Preparation of Aluminum Dross Non-Fired Bricks with High Nitrogen Concentration and Optimization of Process Parameters

Abstract

In order to solve the difficulties in the utilization of aluminum dross resources, non-fired bricks with aluminum dross with high nitrogen concentration as the main raw material were prepared. Three process parameters, including forming pressure, mixing-water amount, and aluminum dross particle size, were subjected to single-factor experiments. Based on the response surface method, a mathematical model was established between the process parameters and the non-fired bricks’ compressive properties, which were subjected to ANOVA. The process parameters were optimized and then verified experimentally. According to the results, the established regression model is able to accurately predict the compressive properties of non-fired bricks. The difference between the experimental value and the model’s predicted value was only 0.36%. The optimal process parameters for aluminum dross to prepare non-fired bricks are as follows: forming pressure is 18 MPa, mixing-water amount is 15% and particle size range is 80–130 mesh. The compressive strength of the prepared non-fired bricks is 24.66 MPa, which meets the requirement of MU20 non-fired bricks in Non-fired Rubbish Gangue Bricks.

1. Introduction

Aluminum dross with high nitrogen concentration refers to aluminum dross with high aluminum nitride content. High-temperature melting operations are essential in processes such as remelting, casting and recycling of metallic aluminum. In the production process, some aluminum will react with nitrogen to generate a large amount of aluminum nitride [1,2,3]. Such aluminum dross generally has a lower fluorine content, which meets the relevant criteria for leaching toxicity. Therefore, the main harmful component of this part of aluminum dross is aluminum nitride [4,5]. The Directory of National Hazardous Wastes (Version 2021) distinguishes between the hazardous properties of different processes of aluminum. Among them, salt slag and aluminum dross generated during the processing and recycling of recycled aluminum and aluminum products are considered to have reactive hazardous properties [6].

At present, the research on the comprehensive utilization of aluminum dross mainly focuses on building materials, refractory materials and other inorganic materials. The processing methods can be divided into wet method and fire method [7]. Fire method uses high temperature for harmless treatment or resource utilization, while wet method generally achieves its purpose with the help of some solvents. In a related study on wet method processing, Xun [8] developed a flocculant prepared from aluminum dross, polyaluminum iron silicate (PAFSC). With the fast sedimentation rate, the water purification effect of homemade flocculant is significantly better than that of flocculant poly aluminum chloride (PAC) sold on the market. Zhou et al. [9] used alkali sintering to convert aluminum in aluminum dross into sodium aluminate, and then prepared hydrogen alumina by hydrolysis to be used as an industrial raw material. David et al. [10] prepared hydrogen by hydrolyzing highly active aluminum dross. In addition, an important byproduct, AlOOH, is obtained. Therefore, aluminum dross hydrogen production technology has great theoretical value.

There are relatively many studies on fire method. Hu et al. [11] studied in detail the process of preparing premelted calcium aluminate slag for molten steel refining from recycled aluminum dross. Yoshimura et al. [12] studied replacing part of calcined alumina with aluminum dross. In this process, aluminum dross can be directly used in the preparation of refractory materials without calcination. It should be noted that the mass fraction of aluminum dross that can be directly added is less than 5%, which contributes relatively little to the total utilization. Zhang et al. [13] studied the feasibility of NaOH solution extraction to recover aluminum from secondary aluminum dross and the subsequent sintering of the extraction residue to prepare MgAl₂O₄ spinel. Benkhelif et al. [14] investigated the production of pure spinel (MgAl₂O₄) powder from waste aluminum dross by a leaching-precipitation-calcination process. EWais et al. [15] used aluminum dross, aluminum sludge and alumina as raw materials to prepare calcium aluminate cement, with various performance indicators reaching relevant international standards. Mandal et al. [16] studied the possibility of mixing fly ash (FA) and red mud (RM) with sawdust in different proportions to prepare thermal insulation bricks. According to the results, these bricks are suitable for thermal insulation in environments with medium and low temperatures (up to 600 °C). Ni et al. [17,18] used plasma spray technology to make aluminum dross into a novel coating, which not only enhanced the strength of the base material, but also treated the aluminum dross cleanly.

The above research shows that the comprehensive utilization of aluminum dross has broad prospects. However, the research studies that consume a large amount of aluminum dross meshes all use the fire method, which requires huge equipment costs and energy consumption, which is difficult to be fully promoted. The aluminum dross based on wet method to prepare non-fired bricks has the advantages of low cost and large consumption of aluminum dross, which can well solve the problem of excessive aluminum dross stock.

Based on previous studies, this paper studies and analyzes the effects of three process parameters on the compressive and flexural properties of non-fired bricks. Based on the response surface method, a mathematical model between the three process parameters (forming pressure, mixing-water amount and aluminum dross particle size) and the compressive properties of non-fired bricks was established for the application of aluminum dross in the preparation of non-fired bricks to provide a theoretical basis.

2. Materials and Methods

2.1. Raw Materials

The aluminum dross used in the test came from Jiangsu Haiguang Metal Co., Ltd. (Suqian, China). More specifically, this is the secondary aluminum dross produced after the primary aluminum dross extracts metal aluminum. The main components are shown in

Table 1. Aluminum dross XRF.

Element	Na	Mg	Al	Si	S	Cl	K
Mass proportion %	24.25	1.86	39.76	1.11	1.88	17.73	4.96

and Figure 1. Among them, the main phases are Al₂O₃, Al, AlN and a large amount of chloride salts. After denitrification by water, the main components become Al₂O₃ and Al(OH)₃. The aluminum dross particle size distribution measured by dry laser particle size analyzer is shown in Figure 2. It can be seen from the figure that the particle size of the aluminum dross used in the test mainly ranges from 1 to 300 μm.

The raw materials used in the experiment included CaSO₄, Ca(OH)₂ (analytical grade, from Tianjin Zhiyuan Chemical Reagent Co., Ltd., Tianjin, China), engineering sand (fineness modulus 2.6, from Jiuqi Building Materials Co., Ltd., Weifang, China) and cement (P.O 42.5 Portland cement, from Zhucheng Yangchun Cement Co., Ltd., Weifang, China).

2.2. Sample Preparation

With aluminum dross as the main raw material, the mix ratio of non-fired bricks was shown in the previous study [19], and the preparation method is shown in Figure 3.

The sample preparation procedure is as follows. Larger impurities in aluminum dross are removed through a 50-mesh standard sieve (Taylor standard sieve is used in the full text). After pouring into the autoclave, ultrapure water is added, and then heated to 100 °C for 24 h. The nitrogen removal and desalination treatment of aluminum dross is completed. The main chemical reaction formula is shown in Equation (1). After drying the pretreated aluminum dross, particle size screening is performed in a vibrating screen through 50 mesh, 100 mesh, 150 mesh, 200 mesh and 250 mesh screen. In the preferred formulation, the other ingredients are dry-added and mixed for 10 min. Then, mixing-water amounts of 10%, 12%, 14%, 16% and 18% are added for wet mixing for 10 min. After pouring into the 40 × 40 × 160 mm mold, the compressive and flexural testing machine completes the brick making operation under the pressure of 9 MPa, 12 MPa, 15 MPa, 18 MPa and 21 MPa, respectively. After demolding, the non-fired bricks are steam-cured for 18 h and naturally sprinkle for 3 days. Thereafter, measurements of the compressive and flexural properties of non-fired bricks are carried out. In order to avoid chance, each group of experiments was performed three times, taking the average value. The physical drawing is shown in Figure 4.

A1N + 3H₂O → A1(OH)₃ + NH₃↑

(1)

2.3. Characterization

According to the Chinese standard GB/T 17671-1999 Test Method for Cement Mortar Strength (ISO Method) [20], the YAW-300C microcomputer-controlled compressive and flexural testing machine produced by Jinan Dongfang Testing Instrument Co., Ltd. (Jinan, China) is used to test non-fired bricks’ compressive and flexural properties. The force loading speed is 2.4 KN/s at 120 s under a certain pressure. A 200 mm vibrating screen of Henan Zhongtai Machinery Co., Ltd. (Zhengzhou, China) is used to classify the pretreated aluminum dross. Gemini SEM 300 field emission scanning electron microscope from Carl Zeiss, Aalen, Germany is used to observe the microscopic morphology of the test samples. Rigaku D/max2550V X-ray diffractometer from Rigaku Corporation (Tokyo, Japan) is used to analyze the phase composition of aluminum dross samples.

3. Results and Discussion

3.1. Influence of Different Forming Pressures on Non-Fired Bricks

Forming is the most important step in the manufacturing process of non-fired bricks, which determines the initial performance of non-fired bricks. According to the previous exploratory experiments, the selected forming pressure levels are 9, 12, 15, 18 and 21 MPa, respectively. Other process parameters are as follows. The mixing-water amount is 10%, and there is sieve blanking with 50-mesh screen. After forming, non-fired bricks with height tolerance within ±0.5 mm are considered acceptable. The compressive and flexural properties of non-fired bricks are shown in Figure 5.

It can be seen from Figure 5 that the compressive and flexural properties of non-fired bricks increase significantly with the increase of forming pressure. When the forming pressure rises to 18 MPa, the change curve obviously slows down. This is the critical pressure for non-fired bricks. Considering that a large amount of air will be discharged during the brick making process, a certain amount of air will be retained. When the air is compressed to a certain extent, its expansion force increases sharply to counteract the pressure of the brick making machine [21,22]. During constant pressurization, the raw material particles of non-fired bricks are displaced. Here, small particles are squeezed into the gaps of larger particles, thereby increasing the density of non-fired bricks and improving compressive and flexural properties. It should be noted that there is a critical point. Based on the comprehensive consideration of compressive and flexural properties, the most suitable forming pressure is 18 MPa.

The XRD analysis results of aluminum dross non-fired bricks under different forming pressures are shown in Figure 6.

It can be seen from Figure 6 that the main phases of aluminum dross non-fired bricks are Al₂O₃, Al(OH)₃, ettringite (AFt), calcium aluminate hydrate (CAH) and calcium silicate hydrate (CSH). The content of the main phases of non-fired bricks under different forming pressures did not change significantly. According to the electron microscope image in Figure 7, there is a large gap between the materials of the non-fired bricks under 9 MPa. The material-to-material clearance of non-fired bricks at 18 MPa is significantly reduced. At 21 MPa, there is no significant difference in the gaps between non-fired bricks.

3.2. Influence of Mixing-Water Amount on Non-Fired Bricks

One of the strength sources of aluminum dross non-fired bricks is the gel-like hydration products formed by the internal hydration reaction [23]. Therefore, studies on the variation of mixing-water amount can help to analyze the strength formation mechanism of non-fired bricks. Through preliminary exploratory experiments, the selected mixing-water amount levels are 10, 12, 14, 16 and 18%. Other process parameters are as follows. The forming pressure is 18 MPa, and there is sieve blanking with 50-mesh screen. Non-fired bricks with height error within ±0.5 mm after forming are considered acceptable. Figure 8 shows the compressive and flexural properties of non-fired bricks measured under this condition.

It can be seen from Figure 8 that when the mixing-water amount increases from 10% to 14%, the compressive strength and flexural strength of non-fired bricks are significantly improved, with a maximum compressive strength of 23.77 MPa. Both the compressive strength and flexural strength of non-fired bricks decreased when the mixing-water amount increased from 14% to 18%. When the mixing-water amount is 18%, it can be clearly observed that there is slurry overflow on both sides of the mold when pressing bricks. Therefore, it can be concluded that excessive mixing-water amount hinders the formation of hydration products, thereby reducing the strength of non-fired bricks. It can be seen from Figure 9 that when the mixing-water amount increases from 10% to 14%, the two hydration products CAH and CSH in the non-fired bricks gradually increase. The amount of hydration products peaked at a mixing-water amount of 14%. The hydration products in the non-fired bricks gradually decreased as the mixing-water amount increased from 14% to 18%. Figure 10 is the microscopic morphology of the sample. At a mixing-water amount of 10%, there are relatively few hydration products on the surface of non-fired bricks. At a mixing-water amount of 14%, the hydration products on the surface of non-fired bricks are the most numerous and intertwine, which increases the strength of non-fired bricks [24]. At a mixing-water amount of 18%, non-fired bricks have a certain amount of hydration products on the surface, which is less than when the mixing-water amount was 14%.

3.3. Influence of Aluminum Dross Particle Size on Non-Fired Bricks

As a building material, the raw material particle size of non-fired bricks can affect the performance of the finished product. This group of experiments takes the particle size of aluminum dross as the research object. According to the detection data of the particle size instrument, the aluminum dross particle size in the test was divided into 50–100 mesh, 100–150 mesh, 150–200 mesh, 200–250 mesh and more than 250 mesh. Other process parameters are as follows. The forming pressure is 18 MPa, and the mixing-water amount is 14%. Non-fired bricks with height error within ±0.5 mm after forming are considered acceptable. Figure 11 shows the compressive and flexural properties of non-fired bricks measured under this condition.

When the aluminum dross particle size was increased from 50–100 mesh to 100–150 mesh, the compressive strength and flexural strength of non-fired bricks reached a maximum value and then decreased rapidly. When the aluminum dross particle size is greater than 250 mesh, the compressive strength and flexural strength of non-fired bricks are relatively low, which is lower than the MU20 strength level in JC/T 422-2007 Non-fired Rubbish Gangue Bricks [25]. Therefore, it can be concluded that non-fired bricks made of aluminum dross with 100–150 mesh particle size have better performance. According to the XRD results in Figure 12, the Al(OH)₃ and AFt in the non-fired bricks phase increase significantly from 200 mesh. When the aluminum dross particle size is larger than 250 mesh, the content of Al(OH)₃ and AFt reaches the peak value. Therefore, the Al(OH)₃ in the shape of cotton generated by the hydrolysis of AlN is distributed in the interval larger than 200 mesh, which reduces the strength of non-fired bricks to a certain extent. For the change of AFt content, active alumina is mainly distributed in the interval larger than 200 mesh, which reacts with Ca(OH)₂ and CaSO₄ to generate excess AFt. The expansive AFt [26] agglomerates together and bursts the non-fired bricks. There is no connection between the particles, which results in a reduction in the strength of non-fired bricks on a macroscopic scale. According to Figure 13a, the cut surfaces of the non-fired bricks are covered with hydration products without obvious particles. According to Figure 13b, there are a large number of independent agglomerates composed of columnar AFt with distinct particles, which are the fragments of the brick body after the expansion of AFt.

3.4. Multi-Factor Optimization Experiment Based on Response Surface Method

Based on the above research, the effects of process parameters A (molding pressure, MPa), B (mixing water, %) and C (particle size of aluminum ash, mesh) on Y in the process of preparing non-fired bricks with aluminum dross were investigated by using Box–Behnken response surface scheme and Y (compressive strength, MPa) as the response value. The specific ranges of the aluminum dross particle size are 50–100 mesh, 100–150 mesh and 150–200 mesh. The experimental factors and level encoding values of the response surface method are shown in

Table 2. Factors and level encoding of response surface development experiment.

Element	Level
	−1	0	1
A (forming pressure/MPa)	12	15	18
B (mixing-water amount/%)	12	14	16
C (aluminum dross particle size/mesh)	50	100	150

. The specific test plan and test data are shown in

Table 3. Experimental design and results of response surface method.

Experiment No.	A	B	C	Compressive Strength/MPa
1	12	12	100	18.91
2	18	12	100	22.13
3	12	16	100	18.85
4	18	16	100	24.04
5	12	14	50	18.90
6	18	14	50	23.14
7	12	14	150	17.44
8	18	14	150	21.34
9	15	12	50	18.96
10	15	16	50	22.07
11	15	12	150	18.10
12	15	16	150	18.81
13	15	14	100	22.42
14	15	14	100	22.54
15	15	14	100	22.92

.

Design Expert (Version number: 11, Stat-Ease in Minneapolis) software is used to process the data in Table 3. The fitted multivariate quadratic regression equation is as follows:

Y = 22.63 + 2.07A + 0.7087B − 0.9225C + 0.4925AB − 0.0850AC − 0.6000BC − 0.4621A² − 1.18B² − 1.96C²

(2)

After calculation, the regression equation of the actual factor is as follows:

Y = −60.15833 + 1.13736A + 7.99771B + 0.230817C + 0.082083AB − 0.000567AC − 0.006000BC − 0.051343A² − 0.295521B² − 0.000784C²

(3)

The variance analysis of the quadratic regression equation is shown in

Table 4. Analysis of variance (ANOVA) for quadratic model.

Source	Sum of Squares	df	Mean Square	F-value	p-Value	Significance
Model	65.84	9	7.32	50.60	0.0002	significant
A	34.24	1	34.24	236.80	<0.0001	significant
B	4.02	1	4.02	27.79	0.0033	significant
C	6.81	1	6.81	47.09	0.0010	significant
AB	0.9702	1	0.9702	6.71	0.0488	significant
AC	0.0289	1	0.0289	0.1999	0.6735	not significant
BC	1.44	1	1.44	9.96	0.0252	significant
A2	0.7884	1	0.7884	5.45	0.0668	not significant
B2	5.16	1	5.16	35.68	0.0019	significant
C2	14.18	1	14.18	98.06	0.0002	significant
Residual	0.7229	5	0.1446	none	none	none
Lack of fit	0.5867	3	0.1956	2.87	0.2690	none
Pure Error	0.1363	2	0.0681	none	none	none
Cor Total	66.57	14	none	none	none	none

. It can be seen from Table 4 that the model is significant (p < 0.05), while the lack of fit is not significant (p > 0.05). It can be seen from

Table 5. Correlation analysis result of experimental model.

Item	Value	Item	Value
Std. Dev	0.3802	R2	0.9891
Mean	20.70	Adjusted R2	0.9696
C.V.%	1.84	Predicted R2	0.8544
		Adeq Precision	22.3974

that the adjusted R² value is 0.9696, the predicted R² value is 0.8544 and the difference is less than 0.2. Therefore, the model is able to accurately predict the experimental results. The model can be used for analysis and optimization of process parameters. In addition, it can be concluded from Table 4 that A, B, C, AB, BC, B² and C² are the significant terms of the model, while AC and A² are insignificant terms. The size of the F value can reflect the degree of influence of the corresponding factors on the response value (Y). Therefore, the influence of different process parameters on the compressive strength of aluminum dross non-fired bricks is as follows: forming pressure > aluminum dross particle size > mixing-water amount.

Figure 14 is a response surface diagram of the interaction between process parameters A (forming pressure), B (mixing-water amount) and C (aluminum dross particle size).

After optimization of the response surface method, the optimal process parameters for aluminum dross to prepare non-fired bricks can be obtained: 18 MPa forming pressure, 15.19% mixing-water amount and 82.61–132.61 mesh aluminum dross particle size. The compressive strength of the final non-fired bricks can reach 24.77 MPa. The plan is revised according to the actual situation. The mixing-water amount was corrected to 15%. The aluminum dross particle size was revised to 80–130 mesh. At this time, the model’s predicted value was 24.75 MPa, and the compressive strength of the obtained non-fired bricks was verified to be 24.66 MPa. The error rate was 0.36%, which is consistent with the prediction results of the response surface model.

4. Conclusions

• There is a positive correlation between the forming pressure and the mechanical properties of aluminum dross non-fired bricks, with a critical value of 18 MPa. When the forming pressure is greater than 18 MPa, the forming pressure continues to increase, which slightly improves the performance of non-fired bricks. The content of the main phases of non-fired bricks under different forming pressures does not change significantly.
• The mixing-water amount has a positive correlation with the mechanical properties of aluminum dross non-fired bricks before 14%, and has a negative correlation after 14%. When the mixing-water amount is 14%, the two hydration products CAH and CSH on the surface of the non-fired bricks are most abundant and intertwine, thereby increasing the strength of the non-fired bricks.
• When the particle size of aluminum dross is 100–150 mesh, the mechanical properties of non-fired bricks reach the maximum value and then decrease rapidly. This is because the active alumina is mainly distributed in the range larger than 200 mesh, which reacts with Ca(OH)₂ and CaSO₄ to generate excess AFt. Since there is no connection between the particles, the performance decreases overall.
• Through the multi-factor test based on the response surface method, the influence of different factors on the compressive strength of aluminum dross non-fired bricks is as follows: forming pressure > aluminum dross particle size > mixing-water amount. The optimal process parameters are as follows: forming pressure is 18 MPa, mixing-water amount is 15% and aluminum dross particle size range is 80–130 mesh. Under this condition, the compressive strength of non-fired bricks is 24.66 MPa, which is only 0.36% different from the model’s predicted value.