Binding and release of odor compounds from textiles: Changing fiber selection for apparel

Abstract

Understanding odor volatiles known to constitute those emanating from the human body and how these interact with textiles is relevant to diverse interested parties because of changing fiber use, a better understanding of apparel life cycles including cleaning and the potential for fiber/textile re-use. This paper extends the application of our proton transfer reaction-mass spectrometry model system on adsorption and release behavior of fiber types typical of next-to-skin uses to include both viscose and other volatile organic compounds in body odor for which release has not previously been studied (hexanoic acid, acetone, cyclohexanone, hexanal, methyl butanoate, ethyl benzene, 1-octanol, decanal, butanoic acid). The current findings confirmed release patterns of different fiber types found in our earlier studies – low release of volatile organic compounds from cotton and wool, but higher release from polyester – and made a new finding of low release from viscose. Adsorption characteristics were different given the different volatile organic compounds analyzed. Viscose showed better adsorption characteristics for primarily polar volatile organic compounds, but was otherwise very similar to cotton.

Keywords

odor properties materials knitting fabrication testing

Environmental issues continue to affect fiber availability, use and care practices, with natural fibers such as cotton, wool and silk often perceived more favorably than synthetics such as polyester and polyamide for next-to-skin applications. This perception is attributable in part to their being of plant or animal origin, and for those of animal origin, their comparative scarcity in relation to global total fiber production.¹ The potential for fiber re-use including up-cycling, and their efficacy of disposal at end of life also contribute to positive perceptions of these groups of fibers. Other cellulose-based fibers such as viscose are manufactured from a variety of readily available plant-based sources (e.g. pine, spruce, wood pulp, bamboo, hemp).^2,3 While of interest for textile end applications, much less is known about how these fibers/textiles such as viscose behave in use, particularly in relation to their interactions with the volatile organic compounds (VOCs) known to be present in human body odor. Results from two investigations which included viscose have been published: in one, t-shirt patches of fabric from viscose were worn, and using human sensory detection responses, were similar to cotton;⁴ and in the second, viscose (spun rayon) standard fabric was spiked with a mix of VOCs, and detection instrumentally showed rather high recovered VOC concentrations, with apparent confounding effects of fabric structure.⁵

Recent direct headspace measurement studies by our group^6,7 using proton transfer reaction-mass spectrometry (PTR-MS), assessed the adsorption characteristics of six VOCs present in human body odor by cotton, wool and polyester fibers/yarns. Phenol and ethanethiol were adsorbed similarly by all three fibers/yarns, wool was more effective in adsorbing cyclohexanone and decanal, and polyester more effective in adsorbing ethylbenzene and methyl butanoate.⁷ In the second study, the adsorption and release behavior of these same three fibers/yarns exposed to a further group of VOCs known to be present in human body odor (dimethyl disulfide, 2-propanethiol, benzaldehyde, nonanal, butanoic acid, 3-methyl-2-hexenoic acid) showed three different patterns: low relative adsorption and low overall release of the volatiles by cotton, high relative adsorption and continuous release of the volatiles by polyester, and high relative adsorption but low overall release by wool.⁶ Stimulated by changes in fiber use, in cleaning-related requirements and other emerging environmental matters (e.g. product re-cycling, concepts of circular manufacture⁸), we extended our investigations of adsorption and release behavior to include viscose rayon and used the same VOCs as in the study by Yao et al.⁷ (i.e. hexanoic acid, acetone, cyclohexanone, hexanal, methyl butanoate, ethyl benzene, 1-octanol, decanal, butanoic acid).

Methods

Materials

Four standard fabrics (James Heal, UK⁹) were used, viscose rayon (SDC Enterprises Ltd) and the three used in our earlier investigations (cotton, wool, polyester). All fabrics were washed and dried according to ISO 6330-2012¹⁰ then cut into 200 mm × 200 mm swatches. As before, to remove effects of fabric structure on adsorption and desorption, the fabrics were disaggregated to yarn form, loosely bundled to ca. 1.0 -g samples and held with a nylon thread. The yarn bundles were conditioned at 20 ± 2℃ and 60 ± 4% relative humidity for at least 24 hours according to ISO 139:2005/AMD1:2011.¹¹ The exact weight of the conditioned yarn bundles was measured prior to exposure to the test protocol.

The VOCs used in this study included compounds used as in the first investigation from our group, but in this earlier study desorption (now recognized as having practical value to end users) was not considered, and an extended range of VOC groups previously analyzed with the current method.^6,7 These VOCs were butanoic acid, hexanoic acid, 1-octanol, acetone sourced from McCormack Chemicals (Ireland) and cyclohexanone, hexanal, decanal, phenol, ethyl benzene, benzyl acetate and methyl butanoate from Sigma Aldrich (Germany). The body odor matrix was developed from the VOCs based on their individual responses to achieve the optimal ratio between detection limit and signal response, targeting signal intensities between 10³ and 10⁴ counts per second (cps). The composition of the body odor matrix is given in Table 1.

Table 1.

Composition of body odor matrix and target m/z used to measure volatile organic compounds

Name	Body odor matrix composition (%)	Target m/z
1-octanol	39.9	111
Phenol	4.9	95
Ethyl benzene	2.4	107
Cyclohexanone	2.7	99
Acetone	0.8	59
Methyl butanoate	1.2	103
Benzyl acetate	2.4	91
Decanal	23.9	97
Hexanal	15.9	101
Butanoic acid	1.2	89
Hexanoic acid	4.9	117

Experimental setup

An experimental setup previously used and described by Richter et al.⁶ was followed, and key points are repeated here for convenience of the reader. A schematic of the experimental setup is given Figure 1. It consisted of a three-glass bottle (C1–C3) system connected through polytetrafluoroethylene (PTFE)-coated spherical joints (11-mm internal diameter; MS19 and FS19 joint size as per SciLabware Limited, UK) with two PTFE valves (V1, V2). The experiment was carried out in an enclosed system conditioned at 33℃ to simulate human skin temperature in a thermo-neutral state.¹² One hundred microliters of the body odor matrix was added to bottle C1. Prior to exposure of the fiber/yarn specimens, a blank was measured. Exposure time was 18 hours for each fiber/yarn type (polyester, wool, viscose, cotton) and the empty control (no yarn/fiber). Adsorption measurements were carried out using a PTR-MS after the equilibrium time. Desorption measurements were carried out at 15, 30, 60, 90, 120, 180, 240, 300, 360 and 1350 minutes (22.5 hours) after the yarn bundle and its adsorbed volatiles had been transferred to a fresh empty bottle (volume 1 L, modified PTFE cap with two access points).

Figure 1.

Experimental setup: odor matrix delivery system for adsorption and release measurements as previously described by Richter et al.⁶

The PTR-MS (INOCON Technologie GmbH, Austria) operation conditions were replicated from the study by Richter et al.⁶ Again, this is repeated here for convenience of the readers: inlet temperature at 110℃, drift tube pressure 2.23 mbar, chamber temperature at 70℃, drift tube voltage at 600 V, field density 136–138 Td and a flow of 60 standard cubic centimeters per minute (sccm). To avoid signal fluctuation due to pressure changes in the sampling bottles, a make-up gas (synthetic dry air; IG BOC, New Zealand) was supplied when sampling at a flow rate of 60 sccm. The PTR-MS was operated in multiple ion detection mode recording 20 cycles of the following masses with a dwell time of 500 ms each. Instrument performance was monitored through the masses of m/z 21.0 (H₃¹⁸O⁺), m/z 30.0 (NO), m/z 32.0 (O₂) and m/z 37.0 (water cluster).

Analysis

Calculation of adsorption was based on the following formula:

\begin{matrix} relative adsorption \\ = \frac{{average ({ncps}_{(control)} - {ncps}_{(blank)}) - average (\frac{ncps (sample) - ncps (blank)}{m_{yarn bundle}})}}{average (ncps (control) - ncps (blank))} \end{matrix}

where ncps is normalized counts per second.

Averaged signal intensities (from the last 15 cycles per measurement) were normalized by correcting for variations in the primary ion (H₃O⁺) and water cluster (H₂O.H₃O⁺) using the isotopologue at m/z 21 and m/z 37, respectively. Blank measurements of each system were subtracted from the sample measurement and normalized to the respective weight of the fiber/yarn bundle for all calculations. Relative adsorption capabilities were calculated as a percentage of the average in a control system without fiber/yarn bundles. All tests were carried out in triplicate. Univariate analysis of variance (ANOVA) was carried out to identify differences among the types of fiber/yarns and changes over time with acceptance of significance level of p ≤ .05.

Results and discussion

The study sought to investigate the adsorption or binding capabilities of different fiber types as well as the release of VOCs. Results for adsorption and desorption are presented separately.

Adsorption

The binding capacities of viscose and the three other fiber types (cotton, wool, polyester) over an exposure time of 18 hours were investigated for 11 different VOCs. The VOCs of acetone, cyclohexanone, hexanal, methyl butanoate and ethyl benzene were not adsorbed in significant amounts by any of the fiber types. In contrast, 1-octanol, butanoic acid, benzyl acetate, phenol, decanal and hexanoic acid were adsorbed in significant amounts. The adsorption results for all VOCs are given in Table 2. VOCs that showed adsorption in significant amounts are displayed in Figure 2 (results expressed as relative adsorption).

Figure 2.

Boxplot of overall relative adsorption of 1-octanol, phenol, benzyl acetate, decanal, butanoic acid and hexanoic acid on four fiber types: polyester, wool, viscose and cotton (n = 3). Letters representing groups of Tukey’s LSD test with ‘a’ indicating not significantly different from control measurements.

Table 2.

Relative adsorption for volatile organic compounds for four fiber types. Results displayed as average with errors calculated as standard deviations (n = 3)

	Polyester	Wool	Viscose	Cotton	ANOVA^a (between fiber effects)
1-octanol	15% ± 5%	16% ± 7%	26% ± 6%	27% ± 5%	F_2,4 = 6.9, p = 0.06
Benzyl acetate	47% ± 4%	53% ± 3%	47% ± 3%	48% ± 3%	F_2,4 = 68.3, p ≤ 0.01
Decanal	19% ± 6%	45% ± 3%	44% ± 4%	46% ± 2%	F_2,4 = 48.9, p ≤ 0.01
Butanoic acid	36% ± 5%	73% ± 1%	81% ± 5%	73% ± 3%	F_2,4 = 143.9, p ≤ 0.01
Hexanoic acid	6% ± 9%	16% ± 7%	34% ± 7%	19% ± 7%	F_2,4 = 5.3, p = 0.015
Phenol	69% ± 1%	74% ± 4%	92% ± 7%	76% ± 3%	F_2,4 = 143.7, p ≤ 0.01
Acetone	2% ± 3%	5% ± 5%	13% ± 8%	11% ± 9%	F_2,4 = 1.3, NS
Cyclohexanone	6% ± 7%	18% ± 7%	19% ± 6%	18% ± 4%	F_2,4 = 4.0, NS
Ethyl benzene	9% ± 5%	0% ± 1%	5% ± 6%	6% ± 7%	F_2,4 = 2.4, NS
Hexanal	9% ± 6%	2% ± 2%	4% ± 5%	5% ± 6%	F_2,4 = 1.1, NS
Methyl butanoate	10% ± 5%	8% ± 8%	19% ± 7%	14% ± 4%	F_2,4 = 2.4, NS

ANOVA calculated based on the average adsorption values of four fiber types and control.

The overall relative adsorption by all fibers for 1-octanol was quite low (relative adsorption below 30%) but viscose and cotton showed significant adsorption (p ≤ 0.01 for both, Tukey’s LSD). Phenol adsorption was significant with high rates for all fiber types. Furthermore, viscose had the highest rate, adsorbing around 92% of the available phenol (p ≤ 0.05, Tukey’s LSD). Benzyl acetate was adsorbed in significant amounts ranging between 47 and 53% for all fiber types, but no difference in the adsorption capacity among the fibers was observed. All fiber types significantly adsorbed decanal, reaching relative adsorptions of around 45% for wool, viscose and cotton. Polyester, however, had a significantly lower adsorption capacity (p ≤ 0.05, Tukey’s LSD), adsorbing only 19%. Following the same trend as decanal but more strongly, a high adsorption of over 70% was found for butanoic acid by wool, cotton and viscose, while polyester showed a significantly lower adsorption (p ≤ 0.05, Tukey’s LSD) of just 47%. Results for hexanoic acid were again indeterminate: the adsorption of hexanoic acid by polyester, wool and cotton was not significant (all below 20%) and significant for viscose with adsorption of around 30% (p ≤ 0.05, Tukey’s LSD).

This study confirmed results from our earlier work on the adsorption of VOCs with a different experimental design.⁷ Comparable to results from this earlier adsorption study, we found the behavior of all fiber types to be similar, including the newly included viscose, only adsorbing the ketones (cyclohexanone and acetone) in small amounts. Also in line with the findings in the study by Yao et al., was that ethyl benzene showed no significant adsorption by wool and cotton (and viscose), but low adsorption by polyester (a similar trend although not significant in this study), and phenol had a very high relative adsorption by all fiber/yarn types. Results differing from the earlier study were observed for methyl butanoate with no significant adsorption in this study but previously a significant adsorption capacity for polyester. Furthermore, decanal was significantly less adsorbed by polyester in the present study than in Yao et al.'s investigation. Differences in the findings between the two studies may be accounted for by the slightly different experimental setup. In the study of Yao et al., the VOCs were introduced through a syringe which can thus be described as delivering the complete VOC load in a short amount of time (seconds). In contrast, in this study the VOCs were introduced to the fiber slowly (over an 18-hour period), and passively. Both in this and an earlier study,⁶ polyester was shown to adsorb VOCs but not hold them for a prolonged period of time. This may indicate that results showed high adsorption capacities for the direct, fast method, and no significant adsorption for the indirect, slow method. Variations among the different VOCs are also relevant.

For viscose, better adsorption characteristics were observed for polar VOCs, well represented through the high adsorption of phenol and low adsorption of 1-octanol, and thus similar to results for cotton. Viscose has a physico-chemical structure similar to cotton, but differences in the manufacturing process of viscose (additives, process conditions) are likely to influence the swelling and water-holding capacity of viscose and thus will affect the binding capacity of polar VOCs in particular.² Kamppuri et al. found viscose to have a higher water adsorption capacity than cotton, giving the larger pore size of viscose fibers as a possible reason for this.¹³ Abu-Rous et al suggested moisture-related properties of viscose in part accounted for the intensity of odor volatiles (and microorganisms) on worn t-shirts.¹⁴ With a possible greater capacity to hold water, the fiber would also have a higher capacity to bind/hold other polar compounds. Overall, viscose showed higher adsorption capabilities for water-soluble VOCs (phenol, butanoic acid, hexanoic acid) compared to the other fiber types tested. Similar differences in the adsorption capacity between butanoic acid and hexanoic acid have been reported by Wang et al. comparing odor retained on a range of polyester/wool blends.¹⁵

Desorption

The desorption behavior of the VOCs was examined in relation to the different fiber types (i.e. those VOCs exhibiting significant adsorption by any of the fibers: 1-octanol, benzyl acetate, decanal, butanoic acid, hexanoic acid and phenol). Results for 1-octanol, benzyl acetate and decanal are shown in Figure 3 and for butanoic acid, hexanoic acid and phenol in Figure 4. Both graphs display desorption over time (Figures 3(b) and 4(b)), and for a better comparison the absolute adsorbed amounts (Figures 3(a) and 4(a)).

Figure 3.

Absolute adsorbed amount (a) of normalized counts per second (ncps) after 18 hours of exposure and the respective release over time (b) from the fiber types for of 1-octanol, benzyl acetate and decanal. Error bars displayed as standard deviations (n = 3).

Figure 4.

Absolute adsorbed amount (a) of normalized counts per second (ncps) after 18 hours of exposure and respective release over time (b) from fiber types for butanoic acid, hexanoic acid and phenol. Error bars displayed as standard deviations (n = 3).

Considering the desorption of 1-octanol, overall significant differences among fiber types over time were observed (F_2,27 = 14.4, p ≤ 0.01). Viscose had a significantly higher initial release compared to the three other fiber types (p ≤ 0.05, Tukey’s LSD); however, a significant change was not observed over time for either viscose or cotton. Polyester and wool, although showing no significant adsorption, showed a steady release over time for 1-octanol (p ≤ 0.05, Tukey’s LSD). Diffusion from the yarn bundle over time is likely, suggesting no interaction between the fiber and the VOC. Although no significant differences among the fibers were observed in adsorption, there were significant differences in the release of benzyl acetate (F_2,3 = 130.9, p ≤ 0.01). Overall, wool was found to have a significantly lower release, and polyester to have a significantly higher release (p ≤ 0.05, Tukey’s LSD) compared to other fiber types. The initial release of benzyl acetate was low for all fiber types, with viscose, cotton and wool exhibiting no significant change over time. For polyester, however, a significant change over time was observed (F_2,9 = 6.4, p ≤ 0.05) with a maxima at 180 minutes (p ≤ 0.05, Tukey’s LSD). This result highlights polyester as a fiber type with weaker binding capabilities compared to other fibers. Manifesting this point were the results for decanal. Decanal overall showed a low initial release amount for all fibers, but overall (F_2,3 = 53.1, p ≤ 0.01) polyester and viscose had significantly higher released amounts (p ≤ 0.05, Tukey’s LSD). While a high adsorption of butanoic acid was observed with cotton, viscose and wool, the observed desorption was low, with no significant change over time for the three fibers (p ≥ 0.05, Tukey’s LSD). Polyester, with a lower adsorption capacity, showed a significantly higher release over time compared to the other fibers (p ≤ 0.05, Tukey’s LSD). Similar to an earlier study, acids tend to be difficult to analyze with the current experimental setup, causing relatively high standard deviations. This was particularly evident for hexanoic acid: its low volatility resulted in very low desorption measurements. While viscose showed significant adsorption, none of the four fibers investigated showed a significant release over time or differences among the fibers (F_2,27 = 0.5, p = 0.97). For phenol, the adsorption was very high for all fibers, but showed overall very low desorption with no significant increase over time (F_2,27 = 0.4, p = 1.00). The overall release of phenol for polyester was significantly higher (p ≤ 0.05, Tukey’s LSD).

In general, polyester remains the fiber that releases a wide range of VOCs (e.g. ethyl benzene, decanal, methyl butanoate, butanoic acid, phenol and benzyl acetate), most are released gradually over time and in higher amounts compared to the other fiber types. Worth noting is phenol was released only initially, and benzyl acetate was released within the first 3 hours. Viscose generally showed a similar release behavior to cotton: for 1-octanol the higher released amounts were due to higher adsorption. Overall, the observed results of desorption confirmed our earlier study on the release of VOCs from different fiber types,⁶ polyester with fast and higher release, while cotton and wool (and now viscose) showed low or no significant release. However, as discussed in an earlier paper,⁶ information on the mechanism of release of VOCs from fibers is still largely unknown. Viscose, as a manufactured cellulose fiber, has a very similar structure to cotton and therefore H-bond forces will be primarily responsible for interactions between the fiber and the VOCs.¹⁶ The more uniform structure of viscose compared to cotton may have resulted in greater adsorption and release for VOCs with higher polarity.

Conclusions

The experimental setup showed good responses for aldehydes and for esters, while it was less successful for the analysis of acids. This study successfully expanded our knowledge on the adsorption and release of VOCs by cotton, wool, polyester and now viscose rayon. Viscose rayon showed better adsorption capabilities for polar VOCs (phenol, butanoic acid and hexanoic acid) and an overall desorption characteristic similar to cotton. Polyester, in contrast to our earlier research, showed overall lower adsorption characteristics, but still the highest release over time for most of the tested VOCs. Wool overall had more modest adsorption characteristics, and still offered limited release of the adsorbed VOCs.

The need to expand our understanding of fiber-related patterns of adsorption and release of VOCs continues, given two critical factors: (1) diversity in the range of fiber sources for textile manufacture in response to increased world fiber demand, environmental issues relating to cleaning, re-use and up-cycling; and (2) the need to understand the comparative behavior of these fibers/textiles in relation to VOCs known to be present in human body odor.

Footnotes

Acknowledgements

The authors thank Dr. L.A. Dunn for undertaking the disaggregation of the standard fabrics to yarn form.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Raechel M Laing

References

Textile Outlook International. Global trends in fibre prices, production and consumption. Text Outlook Int 2018; 191: 71–71.

Roeder T. The viscose process: scope and limitations. In: Ek M (ed) FP1205 COST. Stockholm: COST, 7-9 March 2017.

Kadole

Lavate

Ukey

, et al. Studies on fabrics produced from modified viscose/bamboo yarns. Colourage 2016; 63: 31–38.

Rathinamoorthy

Thilagavathi

Brindha

, et al. Odour control studies on apparel fabrics finished with methanol extract of Terminalia chebula. Fibers Polym 2014; 15: 1669–1676.

Prada

Curran

Furton

. Characteristic human scent compounds trapped on natural and synthetic fabrics as analyzed by SPME-GC/MS. J Forensic Sci Criminol 2014; 1: 1–10.

Richter

Bremer

Silcock

, et al. Textile binding and release of body odor compounds measured by proton transfer reaction – mass spectrometry. Text Res J 2017; 88: 2559–2567.

Yao

Laing

Bremer

, et al. Measuring textile adsorption of body odor compounds using proton-transfer-reaction mass spectrometry. Text Res J 2015; 85: 1817–1826.

Laing

. Natural fibres in next-to-skin textiles: current perspectives on human body odour. SN Appl Sci 2019; 1: 1329–1329.

Introtech Limited. James Heal Consumables 2019. (Accessed May 24, 2019).

10.

International Organization for Standardization. ISO 6330:2012(E) Textiles – Domestic washing and drying procedures for textile testing. Geneva: International Organization for Standardization, 2012.

11.

International Organization for Standardization. ISO 139: Textiles - Standard atmospheres for conditioning and testing. Geneva: International Organization for Standardization, 2011.

12.

Huizenga

Zhang

Arens

, et al. Skin and core temperature response to partial- and whole-body heating and cooling. J Therm Biol 2004; 29: 549–558.

13.

Kamppuri

Vehviläinen

Puolakka

, et al. Characterisation of novel regenerated cellulosic, viscose, and cotton fibres and the dyeing properties of fabrics. Color Technol 2015; 131: 396–402.

14.

Abu-Rous M and Liftinger E. The role of fiber moisture management and surface properties in bacterial growth and odor development on textiles. AUTEX2019 – 19th World Textile Conference on Textiles at the Crossroads. Ghent, Belgium: AUTEX, 2019.

15.

Wang J, Lu X, Wang J, et al. Quantitative and sensory evaluation of odor retention on polyester/wool blends. Text Res J 2019; 89: 2729–2738.

16.

Guthrie JD. The chemistry of lint cotton. In: Kyle W (ed) Chemistry and chemical technology of cotton. New York: Interscience Pubs, 1955, pp.3–13.