Toner-print removal from paper would allow paper to be re-used instead of being recycled, incinerated or disposed of in landfill. This could significantly reduce the environmental impact of paper production and use. Previous work on the subject has explored the applicability of ultraviolet, visible and infrared (IR) lasers under nanosecond pulses for toner removal. This article expands on this work by testing a wider range of ultrafast and long-pulsed lasers. Results from 10 distinct laser set-ups are used to propose an operating window for the toner-removal process. Colour analysis under the L*a*b* colour space, scanning electron microscope examination and attenuated total reflectance–Fourier transform IR spectroscopy measurements of the outcome show that, with the right laser, it is possible to remove toner from paper to enable its re-use. Theoretical models to predict the laser ablation of toner are discussed, and, while imperfect, provide sufficient evidence to support a physical explanation of toner ablation.
1. Introduction: why is paper re-use relevant?
In a highly energy-efficient industry, such as the pulp and paper sector, there is little room for improvements that would yield a considerable reduction of its energy consumption and associated impact on climate change. Reducing paper demand appears unlikely as paper consumption has increased steadily throughout recent decades and this trend is expected to continue in the near future, as suggested by Hekkert et al. (2002). Employing carbon neutral fuels and increasing the energy efficiency of manufacturing processes are widespread policies in the pulp and paper industry, as indicated by the Confederation of European Paper Industries (CEPI 2009). Material recovery through recycling eliminates the forestry step from the life cycle of paper and eradicates emissions arising from incineration or landfill dumping. However, there is modest room for improvement if it is considered that the current European recycling rate of 66 per cent is close to the theoretical maximum rate of 81 per cent estimated by CEPI (2006). What would happen if paper was unprinted and re-used instead of being recycled? This action would remove four steps from the life cycle: forestry, pulping, paper-making and disposal by incineration or landfill. Counsell & Allwood (2007) have estimated that this could translate into a reduction of up to 95 per cent CO2-equivalent (CO2-eq) emissions per tonne of office paper, in comparison with only 76 per cent achieved by recycling.
This article discusses the applicability of an alternative paper re-use method that employs laser radiation to remove toner print from paper. Previous work on the subject by Leal-Ayala et al. (2011) demonstrated that there is a particular combination of laser parameters that can remove certain toner-print types from paper using visible light at 532 nm with long pulses of 4 ns. It is known from laser–matter theory that laser ablation of solid materials follows different physical mechanisms depending on the duration of the laser pulses. Long pulses in the nanosecond or slower time scales first heat up the surface of the target material to its melting point and then to its vaporization temperature. In contrast, during ablation of solid surfaces exposed to ultrafast pulses in the pico- and femtosecond scales, material removal occurs by direct vaporization away from the surface (quick solid–vapour transition with only a brief liquid phase). In this case, thermal conduction into the target can be neglected because of the very short time scales involved. Therefore, smaller pulse lengths can, in theory, diminish the negative thermal effects generated by long pulses on the substrate. The objective of this article is to answer the following questions:
— Can toner print be effectively removed from paper by laser ablation in pulse regimes other than those explored in previous work? And, if so, can these methods produce paper sheets of high enough quality to replace new paper?
— What is the operating window of the laser removal process for the complete range of lasers tested?
— What happens to toner and paper when they interact with ultrafast laser pulses? How is this process different from toner and paper interaction with long-pulsed lasers?
— Can this process be modelled and controlled?
This study starts by reviewing previous work on the subject and explaining the relevant laser–matter interaction theory. The experimental and analytical methodology is summarized in §3. Section 4 presents a series of experimental trials aimed at identifying the most suitable type of laser for toner-print removal from a range of ultrafast and long-pulsed lasers. Analyses of the laser–paper interaction and removal mechanisms are provided in §5 together with a definition of the operating window of the laser process and two ablation depth predictive models. The certainty about the results is discussed in §6 together with the economical and environmental implications of laser removal and possibilities for future work.
2. Background: previous work on toner removal by laser ablation
The idea of using lasers to remove toner print from paper, although not new, has not been fully developed until now. This is evidenced by a survey performed by Counsell & Allwood (2006), who stated that most of the previous research in this area exists only in the form of patents that provide no details about their success. The closest source of literature on the effects that laser radiation might have on cellulose-based paper comes from the artwork restoration field. A detailed review of the literature on this subject has been performed by Leal-Ayala et al. (2011), suggesting that ultraviolet (UV) laser radiation has a degrading effect over both blank and soiled paper, decolouring cellulose fibres through photo-oxidative reactions. In addition, infrared (IR) light is also considered harmful because of the extended damage and discoloration that it produces on paper as a result of photothermal reactions originating from heat transfer from the soil into the substrate. The effects of both UV and IR radiation on paper seem to contrast with those of visible light, which has been successfully used to clean ancient manuscripts and documents.
In theory, paper could be cleaned and re-used without damage to its substrate if a laser fluence level can be found below the ablation and destruction threshold of the paper and above that required to ablate toner. A study by Ihori et al. (2009) has explored this idea by testing long-pulsed IR and visible lasers working at 1064 and 532 nm, respectively, on printed paper samples. They determined that, while the IR laser burned the paper and reduced its moisture, visible radiation with energy fluence of 1.2 J cm−2 was capable of removing toner while maintaining paper whiteness to a level that would allow it to be re-printed. The exact damage and degree of re-usability of paper after treatment is not explicit in their results. It seems that the whiteness level achieved after toner removal was not the same as that of blank un-lased paper.
In the study by Leal-Ayala et al. (2011), five different UV, visible and IR lasers with nanosecond pulses were applied to a range of toner–paper combinations to determine their ability to remove toner. They found that it is possible to remove toner from paper without considerable substrate damage when using visible radiation at 532 nm with 4 ns pulses and energy fluences under 1.6 J cm−2. UV and IR lasers created higher damage to cellulosic paper. The authors suggested the possibility that UV and IR laser radiation act as a catalyst that accelerates paper degradation mechanisms (photochemical and photothermal) as a result of the absorption of high-energy photons and the generation of localized heating. Regarding the ablation mechanism, they considered that the high-energy absorption of the toner results in a localized temperature increase that leads to the thermal degradation of its polymer fraction (for toner samples roughly composed of 50 wt% polyester resin and 50 wt% iron oxide), detaching the rest of its components from paper. This led to the conclusion that toner ablation under long laser pulses at a wavelength of 532 nm is a photothermally dominated process. Based on this assumption, the authors suggested that their results might be improved if the negative thermal effects exerted by laser radiation on paper could be avoided.
As explained by Liu et al. (1997), absorption of laser energy by the target material can be achieved by linear or nonlinear absorption mechanisms. The former includes direct ionization of valence electrons with an ionization potential lower than photon energy while the latter refers to multi-photon and avalanche ionization processes. In the case of opaque materials such as toner, linear absorption dominates under long-pulsed ablation with low intensities while nonlinear absorption becomes more important during high-intensity ultrashort pulse lengths. In both situations, the material is heated as a result of Joule heating. The amount of heat generated is directly related to the pulse duration. Longer pulses lead to a stronger energy transfer and a higher heat generation. During long pulses, there is enough time for the thermal wave to propagate into the material and create a relatively large melted layer while only a small portion of the material is evaporated. Shorter pulse lengths generate higher laser peak powers, increasing the electrons' peak temperatures to the range of a few to tens of electrovolts (1 eV=11 600°K), as explained by Liu et al. (1997), while ion and material lattice temperatures remain relatively low. Owing to this thermal gradient and the short duration of the pulses, electron–ion energy transfer continues after the laser pulse is over, heating the ions to much higher temperatures than do long pulses, vaporizing a large portion of the material subjected to laser radiation. The interaction time is so short that heat has little time to diffuse into the material, reducing the energy losses and the heat-affected volume. The melt layer remains small as most of the heated material is vaporized. These characteristics of ultrashort laser pulses could be beneficial during the ablation of toner from paper and need to be explored given that earlier work on this subject and related paper conservation studies have only dealt with lasers working in the long-pulsed nanosecond regime. In addition, little is known about the physical mechanisms behind toner removal by laser ablation at distinct pulse regimes. The following sections address these knowledge gaps.
A total of 10 laser set-ups were tested in this study: six in the ultrashort-pulsed regime and four in the long-pulsed category. Ultrafast ablation tests were performed on a Fuego Integrated Picosecond MOPA laser (Time-Bandwidth Products) and a Spitfire Pro XP Ultrafast Amplifier (Spectra-Physics). The former runs at a constant pulse width of 10 ps with adjustable wavelengths between 366, 532 and 1064 nm. The Spitfire laser device works at wavelengths of 266, 400 and 800 nm with variable pulse lengths ranging from 20 fs to 3 ps. Long-pulsed tests were performed on a QuikLaze 50ST2 laser (New Wave Research) for tests with 4 ns pulses at 532 nm, an SPI Photonics System for trials at 1064 nm with 40 ns pulses, and a Navigator II YHP40 OEM Laser System for tests at 532 and 355 nm with adjustable pulse lengths between 29 and 75 ns.
Tests were performed on uniform rectangular areas printed with HP LaserJet Q1338A-AC-D black toner on white, uncoated, wood-free (lignin-free) 80 g m−2 Canon copy paper. The exact composition of the paper and fillers is unknown as that information is not commonly released by Canon. It is assumed that a typical sheet of office paper is formed by 70–90% cellulose fibres, while the rest is a combination of fillers (typically clays), whiteners (typically titanium oxide) and additives related to strength and water-absorption properties. The toner composition reported by HP (2004) consists of 40–50 wt% polyester resin, 40–50 wt% ferrous iron oxide and 1 wt% amorphous silica. During each test, the samples were placed on the laser bed in the focal plane of the beam and moved in a raster pattern to generate a rectangular scan across the printed sample surface. The full range of parameters tested in this study for femtosecond, picosecond and nanosecond lasers are contained in the electronic supplementary material. Only one variable was changed at a time while the rest were set at a default value. The variables included wavelength, energy fluence, pulse frequency, pulse width, scan speed, spot area and number of horizontal passes. The best parameters identified for each laser are shown in §4.
All test samples were analysed under an optical microscope in order to select the best results. These were qualitatively evaluated under a scanning electron microscope (SEM) and subjected to a quantitative colour analysis obtained by scanning them with a CanoScan Lide 25 scanner from Canon and converting the resulting images to the LAB colour space using ImageJ software. The L*, a* and b* colour variables from the LAB colour model were measured with the same software. Under this system, L* defines the lightness of the colour (L*=0 for black and L*=100 for white), a* measures green and magenta tones (negative=green, positive=magenta) and b* quantifies the yellow and cyan colours in a sample (negative=cyan, positive=yellow). Every scanned sample was accompanied by two colour references (black and white) which were used to calibrate the process. L*a*b* variables from each colour reference in each sample were measured and compared with the previously determined values to ensure that the scanner repeatability remained at an acceptable level. The scanned sample would be accepted if there were no significant variations in the reference values. Ten measurements were taken from each lased paper sample under study (L*, a* and b* for each variable). The values reported in this study are average values with their corresponding standard deviation.
Further analysis was performed through Fourier transform IR spectroscopy with attenuated total reflectance (FTIR–ATR). This technique can be used to evaluate whether exposure to laser radiation produces any changes in the atomic structure of paper—and hence any damage—by measuring the changes that occur in a reflected IR beam when the beam comes into contact with a paper sample. ATR–FTIR spectra were collected at a 1 cm−1 resolution between 525 and 4000 cm−1 using a Bruker Optics FTIR Tensor 27 coupled with a Pike Miracle ATR single reflection accessory, fitted with a diamond crystal and using an angle of incidence of 45°. A high-pressure clamp was used to ensure good contact between the diamond (active area of 1.8 mm) and the paper samples in order to obtain high-quality spectra. All paper samples were conditioned in accordance to the T402 sp-08 TAPPI standard (standard conditioning and testing atmospheres for paper, board, pulp hand sheets and related products). Pre-conditioning consisted of introducing the samples in a forced ventilation oven at a temperature of 39°C for a 2 h period. The oven was operated in a room at 55 per cent relative humidity (RH) and 23°C. Under these conditions, room air drawn into the oven resulted in an RH between 20 and 30 per cent inside the oven (according to the information given in the T402 sp-08 standard). After pre-conditioning, the samples were exposed to the conditioning atmosphere (55% RH and 23°C) for 4 h in order for them to come into equilibrium with the atmosphere. After this, the samples were transported as quickly as possible to the testing atmosphere where the ATR–FTIR machine was located. The testing atmosphere was measured as 54 per cent RH and 21°C. This procedure ensured that all samples possessed similar humidity levels during testing.
4. Toner removal by ultrafast lasers
This section presents the best results obtained from toner-print removal experiments using UV, visible and IR lasers under pico- and femtosecond pulse regimes. Additional results with long-pulsed lasers in the nanosecond regime are shown in §4b. Further evaluation of the results is presented in §4c.
Real size test samples and microscopic images taken from them are shown in figures 1 and 2. Colour measurements from all lasers under study are also included in figure 2 while the optimum experimental parameters used for each laser are summarized in tables 1 and 2. The results suggest that visible and IR light with 10 picosecond pulses are the most successful in removing toner while minimizing paper damage. The rest of the lasers evaluated inflicted considerable damage on cellulosic fibres.
It can be seen from figure 2 that, although UV, visible and IR lasers (266, 400 and 800 nm, respectively) under femtosecond pulses fully remove toner print from paper, they also inflict considerable damage to the paper substrate. All three lasers left linear marks on cellulose fibres which suggest that a layer of the material was removed. UV radiation at 355 nm with 10 ps pulses does not cut cellulose fibres as the femtosecond lasers did but it produced a dramatic change in their physical appearance without achieving full toner removal.
Better results were obtained with the visible and IR picosecond lasers. Figures 1 and 2 show that IR radiation at 1064 nm with 10 ps pulses successfully removed toner from the sample while the values of all three colour variables remained very close to those of un-lased paper (only a minor decrease of 5 points in L* and increments of 3 and 2 for a* and b*, respectively). Despite this, a comparison of the 1064 nm sample at 10 ps with blank un-lased paper suggests that cellulose fibres lost some of their bonding and appear less densely packed after exposure to the laser treatment. This signifies that laser radiation at 1064 nm has slightly damaged the paper areas under the toner print.
Visible radiation at 532 nm and 10 ps pulses is also capable of achieving full toner removal, while it appears to create no mechanical damage in cellulose fibres (figure 2). Colour analysis reveals that there is little difference between the L* and a* values from this sample and white un-lased paper. However, there is an increment of 6 units in b*, which means that there is minor discoloration or yellowing of the paper after exposure to laser radiation. This yellowing can be appreciated from a naked-eye visual examination of the sample.
(b) Toner removal with long-pulsed lasers
Microscopic images from the best long-pulsed laser removal results are shown in figure 3. The samples shown in figures 3a(i,ii),b(i,ii) were obtained with the experimental parameters previously reported by Leal-Ayala et al. (2011). Although the four lasers shown in figure 3 were able to remove toner print from paper, they all interact in distinct ways with the paper substrate. Sample 3a(i),b(i) suffered significant yellowing after toner ablation with a 1064 nm laser under 40 ns pulses. It can be seen from its SEM image that the fibres lost some of their bonding when compared with blank un-lased paper. The same effect can be appreciated in sample 3a(iii),b(iii) exposed to visible radiation at 532 nm with 29 ns pulses. A very different outcome was obtained with UV light at 355 nm and 75 ns pulses in sample 3a(iv),b(iv). This type of laser damaged the paper fibres considerably, producing a complete change in their morphology. The best result was obtained with visible radiation at 532 nm with 4 ns pulses, shown in sample 3a(ii),b(ii). This laser left no residual toner on the ablated area and no appreciable damage on cellulose fibres. The exact parameters used for this test are summarized in table 3.
The first question established at the beginning of this study can be answered at this point: toner print can be effectively removed from paper by laser ablation in pulse regimes distinct from those previously explored by Leal-Ayala et al. (2011). In particular, visible and IR radiation at 532 and 1064 nm with 10 ps pulses are outstandingly successful in removing toner from paper. The second question—‘Can these methods produce paper sheets of high enough quality to replace new paper?’—deserves careful consideration. Most of the tested lasers generated sufficient physical damage on cellulose fibres to be considered as unsuitable solutions for the problem under study. One of the best results was obtained with visible radiation at 532 nm and 4 ns pulses, in agreement with previous work. Two options stand out from the rest of the tested lasers: IR and visible light at 10 ps. The former leaves the paper with high enough quality to be re-used in terms of appearance, but the mechanical damage that it inflicts on cellulose fibres means that the number of times it could be re-used will be limited by its mechanical resistance to the treatment. On the contrary, visible light at 10 ps leaves cellulose fibres in a good structural shape but it generates a minor whiteness loss. In this case, paper can be re-used but the amount of times this process could be repeated would be limited by its physical appearance. In summary, a conditional positive answer can be given to the second question stated here the best results obtained under visible and IR laser radiation with 10 ps pulses produce paper sheets of high enough quality to replace new paper, as concluded from colorimetric and SEM analyses, suggesting that it might be technically feasible to achieve paper re-use after ultrafast laser ablation of toner print. However, the number of times that paper could be re-used will be limited by the detrimental side-effects of the removal treatment. This needs to be explored in further work.
5. Defining an operating window: laser–matter interaction and paper damage
Figure 4 and table 4 classify all laser set-ups tested in the previous section based on the relation between their wavelength type, pulse length and the damage that they exert on paper. A detailed discussion of this classification is presented in §5a–c.
The threshold fluence for ablation of a certain material will become lower if the pulses are shortened as this will translate into much higher peak powers. This is confirmed by the power and fluence trends in figure 5, which show a more specific view on how the optimum ablation fluence and peak power vary as a function of the pulse length for all lasers tested in the visible spectrum. A clear trend can be appreciated: smaller pulses lead to smaller optimum fluences and higher peak powers. Similar graphs of fluence versus peak power for UV and IR tests are contained in the electronic supplementary material.
Although it would be extremely difficult to test all possible combinations of wavelength, pulse width, fluence and peak power, the information shown in figure 4 is a representative sample of a very wide laser range covering UV, visible and IR radiation in long and ultrashort pulse regimes. Therefore, the data in this figure can be considered as the operating window of the toner-removal process and represent the answer to the second question stated at the beginning of this study. An examination of this operating window suggests that the tested lasers produced varied effects during their interaction with toner and paper. This leads to the third question set in §1: what happens to toner and paper when they interact with ultrafast laser pulses? How is this process different from their interaction with long-pulsed lasers? This section addresses these queries.
(a) Paper damage under ultraviolet radiation and femtosecond ablation mechanisms
UV radiation has a strong negative effect over cellulose under any experimental conditions. This behaviour could be attributed to the creation of photochemical degradation reactions in cellulose as a result of UV light absorption. Kolar et al. (2000) measured the absorption of UV and visible light by cellulose and found that, although most of the light in the visible band is scattered, absorption is greatly increased in the UV region. Photons in this region are more energetic than those on the visible and IR fractions of the light spectrum. Direct photolysis (chemical process by which molecules are broken down into smaller units through the absorption of light) of an atomic covalent bond can occur if the energy of an absorbed photon is higher than that of the bond. Typical bonds in cellulose include C−C, C−O and C−H links. The approximate energies for these bonds are 347, 359 and 414 kJ mol−1 according to Kolar et al. (2000). These energies correspond to UV light wavelengths of 347, 333 and 289 nm, respectively. Given that the shorter the wavelength, the more energetic its photons become, the photon energy at 266 nm used in this study is higher than the energies from the three types of covalent bonds mentioned before. Absorption of this kind of photons could result in direct cellulose degradation by photolysis. In the case of laser tests at 355 nm, this wavelength has an energy level slightly lower than that of C−C, C−O and C−H bonds and photolysis seems unlikely. It is also possible that absorption of UV photons at 266 and 355 nm results in excitation of electrons in a chemical bond leading to ionization events and thus facilitating photochemical degradation reactions, as mentioned by Kolar et al. (2000). Both situations result in considerable deterioration of cellulose fibres.
A similar suggestion has been provided by Kaminska et al. (2004) when exploring the use of a 355 nm laser with 6 ns pulses to remove a mixture of charcoal and dust contamination from hand-made cellulose paper. They concluded that UV radiation produced photochemical reactions in cellulose that led to severe cellulose degradation by photo-oxidation.
The level of impairment in cellulose was more severe for the 266 nm laser with 120 fs pulses, as evidenced from SEM analysis in figure 2. As explained by Liu et al. (1997), shorter pulse lengths generate higher electron peak temperatures, which allows for electron–ion energy transfer to continue after the laser pulse is over, heating the ions to much higher temperatures than long laser pulses and vaporizing a large portion of the material subjected to laser radiation. This effect was repeated for all wavelengths tested in the femtosecond regime (266 nm under 120 fs pulses, 400 nm under 500 fs pulses and 800 nm under 1000 fs pulses), where the elevated peak power magnitudes led to highly localized vaporization, leaving no traces of any heat-affected zone on the remaining cellulose fibres.
Analysis of the samples shown in figures 2 and 3 under ATR–FTIR indicates that there is a considerable reduction in hydrogen bonds from OH groups in cellulose after treatment with UV lasers. This is depicted in figure 6 by the increase in the transmittance (T) at 3344 cm−1, which signifies that hydrogen bonds from OH groups responsible for attaching cellulose molecular chains together are dissociated during the laser ablation process, hence altering the normal structure of cellulose. The femtosecond laser at 266 nm resulted in the highest OH hydrogen bond loss followed by the 355 nm with 10 ps pulses and 355 nm with 75 ns pulses. This suggests that there is a highly localized photochemical degradation effect induced by UV light on paper, in addition to the photochemical reactions discussed before, and also suggests that shorter pulse lengths (and therefore higher peak powers) result in higher photochemical degradation of cellulose. In addition, there is a considerable difference in the band at 1028 cm−1, which corresponds to C−O ether bonds, suggesting a dissociation of bonds as a result of additional photochemical degradation reactions. These factors make UV laser unsuitable for toner removal.
(b) Visible spectrum: from paper yellowing to optimum removal
Even if cellulose scatters most laser light, it is still possible that other components of paper absorb a certain amount of radiation and allow for chemical changes to take place, as suggested by Kolar et al. (2000). Typical office paper contains elements such as hemicelluloses, lignin, fillers, pigments, dyes and sizing agents in addition to cellulose fibres. These materials are also sensitive to laser radiation and can initiate photo-induced reactions in the substrate, leading to photodegradation. The SEM image shown in figure 3 corresponding to a sample ablated under 532 nm with 29 ns pulses shows that cellulose fibres lose their cohesion and appear less densely packed than blank un-lased paper. In other words, the secondary components of paper (fillers, pigments, dyes and so on) have disappeared from the sample. The 29 ns pulses appear long enough for secondary materials to absorb sufficient laser energy to allow photothermal reactions to occur, resulting in ablation of these materials while leaving cellulose untouched owing to its distinct absorption rate and ablation threshold. The same effect is no longer appreciated when the pulse length is reduced to 4 ns into the optimum ablation region shown in figure 4. Reducing the pulse length from 29 to 4 ns results in a slight peak power increase and a considerable toner ablation threshold fluence reduction, as shown in figure 5, suggesting that a power threshold has to be met in order to avoid the negative effects observed under 29 ns pulses while an upper pulse length limit restricts the overall acceptable energy fluence.
Figure 7 indicates that light at 532 nm with 10 ps has a strong interaction with paper. This sample suffered a considerable loss of hydrogen bonds from OH groups responsible for attaching cellulose molecular chains together, represented by the increment in the transmittance at the 3344 cm−1 band. It also suffered the greatest increment in transmittance at the 1028 cm−1 band corresponding to C−O ether bonds, suggesting the existence of photochemical degradation reactions that result in the dissociation of both hydrogen bonds from OH groups and C−O bonds. These reactions could be responsible for the yellowing effect observed on both printed and blank paper after exposure to this kind of radiation. As explained by Liu et al. (1997), the light absorption mechanism changes towards nonlinear absorption at ultrashort pulse lengths with high intensities. This means that mechanisms such as multi-photon absorption and avalanche ionization become more dominant. It is likely that these absorption mechanisms result in photolysis or excitation of electrons in chemical bonds which could facilitate photochemical reactions. In addition, absorption of 532 nm light with 10 ps leads to high electron peak temperatures owing to the high peak power employed. Therefore, both photochemical and photothermal reactions could be induced on cellulose.
It is known (Daniels 1996) that paper degrades naturally through three main reactions: acid-catalysed hydrolysis of cellulose molecules, oxidative degradation induced from atmospheric oxygen and light, and thermal degradation that leads to chemical bond breakage as the temperature is increased. Artificial paper ageing tests have demonstrated that these reactions become more intense if temperature is incremented. Carter (1989) has suggested that natural paper degradation reactions are accompanied by the generation of chromophores (discoloured components) that are responsible for the natural yellowing that occurs on paper under ordinary degradation. Photochemical and photothermal degradation reactions resulting from laser–paper interaction are similar to the natural deterioration mechanisms enlisted by Daniels (1996), so it seems plausible that laser radiation works as a catalyst that accelerates paper degradation as a result of multi-photon absorption and excitation of electrons in chemical bonds (leading to bond dissociation), and the generation of localized heating (as denoted in figure 7), in a way similar to accelerated ageing tests. This could in turn quicken the generation of yellow chromophores and consequently be the reason behind the yellowing observed on paper after laser radiation at 532 nm with 10 ps. All these make this laser unsuitable for toner-print removal.
(c) Infrared radiation: paper discoloration and fibre-bonding loss
Long-pulsed IR light at 1064 nm with 40 ns pulses removed toner print from paper at the cost of damaging and yellowing the paper areas under the print, as seen in figure 3. This effect was not observed over blank areas of the paper also subjected to laser radiation. This leads to the idea that during long IR pulses there is enough time for heat to diffuse from the toner layer exposed to laser radiation onto the paper substrate. The transferred heat works as a catalyst for thermally induced cellulose degradation reactions accompanied by the production of yellow chromophores. A similar effect has been reported by Kolar et al. (2002) when using a 1064 nm laser to remove charcoal powder from purified cotton cellulose. The substrate under the soil became yellowed after laser treatment. This discoloration was attributed to the creation of yellow chromophores in cellulose as a consequence of photothermal degradation driven by heat transfer from the carbon soil to cellulose.
The yellowing under the print is no longer observed when toner is ablated under the same wavelength with shorter pulses of 10 ps, although both cases produce similar mechanical damage on cellulose fibres as can be appreciated from figures 2 and 3. IR light at 1064 nm with 10 ps pulses has a bleaching/whitening effect over paper areas under the print and blank paper. The short interaction time and elevated peak power magnitude lead to a much quicker ablation of the toner layer, leaving no time for heat diffusion between toner and paper. This eliminates the yellowing under the print observed with longer laser pulses. The bleaching of the paper can be attributed to the same phenomenon observed during tests with the 532 nm and 10 ps laser, where nonlinear absorption mechanisms such as multi-photon absorption and avalanche ionization become present. ATR–FTIR spectra of the three best results obtained in this study are shown in figure 8. It can be seen that the spectra from both lasers at 1064 and 532 nm with 10 ps pulses are quite similar at the 3344 and 1028 cm−1 bands, suggesting that both have a strong and comparable interaction with the paper substrate. Therefore, the formation of discoloured chromophores during laser ablation at 1064 nm with 10 ps seems likely, as was the case for laser radiation at 532 nm with 10 ps pulses, leading to bleaching of the substrate.
The third question established in §1 of this study can be answered at this point:
— The high peak powers employed during all femtosecond tests led to elevated substrate temperatures that resulted in strong mechanical damage as a consequence of highly localized vaporization of material.
— UV radiation is not suitable for toner-print removal mostly owing to the creation of photochemical deterioration reactions in cellulose which lead to direct cellulose degradation by photolysis or by excitation of electrons in chemical bonds leading to ionization events.
— The difference between ablation under green light at 532 nm with 4 and 29 ns pulses can be attributed to the fact that the duration of the latter is long enough for secondary materials in paper to become ablated while a reduction of pulse length to 4 ns leads to a higher peak power and lower energy fluence combination which prevents any elements in paper from being ablated. The same wavelength under shorter pulses of 10 ps achieves toner removal but discolours the paper substrate as a result of accelerated chromophore generation, in a way similar to accelerated paper ageing tests.
— The same effect is true for IR radiation under 10 ps pulses. In contrast, an IR laser at 1064 nm with 40 ns pulses damages and yellows paper as a result of heat diffusion from toner to the substrate during the ablation process, facilitating the thermal degradation of cellulose.
In addition, the following topics could be further explored in future work:
— It is necessary to explore whether colour changes observed in the paper substrate after laser radiation could be related to damage of the optical brighteners commonly used in office paper. Future work could include UV–visible fluorescence spectra measurements of the lased samples to verify this.
— The changes seen at the surface of some samples could be related to thermal degradation of CaCO3 or TiO2 minerals commonly used as paper fillers. The fact that the paper is uncoated reduces the potential presence of these minerals to the filler content only. The low percentage of fillers in the paper would suggest that most of the appreciable changes seen at the surface are due to cellulose-related reactions. Despite this, X-ray fluorescence analysis could be used to verify the presence of these minerals in future work.
(d) Ablation depth prediction
This section examines the applicability of two theoretical models to predict the toner ablation depth obtained using the three best lasers identified in this study.
(i) Ablation under a 532 nm laser with 4 ns and 10 ps pulses
Ablation depth prediction for this particular wavelength under 4 ns pulses was explored by Leal-Ayala et al. (2011), who applied a one-dimensional model developed by Ready (1965) to predict the toner ablation rate as a function of energy fluence. Ready's model is based on the assumption that the laser energy absorbed at the surface of an object quickly heats up the material to its melting point and then to the vaporization temperature. At this point, the energy per unit mass deposited in the material becomes larger than the specific heat of evaporation and the laser starts supplying the latent heat of vaporization to a thin layer of the surface, creating a vapour front that is removed immediately. The surface, initially at z=0, moves inward with time, reaching a position z(t) at time t. A schematic view of this process and a detailed theoretical explanation can be found in the electronic supplementary material. Ready's adapted model is described by 5.1 where x represents the effective energy fraction, A is the absorption fraction, tp equals pulse time in seconds, heat flux H is measured as W m−2, ρ stands for material density (kg m−3), C is the specific heat capacity (J(kg°K)−1), Tv is the vaporization temperature (°K) and Lv stands for latent heat of vaporization (J kg−1). This model can predict the ablation depths obtained when varying the energy fluence of a 532 nm laser with 4 ns and 10 ps pulses used to remove HP LaserJet Q1338A-AC-D black toner from paper. The experimental values used by Leal-Ayala et al. (2011) and the ablation depth predictions are shown in table 5 and figure 9, respectively. The power H and the pulse width tp were modified according to the characteristics of the lasers under study while the rest of the parameters remained constant. Considering that the toner under study is mainly composed of 50 wt% polyester resin and 50 wt% iron oxide, a key assumption made is that full toner removal can be achieved by degrading the polyester fraction, as this action would detach the rest of the toner elements from the paper. Assuming a 50 per cent heat flow loss to the substrate (rule of thumb), the remaining energy is distributed between the two material fractions that compose the toner. The polyester resin fraction degrades at a lower temperature than iron oxide, which means that only 25 per cent of the energy is available to detach the toner.
(ii) Ablation under a 1064 nm laser with 10 ps pulses
The model proposed by Ready (1965) fails to predict the ablation rate obtained with a 1064 nm laser working with 10 ps pulses. This could be related to the fundamental difference in the way visible and IR light are absorbed and inter-relate with toner at an atomic level. Although there is no general consensus in the literature regarding IR ablation depth models for polymer composites such as toner, the logarithmic dependence of the ablation depth on the laser pulse fluence, based on the Beer–Lambert absorption law, is well known for the ablation of polymers and has also been demonstrated for ultrafast ablation of metal targets. This is exemplified by the work of Schmidt et al. (2002) and Preuss et al. (1995), respectively. Toner is neither a metal nor a pure polymer, but a polymer composite containing a high fraction of electrically conductive iron oxide. The ablation rate described by the Beer–Lambert law is valid as long as the optical penetration depth exceeds the thermal penetration depth, as this means that the optical effects would govern the ablation process. The use of picosecond pulses allows this method to be employed as very low thermal penetration depths are achieved. The model is described as 5.2 where D stands for ablation depth, α is the absorption coefficient, F is the applied fluence and Fth is the threshold fluence for ablation. The last has been determined experimentally as 0.035 J cm−2. The absorption coefficient α can be obtained by simple curve fitting using equation (5.2) and the depth measurements obtained experimentally as shown in figure 10. The value of the absorption coefficient α has been determined as 4.4×106 cm−1, giving an optical penetration depth of 0.0022 μm. The thermal penetration depth can be estimated from the formula 2√(Dτ), where D refers to thermal diffusivity (estimated at 1.1×10−3 cm2 s−1 for toner) and τ is the pulse length (10 ps), giving a thermal penetration of 0.0021 μm. Since the optical penetration depth is practically the same as this value, it can be said that the Beer–Lambert absorption law is valid to describe the process.
The results shown in figures 9 and 10 demonstrate that the ablation depth is directly determined by the energy fluence employed, regardless of the pulse width value. In any case, the pulse length remains a critical parameter owing to the distinct impacts that it has on the paper substrate under the toner, as discussed before. The data presented in this section confirm that it is possible to control the laser ablation of toner and provide an answer to the fourth question set in §1 of this study. The use of these models for other types of lasers and toners is an area that remains open and needs to be explored in future work.
6. Certainty about best results and future work
Additional research by Leal-Ayala (2012) has looked at the commercial feasibility of laser un-printing by exploring the mechanical damage and ageing stability of paper after laser treatment, and the safety, energy and economic performance of the process. They found that the best laser identified in this study (4 ns pulses with a 532 nm wavelength) does not generate any significant mechanical damage on paper (based on bending, curl and tensile properties) and the ageing stability of paper (tested through accelerated ageing) remained unaffected by the treatment and was comparable to that of blank un-lased paper. Furthermore, it was found that toner ablation results in the formation of nanoparticles and emission of mostly harmless gases that can easily be captured by employing high-efficiency particulate arresting filters and a gas extraction system, ensuring the safety of the process.
Regarding the cost-effectiveness of the laser solution, Leal-Ayala (2012) applied a cost of ownership model to establish the maximum capital costs that would allow this technology to be competitive when compared with recycled paper (considering that a ream of Xerox Supreme Recycled paper has an average retail price of £0.03 per sheet). Assuming that the equipment has a 7 year lifetime, a utilization of 20 per cent, variable costs of £1408 (energy) and a throughput rate of 50 pages per hour, it is estimated that a laser un-printer needs to have a maximum capital cost of £16 800 to compete with the £0.03 per sheet cost of recycled paper. This is considered to be a feasible goal by the authors as they estimate that a purpose-built laser un-printer prototype with the performance characteristics of the QuikLaze 50ST2 used in this study (4 ns pulses with a 532 nm wavelength) could be built for around £19 000 at the present time.
With respect to the environmental benefit of laser removal, Leal-Ayala (2012) estimates that unprinting could translate into CO2-eq savings between 52 and 79 per cent when compared with paper recycling (considering that 1 tonne of recycled paper generates between 650 and 1510 kg CO2-eq per tonne of paper and maximum emissions from laser un-printing could be limited to 310 kg CO2-eq per tonne of paper). These topics define the scope of an ongoing project in this area and future work will further expand on these results and estimates.
The toner-print removal ability of a range of UV, visible and IR lasers working under long and ultrashort pulses has been tested in this study. The following conclusions were obtained:
— Toner print can be effectively removed from paper by UV, visible and IR ultrafast laser ablation.
— Only two ultrafast lasers at 532 and 1064 nm with 10 ps pulses are capable of leaving paper with a high-enough quality to replace new paper. The former resulted in slight yellowing of the paper substrate while the latter produced minor fibre bleaching and cohesion loss.
— UV radiation failed as an appropriate solution to the problem under study. This behaviour could be attributed mostly to the creation of photochemical degradation reactions in cellulose as a result of UV light absorption.
— Green light at 532 nm with 29 ns pulses results in ablation of the secondary materials in paper (lignin, hemicellulose and so on). This effect is no longer appreciated when the pulse length is reduced to 4 ns as this leads to a higher peak power and lower energy fluence combination, which prevents any elements in paper from being ablated. The latter represents the best laser set-up for toner-print removal.
— IR light at 1064 nm with 40 ns pulses damaged and yellowed the paper areas under the print but did not interact with blank paper. The yellowing could be accredited to thermal degradation accompanied by the production of yellow chromophores as a result of heat diffusion from the toner layer to the paper substrate. This effect disappears when the pulse length is reduced towards 10 ps since this leads to a much quicker ablation of the toner layer, leaving no time for heat diffusion between toner and paper.
— Both IR and visible trials using 10 ps pulses resulted in discoloration of the paper substrate owing to the accelerated generation of discoloured chromophores.
— All lasers using femtosecond pulses generated great physical damage on paper owing to the vaporization of highly localized material volumes.
— An optimum operating window for the laser ablation process has been presented for the ablation of HP LaserJet Q1338A-AC-D black toner on white, uncoated, wood-free 80 g m−2 Canon copy paper. Future work should focus on expanding this window by testing a wider range of toner–paper combinations.
— Two ablation models taken from the literature have been used to predict the ablation depths obtained when removing HP LaserJet Q1338A-AC-D black toner from paper. The models compared favourably with experimental measurements, suggesting that the toner-removal process can be controlled.
We would like to thank Johannes Heberle, Peter Bechtold and Ulf Quentin at the Lehrstuhl für Photonische Technologien for their assistance in setting up some of the experiments reported in this article.
- Received October 20, 2011.
- Accepted February 16, 2012.
- This journal is © 2012 The Royal Society