Review: The softness of hygiene tissue

Abstract

The hygiene tissue industry has an extensive global market that is quickly growing. Market research has indicated that softness is one of consumers’ most highly desired properties. For certain hygiene tissue products (specifically bath tissue), this property can influence prices. A better understanding of the science of softness would allow companies to engineer soft tissue more economically and efficiently. Softness is a subjective perception related to physical aspects that make it challenging to express and measure. Human handfeel panel testing, which ranks the specimens through physical tests, has been recognized as the most reliable method to measure tissue softness. Much effort has been expanded in correlating the panel test results with some measurable properties. In this regard, equipment has been recently developed by combining several different mechanical, surface, and acoustic properties to characterize softness. In comparison with panel tests, these instruments (e.g., tissue softness analyzer) have been found to give equivalent softness metrics. A combination of materials selection and manufacturing operations are used to create softer tissue sheets. This paper reviews the sensation of softness as perceived by the human touch, techniques for measuring softness, the influence of fiber on softness, manufacturing techniques, and additives used for softness enhancement.

Full Article

Review: The Softness of Hygiene Tissue

Joel J. Pawlak,* Ryen Frazier, Ramon E. Vera, Yuhan Wang, and Ronalds Gonzalez

The hygiene tissue industry has an extensive global market that is quickly growing. Market research has indicated that softness is one of consumers’ most highly desired properties. For certain hygiene tissue products (specifically bath tissue), this property can influence prices. A better understanding of the science of softness would allow companies to engineer soft tissue more economically and efficiently. Softness is a subjective perception related to physical aspects that make it challenging to express and measure. Human handfeel panel testing, which ranks the specimens through physical tests, has been recognized as the most reliable method to measure tissue softness. Much effort has been expanded in correlating the panel test results with some measurable properties. In this regard, equipment has been recently developed by combining several different mechanical, surface, and acoustic properties to characterize softness. In comparison with panel tests, these instruments (e.g., tissue softness analyzer) have been found to give equivalent softness metrics. A combination of materials selection and manufacturing operations are used to create softer tissue sheets. This paper reviews the sensation of softness as perceived by the human touch, techniques for measuring softness, the influence of fiber on softness, manufacturing techniques, and additives used for softness enhancement.

DOI: 10.15376/biores.17.2.Pawlak

Keywords: Hygiene Tissue; Softness; Softness Measurements; Tissue Softeners; Machine Technology; Fibers; Softness Sensation

Contact information: Department of Forest Biomaterials, North Carolina State University, Raleigh, North Carolina, USA; *Corresponding author: jjpawlak@ncsu.edu

 INTRODUCTION

Hygienic tissue is one of the most important consumer products in the forest products industry. It serves to clean and protect. In 2019, 40.5 million tons of tissue were consumed globally, including 9.3 million tons in North America (RISI 2019). The global tissue market has steadily grown over the years (RISI 2019). Hygienic tissue includes both toilet tissue and paper towel. A minor component of this market includes napkins and facial tissue. Toweling is an area that continues to grow. For toweling, the critical properties associated with higher consumer prices include the strength of the sheet, absorbency, absorption rate, and marketed sustainability (de Assis et al. 2018). For towels, softness is not strongly correlated with consumer pricing. However, at home toilet tissue softness has been shown to correlate with consumer pricing in the North American market (Wang et al. 2019b). The importance of softness provides an incentive for papermakers to characterize it in a routine manner.

The overall hygienic tissue market has a global value of $100 billion despite the relatively small tonnage produced (ca. 40.5 million tons market per year). The overall industry has been showing steady growth globally at a 3% CAGR over the past five years, and regionally the growth rates can be much higher (RISI 2019). This market appears to be an area of opportunity for the future growth of the forest products industry.

The manufacture of hygienic tissue has three broad areas of variables associated with it. These variables include fiber selection, manufacturing technology, and the additive used or applied to the sheet (Gigac and Fišerová 2008). Converting processes are necessary to optimize tissue product properties such as strength, softness, and absorbency. Moreover, the importance of converting should not be discounted in determining the final product the consumer purchases. Multi-ply sheets, embossing patterns, and winding all impact the final product and can be directly observed by the consumer.

In terms of fibers, three main groups of fibers are commercially important. Virgin wood fibers are produced by chemical pulping followed by bleaching. Recycled fibers are mainly derived from mixed office waste (MOW) papers and undergo a cleaning and deinking process before use (FisherSolve International 2017). While significantly lower in tonnage compared to virgin fiber and recycled fiber, non-wood fiber such as bamboo and wheat straw has gained interest, as consumers are willing to pay a premium for tissue made from such fiber material.

Fibers that have not been previously made into paper are called virgin fibers. Most of these fibers are chemically pulped and then bleached, derived from hardwood (angiosperms) and softwood (gymnosperms) trees. These fibers have distinct differences and could be blended to balance tissue strength and softness (FisherSolve International 2017; de Assis et al. 2018). On the other hand, tissue sheets can be layered such that one layer is made of hardwood fibers to give softness, and another layer is made of softwood fibers to provide strength (Boudreau 2013). Softwood fibers typically have twice the aspect ratio (length:width) compared to hardwood fibers. This difference in physical dimensions makes softwood better for enhancing strength (de Assis et al. 2018). The finer hardwood fibers allow for more free fiber ends to occur at the surface of the tissue, adding to the velvety feel of the tissue surface (Wang et al. 2019b). Recycled fibers are recovered from waste paper and undergo processing before being used in tissue. Recycled fibers typically have more fines and are hornified, reducing their ability to bond and creating lower bulk, softness, and water absorbency when compared to virgin fibers (Welf et al. 2005, Banavath et al. 2010). Recycled fiber is most often used in away-from-home market segments, but such fibers can also be found in the at-home segment.

The softness of tissue in the at-home market segment has been found to be one of the most important properties (Hollmark and Ampulski 2004; Wang et al. 2019b). Several strategies exist to enhance the softness, and the tissue industry uses a number of industry-specific manufacturing technologies. Many of these technologies focus on the pressing and drying aspects of tissue manufacturing to prevent densification that occurs in conventional papermaking. Lightly refined or unrefined fibers are used to form the tissue at basis weights typically ranging from 15 to 50 g/m^2, with an average consumer sheet weight being about 40 g/m². The sheets are lightly pressed or not pressed and then typically dried in one of four major types of drying technologies: light drying-crepe (LDC), creped through-air dry (CTAD), creped through-air dry belt (CTADB), and uncreped through-air dry (UCTAD) (Kullander et al. 2012).

Thru-Air Drying (TAD) is a specialized drying technique that uses wet mold tissue with air passed directly through it to create a bulky and soft tissue. The molding process can be used to create areas of softness and strength in a honeycomb-like structure (Valmet 2014). The sheets are dried with air passing through the sheet at a constant pressure drop and temperature that ranges from 100 to 250 °C. Often the TAD dryer is combined with a traditional Yankee dryer that additionally allows for creping and the development of bulk. Uncreped Thru-Air Drying (UCTAD) developed by Kimberly-Clark does not use a Yankee dryer and only TAD (Wendt et al. 1998). The elimination of the creping step can increase the productivity of the machine.

The development of various drying technologies for tissue has been primarily driven by improving the softness of the tissue sheet. Besides drying, other processes strongly influence tissue product performance and properties such as softness. Some of them are creeping and converting processes during papermaking (de Assis et al. 2018, 2020). Creping involves scraping the tissue sheet from the Yankee dryer surface using a creping blade (described in a later section) to create crepe folds in the tissue structure. As a result, softness perception has been demonstrated to increase (de Assis et al. 2020). Moreover, creeping performance will depend on the type of creping blade and creping blade angle (de Assis et al. 2020). On the other hand, converting processes provides finished tissue products with critical functional properties (e.g., brand patterns) that add both value when placed into the commercial market and improve properties such as softness perception (Vieira et al. 2020b)

Panel softness has been the traditional benchmark for softness characterization, but it requires a trained panel and has a good deal of subjectivity associated with it. Thus, many researchers have explored more analytical methods for characterizing softness. This exploration includes using algorithm methods and instrumental softness testers. Understanding the nature of softness also gives insight into the development of softness measurement techniques. Softness is a perception that combines a complex set of inputs, including appearance, mechanical properties, friction properties, vibration characteristics, and sound. Giselher Grüner (Grüner 2016) developed a tissue softness analyzer that is a purpose-built instrument for measuring tissue softness. The method has found a degree of acceptance in the industry due to its ability to reasonably predict panel score softness rankings (Wang et al. 2019b).

This review explores the many aspects of tissue softness. The authors cover many areas, including fiber selection, manufacturing technologies, and softness measurements. The physical nature of the softness sensation is also reviewed to understand better the connection between softness measurement, materials selection, and manufacturing. The review’s goal is to provide a complete discussion of the various aspects of softness.

DEFINING THE PERCEPTION OF SOFTNESS

It is challenging to select the most effective and affordable method to achieve a desired softness level because the property is difficult to quantify (Patterson 2013). Therefore, there is an interest in defining softness in a manner such that it can be evaluated via analytical testing. The property of softness includes several texture perceptions such as velvety, delicate, and bulky (Hollmark and Ampulski 2004; de Assis et al. 2018). A person’s experience and regional differences can affect the softness perception. The softness perception involves a number of senses, including tactile, visual, auditory, and olfactory (Leporte 1970). These sensory inputs are processed in the mind to make a softness evaluation (Gallay 1976). Though “mainly based on hand-felt sensing,” softness can also include auditory and visual aspects (Teng et al. 2011). The complex nature of softness makes it difficult to determine analytically. However, studies have shown that the tactile component shows the best relationship with the overall softness (Gallay 1976).

There are three important anatomical components of a human hand used for softness evaluation: lamellar and tactile corpuscles and Merkel cells (Wang 2019a). The lamellar corpuscle on the human finger touches each free fiber protruding from a tissue’s surface when the human hand moves across the surface and initiates vibrations that have “an optimal sensitivity at 250 Hz” (Wang 2019a).

As one of the most important properties of hygiene tissue (de Assis et al. 2018), softness has been rarely studied in the papermaking field. Softness has been linked to tissue bulk, smoothness, roughness, hardness, stiffness, strength, etc. No single property is directly related to softness, as softness is the interaction of many properties.

Objects can have two types of tangible object properties: “macro-spatial properties, including shape and orientation; and material properties, such as roughness, softness, and temperature” (Kitada et al. 2019). Neuroimaging studies have found that macro-spatial properties and material properties require different network engagement for processing (Kitada et al. 2019). For the property of softness, there have been very few studies of the “neural correlates underlying the perception of object compliance and softness.” It has been found that “tactile perception of softness is based on the spatio-temporal variation of pressure on the skin” (Kitada et al. 2019).

In the paper industry, tissue paper is often defined by physical and mechanical properties. The desired properties include “high softness, low grammage, high bulk, and high liquid absorption capacity” (Vieira et al. 2020a). Softness can be broken down into two major segments, bulk softness, and surface softness. Bulk softness “can be indicated by the elasticity of the sheet” (Ismail et al. 2020) and can be estimated by “measuring the stiffness and the thickness of the sheet” (Raunio and Ritala 2013). Although there is no explicit mention of elasticity as a direct indicator of bulk softness by Ko et al. (2018), they concur that bulk softness can be determined from bulk stiffness and defines the bulk stiffness measurement as “the slope between the two specified points in a load-elongation curve from tensile testing.” Elasticity is inherently involved since, in tensile testing, the initial slope (Young’s modulus) in the stress-strain curve is in the elastic region. It should be noted that the stiffness described here is in-plane stiffness, and it is significantly different from bending stiffness. However, simply measuring the bulk softness is not a comprehensive measurement of overall softness. The softness of the surface “is a complex combination of roughness, friction and elasticity of the surface” (Raunio and Ritala 2013). This complex property of surface softness might be determined from a surface tester that includes several measurements (stiffness, roughness, bulk softness, friction, etc.) described above (Ko et al. 2018).

One review considers softness “a state-of-the-art technology” which “belongs to one of the most protected proprietary areas for tissue and towel manufacturers” (Ko et al. 2018). Softness evaluation is labeled as an art rather than a science because it has not been distinctly defined. Described as a “psychological phenomena which involve many different components that may interact with each other,” softness itself is quite difficult, if not impossible, to isolate from other factors that may contribute to or be dependent on softness (Ko et al. 2018). “Softness is difficult to quantify even with modern equipment that imitates a human hand because it can vary between individuals, markets, and cultures” (Ismail et al. 2020). Despite this variation, specific properties can together influence the perception of softness, including but not limited to “crepe count [number of crepes per centimeter], crepe-to-stretch ratio, sheet density, strength, stiffness, and creping geometry” (Ismail et al. 2020). These properties on their own can be individually measured. They can help determine relative softness, but it is difficult to quantify a universal softness metric with any single property alone. For example, Hollmark (2004) attempted to decouple bulk and surface softness from overall softness but failed because these two properties depend on each other. Likewise, strength, stiffness, and softness typically depend on sheet density and creping geometry. This dependence makes relying on any linear regression analysis questionable for developing a tissue softness model. Therefore, studying the autocorrelation between physical properties is important when creating a softness model.

Properties Affecting the Feeling of Softness

Changing the furnish or the tissue machine operation can change softness properties. Even after the tissue product is made, specific surface treatments can affect softness (Patterson 2013). A change in the crepe count and the height and structure of the crepes affect the quality of the end product and how a human might perceive the feeling of softness on this product. Crepe folds are a strong microstructure generated on the paper web and increase softness feel while stretching the sheet along the machine direction (Raunio and Ritala 2013). Factors such as these can affect the softness of tissue paper. If a paper machine blade becomes worn down, “the integrity of the tissue is altered,” which affects final product softness. One experiment supported the notion that the greatest contributor to softness is this fiber bonding destruction that occurs due to both doctor-blade motion and the addition of creping agents on the Yankee cylinder (Teng et al. 2011). It has also been found that increasing the crepe distance can affect softness perception in end products (Ismail et al. 2020).

The perceptual softness of tissue paper is said to be distinguishable by “hand feel and surface smoothness” (Ismail et al. 2020). The hand feel metrics follow the same pattern as the crepe count, where both are low at the time directly before the exchange of the old doctor blade during the doctoral blade cycle, which corresponds to higher surface smoothness. Near the end of production time, surface smoothness increases slightly “due to the fact that more of the inhomogeneous and broad crepes are considered ‘soft’” (Ismail et al. 2020). In one study, when the tissue structure was less homogeneous, the smoothness of the surface increased, but the perceptual softness on average did not change. The explanation for this is “the irregular peaks stacking together,” such that they form even larger crepes, making for a soft feeling on the surface. A human finger “cannot differentiate roughness below 270 nm in height” (Ismail et al. 2020). Other properties also can influence or help predict softness, including out-of-plane elastic modulus and the presence of surface-extending free fiber ends. The out-of-plane elastic modulus “has been measured to correlate with subjective softness evaluation” (Ko et al. 2018), and it is known that the density of free fiber ends on tissue paper can impact the softness feel, with higher densities typically feeling “softer” (Raunio and Ritala 2013). The reduction in the contact area between the tissue web and a hand that occurs when free fibers are present increases the feeling of softness (Wang 2019a). However, neither out-of-plane elastic modulus nor free fiber density is heavily relied upon for standard softness measurements. It is important to note that these indicators of relative softness are not necessarily measurements of softness itself.

The Brain and Softness

The brain must process electrical signals from more than “17,000 mechanoreceptive units” on a human hand (Wang 2019a). Each fiber is subjected to both pressing and deflective forces upon touching a tissue surface. The forces, in turn, send an impulse to the brain. If the impulse is higher, that indicates a less soft surface (Wang 2019a).

In one study, functional magnetic resonance imaging was used to determine whether certain parts of the brain, specifically the parietal operculum and insula, were used to perceive tactile softness (Kitada et al. 2019). The study took a sample of 56 participants who “estimated perceived softness magnitude using their right middle finger” (Kitada et al. 2019). The stimuli given in the study “had the same shape but different compliances” (Kitada et al. 2019)

The results of the study showed that “activity in the parietal operculum, insula, and medial prefrontal cortex was positively associated with perceived softness magnitude, regardless of the applied force” (Kitada et al. 2019). In the ventral striatum, more softness perception activity occurred in the high-force condition than the low-force condition. From this study, it can be concluded that “a distributed set of brain regions” are required to perceive softness, and the clarity of the softness perception is related to “the magnitude of deformation of an object under an applied force” (Kitada et al. 2019).

It is necessary to know more about the brain networks involved to make substantial leaps in determining softness. According to a study on tactile softness perception in the brain, “the brain networks that are involved in extracting information on compliance or softness perception are still unknown” (Kitada et al. 2019).

Softness as a Vibrotactile Sensation

The perception of softness is contributed by direct static touching and frictional sliding (Di Luca 2014). These two steps generate vibrations at different frequencies and amplitudes, which contribute differently to softness perception (Rust et al. 1994; Okamura et al. 2001; Kuchenbecker et al. 2006; Kobayashi et al. 2008; Kildal 2010, 2012; Okamoto 2010; Takahiro et al. 2010; Porquis et al. 2011; Visell et al. 2011; Giordano et al. 2012; Ikeda et al. 2013).

These vibrations stimulate receptors both on the skin’s surface and deep in the tissues (Di Luca 2014). Meissner and Pacinian corpuscles, which are responsible for fast-adapting (FA) afferents, can respond to either transient or high-frequency mechanical stimuli. Merkel disks and Ruffini corpuscles receive the slow-adapting (SA) afferents that respond to relatively static or low-frequency stimuli (Freeman and Johnson 1982; Vedel and Roll 1982; Ribot-Ciscar et al. 1989; Johnson 2001).

Static Direct Touching

The first major contribution to softness perception is direct vertical skin touching. The direct touching causes deformation of both the sheet and finger. Due to the high viscoelastic nature of the system, the normal force is commonly treated as constant. In this relatively static contact, the effects of FA can be treated as insignificant (Di Luca 2014). The pressure generated at the proximity becomes the first cue of softness. It is believed that broad and gradual pressure and indentation are preferred.

As shown in Fig. 1, the finger was to some extent wrapped into the material for soft material, which resulted in a higher contact area. This implied a lower and broader pressure on the finger for a given force. Such deformation is less likely for a hard object, which results in a narrow and intense pressure on the finger. The broader contact area also triggers larger activated areas, which provide large, slow, and gradual signals to the central nervous system.

Fig. 1. Image based on information and sketch from Di Luca (2014). Contact pressure distributed on a material, deformation from indentation

Roughness and Low-Frequency Vibration on Softness

The surface of a material, especially tissue, is usually non-uniform. Bonds, voids, and fibers create bumps and valleys on the surface, which result in profile roughness. Roughness can be defined as the deviation of the surface from a flat plane. While under contact, the roughness can vary significantly, depending on the mechanical properties of the tissue. For example, certain processes in tissue-making, such as creping, can dramatically increase the roughness while at the same time lowering the z-direction mechanical properties. This structure leads to a rough but highly compressible tissue surface perceived as “soft.” It should be noted that roughness is both a physical property describing the shape of a sheet and a mechanical property that can be measured on the sheet in some lab tests. Roughness is constantly linked with softness due to their similar perceptual process. In general terms, the perception of roughness is part of the softness perception. The surface profile and the mechanical properties combine to create a perception of “roughness” in softness terms. Perceiving “roughness” can be contributed by static touching and low-frequency vibration. Perceiving roughness can be regarded as sensing the spatial difference in the profile and mechanical properties on the surface. When there exists a great deal of mechanical and profile variation on the surface of the samples, the “roughness” can be sensed by direct static touching. Due to the mechanical and profile variation, the contact areas are different locally, which results in different pressures and sensations. In most scenarios, the surface of a tissue or fabric is fine enough that static touching is difficult to differentiate. The “roughness” difference in finer surfaces has to be differentiated by sliding fingers on materials, adding the frictional properties between the finger and tissue.

Hollins et al. (2000) proposed a “duplex model of tactile roughness perception,” which argued that the perception of fine texture by the induced vibration is different from the perception of coarse textures. Fagiani et al. (2012) developed experiments to support the duplex model and further argued that the roughness perception by friction-induced vibration is responsible for the SA mechanoreceptors at 2 to 100 Hz. At these low frequencies, the roughness perception is a function:

• of sample roughness wavelength, when the sample roughness wavelength is much smaller than the fingerprint wavelength.
• of fingerprint wavelength, when the sample roughness wavelength is much larger than the fingerprint wavelength.
• of the ratio of two wavelengths, when the width of the two wavelengths are comparable.

Frictional Sliding and Induced Vibration

As one of the most significant contributors to softness perception, friction-induced vibration provides the signal components related to the relative displacement of the objects at high-frequency, where the frequency bandwidth can overlap that of the vibrotactile sense (Ibrahim 1994; Akay 2002). It is plausible to relate softness perception to friction-induced vibration since the vibration may contribute by both surface and internal characteristics. The physical aspects at the proximity are too complicated when stroking fingers over the surface of samples, which makes it difficult to measure and interpret. Friction-induced vibration results from complex interactions involving contact mechanics, tribology, and non-linear dynamics at the micro-macro levels (Dahl 1976; Akay 2002; Cao et al. 2014). For a given sliding pair system, the dissipation of frictional energy involves four different mechanisms. The first two mechanisms include breaking boundary films between components and deforming the contacting asperities elastically and plastically. In the third mechanism, energy dissipation triggers interaction beyond the interface, which results in a vibration response of the whole system (Cao et al. 2014). The vibration changes the true contact area and force between the two components and forms a closed-loop feedback relationship in the fourth mechanism (Akay 2002; Sheng 2007; Cao et al. 2014). Classic friction-vibration interactions include stick-slip (Van Campen et al. 1998), modal couplings (Kippenberg et al. 2002), vibro-impact (Cao et al. 2014), sprag-slip (Sinou and Jézéquel 2003), and closed-loop interaction (Akay 2002; Cao et al. 2014), which could be taking place during the interactions.

Tactile perception occurs when a surface or material is touched or scanned by a human finger. The interfacial friction that occurs during touching “results in vibrations carried by nerves to the brain, which are interpreted as the level of smoothness” (Ding and Bhushan 2016). The skin is deformed, and friction-induced vibration stimulates human sensory receptors. The texture information becomes an electric potential that nerve fibers send to the brain (Ding et al. 2018). Determination of the shape or texture of a material involves proprioceptors and mechanoreceptors. People choose certain paper or textile products based on fingertip sliding, “because textures like smoothness, glossiness, and naturalness can be sensed” by mechanoreceptors within human skin (Ding et al. 2018). It has been found that a material will be more difficult to identify if a surface has non-periodic roughness. The ability to recognize a vibrational frequency pattern allows for better perception. Further, tactile determination “can be improved by discontinuities of the surface texture within the same sample surface” because a person can perceive the discontinuity when applying the same stimuli (Bartolomeo et al. 2017).

These friction-induced vibrations mentioned are created by relative motion between the finger and the material touched. Important contact parameters include both load and scanning speed. An increase in scanning speed shows a decrease in the friction coefficient for the contact between a finger and a fabric, but “hairier” fabrics show larger variations in the friction coefficient related to scanning speed (Fagiani et al. 2011). More work needs to be done to determine how the magnitude and frequency of spectrum upon fabric touching is changed during tactile scanning (Fagiani et al. 2011).

MEASUREMENT OF SOFTNESS

Due to the complex nature of softness, it is challenging to measure and quantify softness. Measurement of bulk softness is trusted as a proper and accurate measurement, easily measured by the elasticity and thickness of a sheet (Raunio and Ritala 2013). However, softness is not just the bulk, but also the surface softness, which is a complicated measurement that sometimes requires a multi-step evaluation process. The surface softness requires a consideration of the topography of the surface, “particularly the crepe structure and its periodicity” (Ismail et al. 2020). Usually, tissue softness is studied through panel tests “in which people evaluate the softness of tissue paper subjectively” (Raunio and Ritala 2013). However, human panel tests tend to show variability. According to one study, the variability in human perception of softness can be decreased or mitigated by training (Teng et al. 2011), but an instrument that could give a repeatable softness value would be valuable and less time consuming.

Softness is perceived on a tissue surface “when the crepe folds are inhomogeneous, nonperiodic, and long” because a hand “cannot differentiate in the microscale between proper crepe waves and inhomogeneous peaks if they are less than 760 nm in height” (Ismail et al. 2020).

It has proven difficult to create reliable physical test methods for the softness of hygiene tissue papers (Ko et al. 2018). Despite this, there has been much effort in the pulp and paper industry to develop methods that can “be used to predict in-use performance of a consumer product that is also reasonably well-correlated with subjective softness evaluation” (Ko et al. 2018).

Since softness is a human perception, much of the work in developing measurement devices have been done “with the goal of correlation with the rating by softness panels” (Raunio and Ritala 2013). However, this has proven difficult because these devices have often shown a poor relationship with panel test results. This weak correlation has been attributed to two major factors: 1) “the uncertainty of factors affecting the subjective feeling of softness” and 2) “the current devices measure the forces that are not in the same sensitivity scale as what humans perceive” (Raunio and Ritala 2013). It has even been suggested that “objective softness evaluation should be impossible since softness is subjective in nature” (Ko et al. 2018).

Benchmarking / Previous Softness Measurements Models

Several different measurement methods have been designed to attempt to understand tissue paper softness, including internal methods created within companies and external processes, where other instruments are brought to test samples. In-house methods are not well known because companies generally do not publish internal methods. Many methods involve softness modeling, and several different softness models have been developed over the years.

Direct Measures of Softness

Panel testing

The panel test (Fig. 2) has become a widely accepted method for softness evaluation (Institute of Paper Chemistry 1967). This test is also the most common method for the tactile perception of the softness of fabrics, qualitatively measuring the perception of softness feeling (Thieulin et al. 2016). Two principal types of panel tests can be identified. The scoring method involves assigning numerical values of softness to softness references. Panelists are then asked to score the softness of the samples relative to the reference sheets. Two- and three-point reference panels are commonly used. The numerical system may be arbitrary, but it provides relative softness intensity as related to the reference. The ranking method asks the panelist to rank the sheets in order of softness. This panel does not provide a relative softness intensity but only a ranking. The ranking method can be tedious if there are many samples, but panel test methods can be developed to improve the efficiency of this method (Hollmark and Ampulski 2004). The scoring method may add more references to improve accuracy. Too many reference samples can impart a bias to the overall panel, and thus it is important to select the number of references carefully. It also requires more highly trained individuals to get reliable results (Hollmark and Ampulski 2004).

Fig. 2. Drawings of panel testing for a) surface softness and b) bulk softness components

New sensory panel test (N-SPT)

A new type of sensory panel test (N-SPT) similar to a conventional SPT was developed, in which a set of untrained panelists rated, ranked, and compared samples (Ko et al. 2018). This N-SPT test “can generate interval-scale softness evaluation from round-robin paired-comparison tests” (Ko et al. 2018). This numerical scale is linear and continuous, with equal intervals of physical measurements, including length, weight, and temperature. Undoubtedly, such an interval scale of subjective softness data is critical to developing tissue softness models based on physical and mechanical properties. From the results of this new test, several physical softness models were developed, including the “Handle-O-meter, Clark’s Softness Tester, Brown Softness tester, and C.H. Dexter softness tester” (Ko et al. 2018).

Artificial finger

Another quantitative method for softness evaluation is using an artificial finger (Fig. 3). This mechanism can measure the friction coefficient between the finger and the material as well as the “acoustic vibratory level generated by sliding the finger on the bathroom tissue” (Thieulin et al. 2016). The artificial finger was made in an attempt to “quantify the sensation of the tactile quality of bathroom tissues. The intrinsic characteristics of the bathroom tissues cannot explain the softness and the velvetiness felt by the hand feel panel” (Thieulin et al. 2016). This instrument can separate softness and surface texture, both important pieces to the tactile perception. A tribohaptic system was used to measure the friction coefficient and vibratory level (Thieulin et al. 2016). The vibrations of a human finger in contact with tissue was used to define the tribohaptic system. An accelerometer attached to the person’s finger was aligned parallel to the plane of contact to characterize the vibration. The finger’s normal and tangential force measurements were taken underneath the tissue. From this, typical human handling conditions were determined. The measurement conditions include five back and forth movements in the machine direction at a normal force of 0.3 to 0.4 N, a sliding speed of 20 to 30 mm/s, and a 20 mm travel length (Thieulin et al. 2016). From this, an average frictional coefficient can be calculated from the ratio of the friction force to the normal force (Thieulin et al. 2016). The tribohaptic artificial finger mechanism is shown schematically in Fig. 3.

Fig. 3. a) Depiction of artificial finger system with the: 1) accelerometer aligned parallel to the contact plane, 2) force sensors to take measurements underneath the support surface, and 3) a displacement system that can slide back and forth. b) Schematic of the working mechanisms of this artificial finger device. Image courtesy of Thieulin et al. (2016)

The artificial finger allows quick, repetitive, and direct measurements of tactile perception. It was found that the artificial finger could measure vibrations that correlated to the softness evaluated by panel tests. The friction coefficient could be related to the tissue surface texture (Thieulin et al. 2016). The study found that the internal characteristics of the tissue did not make a big difference on the feeling of softness. Additionally, it was found that there was an increase in softness as thickness increased, which suggested that the softness perception is related to the thickness (Thieulin et al. 2016). Additional paper properties were measured and compared to the softness feeling measurements. No significant correlations were determined between other paper properties and softness. This insight can only mean that the feeling of softness does not depend on just one parameter, but several in combination. It was determined that “the acoustic vibratory level was a good marker of perceived softness, and the friction coefficient expressed the velvetiness of the surface” (Thieulin et al. 2016); as the acoustic vibration decreases, softness increases. The feeling of the surface (texture) seems related to the friction coefficient, and the feeling of softness seems to be connected to the level of acoustic vibration (Thieulin et al. 2016).

Prediction Models for Correlating with Softness

Beyond panel scoring and other more direct measures of softness, instrumented and algorithm-based methods can be used to evaluate softness. Instrumented methods are techniques that use instruments specifically designed to assess softness. The algorithm methods are techniques that use measurements from instruments not specifically designed to measure softness. Typically, this involves making multiple tissue property measurements and then correlating them with panel softness or another accepted softness measurement.

The instruments and algorithmic models used to correlate with softness include:

Handle-O-Meter

The Handle-O-Meter became an accepted TAPPI (Technical Association of the Pulp and Paper Industry) test method in 1985 but was withdrawn in 1996. Increasing weight is used to push the sample through a hole with increasing force (Lashof 1960). While this method is repeatable, it did not correlate well with panel test methods (Lashof 1960).

Kawabata KES system

The Kawabata device combines three major measurements: the friction coefficient, deviation in the friction coefficient, and geometric surface roughness to create a softness parameter (Kawabata 2002). The Kawabata KES systems are used for measuring handfeel of textile and non-woven materials, and the FB4 unit, in particular, measures surface roughness and surface friction. This method can be used to characterize both textiles and hygienic tissue paper. The device was able to determine the softness of paper towels with a degree of accuracy, but it did not work well with toilet tissue (Hollmark and Ampulski 2004).

Hollmark bulk softness model

The Hollmark bulk softness model was based on the foundation that bulk softness was not reliably found using bending stiffness as a parameter. Instead, the thickness should be used for determining bulk softness. A stress-strain curve from Young’s modulus is used for the tensile stiffness measurement in Hollmark’s model. This model also emphasizes that although bulk and surface softness are different components of softness, they should not be separated because they are dependent on one another (Ko et al. 2018)

P&G softness model

Another softness measurement model by Procter & Gamble was designed to determine bulk and surface softness. In this model, bulk softness was measured by bulk flexibility (a slope on the load-elongation curve from tensile testing), and the test was regarded as reliable. The surface softness was found to be related to the surface friction, found by using the FB4 surface tester unit of a mechanical testing system by Kawabata. Surface friction is the “mean deviation from the average friction coefficient” and was used in the P&G model as the main surface softness indicator (Ko et al. 2018). Georgia Pacific also created a similar model, which verified the results from the P&G model. This type of model may be sufficient for determining an overall softness measurement because it measures both bulk and surface softness.

Kimberly-Clark softness model

Kimberly-Clark’s softness model claimed that bulk softness could be measured from the bulk stiffness measurement in a tensile test. The surface softness can be sufficiently measured using the surface friction component (Ko et al. 2018). Hence, any of these models may quantify softness more holistically if softness is defined as a combination of bulk and surface softness. The three major global tissue manufacturers of P&G, G-P, and K-C use similar methods in determining the bulk stiffness and the surface friction.

Ultrasound for out-of-plane properties

A method developed by Pan et al. (1989) uses ultrasonic testing, caliper, and basis weight. This study found that the parameters measured correlated well with the panel softness for the limited sample set tested. There were only seven samples characterized, and each was a two-ply sample that was split. Additional work would need to be done to determine whether this method is more widely applicable to hygienic papers, including toweling and tissues made with advice technologies such as through-air drying.

Sled method

The surface friction, creping ratio, and time of service for the creping blade are used in an algorithm developed by Kuo and Cheng (2000) to predict the softness. This method combines both materials properties and operational parameters. It may be most useful in a mill setting, but it does not consider other factors affecting softness, such as converting. The researchers determined that softness increased with the creping ratio and decreased with the time of the creping blade service.

N-SPT algorithms / models

The surface and bulk softness for the N-SPT method was parsed using an algorithm developed by Ko et al. (2017). Three parameters were measured and then correlated with the softness. These parameters were tensile stiffness, surface roughness, coefficient of friction. Using the concept of surface and bulk softness, each of these factors were found to be independent in the research.

For commercial bathroom tissues, the best model is the “2-Parameter model of bulk softness and surface friction equation” (Ko et al. 2018). This model “predicts that approximately 60% of subjective softness comes from the surface friction component and approximately 40% from the bulk stiffness”. The equation for the 2-Parameter model includes bulk softness (BS) and mean deviation from the average friction coefficient (MMD) as follows:

Equation 1:

Equation for 2-parameter model (BS & MMD), n = 0

X = C + mlog BS + l logMMD

X = 3.20 − 0.46 log BS − 0.72 logMMD

where:

C, m, n, l = curve fitting coefficients;

BS = GM_bulk stiffness;

MMD = mean deviation from the average friction coefficient.

Table 1 below summarizes normalized bath tissue softness model data taken from Ko et al. (2018) to support the assertions discussed above.

Table 1. 2-Parameter Model for Normalized Bath Tissue Softness Data

Tissue Softness Analyzer (TSA)

The variety of softness evaluation methods leads to a lack of pervasiveness, but one instrument, described below, has gained more widespread acceptance. A dedicated instrument was developed for the measurement of softness along with an accompanying algorithm. Grüner (2012) developed this instrument, the Tissue Softness Analyzer (TSA9, Emtec, Germany), that uses thickness and basis weight as inputs while simultaneously measuring other parameters. The TSA was developed to mimic the interaction of the hand with the tissue sheet by measuring the light brushing of the surface by mechanical lamella. The equipment was designed specifically for managing the quality of sanitary tissue paper.

Fig. 4. Emtec’s TSA softness measurement device (Paper Technology International 2021)

As noted by Kim et al. (2020), “the TSA converts the vibrations caused by friction of the fabric surface into acoustic spectrums and measures (the) acoustic frequency and sound pressure with indexing smoothness and softness.” The lamellae spin on the surface of the tissue with a constant applied force. The sample is also stretched to evaluate the mechanical compliance of the sample. The spinning lamellae of the fan generate vibrations in the lamellae and the sheet. The intensity of the sound associated with this excitation can be correlated with softness. The Tissue Softness Analyzer device is depicted in Fig. 4.

The TSA records three primary parameters:

1. TS7, also known as the “real softness,” is the amplitude (dB) of the sound spectra peak at a frequency of ~6500 Hz. The TS7 value is associated with the vibrations induced in the lamellae (Grüner).
2. TS750, also described as the “smoothness” or “roughness,” is the amplitude (dB) of the sound spectra peak at a frequency between 200 and 2000 Hz (Grüner). The TS750 peak is believed to correspond to the vibration of the tissue membrane and is mainly thought to be caused by roughness and embossing (Furman and Gomez 2007).
3. Mechanical compliance/stiffness, the D parameter, measures the inplane sample deformation, i.e. in-plane stiffness, when a load of 100 to 600 mN is applied (Grüner).

These three measurements can then be used to calculate other parameters using proprietary algorithms. These other parameters include Handfeel (HF), fTS750, P, H, and E. The Handfeel value is calculated by “combining several measurements of the sample to obtain a global quantification of softness of the papers” (Vieira et al. 2020a). TSA-HF is a compound function. There are numerous algorithms associated with instrument software used to calculate the Handfeel. The sheet caliper and basis weight, as well as the number of plies, is input into the machine, and these parameters are used for the handfeel algorithms. This algorithm gives the instrument the ability to predict the panel softness for various paper types and consumer preferences. Equation 2 below shows the relationship and dependence of measured properties to determine a value for the feeling of softness.

Equation 2:

TSA-HF function

TSA-HF = f (TS-7, TS-750, D, caliper, grammage, and number of plies).

where HF = handfeel

D = stiffness

TS-7 = softness (dB)

TS-750 = surface smoothness

The TSA can be especially beneficial because it is able to “separately index surface smoothness and fiber softness” (Kim et al. 2020), and it mimics a human hand. One study observed non-woven textiles measured and indexed surface smoothness and fiber softness properties in 749 fabrics with this TSA (Kim et al. 2020). The TSA results from another study showed that drape and bending properties are the most influential factors indicating surface smoothness (TS750). The surface smoothness was more correlated with drape than bending properties. The samples of this study were compared to simple mechanical characteristics as well, and it was found that the fiber softness (TS7) had a weak correlation with caliper thickness, Young’s modulus, as well as weight (Kim et al. 2020). The comparison to thickness was made because it has been suggested that the TS7 measurement of bulk softness comes from the sample’s thickness (Kim et al. 2020).

The ranking of the softness and smoothness do correspond with rankings by other methods of direct physical measures such as human handfeel tests, but the values measured using the TSA were “not perfectly consistent with the value of subjective handfeel” (Kim et al. 2020). In other words, the TSA measures the same value independent of equipment user, whereas human handfeel will inevitably differ slightly. A study by Perng et al. (2019) investigated the relationship between hand-felt panel tests and TSA softness measurements (Table 2). It showed high correlations, with R² values ranging from 0.9659 to 0.9945, though only four samples were tested (Perng et al. 2019). In 2021, this study was furthered. The results from the TSA were correlated with those from Hollmark’s softness theory, and it was found that a high correlation (R=0.904) existed between panel-correlated hand-felt softness and the handfeel softness measurement (HF) from the TSA. Still, a relatively lower correlation existed for the respective smoothness measurements (Perng et al. 2021). Therefore, it seems that the TSA is more comparable to the panel tests than other theories because it provides a more robust analysis of overall softness.

Table 2. Correlations of Standard (STD) Samples between TSA-HF and Corrected Panel-HF (CHF). Panel-A, B, C, D. Adapted from (Perng et al. 2019, 2021)

Recently, a modified version of the TSA device, which includes an additional top microphone for measurement (Fig. 5), has been used to distinguish the influence of hardwood and softwood. A study by Prinz et al. (2021) evaluated the influence of four different furnishes on softness properties and assessed the differences in results when using the device with and without a polytetrafluoroethylene (PTFE) film. It was determined that without the film and using the old version of the TSA, some contradictory results were obtained. With the PTFE film and the new two-microphone device, differences that would be expected between hardwood and softwood handsheets were evident (Prinz et al. 2021).

Fig. 5. New modified TSA device, with two microphones and optional film (Prinz et al. 2021)

Indication of Surface Topography

There are several types of imaging that can be used to comment on softness properties. Tissue paper’s structure has been examined on both micro and macroscopic levels using field emission scanning electron microscopy (FESEM), laser scanning confocal microscopy (LSM), X-ray microtomography technologies (XRT) (Ismail et al. 2020), and Shadow-based Imaging (Raunio and Ritala 2013)

Laser scanning confocal microscopy (LSM) imaging

Non-contact measurements of the surface profile using scanning laser microscopes were also found to correlate with panel softness. Furman and Gomez (2007) imaged the surface of six samples and found a strong correlation with panel softness (R² = 0.9183). The projected surface area was found to correlate with panel softness. This technique may be promising, but the limited number of samples prompts questions of how this would perform in the broader application of the technique. The algorithm is relatively complicated and may limit the utility in a wider setting. Figure 6 shows how the laser measures the images in LSM imaging.

Fig. 6. A schematic of the mechanisms involved in LSM imaging, from models from Ismail et al. (2020)

Regarding LSM, surface properties that can be analyzed by this method include “crepe count, waviness, and the average height of the crepes” (Ismail et al. 2020). Additionally, relationships between these properties and final product softness have been found. This technology uses “laser confocal optics to measure the depth of field across a specimen,” as well as two light sources, one laser, and another white light, that can help determine information about the sample’s shape and roughness through image and height data (Ismail et al. 2020). The crepe structure and periodicity can be determined by detecting waves on the sample through this technology. LSM is a non-destructive method, meaning “it does not affect the wave structure and height of the sample” (Ismail et al. 2020).

Field emission scanning electron microscopy (FESEM) imaging

Field emission scanning electron microscopy (FESEM) has been used to study and characterize the planar morphology of tissue papers (Ismail et al. 2020). Detailed surface topography of the tissue samples was possible by imaging when the sample was coated with platinum (to increase conductivity), and 3 to 5 kV of acceleration voltage was applied to them (Ismail et al. 2020). The FESEM is advantageous due to its clear resolution.

X-ray microtomography technologies (XRT) imaging

X-ray microtomography (XRT) uses an X-ray tomograph with a supplementary MATLAB code for structural and wave count analysis (Fig. 7). A UK – Hanatek FT3 precision thickness gauge (UK) was used to measure the average thickness of the paper sample (Ismail et al. 2020).

Fig. 7. Depiction of XRT imaging of a single ply sample a) imaged in the µCT, b) viewed in a 3D model to show fiber orientation, and c) viewed in a 3D MATLAB® model. Image courtesy of Ismail et al. (2020). This image is published under the creative commons attribution 4 license (CC BY 4.0 license) by Springer (http://creativecommons.org/licenses/by/4.0/).

Shadow-based imaging

Another new surface softness evaluation method of tissue paper is an imaging method “based on detecting shadows caused by the free fiber ends” (Raunio and Ritala 2013). Because of the tissue paper’s wavy surface, shadows are difficult to detect on the reflectance image. Therefore, the photometric stereo system was used to estimate the 3D surface information, and “the intensity variations caused by the wavy surface were filtered out” (Raunio and Ritala 2013). Digital images were taken, and the density of surface fibers was measured from these images. This particular method is promising because it showed greater accuracy than some other previous methods, and it could be implemented on a running paper machine (Raunio and Ritala 2013). The mechanism for shadow-based imaging is shown in the drawing below (Fig. 8).

Fig. 8. Drawing based on an image from Raunio and Ritala (2013) depicting the setup of the camera system and polarizers in the shadow-based imaging system

VK analyzer software

VK analyzer software has been used to view and calculate the waviness profile (Ismail et al. 2020). This software is used with various imaging techniques to determine the outer profile of a sample. Therefore, this can indicate the surface and may then be correlated to softness.

The previously described softness measurement methods have limitations, as they may be used only in limited circumstances or have good correlations in limited sample types. When samples vary widely in fiber type, fiber orientation, and moisture content, the reliability of these methods to measure softness is greatly reduced (Hollmark and Ampulski 2004). The procedures can also be time consuming, require specialized equipment, and only be applicable to limited sheet types (Ramasubramanian 2002). However, the methods described elucidate several important properties for indicating softness. These properties include tensile strength, friction characteristics, ultra-sonic characteristics (high frequency vibration/elastic modulus), stiffness, surface profile, surface texture, and sheet thickness. Table 3 details and organizes the previously described methods.

Table 3. Summary of Softness Testing Methods and Models