Dyeing of polyethylene terephthalate fibers with a disperse dye in supercritical carbon dioxide

Abstract

A 2⁴ factorial design of experiments complemented with a central point was performed to examine the influence of operating factors on color strength and color fastness of polyethylene terephthalate (PET) fibers dyed with the Disperse Orange 30 dye in supercritical CO₂. The effects of temperature, pressure, dyeing time and mass ratio between the dye and PET introduced in the dyeing chamber (α ratio) were considered. An additional set of kinetic results of color strength was obtained at the optimum condition in terms of pressure at the already presented temperatures and α ratios. A significant statistical effect of all the investigated factors on the color was observed, but except for the temperature, the influence of the same variables on wash fastness was negligible (p < 0.05). The color results expressed in terms of K/S from 2.4 to 21.8 revealed that the use of supercritical CO₂ as a solvent for the dye is a rapid and reliable alternative procedure for dyeing of PET fibers with the Disperse Orange 30 dye. The results of wash fastness currently obtained (i.e. 4.69 ± 0.18) support the use of ScCO₂.

Keywords

dyes supercritical fluids carbon dioxide textiles polyethylene terephthalate

The use of polyethylene terephthalate (PET) fibers in the textile industry, alone or blended with cotton, has increased substantially around the world.¹ It essentially happens because PET presents several important fabric properties, such as dimensional stability and anti-wrinkling.¹ Nowadays, almost 80% of the worldwide production of PET, which is close to 48 million tons, has been exclusively marketed to the production of fabrics.² In Brazil, where the textile manufacturers are responsible for the fifth largest global fabric production, the scenario is not different.³

The great volume of discharged wastewater is a classical problem of the current technology applied in the textile industry.⁴ It is not exclusively associated with the growing use of such a particular synthetic fiber (PET), but with the expansion of the textile industry in general. The major reason for such a drawback is the volume of water specially used in the conventional dyeing operation of fabrics. A textile process typically consume between 131⁵ and 150 m³ of water per ton of processed fabric.⁶ In the case of PET fibers, drying is not a particularly time-consuming operation.^1,7 However, it represents another major concern in the textile industry since energy is primarily consumed for reducing the moisture content of the fibers early colored by using a wet process.^5,8 In fact, green technologies have been suggested in order to reduce the overall consumption of energy in the textile process to about 40,000 kJ per kg of fabric.⁵

Among the considered alternative processes, the operation of dyeing with disperse dyes in supercritical carbon dioxide instead of water has been particularly investigated.^9–14 The use of ScCO₂ as a solvent for the dye contributes to reduce the mentioned excessive consumption of water typically found in the textile mill, with significant impact on the volume of discharged wastewater.^4,7,15 At this point, it is important to remember that effluents rich in synthetic dyes are not easily treated (advanced oxidative processes are often required) and present several negative pollutant characteristics (reduction of photosynthesis and oxygen solubility are the most common hazards to aquatic organisms) (e.g. Benincá et al.,^16,17). Energy saving by avoiding the operation of drying of fabrics preliminary taken to a water-dye bath is another relevant advantage of using supercritical carbon dioxide.^7,8,14,15

In this framework, the main aim of the current investigation was to evaluate the use of supercritical carbon dioxide for dyeing of PET fibers with the Disperse Orange 30 dye. In particular, the temperature, pressure, dyeing time and mass ratio between the dye and PET introduced in the dyeing chamber (from this point referred to as the α ratio) were varied at two levels. An additional test repeated five times at the same condition (central point, that is, T = 363 K, P = 17.5 MPa, t = 14,400 s, α = 1.5) complemented the adopted 2⁴ factorial design of experiments. Dyeing quality was based on results of color strength and wash fastness, as typically done in dyeing studies.^9,11–13 Kinetic results of color strength were also obtained at the maximum examined pressure by changing the temperature and α ratio.

Materials and methods

Carbon dioxide (min. 99.5%) marketed by Praxair was used without further purification. Disperse Orange 30 (CAS number 5621-31-4, IUPAC name 2-[N-(2-cyanoethyl)-4-[(2,6-dichloro-4-nitrophenyl)diazenyl]anilino]ethyl acetate, MW = 450.27 g mol⁻¹) was purchased from Sinochem Jiangsu I/E Corp (Beijing, China). Its molecular structure is shown in Figure 1. Such a dye has already been used in supercritical dyeing and its solubility in supercritical CO₂ has been measured by other authors.^18–21 The fabric (129.43 g m⁻²) used in all the dyeing tests was provided by Estação da Malha (Apucarana, Brazil).

Figure 1.

Molecular structure of the Disperse Orange 30 dye (CAS number 5621-31-4, IUPAC name 2-[N-(2-cyanoethyl)-4-[(2,6-dichloro-4-nitrophenyl)diazenyl]anilino]ethyl acetate, MW = 450.27 g mol⁻¹).

Commercial textile dyes are not pure,^22,23 but often have dispersing agents (e.g. surfactants) whose role is to increase the dye solubility in water.²³ In fact, the extensive use of such additives is an additional disadvantage of the conventional dyeing process when compared to the dyeing with supercritical CO₂.^23,24 The latter needs no dispersing agents, so the use of ScCO₂ avoids the problem of treating a large volume of wastewater containing such hard-to-destroy additives.^23,24 In this framework, before using the dye for the purposes of PET dyeing, it was solubilized in acetone of analytical grade (Merck) and filtered on a paper filter (28 µm, 80 g m⁻², J. Prolab) where surfactants, as well as other organic and inorganic impurities, were retained. The absorbances of the filtered and unfiltered solutions were measured in the wavelength range from 350 to 550 nm. The dye filtration had a negligible influence on the wavelength of maximum absorbance, which was close to 414 nm in both circumstances.

The dyeing experiments were performed in a cylindrical stainless steel chamber with a volume close to 1.3 × 10⁻⁴m³. It was equipped with a calibrated T-type thermocouple, a pressure transducer (Smar, LD301) and a variable-speed mixer (always operated at 300 rpm) connected to an impeller (WEG, 373 W). A proportional-integral-derivative (PID) controller of temperature (COEL, HW1440) and an externally heating element were responsible for keeping the dyeing medium at the desired temperature. A needle valve (HIP, Model 15-11AF1) was employed to regulate the flow rate of CO₂ pumped into the vessel by a syringe pump (Teledyne, ISCO 500). A more detailed representation of the bench-scale unit and apparatus is shown in Figure 2.

Figure 2.

Schematic of the apparatus used for dyeing of polyethylene terephthalate fibers with the disperse Orange 30 dye in supercritical carbon dioxide. C1: cylinder of CO₂; V1, V2, V3, V4: needle valves; BT: thermostatic bath; PP: syringe pressure pump; DT: display of temperature; DP: display of pressure; R: dyeing vessel, TC: T-type thermocouple; ER: heating resistive elements; M: electric motor; MS: stirrer.

The experiments were basically carried out in order to have a perfect well-known mixture of the dye, fabric and CO₂ in a batch apparatus under controlled conditions of temperature and pressure for the entire time of dyeing. It was essentially done by introducing a known mass of the investigated dye at the bottom of the dyeing chamber, where samples of fabrics early placed in a stainless steel basket were also added. The basket was arranged in a way to avoid any preliminary direct contact between the dye and fabric. The dyeing vessel was closed and an amount of carbon dioxide dependent on the temperature and desired pressure (calculated with the Peng–Robinson equation) was introduced into it. Based on such a condition of temperature and pressure, the mass of dye used in the experiments was always early estimated in order to have a constant molar ratio of dye to ScCO₂ close to 9.3 × 10⁻⁶.

A 2⁴ factorial design of experiments complemented with a central point replicated five times was adopted to evaluate the influence of temperature, pressure, dyeing time and α ratio on color strength and color fastness. It means that a set of 17 pairs of such responses were obtained at 17 different conditions of dyeing in supercritical carbon dioxide. Table 1 summarizes the operating conditions in terms of the low and high values of the four natural investigated variables.

Table 1.

Summary of the investigated operating conditions during the dyeing of polyethylene terephthalate with Disperse Orange 30 in supercritical carbon dioxide

Independent variables	Units	Lowest considered value	Highest considered value
Temperature	K	353	373
Pressure	MPa	15	20
Dyeing time	s	7200	21,600
α ratio	%	1	3

An additional set of experimental results of color strength was obtained to study the kinetics of dyeing. In particular, the dyeing experiments earlier carried out for 7200 and 21,600 s were repeated for 3600 and 18,000 s at the same operating conditions presented in Table 1. The only exception was the pressure, which was always kept at the upper investigated limit, that is, 20 MPa. Based on these further results, four plots of color strength against dyeing time may be built to better comprehend the influence of temperature and α ratio on the kinetics of dyeing.

Color strength of the dyed fabric samples was analyzed by using a spectrophotometer (Konica Minolta, CM-3610d). Reflectance measurements are often applied to determine the color strength to dyes adsorbed in a substrate, so all the dyed fabric samples were directly monitored at the wavelength of maximum absorbance. The wavelength of maximum absorbance was early determined from the recorded absorbance spectra of an Orange 30 acetone solution in the wavelength range from 350 to 550 nm. Based on three readings of reflectance for the same sample at 414 nm, average results of K/S were calculated by using the extensively reported expression of Kubelka–Munk (e.g. van der Kraan et al.,¹¹ Hou et al.¹⁵]). In Equation (1) R is the measured reflectance, while K and S are the light absorption and the scattering coefficients of the dyed fabric, respectively. According to the literature,^13,25 there exists a well-established linear dependence of K/S on the color of a solid surface:

K / S = \frac{(1 - R) 2}{2 R}

(1)

A standard method was applied to determine wash fastness.²⁶ It involved the immersion of the fabric in a polyvinyl chloride (PVC) tube filled with a mass of water 50 times greater than that of the sample, where sodium carbonate and soap powder without softener were added to have a concentration of these chemical equal to 2 and 5 kg m⁻³, respectively. The tube containing the fabric and aqueous solution was sealed and taken to a washing-fastness apparatus (KIMAK, Tubetest A-T2) operated at 50 rpm and at 323 K for 1800 s. It is essentially a water bath equipped with a shaker platform that accommodates simultaneously a set of washing tubes. The fabric sample was finally removed from the PVC tube, rinsed with distilled water and dried at room conditions (T = 295–300 K, relative humidity = 65% ± 5%). The gray change scale was used to determine the color alteration of all the samples treated at the just described washing conditions.²⁷ The results were obtained with a spectrophotometer (Konica Minolta, CM-3610d) in the fastness color range from 1 to 5, where the lower and upper limits indicate most color and no color was lost after washing, respectively.

Results and discussion

Table 2 reports the experimental results of color strength and wash fastness at all the investigated conditions. A statistical model represented by Equation (2) was suggested in order to better understand the influence of the considered factors on the color strength. A Levenberg–Marquardt algorithm for non-linear least-squares was applied to obtain the parameters of this regression model. For a 95% confidence level, the t-test supports the importance of all the main effects and the effect of interaction between the temperature and pressure on K/S. It essentially occurs because the absolute values of the model parameters E_T, E_P, E_t, E_α and E_TP, which are related to the effect of the factors (e.g. Montgomery²⁸), are always higher than their uncertainties (σ = 3.38) (see Figure 3). From a practical point of view, it means that the change in response produced by a modification in the level of the investigated factors is higher than the uncertainty in the measurements of color strength. In Equation (2), E_T is the effect of temperature on K/S, E_P is the effect of pressure on K/S, E_t is the effect of dyeing time on K/S, E_α is the effect of α ratio on K/S and E_TP is the effect of the interaction between temperature and pressure on K/S. Equations (3)–(6) give the relation between the coded (X) and natural (T, P, t, α) variables:

K / S = (10.58 \pm 0.66) + \frac{E_{T}}{2} (X_{T}) + \frac{E_{P}}{2} (X_{P}) + \frac{E_{t}}{2} (X_{t}) + \frac{E_{α}}{2} (X_{α}) + \frac{E_{TP}}{2} (X_{T} X_{P})

(2)

X_{T} = \frac{T - [(T_{low} + T_{high}) / 2]}{[(T_{high} - T_{low}) / 2]}

(3)

X_{P} = \frac{P - [(P_{low} + P_{high}) / 2]}{[(P_{high} - P_{low}) / 2]}

(4)

X_{t} = \frac{t - [(t_{low} + t_{high}) / 2]}{[(t_{high} - t_{low}) / 2]}

(5)

X_{α} = \frac{α - [(α_{low} + α_{high}) / 2]}{[(α_{high} - α_{low}) / 2]}

(6)

Figure 3.

Main (E_T = 5.47, E_P = 5.49, E_t = 3.50, E_α = 6.35) and interaction (E_TP = 3.53, E_Tt = 0.51, E_T_α = 2.24, E_Pt = 0.99, E_P_α = 1.74, E_t_α = 0.33) effects of the investigated variables (T, P, t, α ratio) on the color strength. Shadow bars: significant effects (p < 0.05); white bars: non-significant effects (p < 0.05); dashed line: uncertainty (σ = 3.38) in the statistical model parameters (Equation (2)).

Table 2.

Design matrix for estimating the interaction and main effects of temperature, pressure, dyeing time and α ratio on color strength and color fastness of polyethylene terephthalate fibers

Run	Coded independent variables				Dependent variables
Run	X_T = f(T)	X_P = f(P)	X_t = f(t)	X_α = f(α)	K/S	F (fastness)
1	−1	−1	−1	−1	2.42	4.7
2	+1	−1	−1	−1	2.59	4.9
3	−1	+1	−1	−1	5.18	4.6
4	+1	+1	−1	−1	10.00	4.8
5	−1	−1	+1	−1	4.21	4.5
6	+1	−1	+1	−1	9.82	4.9
7	−1	+1	+1	−1	8.26	4.3
8	+1	+1	+1	−1	10.61	4.8
9	−1	−1	−1	+1	6.15	4.5
10	+1	−1	−1	+1	8.11	4.9
11	−1	+1	−1	+1	8.56	4.3
12	+1	+1	−1	+1	21.47	4.9
13	−1	−1	+1	+1	11.60	4.8
14	+1	−1	+1	+1	11.63	4.7
15	−1	+1	+1	+1	10.22	4.8
16	+1	+1	+1	+1	26.16	4.5
17 ^a	0	0	0	0	13.04 ± 0.03	4.7 ± 0.1

Five replicates.

Figure 3 not only clearly shows this relevant aspect of the statistical analysis, but it also reports the values of the tuned model constants. Because all the main effects are positive, it is known that an increase of temperature, pressure, dyeing time and α ratio always enhance the color of the fabric dyed in ScCO₂. A comparative analysis among E_T, E_P, E_t, E_α and E_TP may lead one to believe that the α ratio has the major impact on the color strength, and that the influence of temperature and pressure are in the same magnitude, while the dyeing time and the interaction between temperature and pressure has a minor, but still important, role on dyeing in ScCO₂. However, a more detailed examination of these results reveals that for a 95% confidence level there is no difference among the effects of the considered factors (|E_i − E_j| _i _≠ _j < σ, e.g. |E_T − E_P| = 0.02 < 3.38, |E_T − E_t| = 1.97 < 3.38).

In Figure 4 is presented a comparison between the experimental and calculated color strength at all the dyeing operating conditions. It confirms the reliability of the statistical model (Equation (2)), which is able to explain approximately 80% of the variability in the response due to changes in the examined factors. A classical response surface plot was currently replaced by an ordinary Cartesian representation, where the color strength is shown in the ordinate and the four different x-axes express the independent variables in a coded scale (i.e. from −1 to +1). The most important aspect to emerge from Figure 4 is that the highest color strength is obtained at the highest values of the investigated factors (X_T = X_P = X_t = X_α = +1, i.e. T = 373 K, P = 20 MPa, t = 21,600 s and α = 3), while the worst operating condition for dyeing with Disperse Orange 30 in ScCO₂ is that involving the set of minimum values for the independent variables (X_T = X_P = X_t = X_α = −1, i.e. T = 353 K, P = 15 MPa, t = 7200 s and α = 1). Such behavior was already expected, since the main effect of the factors on K/S was always positive.

Figure 4.

Experimental (symbols) and calculated (solid line) (Equation (2)) color strength of the dyed fabrics as a function of the investigated factors.

Starting from the best dyeing condition, Figure 4 also evidences a substantial drop of the color strength when the dyeing time is reduced from 21,600 (X_t = +1) to 7200 s (X_t = −1). However, a still more marked reduction of K/S is observed when the dyeing is carried out at the central point (X_T = X_P = X_t = X_α = 0) instead of at the highest values of the examined factors (X_T = X_P = X_t = X_α = +1). From the third highest color strength the responses go down almost linearly as function of combinations of the independent coded variables up to a condition where all of them are equal to −1.

The kinetics of dyeing at the best operating condition in terms of pressure (20 MPa) were experimentally examined at the temperatures and α ratios already defined in Table 1. The plots of color strength as a function of time for all conditions are reported in Figures 5 and 6. A first−order rate model given by Equation (7) was adopted to describe such experimental results of dye adsorption. It is a separable ordinary differential equation whose analytical solution is easily found (Equation (8)). Based on color strengths at equilibrium from Table 2 (runs 7, 8, 15 and 16) the rate constant (β) was tuned to the experimental results by applying the Levenberg–Marquardt algorithm for non-linear least-squares. Both the experimental and calculated results promptly confirm the small influence of time on color strength, already evidenced in Figure 3 for a dyeing time between 7200 and 21,600 s. It essentially happens because the results at 7200 s were already close to those at equilibrium. In fact, the suggested simplified kinetic model was able to reproduce the entire set of curves by involving a constant β:

\frac{d (K / S)}{d t} = β [(K / S) - (K / S) e]

(7)

K / S = (K / S) e [1 - \exp (- β t)]

(8)

Figure 5.

Kinetic of dyeing at P = 20 MPa and T = 353 K for an α ratio close to 1% (squares and dashed line) and 3% (circles and solid line). Symbols: experimental results; lines: calculated with Equation (8) for β = 2.38 × 10⁻⁴ s⁻¹ and (K/S) _e from Table 2 (runs 7 and 15).

Figure 6.

Kinetic of dyeing at P = 20 MPa and T = 373 K for an α ratio close to 1% (squares and dashed line) and 3% (circles and solid line). Symbols: experimental results; lines: calculated with Equation (8) for β = 2.38 × 10⁻⁴ s⁻¹ and (K/S) _e from Table 2 (runs 8 and 16).

In the case of wash fastness, one may rapidly check from Table 2 that the results were less dependent on the considered factors, that is, the measured wash fastness ranges only from 4.3 to 4.9. Based on a standard deviation close to 0.1 (see run 17 in Table 2), the uncertainty in wash fastness is approximately 0.3 for a level of confidence of 95% and a degree of freedom equal to 4. It means that the range of variation of wash fastness in Table 2 is of the same order of experimental uncertainty in wash fastness. An analysis analogous to that already performed for the color strength reveals that except for the dyeing temperature, all the factors have negligible effects on the loss of PET color. The enhancement of fastness properties by increasing the temperature when dyeing PET fibers with different disperse dyes in supercritical CO₂ is confirmed in the literature (e.g. Wang and Lin²⁹). Equation (9) is a linear model build by applying such a statistical approach. The positive effect of temperature (0.24) again indicates that an increase of X_T increases wash fastness. The consistency of the model is verified in Figure 7, where the experimental and calculated results of wash fastness are presented. In spite of the changes in pressure, dyeing time and α ratio, Equation (9) well describes the dependence of F (wash fastness) on temperature:

F = (4.69 \pm 0.03) + (\frac{0.24 \pm 0.18}{2}) (X_{T})

(9)

Figure 7.

A comparison between experimental (symbols) and calculated (solid line) (Equation (9)) wash fastness of the dyed fabrics in different coded temperatures.

The arithmetic mean of all the readings of wash fastness presented in Table 2 is close to 4.7. It is also known that an F equal to 5 is attributed to a fabric whose color was not minimally altered after washing. Based on these information, it is possible to state that 94% of the dye adsorbed during the operation of dyeing in supercritical carbon dioxide was kept after the washing test. It represents a further evidence of the reliability of applying ScCO₂ to replace water during the operation of dyeing of PET fibers in a plant scale. Analogous results of wash fastness obtained at similar dyeing conditions confirm the consistency of the adopted procedure (e.g. Schmidt et al.⁹).

A final remark concerns the well-known effect of supercritical CO₂ on the glass transition temperature (T_g).^14,24,30 In the particular case of PET, T_g is in the range of 353–398 K.¹⁴ However, there are many evidences in the literature to suggest that it may be drastically reduced in the presence of supercritical CO₂ (e.g. Beltrame et al.³⁰): by at least 20–30 K.¹⁴ So, in the current case a T_g from 323 to 378 K is expected, while the dyeing temperature was between 353 and 373 K. In summary, even if a T > T_g was possible and the positive influence of having it on dyeing of textiles is well-established in the literature,^14,24,30 there are insufficient elements to confirm it in the present work.

Conclusion

A 2⁴ factorial design of experiments was proposed in order to investigate the effect of process variables on the color strength and wash fastness of PET fibers dyed with Disperse Orange 30 in supercritical carbon dioxide. The responses were obtained in the range of temperature from 353 to 373 K at 15, 17.5 and 20 MPa for 7200, 14,400 and 21,600 s. The α ratio was changed from 1 to 3%. For a 95% confidence level a significant and positive influence of temperature, pressure, dyeing time, α ratio and the interaction between T and P on color strength emerged from the experimental results of K/S. Based on the set of wash fastness determined by applying the gray scale a statistical model to estimate the effect of the examined factors on such a response was also suggested. It revealed a negligible role of pressure, dyeing time and α ratio on wash fastness, while the effect of temperature was positive and statically important (p < 0.05). Both the results support the practical use of supercritical carbon dioxide for dyeing of PET fibers with the investigated disperse dye. On the whole, the optimal dyeing quality was obtained when the PET fibers were dyed with an α ratio equal to 3% at 373 K and 20 MPa for 21,600 s. However, good results in terms of the examined color parameters may be obtained at the same conditions for dyeing time shorter than 21,600 s and longer than 7200 s.

Footnotes

Funding

This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

References

Zaman

Liu

Xiao

. Hydrophilic modification of polyester fabric by applying nanocrystalline cellulose containing surface finish. Carbohydr Polym 2013; 91: 560–567.

Zou

Reddy

Yang

. Reusing polyester/cotton blend fabrics for composites. Compos B Eng 2011; 42: 763–770.

IEMI. Brasil têxtil: sectorial report of the Brazilian Textile Industry. Instituto de Estudos e Marketing Industrial, São Paulo, 2011.

Atimtay AT and Sikdar SK. Color removal with natural materials for reuse of wastewater in textile industry. Security of Industrial Water Supply and Management Book Series: NATO Science for Peace and Security Series C-Environmental Security, 2010.

Kocabas

Yukseler

Dilek

. Adoption of European Union’s IPPC Directive to a textile mill: analysis of water and energy consumption. J Environ Manage 2009; 91: 102–113.

Liu

. Adoption of European Union’s IPPC Directive to a textile mill: analysis of water and energy consumption. Desalination 2010; 258: 229–232.

Hurren C, Li Q and Wang XA. The impact of textile wet colouration on the environment in 2011. Eco-Dyeing, Finishing and Green Chemistry Book Series: Advanced Materials Research, 2012.

Wang JP and Liu JQ. Energy savings via fast drying in textile industry part I: theoretical discussion and technical strategies. In: second international conference on advanced textile materials & manufacturing technology, Hangzhou, China, 20–24 October 2010, pp.377–381. Hangzhou: Zhejiang University Press.

Schmidt

Bach

Schollmeyer

. The dyeing of natural fibres with reactive disperse dyes in supercritical carbon dioxide. Dyes Pigments 2003; 56: 27–35.

10.

Banchero

Ferri

. Simulation of aqueous and supercritical fluid dyeing of a spool of yarn. J Supercrit Fluids 2005; 35: 157–166.

11.

van der Kraan

Fernandez Cid

Woerlee

. Dyeing of natural and synthetic textiles in supercritical carbon dioxide with disperse reactive dyes. J Supercrit Fluids 2007; 40: 470–476.

12.

Long

Zhao

. Investigations on the level dyeing of fabrics in supercritical carbon dioxide. J Supercrit Fluids 2011; 57: 80–86.

13.

Long

Xiao

. Dyeing of cotton fabric with a reactive disperse dye in supercritical carbon dioxide. J Supercrit Fluids 2012; 69: 13–20.

14.

Banchero

. Supercritical fluid dyeing of synthetic and natural textiles - a review. Color Technol 2012; 129: 2–17.

15.

Hou

Chen

Dai

. Using supercritical carbon dioxide as solvent to replace water in polyethylene terephthalate (PET) fabric dyeing procedures. J Cleaner Prod 2010; 18: 1009–1014.

16.

Benincá

Peralta-Zamora

Cardoso Camargo

. Kinetics of oxidation of ponceau 4R in aqueous solutions by Fenton and photo-Fenton processes. Reac Kinet Mech Cat 2012; 105: 293–306.

17.

Benincá

Peralta-Zamora

Granhen Tavares

. Degradation of an azo dye (ponceau 4R) and treatment of wastewater from a food industry by ozonation. Ozone Sci Eng 2013; 35: 295–301.

18.

Shamsipur

Karami

Yamini

. Solubilities of some aminoanthraquinone derivatives in supercritical carbon dioxide. J Chem Eng Data 2003; 48: 71–74.

19.

Draper

Montero

Smith

. Solubility relationships for disperse dyes in supercritical carbon dioxide. Dyes Pigments 2000; 45: 177–183.

20.

Baek

Kim

Lee

. Density correlation of solubility of C. I. disperse orange 30 dye in supercritical carbon dioxide. Korean J Chem Eng 2004; 21: 230–235.

21.

dos Santos

Mazzer

Machado

. High-pressure phase behaviour of the system (CO₂ + C.I. disperse orange 30 dye). J Chem Thermodyn 2012; 48: 284–290.

22.

Saus

Knittel

Schollmeyer

. Dyeing of textiles in supercritical carbon dioxide. Textil Res J 1993; 63: 135–142.

23.

Liu

Zhang

Liu

. Supercritical CO₂ dyeing of ramie fiber with disperse dye. Ind Eng Chem Res 2006; 45: 8932–8938.

24.

Chang

Bae

Shim

. Dyeing of PET textile fibers and films in supercritical carbon dioxide. Korean J Chem Eng 1996; 13: 310–316.

25.

Fan

Hoskote

Hou

. Reduction of colorants in nylon flock dyeing effluent. J Hazard Mater 2004; B112: 123–131.

26.

ABNT-NBR 13062. Determination of color fastness to washing action. Test number 3. Brazilian Association of Technical Standards, 1994.

27.

ISO 105-A05. Instrumental assessment of change in color for determination of grey scale rating. Tests for color fastness. ISO, 1996.

28.

Montgomery

. Design and analysis of experiments, 5th ed. New York: Wiley, 2001.

29.

Wang

Lin

. Scouring and dyeing of polyester fibers in supercritical carbon dioxide. J Chem Eng Japan 2001; 34: 244–248.

30.

Beltrame

Castelli

Selli

. Morphological changes and dye uptake of poly(ethylene terephthalate) and 2,5-cellulose diacetate immersed in supercritical carbon dioxide. Dyes Pigments 1998; 39: 35–47.