Application of a Phosphazene Derivative as a Flame Retardant for Cotton Fabric using Conventional Method and Supercritical CO2 By Krystal R. Fontenot, Michael W. Easson, Jade Smith, Crista Madison, Sunghyun Nam, Thach-Mien D. Nguyen, SeChin Chang, and Brian D. Condon, USDA-ARS-SRRC

AbstractConventional pad-dry-cure (non-scCO2) and supercritical carbon dioxide (scCO2) application methods were used to study the effectiveness of a newly synthesized phosphazene derivative as a flame retardant on cotton fabric. The 1,1’,4,5-tetrahydrotrispiro[1,3,2-diazaphosphole-2,2’-[1,3,5,2,4,6]triazatriphosphinine-4’,6’’-dibenzo[d,f][1,3,2]dioxaphosphepine-6’,6’’’-dibenzo[d,f][1,3,2]dioxaphosphepine] (2), was synthesized in two steps, in high overall yield, and applied onto cotton fabrics. The flame retardant properties of the treated fabrics were examined by 45° angle flammability test, limiting oxygen index test (LOI), microscale combustion calorimetry (MCC), thermogravimetric analysis (TGA), and scanning electron microscopy (SEM). The results indicate that fabrics treated with 2 are flame resistant; however, non-scCO2 fabrics performed better than scCO2 fabrics in a preliminary durability study.

Key Terms Cotton, Durability, Micro Combustion Calorimeter, Non-halogenated Flame Retardant, Supercritical Carbon Dioxide, Thermogravimetric Analysis

Disclaimer: The use of a company or product name is solely for the purpose of providing specific information and does not imply approval or recommenda-tion by the United States Department of Agriculture to the exclusion of others.

IntroductionFlame retardants (FRs) used in textiles, furniture, electron-ics, and insulation can be divided into two main categories: halogenated and non-halogenated FR.1,2 Halogenated FRs typically contain bromine or chlorine, but due to environ-mental and health concerns, they are becoming restricted in the marketplace.3–5 Current research is shifting towards non-halogenated FRs due to their non-toxic byproducts and lower environmental impact.3,4 Non-halogenated FR, often contain phosphorus, nitrogen, sulfur, and silicon. Phosphorus and nitrogen containing FRs are gaining more attention due to their combined effect of reducing flam-mable gases generated near the cotton surface and retarding fires.3 Molecules that contain both phosphorus and nitrogen exhibit better FR properties and higher char yields.2,3,6,7 For example, phosphazene derivatives impart high thermo-oxi-dative stability and excellent flame retardancy.8 Additionally, phosphazene is easily modified through nucleophilic sub-stitution, offering various synthetic routes that can provide different functionalities, properties, and novel FRs.8

Different methods, including pad-dry-cure and inclusion, have been used to incorporate FRs into cotton textiles.9,10 Conventional pad-dry-cure is extensively used to incorpo-rate FRs into cotton fabric by dipping the fabric into the FR solution, padding to remove excess solvent, and drying and curing the fabric.9–11 Inclusion is an impregnation process achieved when one molecule is non-covalently enclosed

within another molecular structure; however, a drawback of this method is the release of molecules due to the re-swelling of cotton fabric with water.12,13 Given its limitations, inclu-sion is not widely used to incorporate FRs. An alternative method to enclose molecules into fabrics involves the use of a supercritical fluid (SCF).

SCFs are solvents that exist above the specific SCF’s critical pressure and temperature.14 SCFs are inexpensive, envi-ronmentally friendly, non-flammable, sustainable, and are known for their high diffusion in organic matter. Further-more, SCFs can eliminate or reduce the use of organic solvents and water when dyeing or impregnating hydro-phobic or hydrophilic compounds into cotton textiles.15,16 A commonly used SCF is supercritical carbon dioxide (scCO2), which has been used for extractions, impregnations, drying, and dyeing fabrics or polymers.14,17–19 Since 2012, Nike and Adidas have endorsed the use of scCO2 as a water-free pro-cess to dye fabrics.20–22 These dyes contain functional groups that react with the cotton fabric to produce a permanent attachment. Not only does the use of scCO2 eliminate the consumption of billions of gallons of water each year, but it also saves time and energy.22 The supercritical process has also been used to impregnate hydrophilic and hydropho-bic molecules such as D-panthenol, thymol, polymers, and nanoparticles into cotton systems.15,23,24

Current scCO2 impregnation studies have not yielded the levels of durability obtained when cotton fabric is treated

November/December 2014 Vol. 1, No. 6 | 16

AATCC Journal of Research

Use the ACS Style Guide as a reference for formatting yourmanuscript before submitting to AATCC. See:http://www.aatcc.org/pub/aatcc-journal/submissionguidelines/

with molecules having no reactive functional groups, such as the FR reported in this study.23,25,26 Herein, we investigate the use of scCO2 to impregnate FRs into cotton fabric as another avenue to achieve durability. The current research focus includes (1) the synthesis of a phosphazene-based FR that is reacted with 2,2’-dihydroxybiphenyl and 1,2-diamino-ethane, (2) the determination of flammability when cotton fabric is treated with the FR via conventional pad-dry-cure (non-scCO2) and scCO2 methods, and (3) the investigation of whether the scCO2 method can provide durability. The flammability, thermal properties, char, and durability of the fabrics treated by each method were explored.

ExperimentalMaterials Phosphazene and anhydrous sodium sulfate (Na2SO4) were purchased from Sigma Aldrich. Dichloromethane (DCM), methanol (MeOH), acetone, and hexanes were purchased from Fisher Scientific. Anhydrous potassium carbonate (K2CO3) and 1,2-diaminoethane were purchased from Alfa Aesar. Deuterated dichloromethane-d2 and acetone-d6 with 1% trimethylsilane (TMS) were purchased from Cambridge Isotopes Laboratories. All chemicals were used as received without further purification. Cotton fabric (100%) was purchased as twill fabric style 423, with a weight of 258 g/m2, from Test Fabrics, Inc. The fabric was desized, bleached, and was free of all resins and finishes. Nuclear magnetic resonance (NMR) spectroscopy was performed on a Varian 400 MHz spectrometer. 31P-NMR spectra were given in δ (ppm) relative to an external 85% aqueous H3PO4 standard. Liquid chromatograph mass spectrometry (LC-MS/MS) data was collected on an Agilent 6520 QTOF LC-MS/MS instrument. Mathis dryers (W. Mathis AG LTF and Mathis Labdryer LTE) were purchased from Mathis U.S.A. Inc. The supercritical reactor was purchased from Supercritical Fluid Technologies Inc.

SynthesisFR Precursor6’,6’-Dichloro-4’l5,6l5,6’l5-dispiro[dibenzo[d,f][1,3,2]dioxaphosphepine-6,2’-[1,3,5,2,4,6]triazatriphosphinine-4’,6’’-dibenzo[d,f][1,3,2]dioxaphosphepine] (1) synthesis: phosphazene (5.26 g, 14.4 mmol, 1.05 equiv) was added into 150 mL of acetone and allowed to stir until completely dissolved. K2CO3 (4.98 g, 2.50 mmol, 2.50 equiv) was added to the mixture at 0 °C under an argon atmosphere. 2,2’-Dihydroxybiphenyl (5.36 g, 14.4 mmol, 2 equiv) in 50 mL of acetone was added slowly to the reaction mixture over 20 min. The reaction remained at 0 °C for 20 min and then allowed to warm to room temperature (RT). The solution changed from colorless to clear beige during the reaction. The reaction was allowed to stir overnight and was moni-tored by TLC using DCM/hexanes (1:1) as an eluent. When

the reaction was complete, the solid was filtered and the fil-trate was concentrated in vacuo. The product was triturated with hexanes, gravity filtered, and dried under high vacuum. The final product 1 was obtained as a white solid (6.61 g, 77%). Trace amounts of mono-product were observed (Scheme 1). MS (ESI): m/z 573.9816 [M + H]+, calculated for C24H17N3O4P3, 574.9849. The 1H-, 13C-, and 31P-NMR data for 1 are shown in Fig. 1.

FR1,1’,4,5-Tetrahydrotrispiro[1,3,2-diazaphospho-le-2,2’-[1,3,5,2,4,6]triazatriphosphinine-4’,6’’-dibenzo[d,f][1,3,2]dioxaphosphepine-6’,6’’’-dibenzo[d,f][1,3,2]dioxa-phosphepine] (2) synthesis: 1 (4.00 g, 6.96 mmol) was dissolved in 55 mL of DCM and allowed to stir until com-pletely dissolved. The mixture was then chilled to 0 °C and ethylenediamine (0.932 mL, 13.96 mmol) in 10 mL of DCM was added drop wise to product 1 in 55 mL of DCM over 30 min. The reaction was allowed to warm to RT, stirred overnight, and was monitored by thin layer chromatography (TLC) using DCM/MeOH (9:1) and DCM/hexanes (1:1) eluents. The organic layer was extracted using deionized water heated to 33 °C. [Note: the solution was vigorously swirled without capping the separatory funnel.] The organic layer was dried over Na2SO4 and concentrated in vacuo. The final product 2 was obtained as a white solid (3.28 g, 84%) without further purification (Scheme 1). MS (ESI): m/z 561.41 [M + H3O]+, C26H25N5O5P3, 580.1149. The 1H-, 13C-, and 31P-NMR data for 2 are shown in Fig. 2. The crystal structure of compound 2 has been deposited in the Acta Crystallographica Section E with reference number jj2173.

Fabric TreatmentsPad-Dry-Cure Method Cotton twill test fabric 423 was bleached, mercerized, and scoured prior to use. Cotton twill was immersed in a shallow pan containing the solution of 2 in DCM/acetone (1:1) to obtain a percent add-on of 6, 9, 12, 16, and 22 wt %. After soaking for 5 min, the fabric was padded, dried at 100 °C for 5 min using a W. Mathis AG Dryer LTF, and cured at 160 °C for 90 s using a Mathis Labdryer LTE. After removal from the oven, fabrics were allowed to cool to RT, and the percent add-on was calculated using Eq. 1.

Eq. 1

Waftertreatment is the fabric weight after treatment and Woriginal is the original fabric weight.

Supercritical Carbon Dioxide Method The supercritical reactor consisted of ultra high purity carbon dioxide, a series II Prime/Purge, and a high pressure series temperature control reactor. Supercritical treatments were performed by first preparing a 5% co-solvent solution

17 | Vol. 1, No. 6 November/December 2014

AATCC Journal of Research

Scheme 1. Synthesis of compound 2.

Fig. 1. NMR data of compound 1: 1H-, 13C-, and 31P-NMR in d6-acetone.

1H-NMR 13C-NMR

31P-NMR

November/December 2014 Vol. 1, No. 6 | 18

AATCC Journal of Research

of 2 in DCM/acetone (1:1). The fabrics were then soaked in a beaker containing the 5% solution for 1 h. The fabric was removed from the beaker, drip dried, and then wrapped around a metal wire cage, which prevents the fabric from being entangled with the stirrer, and placed into the scCO2 reactor. The remaining solution was then poured into the chamber with a magnetic stirrer and sealed. Ultrapure scCO2 was supplied to the reactor using a gas cylinder. Selected pressures (1500, 1600, 1750, and 2150 psi) were applied at a constant temperature of 90 °C to obtain the desired add-on of 9, 11, 12, 16 and 22 wt %. After removal from the reactor, the fabric was then cured at 140 °C for 3 min using a Mathis Labdryer LTE, allowed to cool to RT, and the percent add-on was calculated using Eq. 1.

45° Flammability Test The 45° flammability test was performed according to ASTM D1230-10 using a 45° angle chamber (Dwyer TC-45).27 The flame exposure time was modified from 1 s, which is standard protocol, to 10 s. Fabrics were cut to 6.4 × 15.2 cm, treated with FR, and conditioned at 21 °C with 65% relative humidity (RH) for 48 h. All specimens were classi-fied as Class I or Class III textiles that do not have a raised fiber surface.

Limiting Oxygen Index (LOI)LOI was performed according to ASTM D2863-13 using a Dynisco Polymer Test Limiting Oxygen Index cham-ber.28 Treated fabrics cut to 6.4 × 12.7 cm (W × L) were conditioned at 21 °C with 65% RH for 48 h prior to use. The average LOI values of four consecutive measurements were reported.

Microscale Combustion Calorimetry (MCC)MCC was performed according to ASTM D7309-13 using a Govmark FAA Micro calorimeter.29 The samples were inves-tigated in an 80% nitrogen and 20% oxygen environment from 75-500 °C at a heating rate of 1 °C /s. Three replicate measurements were conducted and averaged for each sample with fabric weight ranging from 4.5-5.5 mg.

Thermogravimetric Analysis (TGA)TGA was performed on a TA Instruments Q500 under nitrogen. Sample masses of 4.0-8.0 mg were heated from 0–600 °C at a rate of 10 °C/min. Three replicate measure-ments were conducted and their average thermal parameters were reported. Char yield and onset temperature was ana-lyzed using Universal Analysis 2000 software.

Scanning Electron Microscopy (SEM)The control and treated fabrics were examined by SEM (Phillips, XL 30 ESEM) at 1000 and 2000X magnifications. Samples were coated with a gold-palladium alloy using a vacuum sputter coater, and an acceleration voltage of 10 keV and a beam current of 0.5 nA were used.

Preliminary Study of DurabilityPreliminary data for home laundering was obtained by fol-lowing a modified version of AATCC Test Method (TM) 124-2011 for washing.30 Briefly, a beaker containing 500 mL of deionized water was heated to 40 °C. The fabric was then added to the beaker and allowed to stir at 500 rpm for 10 min. This process was repeated 3 to 10 times. After each washing, the fabrics were dried at 110 °C for 10 min, re-equilibrated, and reweighed to determine the remaining percent add-on. Eq. 1 was used to calculate the percent add-on of treated fabrics.

Results and DiscussionFRs are designed to incorporate moieties that reduce textile flammability. These systems typically include phospho-rus, nitrogen, sulfur, silicon, and boron atoms combined in various ways to produce FRs. The composition of 2 is similar to other FR structures known to have good thermal properties such as the phosphazene derivative compounds synthesized by Allcock et al.31 and Sun et al.8 and other phosphazene free derivatives such as commercially avail-able dihydrooxaphosphaphenanthrene oxide (DOPO). Compound 2 has a smaller molecular weight than these phosphazene derivatives.8,31

Synthesis The synthetic route of the FR herein is shown in Scheme 1. Phosphazene was reacted with 2,2’-dihydroxybiphenyl in the presence of acetone and K2CO3 at 0 °C under inert condi-tions according to the protocol published by Carriedo et al.32 The purified precursor 1 was achieved with an 84% yield. The 1H-NMR spectra exhibited aromatic proton peaks from 7.51–7.30 ppm, which are attributed to the 2,2’-dihydroxy-biphenyl components of 1. Compound 1 has six signals that refer to the aromatic carbons at 147.9–121.9 ppm and the ethylene carbons at 75.8 ppm in the 13C-NMR spectra. The 31P-NMR spectra of 1 exhibits an AB2 spin pattern with two resonance signals at 29.42 ppm, as a triplet, and at 20.19 ppm, as a doublet corresponding to the phosphorus atom of phosphazene-bearing chlorine and two 2,2’-dihydroxy-biphenyl groups, respectively.8 NMR spectra of 1 are shown in Fig. 1.

Compound 2 was synthesized by reacting 1 with 1,2-diami-noethane in DCM at 0 °C under inert conditions. The final product 2 was achieved in 90% yield and the overall yield of the two-step reaction was 76%. The characterization of 2 using 1H-NMR spectroscopy shows the aromatic pro-tons from 7.66–7.36 ppm, the amino protons at 3.86 ppm, and the methylene protons at 3.26 ppm. In the 13C-NMR, seven signals are observed for the aromatic and methylene carbons at 149.0–115.6 ppm and 42.8 ppm, respectively. The 31P-NMR of 2 also has an AB2 spin pattern with two

19 | Vol. 1, No. 6 November/December 2014

AATCC Journal of Research

resonance signals at 34.79 ppm, as a triplet, and at 28.80 ppm, as a doublet, corresponding to the phosphorus atom of phosphazene-bearing 1,2-diaminoethane and two 2,2’-dihy-droxybiphenyl groups, respectively.8 NMR spectra of 2 are shown in Fig. 2.

Fabric Treatments Cotton twill fabric was treated with 2 using pad-dry-cure (non-scCO2 method) and scCO2 as the impregnation methods. Add-ons of 6, 9, 12, 16, and 22 wt % were achieved using the non-scCO2 method. A protocol similar to Tsi-optsias et al. was modified for scCO2 treated fabrics and add-ons of 9, 11, 12, 16, and 22 wt % were obtained.16 A co-solvent of DCM/acetone was used to aid in the solubility of 2 in scCO2.

33 Solubility of other supercritical soluble moieties such as siloxane, fluorinated compounds (fluoroalkyl and fluoroether), biphenyl, and phosphazene have been studied

and were found to aid in the solubility of other compounds in scCO2.

34–36 Compound 2 should be soluble in scCO2 with the use of co-solvents, phosphazene, and 2,2’-biphenol—the 2,2’-biphenol structure present in 2 is similar to biphenyl, which is known to be soluble in scCO2. The non-scCO2 and scCO2 fabrics were washed and add-ons of 15–16 wt % W (W indicates washed fabric) and ~1 wt % W were obtained, respectively (see Preliminary Durability Study section).

Flammability The 45° flammability test, limiting oxygen index test (LOI), and microscale combustion calorimetry (MCC) test were performed to determine the flammability of 2. The 45° test evaluates the ability of fabric to resist burning at an angle when exposed to open flames. The control and treated fabrics (non-scCO2 and scCO2) were evaluated using 10 s of open flames; results and images are shown in Table I and

Fig. 2. NMR data of compound 2: 1H-, 13C-, and 31P-NMR in d6-acetone.

13C-NMR1H-NMR

31P-NMR

November/December 2014 Vol. 1, No. 6 | 20

AATCC Journal of Research

Fig. 3, respectively. The control fabric burned for a total of 71 s, which was expected due to the release of flammable gasses upon combustion of cotton fabric. The non-scCO2 fabric with an add-on of 6 wt % burned for a total of 64 s and resulted in a reduced burning time compared to the control. As the add-on increased to 9 wt %, the fabric burned for 5 s and self-extinguished. The 9 wt % add-on may have ignited due to an uneven coating of the fabric. Treated fabrics with an add-on of 9 wt % and higher did not ignite, including the washed fabric with an add-on 16 wt % W. The washed fabric with a ~1 wt % W add-on burned for 60 s, which also resulted in a reduced burning rate compared to the control burning rate of 71 s. It was observed that the non-scCO2 add-on of 6 wt % and the scCO2 add-on of ~1 wt % W burned due to a lower loading level of 2; while treated fabrics (non-scCO2 and scCO2) with a higher add-on did not ignite.

Flame retardants with an LOI value of 27–28% are desirable for cotton textiles. LOI measures the minimum amount of oxygen required for fabric to burn in the presence of flowing oxygen and nitrogen gas, while maintaining a downward candle-like combustion.8 Flame retardants that have an LOI value of less than 21% are flammable, those with an LOI val-

Fig. 3. 45° flammability test of control and treated fabrics.

Table I: Flammability Test Results of Control and Treated Fabrics

# % Add-On (wt %) LOI value (vol %) 45° TestAverage Timea (s)

non-scCO2

1 Control 21 ± 0.50 71 ± 8

2 6 26 ± 0.00 64 ± 15

3 9 27 ± 0.70 5 ± 2

4 12 26 ± 0.50 DNIb

5 16 29 ± 0.00 DNI

6 22 31 ± 0.00 DNI

7 15% W 28 ± 1.1 —

8 16% W — DNI

scCO2

9 9 25 ± 0.00 DNI

10 12 26 ± 0.00 DNI

11 22 29 ± 1.40 —

12 ~1% W 22 ± 1.00 59 ± 6aAverage time after ten seconds and all specimens are Class I and Class II. bDid not ignite (DNI).

21 | Vol. 1, No. 6 November/December 2014

AATCC Journal of Research

ue of 21–26% are slow burning, and those with an LOI value greater than 27% are self-extinguishing.37,38 However, higher LOI values indicate a stronger flame retardancy effect.4 The LOI value of the control was 21%; this indicates the flamma-bility of cotton fabric. Compound 2 was evaluated at various additive loadings for LOI including 6, 9, 12, 16, 22, and 15 wt % W. The LOI values increased for treated fabrics as the add-on of 2 increased. For non-scCO2 fabrics, an add-on of 6 wt % had an LOI value of 26% and was slow burning, while an add-on of 9 wt % and greater had LOI values of 27% and higher and were self-extinguishing. Treated fabrics had similar LOI values at comparable add-on levels; however the non-scCO2 fabrics showed slightly higher values (Table I). Fabrics with an add-on of 9, 12, and 22 wt % (scCO2) had LOI values of 25, 26, and 29%, respectively. For scCO2 fabrics, add-ons of 9 and 12 wt% were slow burning and an add-on of 22 wt% was self-extinguishing. It was shown that polymer systems without phosphazene have lower LOI values (22%), but when phosphazene was incorporated, the LOI values increased to 28–33% depending on the derivative of phosphazene used.31 The LOI values of 2 were analogous to other phosphazene derivatives.

MCC measures the rate at which the heat of fuel gasses are released from a solid under inert conditions.10,39,40 MCC profiles indicate the peak heat release rate (pHRR), total heat release (THR), and the maximum temperature (TMAX) achieved by treated fabrics (Figs. 4 and 5). The HRR pro-files of the control and treated fabrics had one main stage between 300–375 °C. Treated fabrics had sharp heat release peaks that contrasted with the broad heat release peak observed for the control. The control exhibited a signifi-cantly lower pHRR value of 221 W/g, whereas treated fabrics pHRR values ranged between 367–577 W/g. Higher pHRR values indicated an increased amount of heat released during the burning process, which may result from rapid pyrolysis

and combustion of treated fabrics.39 This may account for the sharp HRR profiles of treated fabrics. Generally, as the add-on levels of 2 increased, pHRR values increased, while THR and TMAX values decreased. The non-scCO2 treated fabrics had a pHRR range of 367–577 W/g (6, 9, 12, 16, and 22 wt %) and were similar to the pHRR range of 409–573 W/g (9, 12, and 22 wt %) obtained for scCO2 fabrics. The pHRR values reported were parallel to the pHRR values obtained for aromatic non-phosphazene systems.39 An add-on of 6 wt % had THR and TMAX values of 6 kJ/g and 315 °C respectively. As the add-on increased to 9 wt %, the THR and TMAX values decreased to 5 kJ/g and 322 °C respectively, and remained constant as the add-on increased. These values were lower than the THR and TMAX values obtained for the control fabric, 10 kJ/g and 381 °C, respectively. Supercritical treated fabrics exhibited a similar calorimetric performance. The scCO2 fabrics at 9 wt % add-on had a higher THR value of 6 kJ/g, which remained constant as the add-on increased. While the washed fabrics had different add-ons—15 wt % W (non-scCO2) and 1 wt % W (scCO2)—their pHRR, THR, TMAX values were in range of the treated fabrics. The non-scCO2 washed fabric had a pHRR value of 562 W/g, a THR value of 5 kJ/g, and a TMAX value of 317 °C, which was in the range of MCC values obtained for treated fabrics. The washed scCO2 fabric has an add-on below the 9 wt % thresh-old of 2 and resulted in a pHRR value of 288 W/g, a THR value of 7 kJ/g, and a TMAX value of 312 °C, which was com-parable to the MCC values obtained for the 6 wt % add-on.

The results of 45° flammability test, LOI, and MCC indi-cated that add-ons greater than 12 wt % had comparable results to those with 9 wt % add-ons. These findings imply that the threshold of 2 as an FR was effective at 9 wt % and a higher add-on was not needed for improved flame retar-dancy. The performance of treated fabrics as determined by the 45° flammability test, LOI, and MCC were similar, but

Fig. 4. Heat release rate profiles of control and non-scCO2 treated fabrics. Data in the graph corresponds to pHHR, THR, and TMAX respectively.

Fig. 5. Heat release rate profiles of control and scCO2 treated fabrics. Data in the graph corresponds to pHHR, THR, and TMAX respectively.

November/December 2014 Vol. 1, No. 6 | 22

AATCC Journal of Research

the results of scCO2 treated fabrics were slightly lower. The differences in the results obtained for treated fabrics may be due to the different treatment methods. An 1H-NMR spectrum was taken of the residual solid in the scCO2 cham-ber to determine if it decomposed upon completion of the scCO2 process, but the compound was found to be stable. As a result, it is not clear at this time why scCO2 fabrics had a slightly lower performance.

Thermal Behavior TGA measures the change in mass with respect to tem-perature and was performed to investigate the thermal stability of 2. Char percent yield and onset temperature were determined for the treated fabrics and are summarized in Figs. 6 and 7. As the add-on increased, char yield and onset temperature increased as well. Treated fabrics on average had a 2.5-fold increase in char yield and 1-fold decrease in onset-temperature compared to the control. The thermo-grams of the treated fabrics exhibited a drastic decrease in sample weight between 302–319 °C and plateaued out (Figs. 6 and 7). The char yield (15%) of the control fabric was lower than the char yield (22–39%) obtained for fabrics treated with 2. The onset-temperature of the control, 331 °C, was higher than onset temperatures (302–319 °C) of fabrics treated with 2.

As the add-on of the FR increased from 6 to 9 wt %, increas-es in char yield and onset temperature were observed from 31–37% and 310–312 °C, respectively. These results support an add-on threshold of 9 wt % for 2; lower add-on values did not provide effective flame retardancy. This threshold level was promising given that most FRs require a 20–40 wt % add-on to achieve flame retardancy.41 The lower loading level of 2 was attributed to its number of P-N bonds, and poten-tially to the synergistic effect between these elements.3,6,42

Studies have shown that an increase in the number of P-N bonds leads to higher char yield, a decrease in fuel forma-tion, and a lower additive loading of the FR.7,41 Greater add-ons of 12, 16, 22 wt %, and 15 wt % W, had a char yield of 37–39%, and the on-set temperature ranged from 312–318 °C. The decrease in on-set temperature from 333 °C to 310 °C, when treated, was due to the interference of the FR in the pyrolysis of cellulose to levoglucosan.43

Supercritical-treated fabrics with a 9, 12, and 22 wt % add-on had char yields of 29, 35, and 39% as compared to non-scCO2 treated fabrics with a char yield of 37, 38, and 39%, respectively. The onset temperatures for scCO2 fabrics were 308–319 °C, which were comparable to the onset temperatures of 310–318 °C for non-scCO2 fabrics. The char yields of scCO2 fabrics with an add-on of 9 and 12 wt % were slightly lower than the char yields obtained for non-scCO2 fabrics. The washed fabrics with an add-on of 1 wt % W had a char yield of 22% and an onset temperature of 302 °C, which was below the 9 wt % threshold of 2.

Between the lowest add-on of 1–6 wt % and 9 wt %, an average 1.5 fold increase was observed compared to the that between the highest add-on of 12–22 wt % and 9 wt %, which resulted in a 1 fold increase in char yield. This data further supports that the threshold of 2 as a flame retardant required a 9% add-on. TGA results supported the findings of the MCC data. It was concluded that the higher char yield of treated fabrics indicated a greater amount of carbon or inor-ganic material left behind, resulting in a decreasing THR.39

Compound 2 was compared to mono-methylolmelamine phosphate (MPR1, C4H17N6O13P3), a derivative of the commercially-available flame retardant melamine phos-phate, which is used as flame retardant for cotton textiles.44 In the structure of MPR1 and 2, phosphorus and nitrogen

Fig. 6. TGA thermogram of control and non-scCO2 treated fabrics with char yield and onset temperature values.

Fig. 7. TGA of control and scCO2 treated fabrics with char yield and onset temperature values.

23 | Vol. 1, No. 6 November/December 2014

AATCC Journal of Research

are key elements and have different content: MPR1 (19% and 21%) and 2 (6% and 12%), respectively. A 15 wt % add-on of MPR1 on cotton fabric has a char yield of 38%.44 Although MPR1 had a higher phosphorus and nitrogen content when compared to 2, its thermal behavior in terms of TGA was similar to 2. These findings are encouraging, considering that compound 2 had a lower phosphorus and nitrogen content, but the same char yield as MPR1 when 2 was used in a 9 wt % add-on.

Analysis of Char Layer Generally, burned cotton fabrics treated with an effective FR will generate char that acts as a protective barrier. The for-mation of a char layer is desirable to reduce heat transfer and the amount of volatile gases generated at the flame zone.45,46 The higher the char yield, the better the flame retardant properties, but the amount of char produced is dependent on the composition of the FR.

SEM was used to image control and treated fabrics to exam-ine the unburned and burned sections of the fabrics (Fig. 8). The image of the control fabric (Fig. 8A) showed smooth elongated fibers, and the image of the burned control (Fig. 8B) showed that the fibers were mainly burned to ash. The highest add-on of non-scCO2 (Fig. 8C) and non-scCO2 washed fabrics (Fig. 8E), 22 wt % and 15 wt % W, respec-tively, were examined by SEM. The images of unburned non-scCO2 (Fig. 8C) and non-scCO2 washed fabric (Fig. 8E) displayed diamond shape crystallites of 2 on the surface of the fibers. The images of burned non-scCO2 (Fig. 8D) and non-scCO2 washed fabrics (Fig. 8F) suggests that the fibers were coated with a protective char layer, which acted as a protective barrier for the fabric.

The scCO2 fabrics with add-ons of 12 wt % and 1 wt % W were examined by SEM. The micrographs of scCO2 fabric (Fig. 8G) and scCO2 washed fabric (Fig. 8I) also showed diamond shaped crystallites of 2. Compound 2 was clumped together on the cotton surface and the cotton fibers were swollen, which was indicative of scCO2 treatment (Fig. 8G).23 The burned scCO2 fabric (Fig. 8H) at 12 wt % add-on showed that the fibers were coated with char and were wrin-kled compared to the burned fibers of non-scCO2 fabrics (Fig. 8D). The scCO2 washed (Fig. 8J) fabric with an add-on of 1 wt % W, was below the 9 wt % threshold of compound 2, and showed that the fibers were covered in char, wrinkled, and appeared to be structurally intact when compared to the burned control fabric (Fig. 8B).

The washed fabric of non-scCO2 (Fig. 8E) and scCO2 (Fig. 8I) appeared to form clusters between the elongated fibers, which may have resulted from washing the fabric and the hydrophobic nature of the compound to resist water molecules.47,48 The fibers of non-scCO2 and scCO2 fabrics remained intact and were thermally stable. A higher additive

loading level of 2 generated more char (Figs. 8D and F). The lower loading of scCO2 fabrics as shown in Figs. 8H and J appeared more wrinkled and may have resulted from the supercritical treatment process since it was not observed for non-scCO2 fabrics.

Preliminary Durability Study Fabrics treated using the non-scCO2 (conventional pad-dry-cure) method were washfast, had better durability properties than scCO2 treated fabrics, and gave an average 1.5 fold decrease in the loss of 2 on the fabric. Non-scCO2 treated fabrics with a 23% add-on that were washed ten times yield-ed a final add-on of 15–16 wt % W and had an average 14 fold decrease in 2 on the fabric. Supercritically-treated fabric

Fig. 8. SEM micrographs of unburned and burned fabrics. Unburned fabrics include (A) control, (C) 22% add-on of non-scCO2, (E) 15% W add-on of non-scCO2, (G) 12% add-on of scCO2, and (I) 1% W add-on scCO2. Burned fabrics include (B) control, (D) 22% add-on of non-scCO2, (F) 15% W non-scCO2, (H) 12% add-on of scCO2, and (J) 1% W add-on of scCO2.

November/December 2014 Vol. 1, No. 6 | 24

AATCC Journal of Research

had an initial add-on of 11–16 wt %, but resulted in a final add-on of ~1 wt % after three washes. The scCO2 washed fabric should have a better wash fastness not only due to the hydrophobic nature of 2 but also due to impregnation. Kraft et al. used scCO2 to impregnate water-soluble D-panthenol into cellulosic fibers.15 This study revealed that D-panthenol was impregnated into the fiber after confirmation using Raman spectroscopy, but after one home laundering, the compound was washed out of the fabric.15 They reported further studies would be needed to determine if this behav-ior is normal for D-panthenol or for other compounds as well.15 In an additional study performed by Katayama et al., scCO2 was used to impregnate nanoparticles onto cotton fabric.24 They determined that time, temperature, pres-sure, and depressurization rate of scCO2 from the reactor generated a wrinkled cotton surface, which provided the necessary support (i.e., pits and valleys) for the nanoparticles to embed into.24 After two rinses of the treated cotton fabric, the nanoparticles remained adhered to the surface.24 While similar effects of Katayama’s work were not investigated, it is likely that the surface of the cotton fabric herein did not form the wrinkled surface to the same extent.

ConclusionsThe newly synthesized phosphazene-based FR, 2, was syn-thesized in two steps and easily purified in overall 76% yield. Compound 2 was characterized using MS and NMR, and tested for its FR properties. This phosphazene derivative has an increased number of P-N bonds that may have aided in a lower flame retardant loading of 9 wt %.

Treated fabrics had comparable thermal and flammabil-ity results. LOI, MCC, and 45° flammability test results of treated fabrics established that 2 is a self-extinguishing flame retardant. TGA further supports 2 as a flame retardant with high char yield and lower onset-temperatures that aid in the protection of cotton fabric. The flame retardancy threshold of 2 was determined to be 9 wt %, since higher additive loadings showing parallel results to the 9 wt % add-on. SEM of cotton fabrics showed that the control fabric was destroyed, while treated burned fabrics were coated with a protective char layer. The preliminary durability study of treated fabrics resulted in non-scCO2 fabrics outperforming scCO2 fabrics by seven washes. The non-scCO2 washfastness of 2 was due to its hydrophobic nature. While the differences observed in the performance of the treated fabrics were due to the sample preparation methods, more studies will need to be performed to determine why a lower performance in washfastness was observed for scCO2 fabrics. The poor washfastness may be attributed to the solubility of 2, the extent of the wrinkled surface of the fabric, and the absence of functional groups that can react with the -OH groups of cellulose.

Compound 2 is a good candidate for use as a flame retardant for cotton fabric due to the low cost of starting materials and ease of synthesis. It requires a low add-on, has self-extinguishing properties, and has high char yield that acts as a protective barrier. This flame retardant has a drawback of not being water-soluble, but this phosphazene derivative can be further derivatized to improve water solubility. Other FRs with a similar design may be able to achieve lower additive loadings. Furthermore, the use of 2 can be expanded in fur-ther applications, such as water repellants, considering water does not penetrate the fabric when it is treated with 2.

The mechanism of pyrolysis of the treated fabrics will be further examined by thermogravimetric analysis-Fourier transform infrared spectroscopy (TGA–FTIR), attenu-ated total reflectance-infrared spectroscopy (ATR-IR), and computational studies. These results will be reported at a later date.

AcknowledgementsThis project was financed by the US Department of Agricul-ture. We thank Dr. Casey Grimm for ESI-MS and we thank Drs. Marsha R. Cole and Michael Santiago Citron for their assistance in proofreading.

References1. Shaw, S. D.; Blum, A.; Weber, R.; Kannan, K.; Rich, D.; Lucas, D.;

Koshland, C. P.; Dobraca, D.; Hanson, S.; Birnbaum, L. S. Review of Environmental Health 2010, 25 (4), 261–305.

2. Weil, E. D.; Levchik, S. V. Journal of Fire Sciences 2008, 26 (3), 243–281.3. Li, Q.; Jiang, P.; Su, Z.; Wei, P.; Wang, G.; Tang, X. Journal of Applied

Polymer Science 2005, 96 (3), 854–860.4. Allen, C. W.; Hernandez-Rubio, D. The Use of Phosphazenes as Flame

Retardants. In Phosphazenes: A Worldwide Insight; Nova Science Publishers Inc.: Hauppauge, NY, USA, 2004; pp 485–503.

5. Zhao, X. Journal of the Textile Institute 2010, 101 (6), 538–546.6. Bakos, D.; Kosik, M.; Antos, K.; Karolyova, M.; Vyskocil, I. Fire and

Materials 1982, 6 (1), 10–12.7. Barker, R. H.; Drews, M. J. Flame Retardants for Cellulosic Materials.

In Cellulose Chemistry and Its Applications; Nevell, T. P., Zeronian, S. H., Eds.; Ellis Horwood Limited Inc.: Chichester, UK, 1985; pp 423–454.

8. Sun, J.; Wang, X.; Wu, D. ACS Applied Materials & Interfaces 2012, 4 (8), 4047–4061.

9. Nguyen, T.; Chang, S.; Condon, B.; Slopek, R.; Graves, E.; Yoshioka-Tarver, M. Industrial & Engineering Chemistry Research 2013, 52 (13), 4715–4724.

10. Yoshioka-Tarver, M.; Condon, B. D.; Santiago Cintrón, M.; Chang, S.; Easson, M. W.; Fortier, C. A.; Madison, C. A.; Bland, J. M.; Nguyen, T. D. Industrial & Engineering Chemistry Research 2012, 51 (34), 11031–11037.

11. Beninate, J. V.; Boylston, E. K.; Drake, G. L.; Reeves, W. A. Textile Research Journal 1968, 38 (3), 267–272.

12. Frank, S. G. Journal of Pharmaceutical Sciences 1975, 64 (10), 1585–1604.

13. Zeronian, S. H., Intercrystalline Swelling of Cellulose. In Cellulose Chemistry and Its Applications; Nevell, T. P., Zeronian, S. H., Eds.; Ellis Horwood Limited Inc.: Chichester, UK, 1985; pp138–180.

25 | Vol. 1, No. 6 November/December 2014

AATCC Journal of Research

14. Fleming, O. S.; Kazarian, S. G. Polymer Processing with Supercritical Fluids. In Supercritical Carbon Dioxide; Wiley-VCH: Weinheim, Germany, 2006; pp205–238.

15. Kraft, G.; Muss, C.; Adelwöhrer, Christian, R.; Roder, T. Lenzinger Berichte 2004, 83 (1), 117–121.

16. Tsioptsias, C.; Panayiotou, C., Journal of Materials Science 2011, 46 (16), 5406–5411.

17. Üzer, S.; Akman, U.; Hortaçsu, Ö. The Journal of Supercritical Fluids 2006, 38 (1), 119–128.

18. Ginneken, L. V.; Weyten, H. Particle Formation Using Supercritical Carbon Dioxide. In Carbon Dioxide Recovery and Utilization, Aresta, M., Ed.; Kluwer Academic Publishers: Dordrecht, The Netherlands, 2003; pp123–136.

19. Filardo, G.; Galia, A.; Giaconia, A. Modification of Polymers. In Supercritical Carbon Dioxide. In Carbon Dioxide Recovery and Utilization, Aresta, M., Ed.; Kluwer Academic Publishers: Dordrecht, The Netherlands, 2003; pp197–207.

20. Makower, J. Color It Green: Nike to Adopt Waterless Textile Dyeing, 2012, http://www.greenbiz.com/blog/2012/02/07/color-it-green-nike-adopt-waterless-textile-dyeing (accessed November 2013).

21. Huntsman, J. M. Innovative Dyes and Process Technology from Huntsman Textile Effects and Dyecoo Eliminates the Use of Water in the Textile Dyeing Process, 2012, http://www.huntsman.com/corporate/a/Innovation/DyeCoo%20delivers%20sustainable%20textiles (accessed December 2013).

22. Ferris, D. Nike and Adidas Want to Dye Your Shirt with No Water, 2012, http://www.forbes.com/sites/davidferris/2012/08/30/nike-adidas-want-to-color-your-shirt-with-no-water/ (accessed November 2013).

23. Milovanovic, S.; Stamenic, M.; Markovic, D.; Radetic, M.; Zizovic, I. The Journal of Supercritical Fluids 2013, 84 (1), 173–181.

24. Katayama, S.; Zhao, L.; Yonezawa, S.; Iwai, Y. The Journal of Supercritical Fluids 2012, 61 (0), 199–205.

25. Chang, S.; Condon, B.; Smith, J. Proceedings, Supercritical Fluids International Symposium, San Francisco, CA, 2012; pp 1–4.

26. Halford, B. Chemical and Engineering News 2011, 89 (1), 24.27. Annual Book of ASTM Standards, Vol. 7.01, West Conshohocken, PA,

USA, 2010, pp 262–268.28. Annual Book of ASTM Standards, Vol. 8.01, West Conshohocken, PA,

USA, 2013, pp 1–13.29. Annual Book of ASTM Standards, Vol. 8.03, West Conshohocken, PA,

USA, 2013, pp 1–11.30. AATCC Test Method 124-2011, Smoothness Appearance of Fabrics after

Repeated Home Laundering; AATCC: Research Triangle Park, NC, USA, 2011.

31. Allcock, H. R.; Hartle, T. J.; Taylor, J. P.; Sunderland, N. J. Macromolecules 2001, 34 (12), 3896–3904.

32. Carriedo, G. A.; Fernández-Catuxo, L.; García Alonso, F. J.; Gómez-Elipe, P.; González, P. A., Macromolecules 1996, 29 (16), 5320–5325.

33. Li, Q.; Zhang, Z.; Zhong, C.; Liu, Y.; Zhou, Q, Fluid Phase Equilibria 2003, 207 (1–2), 183–192.

34. Sanli, D.; Erkey, C. The Journal of Supercritical Fluids 2013, 74 (1), 52–60.

35. Erkey, C. Thermodynamics of Mixtures of Metal Complexes. In Supercritical Fluids and Organometallic Compounds: From Recovery of Trace Metals to Synthesis of Nanostructured Materials; Kiran, E., Ed.; Elsevier Inc.: Amsterdam, The Netherlands, 2011; pp 21–40.

36. McHugh, M.; Paulaitis, M. E. Journal of Chemical & Engineering Data 1980, 25 (4), 326–329.

37. Easson, M.; Condon, B.; Yoshioka-Tarver, M.; Childress, S.; Slopek, R.; Bland, J.; Nguyen, T.; Chang, S.; Graves, E. AATCC Review 2011, 11 (6), 60–66.

38. Schartel, B. Uses of Fire Test in Materials Flammability Development. In Fire Retardancy of Polymeric Materials; CRC Press: Boca Raton, FL, USA, 2010; pp 387–420.

39. Benin, V.; Durganala, S.; Morgan, A. B. Journal of Materials Chemistry 2012, 22 (3), 1180–1190.

40. Yang, C. Q.; He, Q. Journal of Analytical and Applied Pyrolysis 2011, 91 (1), 125–133.

41. Yoshioka-Tarver, M.; Condon, B. D.; Easson, M. W.; Nam, S.; Fortier, C. A.; Madison, C. A.; Ingber, B. F. AATCC Review 2012, 12 (5), 52–59.

42. Howell, B. A.; Dumitrascu, A. Comparison of the Impact of Phosphorus and Phosphorus/Nitrogen on the Flammability of Styrenic Oligomers. In Fire and Polymers VI: New Advances in Flame Retardant Chemistry and Science; American Chemical Society: Washington, DC, USA, 2012; pp 235–250.

43. Shafizadeh, F. Thermal Degradation of Cellulose. In Cellulose Chemistry and Its Applications; Nevell, T. P., Zeronian, S. H., Eds.; Ellis Horwood Limited Inc.: Chichester, UK, 1985; pp 266–289.

44. Ayeni, N. A.; Adeniyi, A; Abdullahi, N. N.; Bernard, E.; Ogunleye, A. O. Bayero Journal of Pure and Applied Sciences 2012, 5 (2), 51–55.

45. Xing, W.; Zhang, P.; Song, L.; Wang, X.; Hu, Y. Materials Research Bulletin 2014, 49 (1), 1–6.

46. Liu, Y.; Zhang, Y.; Cao, Z.; Fang, Z. Fire Safety Journal 2013, 61 (1), 185–192.

47. Jindasuwan, S.; Sukmanee, N.; Supanpong, C.; Suwan, M.; Nimittrakoolchai, O.; Supothina, S. Applied Surface Science 2013, 275 (1), 239–243.

48. Kivotidi, S.; Tsioptsias, C.; Pavlidou, E.; Panayiotou, C. Journal of Thermal Analysis and Calorimetry 2013, 111 (1), 475–482.

Authors Krystal R. Fontenot is a research chemist/postdoctoral scientist with the USDA at the Southern Regional Research Center in New Orleans, LA, USA. Her research interests include the application of phosphorus and nitrogen based flame retardants to cotton textiles and study of peptide-cellulose nanocrystal conjugates as elastase sensors for cotton gauze.

Michael W. Easson (corresponding author) is a research chemist with the USDA at the Southern Regional Research Center in New Orleans, La, USA. Easson’s research inter-ests include innovative textile processing methods, biomass conversion to ethanol-based biofuel, and the synthesis and application of organophosphorus flame retardants to cotton fabrics.

Michael W. Easson, USDA-ARS-SRRC, Cotton Chem-istry and Utilization Research,1100 Robert E. Lee Blvd., New Orleans LA 70124, USA; phone +1.504.286.4493; fax +1.504.286.4390; michael.easson@ars.usda.gov.

November/December 2014 Vol. 1, No. 6 | 26

AATCC Journal of Research

DOI: 10.14504/ajr.1.6.3