Environmental Impact of the Reclaimed Sand Addition to Molding Sand with Furan and Phenol-Formaldehyde Resin—A Comparison

Increasingly strict regulations, as well as an increased public awareness, are forcing industry, including the foundry industry, to develop new binders for molding sands, which, while being more environmentally friendly, would simultaneously ensure a high quality of castings. Until recently, binders based on synthetic resins were considered to be such binders. However, more accurate investigations indicated that such molding sands subjected to high temperatures of liquid metal generated several harmful, even dangerous substances (carcinogenic and/or mutagenic) from the benzene, toluene, ethylbenzene and xylenes (BTEX) and polycyclic aromatic hydrocarbons groups (PAHs). An assessment of the most widely used molding sands technologies at present with organic binders (synthetic resins) from the no-bake group (furan no-bake and phenolic-ester no-bake) and their harmfulness to the environment and work conditions is presented in this paper. In the first stage of this research, gases (from the BTEX and PAHs groups) emitted when the tested molds were poured with liquid cast iron at 1350 °C were measured (according to the authors’ own method). The second stage consisted of measuring the emission of gases released by binders subjected to pyrolysis (the so-called flash pyrolysis), which simulated the effects occurring on the boundary: liquid metal/molding sand. The gases emitted from the tested binders indicated that, in both cases, the emission of harmful and dangerous substances (e.g., benzene) occurs, but, of the given binder systems, this emission was lower for the phenolic-ester no-bake binder. The obtained emission factors of BTEX substances show higher values for furan resin compared to formaldehyde resin; for example, the concentration of benzene per 1 kg of binder for furan no-bake (FNB) was 40,158 mg, while, for phenol-formaldehyde no-bake (PFNB), it was much lower, 30,911 mg. Thus, this system was more environmentally friendly.


Introduction
Metal casting involves pouring a molten metal into a hollow mold to produce metal objects. Cores are used in a casting process to form cavities, hole protrusions, recesses and casting products, which are not possible to be shaped by the mold. These molds and cores are generally made of molding There are two primary sources of emissions from resin-based binders: • Evaporation of solvent, byproduct or chemical constituent occurring during mixing, core/mold making and core/mold storage, prior to pouring • Thermal decomposition during pouring, cooling and shakeout operations.
Special attention should be placed on the emission of substances from the BTEX and PAHs groups since many of them are carcinogenic and mutagenic. The aim of these study was not only the determination of the emission of substances in the BTEX and PAHs groups from the molding sands with furfuryl and phenolic-formaldehyde resins under the laboratory scale but also determining the influence of the reclaimed sand addition. All these elements are factored into the overall assessment of the harmful influence of the given molding sand on the environment and employees. Due to this, it is possible to protect nature against hazardous substances [27][28][29][30].

Methods
Investigations of the gas emission in the tested foundry plant were performed according to the original method developed in the Faculty of Foundry Engineering, AGH-UST (Polish Patent, No. P-398 709; 2012). The schematic presentation of the experimental stand is given in Figure 1.

Pyrolysis-Gas Chromatography/Mass Spectrometry (Py-GC/MS)
The analysis was carried out in a platinum coil. Approximately 1 mg of the solid sample was centered in a quartz tube and heated up to 1100 °C (heating ramp of 10 °C/ms) using a (Py) pyroprobe (Pyroprobe 5000, CDS, Analytical Inc., Oxford, PA, USA). The pyrolysis products were separated using Gas Chromatograph on a 30 m × 0.25 mm × 0.25 μm (film thickness) capillary column (Rxi-5MS, Restek, Bellefonte, PA, USA). The flow rate of the carrier gas (He, 99.9999%) was 1 mL/min. A Single Quadrupole (ISQ, Thermo Scientific, Waltham, MA, USA) MS was used to detect the pyrolytic degradation products (scan mode: (30-600) atomic mass units (a.m.u.); electron energy (EI): 70 eV; emission current: 50 μA). The obtained mass spectra were compared with the mass spectra given in the NIST MS Search 2.0 Libera (Chemm. SW, Version 2.0, Fairfield, CA, USA).
• During the first stage of the research, the composition of gases evolving from molding sands, prepared on fresh sand matrices and poured with cast iron with a temperature of 1350 °C, was tested (samples marked: FNB and PFNB) [11]. • During the second stage, the emission of substances evolving from molding sands, prepared on fresh sand matrices with various fractions of a reclaim and poured with cast iron of a temperature of 1350 °C, was measured (samples marked: FNBXRYFS, where XR is the percent of reclaim fraction and YFS is the percent of fresh sand fraction). • During the third stage, the "flash pyrolysis" of hardened resins was performed by Pyrolysis-Gas Chromatography-Mass Spectrometry (Py-GC/MS) technique at a temperature of 1100 °C. This experiment simulated processes occurring directly on the boundary of molding sand and liquid alloy [13,30]. The analysis was carried out in a platinum coil. Approximately 1 mg of the solid sample was centered in a quartz tube and heated up to 1100 • C (heating ramp of 10 • C/ms) using a (Py) pyroprobe (Pyroprobe 5000, CDS, Analytical Inc., Oxford, PA, USA). The pyrolysis products were separated using Gas Chromatograph on a 30 m × 0.25 mm × 0.25 µm (film thickness) capillary column (Rxi-5MS, Restek, Bellefonte, PA, USA). The flow rate of the carrier gas (He, 99.9999%) was 1 mL/min. A Single Quadrupole (ISQ, Thermo Scientific, Waltham, MA, USA) MS was used to detect the pyrolytic degradation products (scan mode: (30-600) atomic mass units (a.m.u.); electron energy (EI): 70 eV; emission current: 50 µA). The obtained mass spectra were compared with the mass spectra given in the NIST MS Search 2.0 Libera (Chemm. SW, Version 2.0, Fairfield, CA, USA).

•
During the first stage of the research, the composition of gases evolving from molding sands, prepared on fresh sand matrices and poured with cast iron with a temperature of 1350 • C, was tested (samples marked: FNB and PFNB) [11].

•
During the second stage, the emission of substances evolving from molding sands, prepared on fresh sand matrices with various fractions of a reclaim and poured with cast iron of a temperature of 1350 • C, was measured (samples marked: FNBXRYFS, where XR is the percent of reclaim fraction and YFS is the percent of fresh sand fraction).

•
During the third stage, the "flash pyrolysis" of hardened resins was performed by Pyrolysis-Gas Chromatography-Mass Spectrometry (Py-GC/MS) technique at a temperature of 1100 • C. This experiment simulated processes occurring directly on the boundary of molding sand and liquid alloy [13,30]. To determine the influence of reclaimed sand added to matrices of tested molding sands on the amount and kind of emitted gases, molding sands containing 100%, 50% and 0% of reclaimed sand were prepared and marked as follows: • Molding sand with furan resin: FNB100FS (100% fresh sand), FNB50R50FS (50% reclaimed sand + 50% fresh sand) and FNB100R (100% reclaimed sand).
All samples were subjected to the same research procedure. The analysis of BTEX was carried out using the gas chromatography method with the application of a flame-ionizing detector (FID) (TRACE GC Ultra Thermo Scientific, Waltham, MA, USA).
Substances from the PAHs group were analyzed using the gas chromatography technique (FOCUS GC) coupled with MS ISQ Thermo Scientific (GC/MS, Waltham, MA, USA).

Investigations of the Gases Emitted from Molding Sands Prepared on the Fresh Sand Matrix
The obtained results of the research (Tables 3 and 4) conducted according to the methodology developed by the authors [29] showed that for both BTEX and PAHs emission from the FNB molding sand is greater than from the PFNB molding sand. This is consistent with the research presented in [10]. The results of analyses of gases from the BTEX group emitted from FNB and PFNB molding sands, prepared on the fresh sand matrix, are shown in Table 3. The total volume of gases emitted from the PFNB molding sand was more than 20% greater than from FNB molding sand.
Among the BTEX gases emitted from both molding sands, benzene was predominant, constituting more than 90%. Moreover, despite the fact that the volume of gases emitted from the FPNB sand was higher than from FNB sand, emission of BTEX, including carcinogenic benzene, recalculated for 1 kg of molding sand, was lower by approximately 25%. The process of gases emission reached the maximum speed after 100 s, while the emission ended after approximately 250 s from pouring liquid metal into the mold.
Both tested molding sands under a high temperature influence also emit substances from the PAHs group (Table 4). Emissivity of these substances from FNB molding sand is approximately 20% higher than from PFNB molding sand. The main PAHs compounds emitted from FNB sand are fluoranthene, pyrene and phenanthren, while from PFNB sand they are pyrene and acenaphtylene. The concentration of benzo(a)pyrene, a highly carcinogenic substance, is very low for both binders.
"Flash" pyrolysis simulated the pouring temperature of casting alloys (from nonferrous alloys through cast iron to cast steel). Investigation of compounds formed during the "flash" pyrolysis was conducted by means of the coupled equipment consisting of Py-GC/MS. Pyrolysis was carried out at 1100 • C. Analytical data collecting information from the Py-GC/MS techniques can be found in Table 5. Figure 2a,b shows chromatograms obtained for FNB and PFNB, respectively, at temperature of 1100 • C.
The chromatograms obtained for both resins are very similar in terms of qualitative analysis (type of evolved gases), while differences occur in the concentration range. In both cases, mainly gases from the BTEX group and their derivatives (one, two and three methylene) and phenol with its derivatives are emitted. Gases released from FNB resin contain 50% toluene, while it is negligible in gases emitted by the PFNB resin. Gases released from the FNB resin are also present: SO 2 originated from the p-toluenesulfonic acid (PTSA) (hardener) and nitrogen compounds (probably introduced at the resin production stage).   gases from the BTEX group and their derivatives (one, two and three methylene) and phenol with its derivatives are emitted. Gases released from FNB resin contain 50% toluene, while it is negligible in gases emitted by the PFNB resin. Gases released from the FNB resin are also present: SO2 originated from the p-toluenesulfonic acid (PTSA) (hardener) and nitrogen compounds (probably introduced at the resin production stage).

Investigations of the Gases Emitted from Molding Sands Prepared on the Reclaimed Sand Matrices
An addition of the reclaimed sands to matrices of both molding sands caused a significant increase in the volume of emitted gases. For the FNB sand, this volume doubled (when the matrix was made of 100% of reclaimed sand), while for the PFNB sand the increase was lower (70%) (Tables 2 and 6). The mechanical reclamation process of spent molding sand does not fully remove the hardened binder from sand grains. Thus, by adding the reclaimed sand to sand matrices, additional amounts of inactive binder are introduced into molding sands, increasing the volume of emitted gases and constituting a higher

Investigations of the Gases Emitted from Molding Sands Prepared on the Reclaimed Sand Matrices
An addition of the reclaimed sands to matrices of both molding sands caused a significant increase in the volume of emitted gases. For the FNB sand, this volume doubled (when the matrix was made of 100% of reclaimed sand), while for the PFNB sand the increase was lower (70%) (Tables 2 and 6). The mechanical reclamation process of spent molding sand does not fully remove the hardened binder from sand grains. Thus, by adding the reclaimed sand to sand matrices, additional amounts of inactive binder are introduced into molding sands, increasing the volume of emitted gases and constituting a higher risk to the environment. The parameter LOI can be a measure of the binder remaining on the reclaimed sand grains. The dependence of the LOI of the given molding sand on the reclaimed sand fraction in its matrix is presented in Figure 3. Both the BTEX emission and the LOI values are proportional to the fraction of reclaimed sand (Figure 3). The LOI for the tested molding sand should be ≤2; this level of organic substances can be obtained by thermal reclamation [23]. However, high costs of energy carriers lead foundries to mainly apply mechanical reclamation, which is less efficient. Not totally removing binders from sand grain surfaces can be the reason for casting defects, due to emission of too high amounts of gases and environmental contamination.
The results of emission of substances from the BTEX group from molding sands having matrices with the reclaimed sand fraction 0%, 50% and 100% under a high temperature influence are shown in Table 6. Total substitution of fresh sands by reclaimed sands in FNB molding sand caused a threefold increase in the concentration of BTEX substances (fresh sand, 336.6 mg per 1 kg molding sand; 100% reclaimed sand fraction, 1057 mg per 1 kg molding sand).
As far as substances from the PAHs group are concerned, PFNB molding sand also releases less of these substances than FNB. The amounts of PAHs emited from PFNB molding sand are at the same level, regardless of the reclaimed sand fraction in a molding sand (Table 7). Table 6. Concentration of BTEX formed during thermal decomposition of molding sand (FNB and PFNB) with reclaimed sands [32,33].

Sample Code
Volume risk to the environment. The parameter LOI can be a measure of the binder remaining on the reclaimed sand grains. The dependence of the LOI of the given molding sand on the reclaimed sand fraction in its matrix is presented in Figure 3. Both the BTEX emission and the LOI values are proportional to the fraction of reclaimed sand (Figure 3). The LOI for the tested molding sand should be ≤2; this level of organic substances can be obtained by thermal reclamation [23]. However, high costs of energy carriers lead foundries to mainly apply mechanical reclamation, which is less efficient. Not totally removing binders from sand grain surfaces can be the reason for casting defects, due to emission of too high amounts of gases and environmental contamination. The results of emission of substances from the BTEX group from molding sands having matrices with the reclaimed sand fraction 0%, 50% and 100% under a high temperature influence are shown in Table 6. Total substitution of fresh sands by reclaimed sands in FNB molding sand caused a threefold increase in the concentration of BTEX substances (fresh sand, 336.6 mg per 1 kg molding sand; 100% reclaimed sand fraction, 1057 mg per 1 kg molding sand). Table 6. Concentration of BTEX formed during thermal decomposition of molding sand (FNB and PFNB) with reclaimed sands [32,33]. As far as substances from the PAHs group are concerned, PFNB molding sand also releases less of these substances than FNB. The amounts of PAHs emited from PFNB molding sand are at the same level, regardless of the reclaimed sand fraction in a molding sand (Table 7).  However, for the FNB molding sand, the amount of emitted PAHs significantly increases when the reclaimed sand fraction in the matrix increases. When the matrix contained 100% reclaimed sand, the emission of substances from the PAHs group was double that when the matrix contained only fresh sand. The main component of released PAHs was naphthalene, which in a critical case constituted nearly 70% all PAHs. Due to the carcinogenic and/or mutagenic influence of substances occurring in these gases (benzene, toluene, formaldehyde and benzo(a)piren), which are either the initial components or products of reactions occurring under the high temperature influence in tested binders, the intensive research leading to the elimination or limitation of such components is being carried out, e.g. phenol free, formaldehyde free, furfuryl alcohol free and sulfuric acid free.

Volume of
The results presented in this paper only concern the tested binder systems and should not be generalized. Investigations of new systems are necessary to properly assess their influence on the environment.

Conclusions
An assessment of the harmfulness for the natural and work environment created by gases from BTEX and PAHs groups released at high temperature from two molding sands with organic resin binders was performed. The most widely applied binders in the production of molds are furan no-bake acid catalyzed (FNB) and phenolic esters no-bake (PFNB). The following conclusions can be drawn:

•
Molding sands with the PFNB binder release a low amount of substances from the BTEX group (by up to 25%) than molding sands with the FNB binder. In both cases, benzene constituted more than 90%.

•
Molding sands with the PFNB binder release nearly 50% less substances from the PAHs group than molding sands with the FNB binder.

•
The fresh sand substitution by a reclaimed sand in the matrix causes a significant increase (up to threefold) of the released gases from both groups. In addition, BTEX emission from the PFNB binder is twice lower than that from the FNB binder. The reclaimed sand addition to the PFNB sand influences the emission of PAHs only to a small degree.

•
The LOI parameter can be useful in the case of these molding sands-for assessing amounts and approximated composition (e.g., benzene content) of evolving gases.