; VERY HIGH GROUP II/II* ENDPOINTS; AVOID.
POTENTIAL HAZARD FOR GRP I HUMAN AND ECOTOX ENDPOINTS; KNOWN
HIGH HAZARD FOR GRP II/II* HUMAN ENDPOINTS; HIGH PHYSICAL HAZARD.
REASONABLE SUSPICION FOR CONCERN; MORE RESEARCH IS
NECESSARY.
SUITABLE SUBSTITUTION BASED ON AVAILABLE DATA.
NO DATA AVAILABLE.
Figure 18. Health evaluation framework weighting scheme.
While we tried to be comprehensive in our evaluations, we could not study everything. It is

possible that a proposed strategy that appears to offer limited hazard reduction may actually
be a strong alternative, but the right candidate chemicals may have yet to be identified and
evaluated.
Baseline evaluation
Health evaluations for the baseline chemicals are shown in Figure 19 below. While we
evaluated many chemicals currently used in textile processing, we targeted the last four
chemicals for substitution, which are currently used in LS&Co.’s cotton crosslinking systems.

20
Figure 19. Health evaluation data for the baseline wet processing technologies.
Formaldehyde is most notable among the chemicals listed above; it is both well studied and
highly toxic. While formaldehyde is cheap, versatile, and high-performing, it is also a known
carcinogen, acutely toxic, and a known skin and respiratory sensitizer. PFBS, a common
DWR, is another chemical of concern. Most notably from a health perspective, the PFBS
DWR has demonstrated reproductive and developmental toxicity, and is highly acutely toxic
to aquatic life. LS&Co. has partially phased out PFBS and has substituted it with a paraffin-
based alternative in many of its products. Finally, diisocyanates, which are also widely used in
water repellency technologies, are respiratory sensitizers and are possibly carcinogenic,
among other hazard concerns.
Each of these chemicals is integral to textile crosslinking, both for LS&Co. and across the
textile industry. As the demand and expectation for hazard transparency and reduction gains
prevalence, however, safer alternatives for these chemicals will become increasingly
necessary.
Comparing within strategies: disulfide bonds evaluation

As Figure 20 shows, the disulfide bond strategy integrates chemicals that appear, overall, to
pose reduced hazard relative to the baseline technology, particularly for the Group I Human
endpoints and acute mammalian toxicity. Additionally, unlike many of the baseline chemicals,
many of the hazards for which these chemicals have been listed are more easily mitigated [i.e.
skin and eye irritation]. That said, most of these compounds are not as well studied as the
baseline chemicals, which significantly limits the conclusions that can be drawn from this
comparison.

21
Figure 20. Health evaluation data for the proposed disulfide bond crosslinking strategy.
The table does indicate that some candidates are more favorable than others of comparable
function, both on the basis of data availability and on based on the hazard levels those data
indicate. For example, the evaluation clearly favors further research on oxygen’s feasibility as
a thiol oxidant in textiles.
Comparing within strategies: amine/imine bonds evaluation

Like the previous evaluation, the amine/imine bond health evaluation is plagued by data
gaps [Figure 21]. The table can, however, be used to create a relative hazard ranking, both
among candidate chemicals within the strategy, and, to some extent, relative to the other
crosslinking and binding strategies.
Figure 21. Health evaluation data for the proposed imine/amine bond crosslinking strategy.
For example, the results differentiate and prioritize among the proposed catalysts and the
reductants. While the data gaps cannot be ignored, hydrogen emerges as a promising starting
point for an imine reductant relative to sodium borohydride; similarly, Raney nickel is
quickly singled out as a low priority catalyst for further feasibility and performance research,
while platinum may be a reasonable candidate, as it has shown reduced hazard for more
easily mitigated hazard endpoints.

22
Comparing within strategies: poly[carboxylic acids]

The data gaps for the poly[carboxylic acid] strategy make hazard nearly impossible to
interpret, even on a relative basis [Figure 22]. That said, more so than any of the other
strategies, poly[carboxylic acid] crosslinking in cellulose is well studied within the textile
industry; therefore, it is possible that some proprietary hazard data may already exist, or data
may become available as poly[carboxylic acid]-based systems are developed and more widely
used.POLY CARBOXYLIC ACIDS
&#'()*+*,("-. &#'()*++*/*++0*,("-. %*1': ;-1% ),<2
21 .
!,%"+!-? ;(.!1+'. ! " # $ % -1 345 678 345 678 292 29# +62 +6% -- !- ) = #> ;
!+1#+!*-!+$ !-#=':<?+!*-!+$
2(!!+.+!*-!+$ !-#=':<?+!*-!+$
"-?+!*-!+$ !-#=':<?+!*-!+$
!<21-"+.%*,!5 1,+'?-1+.&*-&%.1
.@,<$#':<2(!!+.-"+$% !-1-?<21
.@AB@$+"%1,<?-"+.'CCCD !-1-?<21
1,+'?-1%$*#%2+. #%2+.
Figure 22. Health evaluation data for the proposed poly(carboxylic acid) crosslinking strategy.
Comparing within strategies: polymer thiolation evaluation

More so than many of the other strategies, the chemicals required of the polymer thiolation
strategy show great, well-supported improvement relative to the baseline chemicals for
Group I Human endpoints [Figure 23]. There are many promising compounds in several of
the functional categories, such as the cysteine-based thiolating agents.
Figure 23. Health evaluation data for the proposed polymer thiolation strategy.

23
Furthermore, we were able to find some data describing hazard impacts for the thiolated
polymer; thus, we can begin to understand not only the inputs and processing chemicals
required to implement the technology, but also the impacts of the resulting chemical
products.
Comparing within strategies: polymer coating evaluation

Figure 24 shows the health evaluation data summary for the polymer coating strategy. Based
on the health evaluation shown, this strategy appears less promising than the others, as few
of the chemicals demonstrate clear improvement over existing technologies at LS&Co. Even
comparing candidate chemicals of similar function, it is difficult to identify chemicals that do
not pose significant hazard in at least one of the Group I or Group II Human endpoints.
Figure 24. Health evaluation data for the proposed polymer coating strategy.
RECOMMENDATIONS
Based on our assessment of biologically-inspired crosslinking strategies for cellulose, we

have several recommendations for Levi Strauss & Co. First is that we recommend that non-
traditional crosslinking strategies be considered. Nature employs various bonds in
conjunction with one another to achieve crosslinking, and there is room for the textile
industry to emulate this through the adoption of non-traditional bonds and piece-wise
crosslinking techniques.

24
This report identifies several non-traditional bonding alternatives that are inspired by nature.
We have additionally outlined two evaluation tools – a technical evaluation framework and a
health evaluation framework – to frame the validity of each bonding strategy.
Of the five strategy combinations we considered, imine/amine bond crosslinking with

poly(ethylene terephthalate) weaving is the most promising solution from a technical
standpoint, but any of the combinations are potentially worth pursuing. From a health
standpoint, the most promising solution is the thiolated polymer coating, thus highlighting
the complex and nuanced interactions and tradeoffs between technical considerations and
health considerations.
Moving forward, we recommend that Levi Strauss & Co. use these two frameworks and the
information contained in this report to further pursue one of the alternatives suggested, or
to pursue another proposed solution.
BIBLIOGRAPHY
(1) IARC Monograph Volume 88; IARC: Lyon, France, 2006.

(2) Toluene Diisocyanate [TDI] And Related Compounds Action Plan; U.S. Environmental
Protection Agency, 2011.
(3) AskNature. What Is Biomimicry?
(4) Cattermole, A. Sustainability at Levi’s, 2013.
(5) Levi’s Website-US http://us.levi.com/home/ (accessed Nov 20, 2013).
(6) Why is recycled polyester considered a sustainable textile?
http://oecotextiles.wordpress.com/2009/07/14/why-is-recycled-polyester-
considered-a-sustainable-textile/ (accessed Nov 15, 2013).
(7) Hee, M. C. Vinylon and CNC? What are they good for?
http://www.dailynk.com/english/read.php?cataId=nk01300&num=6136.
(8) Gupta, B.; Anjum, S.; Ikram, S. Preparation of Thiolated Polyvinyl Alcohol Hydrogels.
J. Appl. Polym. Sci. 2013, 129, 815–821.
(9) Eastman, S. A.; Lesser, A. J.; McCarthy, T. J. Quantitative Poly(vinyl Alcohol)
Modification in Ionic Liquids: Esterification and Urethanation with Low Surface
Tension Producing Reagents. Macromolecules 2010, 43, 4584–4588.
(10) Gruber, P.; Kolstad, J.; Ryan, C.; Hall, E.; Eichen, R. Paper Having a Melt-stable
Lactide Polymer Coating and Process for Manufacture Thereof. 5,475,08, December
12, 1995.
(11) Heisey, C. L.; Wightman, J. P.; Pittman, E. H.; Kuhn, H. H. Surface and Adhesion
Properties of Polypyrrole-Coated Textiles. Text. Res. J. 1993, 63, 247–256.
(12) Gregory, R. V.; Kimbrell, W. C.; Kuhn, H. H. Electrically Conductive Non-Metallic
Textile Coatings. J. Ind. Text. 1991, 20, 167–175.
(13) Semba, T.; Kitagawa, K.; Ishiaku, U. S.; Hamada, H. The Effect of Crosslinking on
the Mechanical Properties of Polylactic Acid/polycaprolactone Blends. J. Appl. Polym.
Sci. 2006, 101, 1816–1825.
(14) Wu, J.; Zhou, D.; Too, C. O.; Wallace, G. G. Conducting Polymer Coated Lycra.
Synth. Met. 2005, 155, 698–701.

25
(15) Harifi, T.; Montazer, M. Past, Present, and Future Prospects of Cotton Cross-linking:
New Insight into Nano Particles. Carbohydr. Polym. 2012, 88, 1125–1140.
(16) Choi, H.-M.; Welch, C. M.; Morris, N. M. Nonphosphorus Catalysts for
Formaldehyde-Free DP Finishing of Cotton with 1,2,3,4-Butanetetracarboxylic Acid:
Part II: Sodium Salts of Fumaric, Maleic, and Itaconic Acids. Text. Res. J. 1994, 64,
501–507.
(17) Welch, C. M. Tetracarboxylic Acids as Formaldehyde-Free Durable Press Finishing
Agents Part I: Catalyst, Additive, and Durability Studies1. Text. Res. J. 1988, 58, 480–
486.
(18) Schramm, C.; Rinderer, B. Influence of Additives on the Formation of Unsaturated
PCAs Produced During Durable-press Curing with Citric Acid. Color. Technol. 1999,
115, 306–311.
(19) Morris, C. E.; Morris, N. M.; Trask-Morrell, B. J. Interaction of meso-1,2,3,4-
Butanetetracarboxylic Acid with Phosphorus-Containing Catalysts for Esterification
Cross-Linking of Cellulose. Ind. Eng. Chem. Res. 1996, 35, 950–953.
(20) Yang, C. Q.; Xilie Wang. Formation of Cyclic Anhydride Intermediates and
Esterification of Cotton Cellulose by Multifunctional Carboxylic Acids: An Infrared
Spectroscopy Study. Text. Res. J. 1996, 66, 595–603.
(21) Ibrahim, N. A.; Abo-Shosha, M. H.; Gaffar, M. A. Eco-Friendly Durable Press
Finishing of Cellulose-Containing Fabrics. J. Appl. Polym. Sci. 2001, 84, 2243–2253.
(22) Fouda, M. M. G.; El Shafei, A.; Sharaf, S.; Hebeish, A. Microwave Curing for
Producing Cotton Fabrics with Easy Care and Antibacterial Properties. Carbohydr.
Polym. 2009, 77, 651–655.
(23) Yao, W.; Wang, B.; Ye, T.; Yang, Y. Durable Press Finishing of Cotton Fabrics with
Citric Acid: Enhancement of Whiteness and Wrinkle Recovery by Polyol Extenders.
Ind. Eng. Chem. Res. 2013, 52, 16118–16127.
(24) Vargantwar, P. H. Preparation of Ionic Cellulose for Wrinkle Resistant Fabrics, North
Carolina State University, 2007.
(25) Chen, D.; Yang, C. Q.; Qiu, X. Aqueous Polymerization of Maleic Acid and Cross-
linking of Cotton Cellulose by Poly (maleic Acid). Ind. Eng. Chem. Res. 2005, 44, 7921–
7927.
(26) Peng, H.; Yang, C. Q.; Wang, X.; Wang, S. The Combination of Itaconic Acid and
Sodium Hypophosphite as a New Cross-Linking System for Cotton. Ind. Eng. Chem.
Res. 2012, 51, 11301–11311.
(27) Peng, H.; Yang, C. Q.; Wang, S. Nonformaldehyde Durable Press Finishing of Cotton
Fabrics Using the Combination of Maleic Acid and Sodium Hypophosphite. Carbohydr.
Polym. 2012, 87, 491–499.
(28) Hashem, M.; Elshakankery, M. H.; El-Aziz, S. M. A.; Fouda, M. M. G.; Fahmy, H. M.
Improving Easy Care Properties of Cotton Fabric via Dual Effect of Ester and Ionic
Crosslinking. Carbohydr. Polym. 2011, 86, 1692–1698.
(29) Thordarson, P.; Droumaguet, B.; Velonia, K. Well-defined Protein–polymer
Conjugates—synthesis and Potential Applications. Appl. Microbiol. Biotechnol. 2006, 73,
243–254.
(30) Perrio-Huard, C.; Aubert, C.; Lasne, M.-C. Reductive Amination of Carboxylic Acids
and [11C] Magnesium Halide Carboxylates. J Chem Soc Perkin Trans 1 2000, 311–316.
(31) Gilbert, H. F. [2] Thiol/disulfide Exchange Equilibria and Disulfide Bond Stability.
Methods Enzymol. 1995, 251, 8–28.

26
(32) Witt, D. Recent Developments in Disulfide Bond Formation. Synthesis 2008, 16,
2491–2509.
(33) Kakizawa, Y.; Harada, A.; Kataoka, K. Environment-Sensitive Stabilization of
Core−Shell Structured Polyion Complex Micelle by Reversible Cross-Linking of the
Core through Disulfide Bond. J. Am. Chem. Soc. 1999, 121, 11247–11248.
(34) Zelikin, A. N.; Quinn, J. F.; Caruso, F. Disulfide Cross-Linked Polymer Capsules: En
Route to Biodeconstructible Systems. Biomacromolecules 2006, 7, 27–30.
(35) Shu, X. Z.; Liu, Y.; Luo, Y.; Roberts, M. C.; Prestwich, G. D. Disulfide Cross-Linked
Hyaluronan Hydrogels. Biomacromolecules 2002, 3, 1304–1311.
(36) Evans, B. J.; Doi, J. T.; Musker, W. K. Fluorine-19 NMR Study of the Reaction of P-
fluorobenzenethiol and Disulfide with Periodate and Other Selected Oxidizing Agents.
J. Org. Chem. 1990, 55, 2337–2344.
(37) Lehrer, S. S. Intramolecular Crosslinking of Tropomyosin via Disulfide Bond
Formation: Evidence for Chain Register. Proc. Natl. Acad. Sci. 1975, 72, 3377–3381.
(38) Arterburn, J. B.; Perry, M. C.; Nelson, S. L.; Dible, B. R.; Holguin, M. S. Rhenium-
catalyzed Oxidation of Thiols and Disulfides with Sulfoxides. J. Am. Chem. Soc. 1997,
119, 9309–9310.
(39) Shaabani, A.; Lee, D. G. Solvent Free Permanganate Oxidations. Tetrahedron Lett. 2001,
42, 5833–5836.
(40) Dreyer, D. R.; Jia, H.-P.; Todd, A. D.; Geng, J.; Bielawski, C. W. Graphite Oxide: a
Selective and Highly Efficient Oxidant of Thiols and Sulfides. Org. Biomol. Chem. 2011,
9, 7292.
(41) Lenardão, E. J.; Lara, R. G.; Silva, M. S.; Jacob, R. G.; Perin, G. Clean and Fast
Oxidative Transformation of Thiols to Disulfides Under Solvent-free Conditions.
Tetrahedron Lett. 2007, 48, 7668–7670.
(42) Joshi, A. V.; Bhusare, S.; Baidossi, M.; Qafisheh, N.; Sasson, Y. Oxidative Coupling of
Thiols to Disulfides Using a Solid Anhydrous Potassium Phosphate Catalyst.
(43) Dhakshinamoorthy, A.; Alvaro, M.; Garcia, H. Aerobic Oxidation of Thiols to
Disulfides Using Iron Metal–organic Frameworks as Solid Redox Catalysts. Chem.
Commun. 2010, 46, 6476.
(44) Saxena, A.; Kumar, A.; Mozumdar, S. Ni-nanoparticles: An Efficient Green Catalyst
for Chemo-selective Oxidative Coupling of Thiols. J. Mol. Catal. Chem. 2007, 269, 35–
40.
(45) Corma, A.; Ródenas, T.; Sabater, M. J. Aerobic Oxidation of Thiols to Disulfides by
Heterogeneous Gold Catalysts. Chem. Sci. 2012, 3, 398.
(46) Bernkop-Schnürch, A.; Clausen, A. E.; Hnatyszyn, M. Thiolated Polymers: Synthesis
and in Vitro Evaluation of Polymer–cysteamine Conjugates. Int. J. Pharm. 2001, 226,
185–194.
(47) Marschütz, M. K.; Bernkop-Schnürch, A. Thiolated Polymers: Self-crosslinking
Properties of Thiolated 450 kDa Poly (acrylic Acid) and Their Influence on
Mucoadhesion. Eur. J. Pharm. Sci. 2002, 15, 387–394.
(48) Bernkop-Schnürch, A. Thiolated Polymers—thiomers: Synthesis and in Vitro
Evaluation of Chitosan–2-iminothiolane Conjugates. Int. J. Pharm. 2003, 260, 229–237.
(49) Leitner, V. M.; Walker, G. F.; Bernkop-Schnürch, A. Thiolated Polymers: Evidence
for the Formation of Disulphide Bonds with Mucus Glycoproteins. Eur. J. Pharm.
Biopharm. 2003, 56, 207–214.

27
(50) Braun, M.; Menges, M.; Opoku, F.; Smith, A. M. The Relative Contribution of
Calcium, Zinc and Oxidation-based Cross-links to the Stiffness of Arion Subfuscus
Glue. J. Exp. Biol. 2012, 216, 1475–1483.
(51) Emerson, W. S. The Preparation of Amines by Reductive Alkylation. Org. React. 1948,
174–202.
(52) Abdel-Magid, A. F.; Carson, K. G.; Harris, B. D.; Maryanoff, C. A.; Shah, R. D.
Reductive Amination of Aldehydes and Ketones with Sodium Triacetoxyborohydride.
Studies on Direct and Indirect Reductive Amination Procedures. J. Org. Chem. 1996,
61, 3849–3862.
(53) Bhattacharyya, S. Reductive Alkylation of Dimethylamine Using Titanium (IV)
Isopropoxide and Sodium Borohydride: An Efficient, Safe, and Convenient Method
for the Synthesis of N, N-dimethylated Tertiary Amines. J. Org. Chem. 1995, 60, 4928–
4929.
(54) Uday Kumar, N.; Sudhakar Reddy, B.; Prabhakar Reddy, V.; Bandichhor, R. Iron
Triflate Catalyzed Reductive Amination of Aldehydes Using Sodium Borohydride.
(55) Gribble, G. W. Sodium Borohydride in Carboxylic Acid Media: a Phenomenal
Reduction System. Chem. Soc. Rev. 1998, 27, 395–404.
(56) Reddy, M. M.; Kumar, M. A.; Swamy, P.; Naresh, M.; Srujana, K.; Satyanarayana, L.;
Venugopal, A.; Narender, N. N-Alkylation of Amines with Alcohols over Nanosized
Zeolite Beta. Green Chem. 2013, 15, 3474.
(57) Ehrenfreund-Kleinman, T.; Gazit, Z.; Gazit, D.; Azzam, T.; Golenser, J.; Domb, A. J.
Synthesis and Biodegradation of Arabinogalactan Sponges Prepared by Reductive
Amination. Biomaterials 2002, 23, 4621–4631.
(58) Van Horn, B. A.; Wooley, K. L. Toward Cross-Linked Degradable Polyester Materials:
Investigations into the Compatibility and Use of Reductive Amination Chemistry for
Cross-Linking. Macromolecules 2007, 40, 1480–1488.
(59) Dolle, R. E.; Gribble, A.; Wilkes, T.; Kruse, L. I.; Eggleston, D.; Saxty, B. A.; Wells, T.
N.; Groot, P. H. Synthesis of Novel Thiol-Containing Citric Acid Analogs. Kinetic
Evaluation of These and Other Potential Active-Site-Directed and Mechanism-Based
Inhibitors of ATP Citrate Lyase. J. Med. Chem. 1995, 38, 537–543.
(60) Jarowicki, K.; Kocieński, P. Protecting Groups.
(61) Cassel, N. S. Non-Woven Fabrics. 2,949,386.
(62) Cattermole, A. Personal communication, 2013.
(63) GreenScreen for Safer Chemicals Chemical Hazard Assessment Procedure; Clean Production
Action, 2013.
(64) GreenScreen for Safer Chemicals Version 1.2 List Definitions; Clean Production Action, 2011.
(65) GreenScreen for Safer Chemicals V1.2 Benchmarks; Clean Production Action, 2011.
(66) Oke, M.; Ching, R. T. Y.; Carter, L. G.; Johnson, K. A.; Liu, H.; McMahon, S. A.;
White, M. F.; Bloch, C.; Botting, C. H.; Walsh, M. A.; et al. Unusual Chromophore
and Cross-Links in Ranasmurfin: A Blue Protein from the Foam Nests of a Tropical
Frog. Angew. Chem. Int. Ed. 2008, 47, 7853–7856.
(67) Cooper, A.; Kennedy, M. W. Biofoams and Natural Protein Surfactants. Biophys. Chem.
2010, 151, 96–104.

28
APPENDICES
Appendix A: 12 Examples of Crosslinking Within Nature*
*These descriptions were prepared by the Biomimicry 3.8 Institute
1. Blue Sea Mussel (Mytilus) – Mussel Foot Protein 1, Fe3+

Overall Composite Material: Mussel byssus thread cuticle
Main Characteristic of Composite: Hard, extensible, resists abrasion, resists biodegradation
Material Being Crosslinked: Mussel foot protein 1 (mfp-1)
Crosslinker/Binder: Catecholato-Fe3+ coordinate complexes
Figure 1. Mussel with numerous byssal threads anchoring it to a rock at low tide.
Introduction
Blue sea mussels live in near-shore coastal waters and feed on suspended plankton. Given the constant
onslaught of waves and current in their turbulent environment, these mollusks produce numerous threads that
securely anchor them to large rocks and other stable objects. To produce these threads, the mollusk foot
secretes a fiber with a collagen-based core and coats it in a 2-5 micrometer protective cuticle made
predominately of the muscle foot protein (mfp)-1. In some cases, the thread as a whole is able to withstand
>100% extension and return nearly to its original strength. The cuticle can be up to 5 times harder than the
fibrous core yet retains nearly the same breaking strain.
Crosslinking Activity
The architecture of the cuticle is one of large granules dispersed in a matrix (fig 2). Both the granules and the
surrounding matrix are rich in inorganic ions, notably iron III (Fe3+), and mfp-1 proteins characterized by the
amino acid 3,4-dihydroxyphenylalanine (dopa); dopa molecules are decorated with catechol side rings. The
dopa catechol rings from two or more mfp-1s form a coordinate bond crosslink when each hydroxyl oxygen
on the catechol ring donates a nonbonding electron pair to an iron (III) atom (fig 3). This dipolar type of
bonding is much more forgiving of being hyperextended and reformed than a traditional covalent bond in
which both atoms in the bond donate electron pairs. Though not even half as strong as ordinary covalent
bonds, the coordinate bonds that link multiple mfp-1 molecules to one Fe3+ atom can be broken and reversibly
repaired hundreds of times. The matrix is composed of loosely packed crosslinked mfp-1 proteins sparsely
scattered throughout; in contrast, the granules contain dense clusters of the crosslinked mfp-1s.

29
Figure 2: TEM micrograph of byssal cuticle granules. Dense cross-linking in the granules provides hardness,
whereas the less cross-linked matrix provides extensibility. (Harrington 2010)
Figure 3: Catechol-iron complex
At rest, the mfp-1 proteins exhibit a random coil structure with few short bent helices throughout. These
random coils and the dopa-iron (III) complexes complement each other’s performance characteristics to
achieve impressive qualities.
During strain up to 30%, the randomly coiled mfp-1s of the cuticle (both the matrix and granules) unravel and
stretch out taught. Past 30% strain, stress is absorbed by the sacrificial breaking of catecholato-Fe3+ complexes.
Due to the much higher density of the iron complexes in the granules, they are able to retain bond integrity
even when strain has started to create microtears in the more iron-poor matrix. When the strain is relaxed, the
granules have maintained their structural strength. Microtears are channeled through gaps between the granules
and don’t threaten the cuticle as a whole.
This model of sequential energy absorption and sacrificial yet self-repairing breakages explains the incredible
resilience of the mussel byssus cuticle. This evolved strategy has helped propel mussels of the Mytilus genus into
highly successful organisms capable of colonizing environments in even the toughest environments throughout
earth’s oceans.

30
Figure 4: SEM micrograph of byssal thread torn by extreme strain (Harrington 2010)
References
Harrington, Matthew J., Admir Masic, Niels Holten-Andersen, J. Herbert Waite, and Peter Fratzl. 2010. “Iron-
Clad Fibers: A Metal-Based Biological Strategy for Hard Flexible Coatings.” Science 328 (5975) (April 9): 216–
220. doi:10.1126/science.1181044.

31
2. Chinese Soft-Shelled Turtle (P. sinensis) - Pelovaterin

Overall Composite Material: Turtle eggshell
Main Characteristic of Composite: Soft, pliable, macro-porous, antimicrobial
Materials Being Crosslinked: Calcium carbonate crystals (aragonite and vaterite)
Crosslinker/Binder: Pelovaterin protein
Figure 5: Chinese Soft-Shelled Turtle
Introduction
Chinese soft-shelled turtles (Pelodiscus sinensis) lay their eggs in the loose soil and sandy shores of many common
bodies of water. Unlike bird eggs with their hard shells of dense calcium carbonate (CaCO3) crystals, P. sinensis
eggs are soft and more sparsely mineralized; they also lack a cuticular layer and so depend entirely on the
mineral shell and thin underlying membrane to carry out all of the necessary functions of an eggshell. A small
protein called pelovaterin serves both to mediate the nucleation of the carbonate mineral and to regulate the
form of the crystal structure. It also interacts with certain bacterial cell membranes in an antimicrobial capacity.
Pelovaterin is a 42 amino acid peptide with a unique sequence not found in any other organism. Three internal
disulfide bridges and hydrogen bonding between three anti-I : K:EE>E LA>>M LMA: MI KH=N<>: ?HE=>= ! LMKN<M
NK>
<EHL>ER K>L>F ; EBG@ : G: GM
BF B<KH; B:EI >I M
B=> <: EE>= ANF : G -defensin 3 (fig 2). Although these two compounds
have a similar 3D shape, a key difference is pelovaterin's extremely hydrophobic core compared to defensin’s
hydrophilic core.
Figure 6. Structure similarities to

-defensin 3 (HBD3 at left), and
pelovaterin (right).
(Lakshminarayanan, 2008)
Entropic driving forces cause dozens to hundreds of individual pelovaterin molecules to coalesce into micelles
(detergent-like spherical structures with a hydrophilic exterior and an oily, hydrophobic interior). The small
hydrophilic N-terminal tails of the proteins extend out into the solution while the larger hydrophobic regions
collect together into a sphere.

32
The two aspartic acid residues at the N-terminus permit the nucleation of calcium carbonate crystals and
provide mineral binding potential owing to their flexible location. The micelle provides a much more stable
nucleation site for calcium carbonate crystals than free-floating individual proteins could. As the concentration
of pelovaterin in solution increases, micelles grow and the number of nucleation sites increases.
EHL>ER I : <D>= F M AB<D : <B<NE:K <: E<BNF <: K; HG: M

> <KRLM
: E G>>=E>L K:=B: M
> ?KHF >: <A GN<E>: M
BHG LBM>
Pelovaterin regulates the phase of the nascent carbonate minerals to produce either aragonite or vaterite
crystals as opposed to calcite, the form found in avian egg shells. This is thought to be the result of
thermodynamic stability inversions mediated by the hydrophobic-hydrophillic interface of the micellar surface.
When crystals from many neighboring micelles meet, they form a mineral film that becomes the eggshell (fig.
3). While the pelovaterin peptide is the major intracrystalline component in the fully-formed eggshell, it's not
clear if there is an actual chemical bond, even a weak one, linking the organic and mineral faces or if the
peptide stays in place because it's now wedged between the crystals.
Figure 7. Ultrastructure of turtle eggshell. (A) Cross section of the whole turtle eggshell. The white arrow
indicates the shell membrane. (B) Cross section of the bleach-treated eggshell. Note that the membrane is
completely removed after bleach treatment. (C) Magnified image of central portion of crystalline layer. (D)
Crystalline layer that is attached to the shell membrane. Note the acicular needles of aragonite radiate from the
center. (Lakshminarayanan, Fig 1, 2008)
Antimicrobial activity: Defensin is a cationic membrane-active peptide that uses electrostatic charge
differential to interact with, and eventually destroy, the integrity of lipid bilayer membranes of many microbes
in a non-specific manner. Pelovaterin interacts with membranes as well, but, having a net charge of only -1,
does so for very different reasons; its core is extremely hydrophobic (in stark contrast to defensin’s hydrophilic
core), which enables it to embed in the hydrophobic microenvironment of lipid bilayers rather than on the
surface. Although pelovaterin shows nowhere near the antimicrobial spectrum breadth of defensin, it’s a
fascinating quirk of evolution that it adopted such a similar 3D structure and function to defensin despite using
a totally different mechanism. Pelovaterin’s interactions extend well beyond lipid membranes, however.
References
Lakshminarayanan, Rajamani et al. “Structure, Self- LL>F ; ER : G= ! N: E/ HE> H? : -Defensin-like Peptide

from the Chinese Soft-Shelled Turtle Eggshell Matrix.” Journal of the American Chemical Society 130.14 (2008):
4660–4668.

33
3. Chiton Tooth – Mg2+ and Na+
Overall Composite Material: Chiton tooth
Main Characteristic of Composite: Hard, crack resistant, wear resistant
Materials Being Crosslinked: Semi-crystalline α-chitin polysaccharide (poly-ß-1,4-N-acetylglucosamine),

aspartate-rich proteins, and magnetite (Fe3O4) crystals
Crosslinker/Binder: Mg2+ and Na+ ions
Figure 8: Chiton
Introduction
Chitons are small marine mollusks that graze on algae and other microbes present on hard surfaces in their
environment. They scrape these surfaces with a mouth organ called a radula (rasping tongue), which contains
two rows of extremely hard teeth (fig. 2). The radula acts as a conveyer belt on which teeth are present in
progressively more developed forms. Fully mature teeth at the tip of the radula are used and lose integrity until
they are shed off. At the other end of the conveyer belt, an organic scaffold mediates the biomineralization of
new teeth. This progressive system allows the chiton to wield ‘fresh’ teeth at all times.
Figure 9: Chiton radula
The organic scaffold is made of α-chitin polysaccharide (poly-ß-1,4-N-acetylglucosamine) strands arranged in

fiber bundles of 5-10 nm diameter. Chitin is present in both crystalline and amorphous states. The core of the

34
fiber is primarily composed of crystalline chitin while amorphous chitin loops extend from its surface to give
the structure a rough appearance.
Figure 10: Chitin fiber (Gordon, 2011)
Bound to the surface and embedded in the core of the chitin fiber are numerous proteins. These probably
feature chitin-binding domains that hydrogen bond with the polysaccharide. Most appear to be aspartate-rich
proteins that have ion binding sites specific primarily for either Mg2+ or Na+ (metal binding sites may be of the
DEAD box motif variety). Aspartate-rich proteins that selectively bind the different ions segregate to different
regions of the fiber; sometimes only a few micrometers separate the different clusters. This probably relates to
how the different protein-ion complexes regulate biomineralization and accounts for an additional level of the
structural hierarchy.
Mg2+ and Na+ act as the main interface between the organic scaffold and the magnetite mineral. Carboxyl,
amine, and hydroxyl side chains of the aspartate-rich proteins also probably form electrostatic bonds with the
iron cations and oxygen anions of the mineral. In addition, free Mg2+ probably increases the stability of the
magnetite-precursor ferrihydrite in solution and prevents biomineralization. Binding the Mg2+ would destabilize
the soluble ferrihydrite and cause biomineralization of the magnetite on the scaffold.
After complete biomineralization, the magnetite crystal lattice completely occludes the organic scaffold. The
rough interface gradient between the scaffold and the mineral probably increases adhesion, toughness, and the
number of sacrificial bonds between the structures.
References
Gordon, Lyle M., and Derk Joester. “Nanoscale Chemical Tomography of Buried Organic-inorganic Interfaces
in the Chiton Tooth.” Nature 469.7329 (2011): 194–197.
Tao, Jinhui et al. “Magnesium-aspartate-based Crystallization Switch Inspired from Shell Molt of Crustacean.”
Proceedings of the National Academy of Sciences (2009)

35
4. Deep Sea Sponge (Euplectella sp.) – Silicatein
Overall Composite Material: Sea sponge skeleton
Main Characteristic of Composite: Hard, flexible, crack resistant fibers in rigid construction
Materials Being Crosslinked: Amorphous hydrated silica
Crosslinker/Binder: silicatein, silintaphin, and galectin proteins
Figure 11: Euplectella sp. Deep sea sponge
Introduction
Euplectella sp. sponges live on the deep sea floor. Exposed to constant currents, they’ve evolved the capacity to
produce extremely rigid skeletons to maintain their structural integrity. Despite belonging to the Porifera
phylum, often regarded as the simplest animals, these sponges produce a remarkable biosilica (i.e., glass)
composite skeleton that features complementary, hierarchically layered organization (fig. 1).
Glass, though very hard, is a poor structural material due to its brittleness. Applied stresses collect at any point
of surface defect and the material fails catastrophically once a crack forms. The bottom-level architecture of the
skeleton, like bullet-proof glass, consists of concentric sheets of hydrated silica nanospheres (50 to 200 nm in
diameter) glued together by an organic matrix forming a spicule about 50 microns in diameter. This strategy
prevents cracks from propagating deep into the spicula and provides multiple, energy absorbing failure points
before the entire structure yields (fig. 2).
This laminated construction of alternating organic/silica layers underlies the strength of the entire material as a
whole. Successively higher tiers of organization require the fundamental integrity of this most basic building
block. The sponge has therefore evolved extremely efficient mechanisms for forming and gluing together the
biosilica sheets.

36
Figure 12. Hierarchical construction of the Euplectella sp. skeleton. A) silica nanoparticles deposited on organic
axial filament B) alternating layers of organic and silica lamina to form spicule C) many spicula bundled into a
fiber D) horizontal, vertical, and diagonal bundles form checkerboard cage of skeleton E) architecture of node
between bundles F) ridges surrounding anchor structure G) flexible anchor attached to inflexible silica rod
(Aizenberg, 2005)
Figure 13. SEM micrographs of the lamina of the spicula. From left to right, the images depict: an undamaged
spicule; a damaged spicule with organic matrix visible between separated biosilica sheets; a severely damaged
spicule with a crack propagating through successive layers of silica; a severely damaged spicule. (Aizenberg,
2005)
Silica is produced by the enzyme silicatein in sponge cells called sclerocytes. The mechanism depicted in figure
4 shows how the active site histidine and serine side chains, bound by a hydrogen bridge to increase
nucleophilicity, interact with a free silicate ion to yield a trisiloxane ring. This compound is extremely reactive
and readily self-assembles to produce amorphous silica nanoparticles. Silicateins have a tendency to form flat
oligomer complexes comprised of multiple isoforms. These complexes direct the spatial assembly of the
nascent silica nanoparticle into solid pieces of basically uniform structure; the complexes remain attached to the
silica due to the clusters of serine hydroxyls present on silicatein- # B@ NK> =>I B<M LM
A>I KH@ K>LLBHGHf material
through a sclerocyte during production of biosilica.

37
Figure 14. Cartoon of the mechanism by which the active site serine and histidine of silicatein produce a reactive
trisiloxin ring (Wang, 2012)
Figure 15. Cartoon of a sclerocyte absorbing silicate, transporting it into the silicasome vesicles, forming the
basic silica nanoparticle, and transporting the nanoparticle to the extracellular space for incorporation into the
nascent biosilica layer on the spicule. (Wang, 2012)
Proteins released from the sclerocytes called silintaphins deliver calcium ions to proteins called galectins in the
extracellular space. In the presence of Ca2+, galectins aggregate into glue like sheets that are capable of adhering
to silicatein-coated silica nanoparticles (fig. 5). Sequential release of silica nanoparticles and silintaphins allows
the sclerocytes to regulate the formation of the concentric lamina in the growing spicula.
According to certain structural engineering theory, evolution has optimized the architecture of the reinforced
cage structure seen in these Euplectella sponges. Each spicula-bundle node has six connections (2x horizontal,
2x vertical, 2x diagonal); this produces a “checkerboard” reinforcement pattern in which alternating square
sectors are diagonally reinforces instead of every one (8 connections/node) (fig. 6). Six connections is
considered sufficient for a 2-dimensionally reinforced structure to be maximally stable. Eight connections
would confer no additional strength to the skeleton. In the sea, silicate is a scarce compound and this
architecture helps reduce waste without loss of structural integrity.

38
Figure 16. Cartoon of the major actors and components in the lamina forming process (Wang, 2012)
Figure 17. SEM micrograph of reinforced biosilica skeleton
References
Aizenberg, Joanna et al. “Skeleton of Euplectella Sp.: Structural Hierarchy from the Nanoscale to the
Macroscale.” Science 309.5732 (2005): 275–278.
Müller, Werner E. G. et al. “Self-healing, an Intrinsic Property of Biomineralization Processes.” IUBMB Life
65.5 (2013): 382–396.
Wang, Xiaohong, Ute Schloßmacher, Klaus Peter Jochum, et al. “Silica-protein Composite Layers of the Giant
Basal Spicules from Monorhaphis: Basis for Their Mechanical Stability.” Pure and Applied Chemistry 82.1 (2010):
175–192.
Wang, Xiaohong, Ute Schloßmacher, Matthias Wiens, et al. “Silicateins, Silicatein Interactors and Cellular
Interplay in Sponge Skeletogenesis: Formation of Glass Fiber-like Spicules.” FEBS Journal 279.10 (2012): 1721–
1736.

39
5. Flax Stem – Arabinogalactan protein (AGP)
Overall Composite Material: Flax stem fiber
Main Characteristic of Composite: Tough, resists compression and tension, water resistant, and
antimicrobial
Materials Being Crosslinked: Cell wall polysaccharides (cellulose, hemicelluloses, and pectic galactans)
Crosslinker/Binder: Arabinogalactan protein AGP
Figure 18: Flax plant (left); flax fibers (right)
Introduction
Plant fibers are highly specialized cells that serve primarily mechanical functions. To fulfill that function, fiber
cells tend to be relatively long (up to several centimeters) and have relatively thick cell walls. Fiber cells can be
categorized according to the polyphenolic (i.e. lignin) content of their cell walls. Wood fibers are lignin-rich
while crop stems such as flax, hemp, ramie, and nettle are considered non-lignified fibers. The elementary
fibers (single cells) of flax and hemp consist of layers with a central lumen (fig 2). In flax stems, fiber cells
present in the phloem layer contribute to their extreme toughness.
Figure 19: Cross section of a flax stem (1. Pith, 2. Protoxylem, 3. Xylem I, 4. Phloem I, 5. Sclerenchyma (bast
fibre), 6. Cortex, and 7. Epidermis.) (source: http://en.wikipedia.org/wiki/Phloem)

40
Flax phloem fibers owe their strength primarily to cellulose, a glucose polymer that crystalizes into bundles
(cellulose microfibrils) via dense hydrogen-bonding between adjacent strands. These cellulose microfibrils are
largely arranged axially with the fiber as a whole and are crosslinked into a network by other polysaccharides,
mostly of the pectic and hemicellulosic varieties (fig. 3). Chief among the crosslinking pectins in flax fibers are
pectic galactans which consist of an RG-1 (Rhamnogalacturonan I) -like pectin backbone with many galactan
branches (fig. 4)
Figure 20. Typical plant cell wall with cellulose microfibrils crosslinked by hemicelluloses and pectins. Although
this illustration shows cellulose fibers in alternating orientations, flax stem cellulose fibers run along the
longitudinal axis of the plant. (middle lamellae attach cells to each other) (Pingry School)
Figure 21: Typical pectic galactan from flax fibers (Rha= Rhamnogalacturonan I, Gal=galactans) (Gorshkova,
2006)
Pectic galactan forms a network of non-covalent bonds with cellulose microfibrils via hydrogens bonds from
its numerours hydroxyl groups. Altogether, the carbohydrate component of flax fiber is sufficient to impart the
material with most of its structural properties. However, there are certain glycoproteins also implicated as
crosslinking agents, particularly a group of proteins with a promiscuous binding repertoire called
arabinogalactan proteins (AGPs).
AGPs are membrane surface proteins bound to the lipid bilayer. During cell wall thickening phases, AGPs are
transported to key positions on the cell membrane (fig. 5) in concert with cellulose synthase enzymes to
reinforce the cell wall. AGP tightly associates with cellulose microfibrils and pectins. It tends to associate with

41
pectins rich in galacturonic acid (like pectic galactans) and hemicelluloses rich in glucose. Its large, branched
glycans probably forms crosslinks with these carbohydrates by hydrogen-bonding between the AGP glycosyl
chain hydroxyl groups and those of the structural fibers. It has also been hypothesized that the arabinogalactan
side-chains form peroxide-mediated covalent bonds with structural polysaccharides.
Figure 22. AGP bound to lipid bilayer by GPI anchor. Miscellaneous common glycosylation chains shown O-
linked to AGP hydroxylproline residues (Nguema-Ona, 2013)
The lipid anchor is a glycosylphosphatidylinositol (GPI) in which the C-terminal is linked by a

phosphoethanolamine to a short oligosacharride that is finally linked to an inositylphosphoceramide lipid
(GPI). This lipid tail embeds in the lipid cell membrane and tethers the entire glycoprotein to the surface. Post-
translational transformation of proline residues on AGP into hydroxyproline allows for O-glycosylation of the
protein. The polysaccharide chains vary greatly in size and composition among different AGPs even within the
same tissue; however, they mostly consist of a galactan backbone with various arabinose and galactose
branches.
Figure 23: Stained AGP-containing cells in 50-day old flax fiber phloem wall (Gorshkova, 2006)
Though AGPs contribute to less than 0.5% of the total flax fiber mass, their selective high-level release at
specific phases of cell wall growth (fig. 6) suggests an important function. Besides forming crosslinks between
structurally significant carbohydrates, their potential ability to link the cell wall to the lipid membrane has been
hypothesized as possibly significant.

42
References
Arabinogalactan-Proteins Projects  : Plant Cell Biology Research Centre  : The University of Melbourne. Web. 14
Aug. 2013.
Girault, Raynald et al. “Identification and Partial Characterization of Proteins and Proteoglycans Encrusting the
Secondary Cell Walls of Flax Fibres.” Planta 211.2 (2000): 256–264. link.springer.com.libproxy.cc.stonybrook.edu.
Web. 13 Aug. 2013.
Gorshkova, T. A., L. V. Kozlova, and P. V. Mikshina. “Spatial Structure of Plant Cell Wall Polysaccharides and
Its Functional Significance.” Biochemistry (Moscow) 78.7 (2013): 836–853. link.springer.com.libproxy.cc.stonybrook.edu.
Web. 8 Aug. 2013.
Gorshkova, Tatyana, and Claudine Morvan. “Secondary Cell-wall Assembly in Flax Phloem Fibres: Role of
Galactans.” Planta 223.2 (2006): 149–158. link.springer.com.libproxy.cc.stonybrook.edu. Web. 8 Aug. 2013.
Gurjanov, Oleg P. et al. “Polysaccharides, Tightly Bound to Cellulose in Cell Wall of Flax Bast Fibre: Isolation
and Identification.” Carbohydrate Polymers 72.4 (2008): 719–729. ScienceDirect. Web. 8 Aug. 2013.
Mikshina, Polina V. et al. “Structural Details of Pectic Galactan from the Secondary Cell Walls of Flax (Linum
Usitatissimum L.) Phloem Fibres.” Carbohydrate Polymers 87.1 (2012): 853–861. ScienceDirect. Web. 8 Aug. 2013.
Nguema-Ona, Eric et al. “Arabinogalactan Proteins in Root–microbe Interactions.” Trends in Plant Science 18.8
(2013): 440–449. ScienceDirect. Web. 14 Aug. 2013.
Qian, Ke-Ying et al. “Structural Elucidation of Rhamnogalacturonans from Flaxseed Hulls.” Carbohydrate
Research 362 (2012): 47–55. ScienceDirect. Web. 8 Aug. 2013.

43
6. Human Cells – Filamin B
Overall Composite Material: Human cytoskeleton
Main Characteristic of Composite: Rigid, flexible, dynamic
Materials Being Crosslinked: Filamentous actin (F-actin)
Crosslinker/Binder: Filamin B homodimers
Figure 24: Actin filaments (stained red), and the nucleus (stained yellow). (source:
http://itg.beckman.illinois.edu/technology_development/web_atlas/structures/nucleus_actin/)
Introduction
Rather than being amorphous blobs, cells build a cytoskeleton that allows them to tightly regulate their 3-
dimensional shape and rigidity depending on their demands and functions (fig 1). Underlying this regulation is a
matrix of filaments made of long protein polymers known as filamentous actin (F-actin) (fig. 2). This inter-
linking greatly increases rigidity of the cell, supports the cell membrane, regulates key functional molecules, and
in some cases, acts as a trackway for cargo transport by motor proteins. The filaments are densely crosslinked
by actin binding proteins which direct the formation of different kinds of F-actin composites. A particularly
interesting example of an actin binding protein is filamin B.
Figure 25. Actin filament made of actin subunits (U.S. National Library of Medicine)

44
Two filamin monomers dimerize in an antiparallel fashion

(i.e., one laying head to tail, the other tail to head) to produce
a long and highly flexible protein link (fig. 3). Separated from
each other by a long chain (ROD domain) of 24 repeated
immunoglobin-like domains (~95 residues each) are the two
N-terminal actin binding domains (ABD). Electrostatic
forces between C-terminal 24th repeats of adjacent filamins
induce dimerization, which is critical to the functionality of
the molecule. A single filamin dimer is able to crosslink F-
actin strands up to 100 nanometers away.
The human filamin B ABD is a 242 residue segment

containing two calponin-homology domains designated CH1
and CH2 (fig. 4). Both appear to be critical to the
functionality of the linker. In consensus with the ABDs of
many other actin binding proteins, filamin B’s ABD features
three actin binding sites (ABS); the first two are located on
CH1 while the third is located on CH2. Each calponin-
homology domain constitutes four alpha-helices.
Interestingly, the ABSs are found on non-aligned surfaces of
the ABD. It is likely that the domain folds on contact with
actin to induce binding. The residues involved in the
crosslink have yet to be precisely defined but probably
engage in non-specific electrostatic interactions with the
! negatively charged actin; a wide variety of binding
conformations have been observed experimentally.
Figure 26. Filamin dimer; note that this
illustration depicts a filamin with six
immunoglogin-like domain repeats instead Perhaps most fascinating about filamin is the great variety of
of the 24 in human filamin B (Popowicz, cytoskeletal structures in can induce depending on its
2004) concentration relative to F-actin (fig. 5). At lower
concentration, filamin forms sparse crosslinks between F-
actin to make a weakly crosslinked matrix. Intermediate
concentrations induce formation of bundled F-actin within the matrix. Increasing the concentration further
completely transforms the network into a bundled one. Extremely high concentrations cause the bundles to
cluster. The precise mechanism underlying these transitions in not yet understood but is obviously related to
the increased crosslink density between actin filaments. The diverse array of cytoskeletal conformations
directed by filamin B, just one of many dozen different actin binding proteins each with their own
characteristics, enables to cell to produce a range of important structures.
Figure 27. Human filamin B

actin binding domain featuring
two calponin-homology
domains (CH1 and CH2) and
three actin binding sites: ABS1,
ABS2, and ABS3. (Sawyer, 2009)

45
Figure 28. Actin structures formed at different relative filamin concentrations (Lieleg, 2010)
References
Lebart, M. C. et al. “Characterization of the Actin Binding Site on Smooth Muscle Filamin.” Journal of Biological
Chemistry 269.6 (1994): 4279–4284. Print.
Lieleg, Oliver, Mireille M. A. E. Claessens, and Andreas R. Bausch. “Structure and Dynamics of Cross-linked
Actin Networks.” Soft Matter 6.2 (2010): 218. CrossRef. Web. 19 Aug. 2013.
Popowicz, Grzegorz M. et al. “Molecular Structure of the Rod Domain of Dictyostelium Filamin.” Journal of
Molecular Biology 342.5 (2004): 1637–1646. ScienceDirect. Web. 20 Aug. 2013.
Sawyer, Gregory M. et al. “Disease-associated Substitutions in the Filamin B Actin Binding Domain Confer
Enhanced Actin Binding Affinity in the Absence of Major Structural Disturbance: Insights from the Crystal
Structures of Filamin B Actin Binding Domains.” Journal of Molecular Biology 390.5 (2009): 1030–1047.

46
7. Human Joint – Fibulin-2
Overall Composite Material: Articular Cartilage
Main Characteristic of Composite: Springy, turgid, compression resistant
Material Being Crosslinked: Aggrecan (chondroitin sulfate proteoglycan 1; primary isoform: 2316 residues,
>2500 kDa; Core Protein <250 kDa + 100-150 glycosaminoglycan chains)
Crosslinker/Binder: Dimeric Fibulin-2 (X-shaped dimer conformation)
Figure 29: Articular cartilage on human elbow synovial joint
Introduction
Networks consisting of polysaccharide (hyaluronan) and protein (aggrecan) are critical components of articular
cartilage; they give the material much of its compressive load resistance. Entrapped within the collagen fibril
complex, aggrecan-hyaluronan networks compress to about one seventh of their volume generating high
turgor, elasticity, and high compression resistance. The substantial negative charge of the aggrecan-hyaluronan
complex is responsible for high water uptake of articular cartilage which can be released upon pressure, serving
as a joint lubricant during motion.
Aggrecan (fig 2) plays an essential role in the formation of the aggrecan-hyaluronan network. Aggrecan proteins
are crosslinked to each other by the fibulin-2 dimer forming an aggregan network which, in turn, binds directly
to hyaluronan (fig 3).
Figure 30: Modular structure of the proteoglycan, aggrecan. The central glycosaminoglycan (GAG) shaft consists
of chondroitin/dermatan sulfate (CS), and keratansulfate (KS). To the left of the GAG shaft is the N-terminal
domain that binds to hyaluronan (HA). The HA binding site consists of an immunoglobulin-like repeat followed
by two proteoglycan tandem repeats. To the right of the central shaft is the C-terminal that binds to Fibulin-2
dimer linking aggrecan molecules together. (sources: von der Mark 2013, and Aspberg 2012)

47
Figure 31: Hyaluronan (formerly called hyaluronic acid, HA), an anionic, non-sulfated glycosaminoglycan
The molecular basis for this compressive load resistance functionality of the aggrecan-hyaluronan network is
the extremely high fixed charge density of the many glycosaminoglycans (mostly chondroitin sulfate) that
branch out from the long central shaft of aggrecan (fig 2). These create high osmotic swelling pressures for the
cartilage as a whole. Flanking either side of aggrecan’s central glycosaminoglycan (GAG) shaft are two globular
domains responsible for binding the proteoglycans to other matrix elements to direct the supramolecular
organization of the tissue. The N-terminal globular domain is responsible for binding to the hyaluronan
network of the cartilage while the C-terminal globular domain consists of several binding motifs including a c-
type lectin-like domain (CLD); several ligands of this C-terminal have been documented including fibulin-2.
Fibulin-2
Fibulin-2 (fig 5) binds numerous extracellular matrix proteins including several lectican proteoglycans like
aggrecan. The binding is mediated through the CLD domain on aggrecan's C-terminal. This binding has been
shown to be specific since fibulin-2 will not bind the related CLD appearing on other lectican proteoglycans
such as neurocan. Of unknown importance is the observation that only the X-shaped dimer of fibulin-2 (fig 6)
is involved in binding with aggrecan.
Fibulin-2 Structure
Figure 32: Fibulin-2 monomer
Fibulin-2 consists of 4 domains. The 400 residue N-terminal domain (N) is comprised of a 150 residue cysteine
(Cys)-rich region (Na) at its own N-terminal followed by an entirely cysteine-free region (Nb). Following the N
domain is domain 1 (I) which consists of three tandem anaphylatoxin-like (AT) modules. Each AT module is
~40 residues in length and contains 4-6 cysteine residues which form 2-3 disulphide bridges amongst
themselves which stabilize the compact alpha-helical structure of the modules. The long central shaft domain
(II) consists of 11 tandem EGF-like modules (10 of which are of calcium-binding variety; cbEGF). These
EGF-like modules are also ~40 residues in length and form 3 internal disulphide bonds to stabilize their
structure. Finally, the C-terminal domain (III) is ~140 residues in length and contains only 2 additional
cysteines. It is of the eponymous fibulin-type carboxyl terminus (FC) variety of domains.
Fibuliun-2 forms disulphide-linked dimers between un-bridged cysteines in the central AT module of domain I.
Mutations which remove/replace that active Cys residue prevent covalent dimerization but do not prevent
dimerization completely; there is affinity between domain II and domain N of different fibulin-2 molecules.
The covalent dimerization involves anti-parallel (i.e., opposite orientation) fibulin-2 monomers linked together

48
by their domain Is. The dimer stretches on in both directions from this central junction for a total length of 70-
80nm. The non-covalent N-II interactions can form or break and give the dimer one of 3 structures: rod, Y,
and X.
Figure 33: Fibulin 2 dimer (Timpl 2003)
Calcium binding in domain II of Fibrulin-2 is thought to impart this crosslink with rigidity and robustness.
The cbEGF-like modules that make up the central shaft of fibulin-2 are probably permanently saturated with
calcium due to their binding affinity and the physiological extracellular calcium concentration of ~1 mM.
Calcium is ligated by several regions of cbEGF including a D-X-D/N-E region prior to the first cysteine
residue (where X is any amino acid), a loop created between Cys2 and Cys4, and in the extremely hydrophobic
intermodular contact regions. This binding within the intermodular junctions is thought to impart the long
shaft (domain II) with it high rigidity.
References
Alcendor, Donald J. et al. “KSHV Regulation of Fibulin-2 in Kaposi’s Sarcoma: Implications for
Tumorigenesis.” The American Journal of Pathology 179.3 (2011): 1443–1454. ScienceDirect. Web. 29 July 2013.
Aspberg A. 2012. The different roles of aggrecan interaction domains. Journal of Histochemistry &
Cytochemistry;60(12):987-996.
Baird, Brandi N. et al. “Fibulin-2 Is a Driver of Malignant Progression in Lung Adenocarcinoma.” PLoS ONE
8.6 (2013): e67054. PLoS ONE. Web. 29 July 2013.
Fujimoto, Norihiro et al. “Extracellular Matrix Protein 1 Interacts with the Domain III of fibulin-1C and 1D
Variants through Its Central Tandem Repeat 2.” Biochemical and Biophysical Research Communications 333.4 (2005):
1327–1333. ScienceDirect. Web. 29 July 2013.
Law, E. W. L. et al. “Anti-angiogenic and Tumor-suppressive Roles of Candidate Tumor-suppressor Gene,

Fibulin-2, in Nasopharyngeal Carcinoma.” Oncogene 31.6 (2012): 728–738. www.nature.com.libproxy.cc.stonybrook.edu.
Web. 29 July 2013.
Obaya, Alvaro J. et al. “The Dual Role of Fibulins in Tumorigenesis.” Cancer Letters 325.2 (2012): 132–138.
ScienceDirect. Web. 29 July 2013.
Olin, Anders I. et al. “The Proteoglycans Aggrecan and Versican Form Networks with Fibulin-2 through Their
Lectin Domain Binding.” Journal of Biological Chemistry 276.2 (2001): 1253–1261.
www.jbc.org.libproxy.cc.stonybrook.edu. Web. 29 July 2013.
Segade, Fernando. “Molecular Evolution of the Fibulins: Implications on the Functionality of the Elastic
Fibulins.” Gene 464.1–2 (2010): 17–31. ScienceDirect. Web. 29 July 2013.

49
Stattin, Eva-Lena et al. “A Missense Mutation in the Aggrecan C-type Lectin Domain Disrupts Extracellular
Matrix Interactions and Causes Dominant Familial Osteochondritis Dissecans.” The American Journal of Human
Genetics 86.2 (2010): 126–137. ScienceDirect. Web. 29 July 2013.
Timpl, Rupert et al. “Fibulins: a Versatile Family of Extracellular Matrix Proteins.” Nature Reviews Molecular Cell
Biology 4.6 (2003): 479–489. www.nature.com.libproxy.cc.stonybrook.edu. Web. 29 July 2013.
Vega, S. de, T. Iwamoto, and Y. Yamada. “Fibulins: Multiple Roles in Matrix Structures and Tissue Functions.”
Cellular and Molecular Life Sciences 66.11-12 (2009): 1890–1902. link.springer.com.libproxy.cc.stonybrook.edu. Web. 29
July 2013.
Von der Mark, Klaus, and Jung Park. “Engineering Biocompatible Implant Surfaces: Part II: Cellular
Recognition of Biomaterial Surfaces: Lessons from Cell–matrix Interactions.” Progress in Materials Science 58.3
(2013): 327–381. ScienceDirect. Web. 29 July 2013.
Yanagisawa, Hiromi, and Elaine C. Davis. “Unraveling the Mechanism of Elastic Fiber Assembly: The Roles of
Short Fibulins.” The International Journal of Biochemistry & Cell Biology 42.7 (2010): 1084–1093. ScienceDirect. Web.
29 July 2013.
Yi, Chun-Hui et al. “Loss of Fibulin-2 Expression Is Associated with Breast Cancer Progression.” The American
Journal of Pathology 170.5 (2007): 1535–1545. ScienceDirect. Web. 29 July 2013.

50
8. Malaysian Tree Frog (P. leucomystax) – bis(LTQ), Zn2+ complex
Overall Composite Material: Foam nest
Main Characteristic of Composite: Viscous, strong, long-lasting
Materials Being Crosslinked: Ranasmurfin protein
Crosslinker/Binder: bis(lysine tyrosyl quinone) (N-linked indophenol bond), Zn2+ complex
Figure 34: A mating pair of Malaysian tree frogs and their foam nest
Introduction
The common Malaysian tree frog, Polypedates leucomystax, unlike other foam-nest producing frogs, makes its nest
on dry surfaces. These nests are positioned so that emerging tadpoles drop into a body of water below. In
further contrast with other foam-nest frogs, P. leucomystax foam depends on its viscosity for stability rather than
the presence of biosurfactancts (i.e. detergents).
One of the key features of foam nest composite is the crosslink observed when two ranasmurfin protein
molecules form a homodimer (fig 2). Ranasmurfin monomers (113 residues) exhibit a novel tertiary structure
comprised of several alpha-helices stabilized by three internal disulfide bonds.
Figure 35. 3D structure of ranasmurfin dimer. Monomers are color-coded cyan and salmon. LTQ and bis(LTQ)
groups depicted as yellow stick structures. N-terminal in blue, C-terminal in red (Oke et al., 2008)

51
The protein features an unusual post-translational modification of lysine-31 and tyrosine-2, dubbed the lysine
tyrosyl quinone (LTQ) linkage (fig 3), in which the tyrosine-2 side-chain phenol ring is transformed into a
catechol ring (i.e. dopa) and forms an ortho bond with a side-chain nitrogen from lysine-31. The dopa hydroxyl
groups of the crosslink ring form a bifurcated hydrogen bond with the hydroxyl side-chain of serine-5 and a
water molecule. This unusual internal protein link probably serves to impart strength and stability to the
dimerizing crosslink which incorporates the adjacent lysine-30.
Figure 36. Diagram of the LTQ bond (left) and hydrogen bonds (dashed-lines) formed between the catchol ring
hydroxyls, the serine-5 hydroxyl, and water (right) (Oke et al., 2008)
Lysine-30 forms a post-translationally modified bond with tyrosine-108 not unlike that found in the LTQ bond
(fig 4). However, unlike the LTQ bond, the Lys-30/Tyr-108 ring crosslinks to that of another ranasmurfin
protein via an N-linked indophenol bond (bis-LTQ). This creates an extended aromatic electron system both
between disparate points within each monomer (internal crosslink) and between the monomers themselves
(external crosslink) that imparts strong covalent stabilization to the dimer. Further strengthening the bond is a
single zinc atom coordinated by the oxygen groups of each monomer’s Lys-30/Tyr-108 ring and e-nitrogens
from imidazole rings of each monomer’s His-112. The planes of the rings in the bis-LTQ bond are offset by
about 27° from each other to produce a tetrahedryl chelation structure about the metal.
Figure 37. Diagram of bis-LTQ bond (left) and coordination bonds (dashed-lines) between the indophenol’s two
oxygens and nitrogen, each monomer’s His-112, and a zinc atom (Oke et al., 2008)
This crosslink is thought to be part of a large network in the foam that gives the nest long-term strength and
stability. Tadpoles take about four days to gestate and the foam gives the embryos a safe microenvironment
within which to develop.
References
Cooper, Alan, and Malcolm W. Kennedy. “Biofoams and Natural Protein Surfactants.” Biophysical Chemistry
151.3 (2010): 96–104.
Oke, Muse et al. “Unusual Chromophore and Cross-Links in Ranasmurfin: A Blue Protein from the Foam
Nests of a Tropical Frog.” Angewandte Chemie International Edition 47.41 (2008): 7853–7856.

52
9. Pearl Oyster (P. fucata) – Pif Protein Complex
Overall Composite Material: Pearl oyster shell nacre
Main Characteristic of Composite: Hard, tough, smooth, high luster
Materials Being Crosslinked: Aragonite (CaCO3) crystal platelets and chitinous matrix
Crosslinker/Binder: Pif complexes (Pif 80 and Pif 97 dimerized by disulfide bond)
Figure 38: Pearl oyster nacre. A) SEM micrograph of interface between outer, prismatic calcite layer (top) and
planar nacre layer (bottom). B) Photograph of nacre (Suzukie 2009)
Introduction
Oysters have extremely soft and delicate bodies. To protect themselves, mollusks produce a hard composite
shell capable of withstanding most any insult their environment or predators may deliver. The outer surface,
exposed to the oyster's marine environment, is rough and visually dull. But the inner face of the shell, the
interface between the oyster’s soft body and the shell itself, is composed of an extremely smooth, shiny,
iridescent, and hard composite material called nacre.
While the exterior portion of the shell is composed of tightly stacked elongated prisms of calcite oriented
perpendicular to the shell surface (calcite is the most thermodynamically stable crystal polymorph of CaCO3),
nacre is composed primarily of flat aragonite plates oriented parallel to the surface (aragonite is the crystal
polymorph of CaCO3 stable in marine waters). The mineral crystals are layered with insoluble organic matrices
sandwiched between them; they mediate crystal formation and add resilience to the finished composite.
The building blocks of the nacre, nanoparticles of amorphous CaCO3 as well as Ca2+ and HCO3- are released
by mantle epithelial cells and are mineralized onto a chitinous organic scaffold by a protein complex called Pif.
The Pif complex not only mediates mineralization of the aragonite plates in the precise orientation required for
nacre, it is also responsible for creating a lasting crosslink between the plates and the organic matrix to yield an
extraordinarily tough composite.

53
Figure 39. Pif Complex (Suzuki 2009)
The active Pif complex is composed of two proteins (Pif 97 and 80) which are bound together by disulfide
bonds. Pif 97 (525 residues) has two primary domains, a von Willebrand type A (VWA) domain for mediating
interaction with proteins and a Peritrophin-A type chitin-binding domain. Pif 80 (460 residues) provides the
aragonite-binding functionality to the complex. Pif 80 contains a high proportion of charged amino acids
(28.5% Aspartic Acid, 18.9% Lysine, 10.9% Arginine). It contains 17 repeats of an Asp-Asp-Arg-Lys motifs
scattered throughout its length and an Asp-Glu-Asp cluster at its core. Poly-Asp in particular has been shown
to mediate aragonite formation. Aspartic acid is known to regulate crystal polymorphs in a variety of organisms
independent of the mineral used. For example, aspartate-rich proteins are involved in the formation of calcium
phosphate biominerals of bone and teeth and the amorphous silicate cell walls of diatoms.
During nacre synthesis, the oyster secretes Pif complexes from its mantle epithelium into the extrapallial space.
Pif 97, with its chitin-binding domain, binds to chitin microfibrils and then associates with other proteins to
form a solid lamellar sheet upon the chitin. The Pif 80 part of the complex then concentrates the calcium
carbonate nanoparticles into properly oriented aragonite plates through its many aspartic acid residues. The
oyster will then synthesize an additional chitinous network below the finished crystals and the process will be
repeated to produce the next layer.
References
Kröger, Nils, "The Molecular Basis of Nacre Formation." Science Vol 325. (2009): 1351
Suzuki, Michio et al. “An Acidic Matrix Protein, Pif, Is a Key Macromolecule for Nacre Formation.” Science
325.5946 (2009): 1388–1390.
Zhang, Gangsheng, and Jun Xu. “From Colloidal Nanoparticles to a Single Crystal: New Insights into the
Formation of Nacre’s Aragonite Tablets.” Journal of Structural Biology 182.1 (2013): 36–43.

54
10. Slug (Arion subfuscus) – Ca2+, Fe2+, Mg2+
Overall Composite Material: Defensive glue
Main Characteristic of Composite: Elastic, adhesive, stiff, tough
Materials Being Crosslinked: Slug mucus proteins and polysaccharides
Crosslinker/Binder: Divalent metals such as Ca2+, Fe2+, and Mg2+
Figure 40: The Slug (Arion subfuscus)
Introduction
Slugs are extremely vulnerable creatures. To compensate for their lack of conventional defensive assets like
teeth, claws, or shells, many have evolved extremely effective chemical defenses. When physically disturbed, the
air-breathing terrestrial slug, A. subfuscus, secretes specialized mucus packets that rupture releasing a viscous
fluid. The fluid could flow off the slug's back, but instead sets within less than a minute into a tough, potent
elastic-adhesive gel. This glue acts as a defense against insects and other small predators.
This tough gel contains more than 95% water; it achieves its defensive properties by forming an interconnected
network of proteins and polysaccharides through multiple cross-linking mechanisms.
A. subfuscus adhesive mucus is essentially identical in composition to its relatively non-adhesive locomotive
mucus, a complex mixture of polysaccharides and proteins, but differs in that adhesive mucus contains many
critical metal ions and specific metal-binding proteins. These divalent metal ions contribute to the formation of
two kinds of crosslinks between, and among, proteins and carbohydrates: (1) direct crosslinking with metal
atoms, and (2) an oxidative reaction caused by the metals in which proteins link together via imine bridges.
Direct Crosslinking With Metal Atoms
The carbohydrate portion of the adhesive mucus contains relatively high concentrations of sulphate probably
due to the presence of heparan sulphate-like glycosaminoglycan carbohydrates, and it likely contains many
carboxylated polysaccharides. The protein portion contains a relatively large number of carboxyl and phosphate
groups.

55
Iron, magnesium, and calcium are all strong ligands of these groups. Electrostatic attraction and eventual
coordination of these metals by the many instances of these functional groups on the many tangled proteins
and carbohydrates in the mucus produces strong crosslinking and accounts for the great part of the adhesive
mucus’ rapid gelling. In fact, removal of the metals causes a fifteen-fold reduction in the strength of the glue.
Experiments designed to remove iron from the glue to test its effects initially concluded that eliminating iron
did not have an effect on the material properties. Nonetheless, this is probably due to iron being so tightly
chelated in the matrix that removal was impossible. It may be one of the most significant metal crosslinks. The
primary cross-links controlling gel stiffness appear to be direct cross-links involving calcium. Calcium likely also
plays a roll as an electrostatic crosslink with sulfate (they are the most abundant ions in the gel), but this may be
a premature interpretation of data.
Imine Bridge Formation
Metals ions also indirectly contribute to additional crosslinking activity. Iron and copper facilitate the formation
of carbonyl groups on protein side chains which readily react with the primary amines of lysine side-chains on
other proteins to form imine bonds (fig. 2).
Figure 41 @ -amino group (i.e. lysine side chain) reacting with a carbonyl group to form an imine bond
Interfering with imine bonds causes a 33% reduction in glue stiffness. Interestingly, the imine bond is one of
the few easily reversible covalent bonds. Impeding reversibility actually increased the strength of the glue 40%
versus natural control. However, the ready reversibility of this bond and all other crosslinking bonds described
here may be of functional significance. After all, the adhesive mucus must rapidly set to deter imminent threats
to the slug but long-term adhesion is not needed once the threat is gone and most likely deleterious as dust
particles and other debris would accumulate on the surface of the slug. The bonds may also serve as sacrificial
shock-absorbers that improve extensibility and self-healing of defensive coating after physical stress.
References
Braun, M. et al. “The Relative Contribution of Calcium, Zinc and Oxidation-based Cross-links to the Stiffness
of Arion Subfuscus Glue.” The Journal of Experimental Biology 216.8 (2013): 1475–1483.

56
11. Vineyard Snail (C. virgate) – Epiphragmin hydrophobic effects
Overall Composite Material: Epiphragm
Main Characteristic of Composite: Hard, water resistant, adhesive
Materials Being Crosslinked: Epiphragmin proteins
Crosslinker/Binder: Hydrophobic interactions between coiled-coils of adjacent epiphragmins (forming coiled

coils of epiphragmin homo-oligomers)
Figure 42: Snail shell with epiphragm completely sealing the shell opening to prevent water loss and infection
Introduction
The vineyard snail, Cernuella virgata, synthesizes a specialized adhesive mucus structure called an epiphragm,
which completely seals the shell aperture and prevents water loss during dry weather inactivity. Snails produce
two kinds of foot mucus: trail and adhesive. Trail mucus lubricates the animal’s locomotion while adhesive
mucus helps anchor the animal to substratum during times of inactivity. The two types of mucus are much the
same in terms of chemical composition but differ with respect to the addition of certain glue proteins in the
adhesive form that are thought to stiffen the material.
Unlike many other molluscan glue proteins, which rely on covalent crosslinking between glue proteins and
carbohydrates in the mucus, the primary glue protein identified from C. virgata epiphragm, dubbed
epiphragmin, exhibits adhesive/cohesive interactions due to hydrophobic self-assembly.
Figure 43: Annotated primary structure of epiphragmin. (Li, 2007)
Epiphragmin (752 residues; ~86 kDa) is the predominant protein in the protein-based epiphragm secretion of
C. virgata. The N-terminal region begins with a 22 residue signal peptide (SP) domain followed by a ~145
residue section with a sequence homologous only to proteins of unknown function (fig. 2). The central domain
shows high sequence similarity to the alpha-helical coiled-coil domains of myosin II and AglZ. A coiled-coil
(fig. 3) is a protein structural motif in which multiple alpha-helices closely associate into one “super coil” much
like the individual strands of a rope coil around each other to produce the final material. Coiled-coils usually

57
self-assemble due to an abundance of hydrophobic amino acid side chains exposed on sections of the coil. In
water, the individual helices passively aggregate into coils so that as many of their hydrophobic residues are
concealed as possible. The C-terminal domain of epiphragmin shows high sequence homology with fibrinogen-
related domains (FReDs). It is likely that the epiphragmin FReD binds with other molecules in the mucus,
particularly Ca2+ ions, to further enhance crosslinking activity, though information on that topic is scarce.
Figure 44: Coiled coils
During synthesis of the epiphragm, it is likely that the presence of high concentrations of epiphragmin causes
the rapid formation of huge epiphragmin homo-oligomers that effectively harden the solution. In other words,
the coiled-coil domains of many epiphragmin molecules further aggregate with themselves by the same
hydrophobic exclusion forces that produced the original coiled-coils themselves. Combined with binding to
other molecules mediated by the FReD, the material is sufficiently crosslinked to act as a robust structural
material. Though it is tempting to think of this kind of non-specific hydrophobic interaction as inherently
weak, this very same kind of aggregation is observed in the coiled-coil assemblies of alpha-keratin in
mammalian hair, nails and hooves.
References
Li, Dongmei, and Lloyd D. Graham. “Epiphragmin, the Major Protein of Epiphragm Mucus from the
Vineyard Snail, Cernuella Virgata.” Comparative Biochemistry and Physiology Part B: Biochemistry and Molecular Biology
148.2 (2007): 192–200.
Pawlicki, J. M. et al. “The Effect of Molluscan Glue Proteins on Gel Mechanics.” Journal of Experimental Biology
207.7 (2004): 1127–1135.
Smith, Andrew M. “The Biochemistry and Mechanics of Gastropod Adhesive Gels.” Biological Adhesives. Ed.
Andrew M. Smith & James A. Callow. Berlin, Heidelberg: Springer Berlin Heidelberg, 2006. 167–182.

58
12. Woody plants – Xyloglucan hemicellulose

Overall Composite Material: Wood
Main Characteristic of Composite: Resists compression and tension; serves as a barrier to temperature
fluctuations, water, fire, and microbes.
Materials Being Crosslinked: Cellulose microfibrils
Crosslinker/Binder: Xyloglucan hemicellulose
Figure 45: Wood's functionality at play while living (tree) and dead (house) (NYTimes)
Introduction
Wood – we take it for granted but it's quite a remarkably multifunctional, robust material. It not only supports
tall, free standing structures exposed to compressive and tension forces, it also serves as a barrier to
temperature fluctuations, water, fire, and microbes.
Wood's incredible mechanical properties derive from a complex network of polysaccharides (cellulose,
hemicellulose, and pectin) and proteins in plant cell walls; about 30 percent of secondary walls surrounding
mature cells also contain phenolic compounds, primarily lignin. Cellulose, an enormous homopolymer
composed purely of (1 4) linked D-glucose monomers (Fig. 2), aggregate into extremely sturdy microfibrils
due to the incredible number of hydrogen bonds that develop between the hydroxyl groups of tightly packed
cellulose chains (Fig. 3a). Tangled layers of cellulose microfibrils are thought to account for the vast majority of
wood’s structural strength (Fig. 3b). However, similar polymers known as hemicelluloses form important
crosslinks between the cellulose microfibrils that increase the performance of the material.
Figure 46. ß1-4 glycosidic linkage between the C1 and C4 carbons of two D-glucose molecules
Hemicelluloses, like cellulose, are based on an equatorial (14) linked polysaccharide chain. Unlike cellulose,
hemicellulose backbones are not necessarily exclusively based on glucose (may include mannose, xylose, etc)

59
and/or are branched and exhibit numerous substituted sugar side chains. These differences are thought to
impart various new functionalities (Fig 4).
Xyloglucan (XyG), a hemicellulose present in wood, features a core backbone of (1 4) linked D-glucose in
an unsubstituted form (monomer code: G), and a more prevalent form with an (1 6) linked D-xylose side
chain (monomer code: X). The most common XyG sequences are XXGG and XXXG depending on the
plant’s species. The xylose side chain of the X monomer can be further substituted at O-2 with a -linked D-
galactose (monomer code: L). The galactose side chain of the L monomer can once again be substituted at O-2
with -linked L-fucose (monomer code: F). The extremely versatile side chain repertoire of this hemicellulose
makes it a powerful crosslinking unit in cell walls. A ‘sticky network’ model is currently favored as the best
explanation for how XyG contributes to cellulose microfibril tethering in cell walls.
Figure 47a (left). Adjacent cellulose strands forming numerous hydrogen bonds (dashed-lines).
Figure 3b (right). Cartoon of major cell wall macromolecules.
Figure 48. An example of an XyG polymer segment featuring a (1 4) linked D-glucose backbone with variable
substituted (X) and unsubstituted (G) (1 6) linked D-xylose side chains. Also shown are additional side chain
substitutions (F & L) (Scheller 2010)
In the sticky network, tethering glycans like XyG crosslink distal cellulose microfibrils by two means. The first
involves parts of XyG strands becoming physically trapped within the cores of cellulose microfibrils during
crystallization of the latter. The second type of crosslink involves intense hydrogen-bonding between the side
chains of XyG and exposed hydroxyl groups on the surface of the cellulose microfibrils; XyG adopts a flat
conformation to increase the available area for hydrogen-bonding.
XyG also forms covalent crosslinks with certain pectins (another type of structural polysaccharide present in
cell walls featuring non-equitorial [axial] backbones) like RG-1 in the cell wall (Fig 5). This bond possibly
manifests as an RG-1 galactan branch forming from an XyG type L monomer galactin side-chain. The
crosslink may form as part of initial synthesis in the Golgi apparatus.
References

60
Burton, Rachel A., Michael J. Gidley, and Geoffrey B. Fincher. “Heterogeneity in the Chemistry, Structure and
Function of Plant Cell Walls.” Nature Chemical Biology 6.10 (2010): 724–732.
Cosgrove, Daniel J. “Expansive Growth of Plant Cell Walls.” Plant Physiology and Biochemistry 38.1–2 (2000):
109–124.
Dyson, R.J., L.R. Band, and O.E. Jensen. “A Model of Crosslink Kinetics in the Expanding Plant Cell Wall:
Yield Stress and Enzyme Action.” Journal of Theoretical Biology 307 (2012): 125–136.
Scheller, Henrik Vibe, and Peter Ulvskov. “Hemicelluloses.” Annual Review of Plant Biology 61.1 (2010): 263–289.

61
Appendix B: Interaction of Bonding Themes in One Organism
Ranasmurfin, a blue protein from the foaming nest of the Malaysian tree frog, has a unique,
chromophoric crosslink. It combines several bonding strategies to create a stable,
biocompatible foam environment for developing eggs and embryos:66
Covalent bonding:
3 disulfide bridges stabilize each alpha
helix
Non-covalent bonding:
Hydrogen bonds impart strength and
stability to the dimerizing crosslink
Ancillary-metal bonds:
Tetrahedral chelation structures
around zinc ions stabilize the tertiary
structure of the protein
3D structure of the ranasmurfin protein67

62
Appendix C: Why Not Metals?

The majority of metals that appear in the biological examples we investigated are Group 1
and 2 cations (Na+, Mg+2, Ca+2) or transition metal cations (Fe+2, Fe+3, Zn+2). They serve to
either link proteins or other materials together through binding to a common metal ion, or
to enhance already existing structure by providing additional binding.
The main benefit of metal use is that the addition of a metal salt to a resin or fabric could be
a straightforward way to “turn on” and control crosslinking. We envisioned application of a
resin to fabric in the finishing step, as is done now, and then crosslinking that resin later by
adding the appropriate metal, once the fabric has been cut and sewn into a garment.
There are two main drawbacks to using metals, however, that led us to decide not to pursue
them further. First, coordinate covalent bonds with metals are weak and susceptible to
disruption by EDTA or other water softeners found in detergents. Second, many transition
metals, such as iron, produce complexes with vibrant colors that might not be desirable in
the finished product.
We did not evaluate metal-containing bonds as a primary strategy, although we are still
considering them as a potential supplement to either covalent or non-covalent bonds. Metals
are also used in the biological examples to catalyze other reactions; we are considering them
for this use.

63
Appendix D: Technical Evaluation Framework – Detailed Description
Detailed descriptions of the eight technical performance factors are listed below.
1. Degree of additional research necessary

- Red: There is no precedent in textiles or we anticipate many hurdles
- Yellow: There is some precedent in textiles or we anticipate few hurdles
- Green: Ample research has been performed but optimization may be required
2. Disruption of chemical supply infrastructure (i.e., synthesis of new materials)
- Red: strategy requires special or new manufacture of chemicals
- Yellow: chemicals have limited availability or can be easily and cleanly made in situ
- Green: chemicals are widely used industrially (not necessarily in the textile industry)
3. Disruption of fabric application infrastructure (i.e., equipment and materials used in the
step where resin is applied)
- Red: entirely new technology or techniques required, such as anhydrous conditions,
airless conditions, high pressure gases, or other specialized equipment
- Yellow: longer times required; limited or advanced technology (such as lasers);
modification of existing technology (such as using washing machines but making
sure they’re acid-resistant)
- Green: required infrastructure or techniques are already being used in the textile
industry
4. Disruption of crosslinking step infrastructure (i.e., equipment and materials used to
perform the crosslinking step)
- Red: significant change in equipment; time or space intensive
- Yellow: new reagents or catalysts are required; longer times required; crosslinking
requires solvent
- Green: heat-cured or other minimal curing (such as air-cured)
5. Controllable crosslinking
- Red: difficult to have “delayed control” or difficult to get reaction to go at all
- Yellow: may need special conditions or additional chemicals to control (reversing
preliminary reactions, using protecting groups, or forcing reaction to go with harsh
conditions)
- Green: crosslinking easy to “turn on” by adding catalyst, reagent or heat
6. Resilience of cellulose linkage or crosslinker during manufacturing process
- Red: chemicals are likely to get washed off or deactivated using current practices;
would need modification of other steps
- Yellow: chemicals may get washed off or deactivated, but unknown or hard to
predict
- Green: no foreseeable problems
7. Resilience of cellulose linkage or crosslinker during consumer laundry
- Red: crosslink or linkage to cellulose is likely to be disrupted by water, alkaline pH,
turbulence, chelators, surfactants

64
- Yellow: crosslink or linkage to cellulose may be disrupted but is not likely

8. Other effects on fabric (fibers, “hand”, dyes, other finishes, etc.) during process
- Red: strategy requires acid catalysis (cellulose degradation), bleaching agents (removal
of dye), fiber coatings (may make dye uptake difficult), or excessive use of something
that might change “hand” (chemicals or fibers)
- Yellow: possible need for acid catalysis, bleaching agents, fiber coatings, or excessive
use of something that might change “hand”, but not required

65
Appendix E: Detailed Technical Evaluation Results
Each strategy was evaluated using these eight parameters, incorporating information from
the literature and from textile industry experts (Figure 14). Gray squares in the table indicate
parameters that are not applicable to the specific strategy – for example, “disruption of
fabric application process” only applies to strategies that deal with binding to cellulose
(polymer weaving or coating, polycarboxylic acids). Similarly, “disruption of crosslinking
step” and “controllable crosslinking” only apply to the crosslinking strategies (disulfide
bonds, imine & amine bonds). In some instances where the specific materials or chemicals
chosen will change the ultimate result, we have given multiple ratings.
Poly(Ethylene Terephthalate) Weaving
- Controllable crosslinking and effects on fabric: As PET weaving is already used

in the textile industry and involves no additional chemicals, the disruption is
anticipated to be low. However, we are unsure of how accessible the carbonyl
groups on the PET would be for crosslinking. Additionally, more PET may need
to be blended into the fabric in order to make the crosslinking effective, which
would affect the hand.
Polymer Fiber Coating
- Degree of additional research: While polymers are commonly applied in the fiber
form during sizing, they are also stripped off later in the desizing process. We
propose to leave a polymer on the fibers so that it can later be used for
crosslinking, which to our knowledge is not currently done.
- Disruption of chemical supply: This varies widely depending on the specific
polymer chosen. Some are very common in industry; some may not yet exist.
- Durability during manufacturing process and in consumer laundry: If a polymer
is applied to the fibers during the sizing step for later crosslinking, it will need to
survive desizing, which includes enzymes, hot water, and potentially
concentrated sodium hydroxide (reference?). Depending on the water-solubility
of the polymer, this may be possible, but it would be tricky. However, we feel
that if the polymer survives manufacturing and manages to be crosslinked, it is
more likely to survive consumer laundry.
- Effects on fabric: Applying a polymer coating in the fabric form could inhibit
dye uptake later on, although if the fibers are dyed instead of the fabric, as is the
case with Levi’s jeans, this may be less of an issue.
Polymer Fabric Coating

66
- Degree of additional research: Polymers are commonly applied in the fabric form,
so we anticipate this to be a relatively straightforward strategy to optimize,
although it may depend on the specific polymer.
- Disruption of chemical supply: As stated above, this varies widely depending on
the specific polymer chosen.
- Durability during manufacturing process and in consumer laundry: It is not clear
how well the coating would penetrate the fabric and thus how well it would
“stick” to the fabric. The durability of the polymer in consumer laundry will
depend on this penetration and also on the water-solubility of the polymer.
- Effects on fabric: Applying a polymer coating could change the hand, depending
on the specific polymer and how much is applied.
In Situ Polymerization
- Degree of additional research: In situ polymerization has been demonstrated on

textiles for specialty applications such as conductive fabrics, but has not been
used on this wide a scale.
- Disruption of fabric application process: Of all the fabric treatments, this is the
only one that would likely require new equipment, in order to facilitate the in situ
polymerization. In addition, the in situ polymerization may be slower than
current fabric treatments.
- Durability during manufacturing process and in consumer laundry: Good fiber
penetration has been demonstrated with in situ polymerization, so we anticipate
good durability during manufacturing, but as with the polymer coating, the
durability in consumer laundry will depend on the specific polymer.
- Effects on fabric: As with the polymer coating, the hand could change depending
on the polymer and how much is generated.
Polycarboxylic Acids
- Disruption of chemical supply: As mentioned above, thiolated polycarboxylic acids

are almost unknown, but we propose to add a thiol in situ using readily available
reagents.29 The disruption from crosslinking to the carboxylic acid directly with a
diamine would depend on the diamine.
- Effects on fabric: As mentioned in detail in the previous section, there are numerous
known problems with polycarboxylic acids. The two major effects on the fabric are
loss of tensile strength15 and fabric yellowing.18 However, here we reiterate again that
a “red” ranking can merely represent that we are aware of many of the potential
problems, rather than indicating more problems in a world of perfect information.
Since this is a well-studied technology, many of the pitfalls have been well-elucidated.
However, it still receives a “red” rating here since acids are inherently part of the
very nature of the strategy – thus, they are required.

67
Disulfide Bonds
- Degree of additional research: To our knowledge, disulfide bonds have never been
used for textile crosslinking, although there are examples of disulfide bond
crosslinking in other industrial applications.34,35,46
- Disruption of chemical supply: Thiolated polymers are somewhat developed, with

examples of thiolation of poly(methacrylic acid), poly(acrylic acid),
carboxymethylcellulose, and chitosan.48 Nevertheless, they are not widely available.
Thiolated polycarboxylic acids are almost unknown; we were only able to find one
example of thiolated citric acid.59 However, the proposed in situ addition of the thiol
to the already-linked carboxylic acid would use readily available reagents.29
- Crosslinking disruption and controllability: New reagents and catalysts will be

required for disulfide bond crosslinking. Depending on the reactivity of the thiol, it is
possible that the disulfide linkage will form in air between the time the resin is
applied to the fabric and the time the permanent crease is set in the garment. It is
possible to use protecting groups60 or reducing agents48 to prevent the thiols from
reacting prematurely, but this adds additional expense and chemical exposure
potential. If disulfide bonds were used to link a DWR to cellulose, on the other hand,
this would be a less important consideration.
- Durability in consumer laundry: Disulfide bonds can be cleaved at more alkaline pH,
such as those found in commercial detergents.
- Effects on fabric: Some of the oxidants commonly used (hydrogen peroxide,

potassium permanganate) are also common bleaching agents already used in the
textile industry. Therefore, an oxidant that is not also a bleaching agent would have
to be used. In addition, some of the oxidants work better in acidic pH, which could
be detrimental to the cotton fabric. The most obvious choice is to use either air or
oxygen, potentially at high temperatures, but this would need to be tested along with
different possible catalysts. Our recommendation would be to examine some of the
“greener” options listed in the previous section, in particular the ones that work
solvent-free, as the ability to set the permanent crease without solvent would be quite
useful.
Imine & Amine Bonds
- Degree of additional research: Imine bond crosslinking is quite common, and has
been explored at least a bit for non-woven fabrics.61
- Disruption of chemical supply: This would depend on the specific diamine

crosslinker chosen.

68
- Disruption of crosslinking step: Using hydrogen gas will possibly require new
equipment to safely handle the gas, as well as the ability to recover the necessary
precious metal catalysts.52 The technical framework does not specifically evaluate
health effects, but the dangers of sodium cyanoborohydride are so great that we
cannot recommend using it.52–54 Sodium borohydride is probably the best choice
from a technical standpoint, as it is effective, has fewer health risks than sodium
cyanoborohydride, and does not require specialized equipment.
- Other considerations: A reaction that requires an acid catalyst55 would be

problematic for the cotton fabric, but this can almost certainly be avoided. Having a
diamine linker means we run the risk of having both amine groups react with one
cellulose chain, instead of crosslinking two chains. Using a shorter linker may
prevent this. In addition, there is a risk that the crosslinker may cleave the chains in
the polymer coating or PET fibers.58


Levis Green Technology

Uploaded by

Copyright:

Available Formats

Levis Green Technology

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Levis Green Technology

Uploaded by

Copyright:

Available Formats

NEW APPROACHES IN

GREENER SOLUTIONS 2013

The Current Problem 1

Appendix A: 12 Examples of Crosslinking Within Nature* 28

*These descriptions were prepared by the Biomimicry 3.8 Institute

FIGURES AND TABLES

Figure 1a and b. Existing crosslinking technologies 1

Figure 4. Polyethylene terephthalate 6

Figure 5. Poly(vinyl alcohol) 7

Figure 6. Thiolated poly(vinyl alcohol) 8

Figure 7a-d. Potential compounds for non-covalent crosslinking applications 8

Figure 8. Common naturally occurring poly(carboxylic acids) 9

Figure 9. Cyclic anhydride intermediate 9

Figure 10. In situ thiolation of poly(carboxylic acid) 10

Figure 11. Disulfide linkages 11

Figure 12. Imine bond formation 12

Figure 13. Reductive amination 12

Figure 14. Technical evaluation of each strategy 15

Figure 15. Technical evaluation of combined strategies 15

Figure 16. Health evaluation process 18

Figure 17. Comparison of health evaluation framework to GreenScreen 19

Figure 18. Health evaluation weighting scheme 19

Figure 19. Baseline health data 20

Figure 20. Disulfide bond health data 21

Figure 21. Imine/amine bond health data 21

Figure 22. Poly(carboxylic acid) health data 22

Figure 23. Polymer thiolation health data 23

Figure 24. Polymer coating health data 23

Table 1. Examples of crosslinking within each of the four bonding themes 2

Table 2. Eight technical evaluation factors 14

Table 3. GreenScreen Benchmark 1 criteria 17

The Current Problem

N DWR R DWR RESPIRATORY TRACT IRRITATION

Figure 1a. DMDHEU, the most N

Inspiration from Nature

Table 1. Examples of crosslinking within each of the four themes

Green Team Final Report

Translation to Textile Crosslinking

Green Team Final Report

Green Team Final Report

TEXTILE PROCESSING FOR DUMMIES

STEP SUBPROCESS OVERVIEW CHEMICAL INPUTS

WAX REMOVAL_ hydrophobic coating is removed

STEP SUBPROCESS OVERVIEW CHEMICAL INPUTS

WARPING yarns are later woven through in the weaving step.

For Levi’s_ indigo dye applied

25-30% caustic soda

STEP SUBPROCESS OVERVIEW CHEMICAL INPUTS

For Levi’s_ DWR properties

Lifecycle performance_ 10-30

Green Team Final Report

Linkages to Cellulose: Non-Covalent and Structural Bonds

Figure 4: PET, the most common polyester in textiles.6

Green Team Final Report

Figure 5. PVA, a water-soluble polymer frequently used in the textile industry.

Green Team Final Report

Figure 6. Thiolated PVA could be used with disulfide bond crosslinking.

Green Team Final Report