Extended wet pressing at elevated temperature enables enhanced dewatering for tissue and linerboard

2.1 Materials

The fibre material of tissue samples (grammage 20 g/m²) was fresh BSKP pulp (mix of pine and spruce) from Metsä Fibre’s Äänekoski mill with a typical fibre length of 2.0 ± 0.1 mm and a Canadian standard freeness (CSF) of 700 ± 50 ml. The old-corrugated-container (OCC) type recycled pulp used for making linerboard samples (grammage 150 g/m²) was obtained by hot disintegrating a brown cardboard box. In both cases, wet laboratory sheets were prepared using a standard Lorentzen & Wettre forming device (Figure 1). After sheet forming, the solids content of the samples was adjusted to 20 ± 1 % by couching them against blotter papers. To prevent evaporation, and to protect wet sheets from damage prior to pressing, the sheets (tissue supported by a blotter paper) were placed between plastic plates and put inside a plastic bag in cold storage.

Figure 1:

Preparation of the samples: a) Lorentzen & Wettre forming device used for laboratory sheet preparation. b) Adjusting the solids content to 20 % by couching. c) Final wet tissue sample.

2.2 Experimental setup and procedure

The pressing of a wet sample took place in a modified MTS device (Hii et al. 2012; Järvinen et al. 2018) between heated top and bottom metal plates. The MTS device included a pressure sensor of 100 kN and the Dasylab control software for producing the desired pressure pulse. Based on surface temperature measurements with a contact thermometer (Testo 925), the top plate reached a temperature close to the target value (100 °C for extended pulses), although it oscillated by an amplitude of ca. 5 °C due to the intermittent operation of the on/off controller. The grooved roll surface on the bottom plate was covered by a typical commercial tissue (2.0 mm thick, 1,060 g/m²) or lightboard (3.5 mm thick, 1,110 g/m²) felt (Tamfelt Ltd), and the wet circular web sample was laid on the felt prior to pressing. Although a temperature of 70 °C was applied to the grooved bottom plate, the felt would not reach more than 40–50 °C due to insufficient thermal contact between the plate and the felt. The sample diameter was either 53 or 90 mm, depending on the target peak pressure. The felt diameter was 60 mm for tissue and 140 mm for board. The solids content of the sample was checked just before the pressing operation by comparing its weight to a previously measured average dry weight of products of the same grade. The obtained values for the board samples turned out to be very close to the target, with an average and standard deviation of 20.0 ± 0.2 %. However, some of the thin tissue samples had slightly higher solids content than the target. In such a case, the solids content was lowered by spraying a suitable amount of water on the sample a couple of minutes before the test, so the moisture had sufficient time to stabilise inside the sample. The sprayed sample was subsequently dried slowly until the desired solids content of 20 % (19.6 ± 0.5 %) was achieved. The moisture content of the felt was followed in a similar fashion. During the pressing experiments, the felt was kept moist (containing 500 g/m² of water for tissue and 800 g/m² of water for linerboard), because this seemed to improve water removal from a sample.

In our experiments, thin aluminium foil (thickness 10 µm) was placed on the top of the sample (Figure 2a) and attached to the top metal plate on one edge with adhesive tape (Figure 2b). The aluminium foil, which adhered to the sample during the pressing, served several purposes: It prevented the sample from adhering to the top plate and being heated by conduction after the pulse; it shielded the sample from radiative heating from the top plate by its reflectivity; and it protected one side of the sample from losing moisture by evaporation before the weighing of the sample on a balance (Figure 2c). The transfer of the pressed foil-sample system from the MTS device to an analytic balance took about 2 s, after which the final solids content (SC) was determined based on the measured weight:

where M _fs is the foil-sample weight, M _f is the foil mass, and M _d is the dry weight of the tested sample as obtained afterwards according to the standard ISO 638-1:2021. The shoe-nip type of pressing for board was always carried out with a felt on both sides of the board sample, and in this case, the aluminium foil was not applied. The pressing experiments were carried out in a laboratory climate of 21 °C temperature and 50 % relative humidity.

Figure 2:

Different experimental steps: a) Cutting a wet sample and shielding it from one side with aluminium foil. b) Wet pressing of the sample against a moist felt. After removing the load, the sample-foil system detaches immediately from the top hot metal surface and from the felt. c) Rapid transfer of the pressed sample to the analytical balance. During this step, the foil blocks moisture evaporation through the upper sample surface.

To observe the thermal transfer through the aluminium foil, we included a thermocouple wire (cross section size 0.25 mm × 0.5 mm) between the sample (5–10 mm inside from the edge) and the felt in tests for both the tissue and the board. This allowed us to monitor the temperature of the bottom sample surface. Moreover, a small set of comparison measurements without the aluminium foil were carried out for both tissue and board to assess the effect of the foil and to support our model analysis of moisture evaporation after pressing.

2.3 Studied pressing conditions

In the case of tissue, a 150 ms extended pulse was compared against a short 5 ms pulse mimicking a conventional roll nip. In the case of linerboard, a 540 ms extended pulse was compared against a 42 ms pulse mimicking a shoe nip with felt on both sides on a slower machine (Table 1). The different pulse lengths reflect the different running speeds of tissue and linerboard machines. For linerboard, the extended pulse was preceded by a shoe nip pulse in order to simulate a setup where preliminary dewatering is performed by a shoe nip. The comparison measurements included pressing with either one or two consecutive shoe nip pulses. The extended pulses consisted of several phases with different pressure levels, simulating different phases in a steel belt press. Initially, a low-pressure phase (80 ms for tissue, 250 ms for board) simulated the steel belt tension against the grooved pressing roll on a production machine. Based on the applied tension relative to the roll radius, this pressure was estimated to be 0.15 MPa. For tissue, the initial pulse was followed by a 70 ms main pulse. For board, the initial pulse was followed by a 220 ms intermediate phase at 1 MPa, followed by a 70 ms main pulse. The applied temperatures of the top (100 °C) and bottom (70 °C) plates were equal for the extended pulse for both sample types, but their peak pressures slightly differed (Table 1). Similarly, the peak pressure for the shoe nip (6 MPa) was higher than for the roll nip (3 MPa).

The pressing force and the position of the top plate were recorded with 0.1 ms intervals. Typical pressure curves for the different pulse types are shown in Figure 3. The number of parallel measurements for tissue was 12 for the extended pulse and 9 for the roll-nip pulse. For linerboard, the shoe + extended pulse was repeated 10 times, the shoe nip pulse 15 times and the double shoe nip 5 times.

Figure 3:

Typical pressure (blue circles) and temperature (orange triangles) curves for the extended pulses for (a) tissue and (b) board. As the temperature sensor detached from the sample at the end of the pulse, temperature development after the pulse is not shown. The recorded maximum temperature at the end of the pulse was slightly lower than the target top pressing temperature because of the included aluminium foil in the experiments.

3.1 Wet pressing of tissue

Pressure and temperature measurements indicated that the target levels of Table 1 for the wet pressing parameters were achieved quite well in the experiments. In general, the reproducibility of the extended pulse (Figure 3a) was good. The actual pulse lengths for tissue were 150–175 ms, the target being 150 ms. The measured nominal peak pressures were 7.9–8.4 MPa (target 8 MPa). The actual pressure on the sample could be slightly lower than the nominal value towards the end of the pulse if the sample sinks into the felt (which is slightly larger than the sample) and the felt around the sample carries some of the load. According to the temperature measurement, sample heating began already during the initial 0.15 MPa pressure pulse. Immediately after the main pulse, the measured temperature of the bottom surface of the sample was about 80 °C (Figure 3).

Figure 4 shows the obtained tissue solids content after wet pressing for all parallel tests. The average value was 83 ± 2 % (error estimate given as the 95 % confidence interval for the mean), with 3 % standard deviation among the individual measurements (12 replicates). This result is slightly affected by moisture evaporation through the bottom sample surface during the 2-s delay before the sample weighing. In a later section, we give a quantitative estimate of this effect based on a model analysis.

Figure 4:

Measured tissue sample solids content after wet pressing for different tests using the extended pressure pulse (blue dots) and the short roll-nip pulse (orange dots). The dashed lines give the averages of the parallel measurements.

The results with the extended pulse were compared with the much shorter roll-nip pulse. Unfortunately, because of the limitations of the MTS device, it was challenging to sufficiently repeat the short 5-ms pulse in the experiments. This led to a large scatter of the measured solids content values, as shown in Figure 4. However, all measured values were clearly below the level found with the extended pulse, with the average level of 48 ± 8 %, even having equal plate temperatures to the extended pulse. As seen in Figure 3, heating up a tissue sample required at least 100 ms, so with the 5 ms pulse, the temperature change was very limited. Furthermore, the peak pressure of the roll-nip pulse was less than half of that for the extended pulse.

In the current typical tissue production, the roll-nip pulse is limited by the Yankee cylinder deflection (Edwards et al. 2008). When replacing the Yankee cylinder with an extended pulse at high temperature, there is a deflection compensated roll instead of the Yankee cylinder enabling a much higher load (around 1,500–3,000 kN/m) than what is possible with Yankee (max c.a. 150 kN/m).

3.2 Wet pressing of linerboard

We found the realised shoe press pulses to be longer than the target at the root of the pulse due to two soft felts, 65 ± 9 ms, whereas the peak pressure 5.85 ± 0.04 MPa was close to the target (Table 1). The extended pressure pulses were on the mark with a duration of 570 ± 12 ms and a peak pressure of 5.97 ± 0.02 MPa. The main pulse had a stepwise shape as shown in Figure 3b simulating the increasing pressure levels that can be arranged in an extended pressing nip on a production scale. In addition, the applied temperature during the extended pressing was roughly double of the shoe-nip temperature as indicated in Table 1.

The measured average solids content, achieved with the combination of shoe-nip and extended pulses, was 62 ± 2 % (95 % confidence), with 3 % standard deviation of 10 parallel measurements (Figure 5). This exceeded the solids content obtained with single (40 ± 2 %) or double shoe nips (47.6 ± 0.3) by as much as 14–22 % points.

Figure 5:

Measured linerboard sample solids content after wet pressing for different tests using the extended pressure pulse (blue dots), the single shoe-nip pulse (orange dots) and the double shoe-nip pulse (grey dots). The dashed lines give the averages of the parallel measurements.

3.3 Model estimate on the amount of evaporation during the sample transfer

Due to the elevated initial temperature of the sample during the transfer to the analytical balance (taking about 2 s), evaporation from the side not covered by the aluminium foil caused a meaningful increase in the solids content at least in the case of the tissue sample. As the amount of evaporation was not experimentally measurable, we estimated it with the following model, calibrated with an evaporation test in moving air at room temperature.

In the model, we considered vapour transfer through the air boundary layer on the side of the sample not covered by aluminium foil. The mass transfer flux J_v through the boundary layer is governed by the concentrations of the vapour on either side of the layer and a mass transfer coefficient γ:

Here M_H2O is the molar mass of water, R is the molar gas constant, T_web and T_ext are the temperatures of the web and the surrounding air, respectively. P_v,web=ϕ_web(M,T_web)P_s(T_web) is the partial pressure of vapour in the web pores given by the desorption, where ϕ_web is the relative humidity of the web pores given by the desorption isotherm (in practice, ϕ_web is close to unity), and P_s(T) is the saturation vapour pressure. The moisture ratio of the web is denoted by M. The external vapour pressure is given by P_v,ext=ϕ_extP_s(T_ext), where ϕ_ext is the external relative humidity. As P_s increases sharply with increasing temperature, the flux is very sensitive to the temperature difference across the boundary layer. Therefore, the evaporation is initially fast until the web cools down. Heat is lost from the web by both direct heat convection across the boundary layer as well as by the heat carried by the escaping vapour. The fraction of heat carried away by vapour strongly affects the mass loss before reaching the ambient temperature. For air, the ratio between γ and the heat transfer coefficient κ is given by the Lewis relation (Çengel and Ghajar 2015)

where ρ_ac_p is the volumetric heat capacity of the moist air. The heat flux across the boundary is given by

where h_v(T) is the enthalpy of the vapour leaving the web. The factor of two accounts for the two sides available for heat convection. Finally, we need to write down the heat equation for the web and aluminium foil:

where ρ f A and ρ al A are the area density of the dry fibres and aluminium foil, respectively, and c_p,f, c_p,w and c_p,al are the specific heats of the dry fibres, water and aluminium, respectively. We neglect the small amount of heat stored in the vapour within the web for simplicity. Combining Equation (5) with the mass balance equation ρ f A M ˙ = J v , we can solve the evolution of M and T with time (Figure 6).

Figure 6:

The simulated solids content (blue curves) and the temperature (red curves) of tissue (solid lines) and linerboard (dashed lines) during the transfer of the sample from the press to the analytical balance, considering moisture loss due to evaporation. The temperature of the thermocouple at the end of the pulse is shown by red dots. The measured solids content after the transfer to the balance (taking 2 s) is shown by blue dots. Filled dots stand for tissue, open dots for board.

To fix the heat transfer coefficient using Equation (3), we exposed a sample to moving air by manually waving it in the air (after pressing, with aluminium foil) and recorded the mass loss as a function of time. Fitting to the data yielded γ ≈ 0.015 kg s⁻¹ m⁻², corresponding to κ ≈ 20 W K⁻¹ m⁻² based on Equation (3). The same κ was used on either side.

The parameters for the drying simulations are shown in Table 2. The initial moisture contents were fitted to approximately yield the measured average solids content of 83 % and 62 % (tissue and board, Figures 4 and 5) at the end of the 2 s simulation. Initial sample temperatures were conservatively assumed to be slightly higher than the measured values after the pulse (Figure 3), as the temperature sensor was under the sample and thus possibly slightly cooler than the average sample temperature. Our results indicate a 5-percentage-point increase in the solids content in the tissue and a 1.5-percentage-point increase in the linerboard (Figure 6). These values can be subtracted from the measured values to obtain the estimated solids content immediately after the pulse.