Waste Glass (WGs) and Coir Fiber (CF) are not widely utilized, even though their silica and cellulose content can be used to create construction materials. This study aimed to optimize mortar compressive strength using Response Surface Methodology (RSM). The Central Composite Design (CCD) was applied to determine the optimization of WGs and CF addition to the mortar compressive strength. Compressive strength and microstructure testing with Scanning Electron Microscope (SEM), Fourier-transform Infrared Spectroscopy (FT-IR), and X-Ray Diffraction (XRD) were conducted to specify the mechanical ability and bonding between the matrix, CF, and WGs. The results showed that the chemical treatment of CF produced 49.15% cellulose, with an average particle size of 1521 µm. The regression of a second-order polynomial model yielded an optimum composition consisting of 12.776% WGs and 2.344% CF with a predicted compressive strength of 19.1023 MPa. C–S–H gels were identified in the mortars due to the dissolving of SiO_{2} in WGs and cement. The silica from WGs increased the C–S–H phase. CF plays a role in preventing, bridging, and branching micro-cracks before reaching maximum stress. WGs aggregates and chemically treated CF are suitable to be composited in mortar to increase compressive strength.

Waste management is a severe environmental issue requiring proper handling. Efforts have been made to utilize waste glass (WGs) and natural fibers as reinforcement for construction materials. Therefore, it is necessary to further study the effects of the two composites on the mortar compressive strength. Many studies have been carried out on the replacement of aggregates utilizing wastes, such as glass waste [

Meanwhile, the use of natural fiber as reinforcement in cement-based construction materials has continued to increase in recent years to overcome the limitations of toughness and tensile strength [

In developing these two materials composites (WGs and CF), it is necessary to know their optimal values so that the mortar can provide the best compressive strength. Response Surface Methodology (RSM) is a suitable tool for this purpose. It is a robust technique for systematically designing and analyzing different experiments. Models are developed to predict the outputs of the experiment, including statistical approaches to determine the relationship between independent input and dependent output variables [

Researchers developed RSM to optimize mechanical strength with several variables [

Althogh RSM is successfully used for the modeling and optimization of concrete. Not many studies have examined the optimal use of WGs and CF in cement-based composites. Therefore, this study aimed to determine the percentage utilization of WGs and CF in mortar compressive strength optimization. The urgency of establishing the relationship between influencing factors and influenced parameters as output and input variables are significant. Before RSM was conducted, CF was chemically treated to obtain cellulose. The percentage of WGs used was 5% to 15%, and the CF was 1% to 3%. In this study, experiments were implemented using Central Composite Design (CCD) because it has a good correlation between experimental and predicted values of responses. The microstructural properties of CF and WGs and CF-based mortar were determined using Scanning Electron Microscopy (SEM), X-ray Diffraction (XRD), and Fourier Transform Infrared (FT-IR).

The chemicals used to obtain cellulose from CF were sodium hydroxide (NaOH, 97%, Merck) and hydrogen peroxide (H_{2}O_{2}, 50%, Merck). Natural fine aggregates from local suppliers were used to produce mortar. OPC Type I, locally produced, was used in this study. The WGs aggregates were obtained from a grinding machine with a final particle size grading below 600 µm.

Chemical composition | OPC (%) | Sand (%) | WGs (%) |
---|---|---|---|

SiO_{2} |
21.74 | 63.16 | 59.5 |

Al_{2}O_{3} |
3.2 | 15.72 | 3.94 |

Fe_{2}O_{3} |
3.6 | 3.13 | 6.98 |

TiO_{2} |
– | 0.66 | 0.08 |

CaO | 65.5 | 1.77 | 7.09 |

MgO | 1.34 | 0.25 | 0.31 |

K_{2}O |
0.37 | 1.64 | 0.49 |

MnO | 1.34 | – | – |

SO_{3} |
2.8 | – | – |

Na_{2}O |
0.37 | – | 9.86 |

Aggregates types | Size |
SSD particle density (kg/m^{3}) |
Specific gravity | Water |
---|---|---|---|---|

Sand | 60 to 1180 | 2520 | 2.55 | 0.70 |

WGs | 60 to 200 | 2401 | 2.42 | 0.34 |

The extraction of CF is shown in _{2}O_{2} solutions at 50°C for 1.5 h in a ratio of 1:20 (g/mL). The fibers were washed with distillate water and dried to reach a constant weight. The chemical composition of coconut fiber (coir) before and after chemical treatment was determined using Chesson data [

In studying the effect of CF and WGs contents on the compressive strength of mortar, thirteen mortars were prepared according to the experiment design. The mix ratio of OPC and sand for mortar production was 1:2, and the volume of OPC in each sample was replaced by 0% to 3% CF. WGs is a partial substitute for sand with 0% to 15%. The water-to-binder ratio used was 0.4. Specimens were produced by placing all ingredients in a Hobart mixer and stirring for approximately 15 min. The mixture was then poured into a 50 mm × 50 mm × 50 mm cube mold. After 24 h, the specimens were removed from the mold and labeled according to their respective codes. The compressive strength test was conducted according to ASTM C-109 [

The microstructures and surface morphologies were examined by SEM, EVOMA 15, ZEISS, from Germany. The specimens were imaged with an accelerating voltage of 5 kV at a working distance of 8 mm. ImageJ software was used to determine the CF length after bleaching treatment. In addition, FT-IR spectra were measured using IRPrestige-21, Shimadzu, Kyoto, Japan. The spectra were obtained using 25 scans at a resolution of 4 cm^{−1} for each sample, with scanning ranges ranging from 4000 to 400 cm^{−1}, and analyzed with Spectrum software. XRD spectra were measured using Shimadzu XRD-7000, Japan. The XRD scans were performed at 10 to 50° 2theta with a scan speed of 0.5 s/step.

CF and WGs-based mortar optimization was designed using CCD-based RSM. The independent variables were WGs concentration (X_{1}) and CF concentration (X_{2}). Based on Design Expert software using the CCD method, the experimental design was produced with 13 experimental runs, with a center point of five runs.

Variables | Range and level | ||
---|---|---|---|

−1 | 0 | 1 | |

WGs | 5 | 10 | 15 |

CF | 1 | 2 | 3 |

Quadratic models were used to estimate the response surfaces for the compressive strength of mortar, calculated according to the 2^{nd}-degree polynomial

where

Run | Variable codes | Extraction variables | ||
---|---|---|---|---|

X1 | X2 | WGs (%) | CF (%) | |

1 | 1 | 1 | 15 | 3 |

2 | −1 | 1 | 5 | 3 |

3 | 0 | 0 | 10 | 2 |

4 | 0 | 1.41 | 10 | 3.41421 |

5 | 0 | 0 | 10 | 2 |

6 | −1 | −1 | 5 | 1 |

7 | 1 | −1 | 15 | 1 |

8 | 0 | 0 | 10 | 2 |

9 | 0 | −1.414 | 10 | 0.585786 |

10 | −1.414 | 0 | 2.92893 | 2 |

11 | 1.414 | 0 | 17.0711 | 2 |

12 | 0 | 0 | 10 | 2 |

13 | 0 | 0 | 10 | 2 |

Treatment | Cellulose (%) | Hemicellulose (%) | Lignin (%) | Ash (%) |
---|---|---|---|---|

Raw | 23.45 | 32.56 | 20.37 | 15.41 |

Alkali treated | 39.56 | 20.43 | 19.34 | 10.65 |

Bleached | 49.15 | 18.84 | 10.33 | 9.03 |

_{2}O_{2} solutions at 50°C is 21.43%. The results indicated the possibility of efficiently recovering cellulose. This also showed that hemicellulose and lignin had been hydrolyzed.

Chemical treatment has changed CF morphology.

The color changes in CF in _{2}O_{2} successfully removing non-cellulose materials and other impurities such as lignin, hemicellulose, pectin, and others [

The longest CF is in the range of 500–3000 µm. The average fiber length is about 1521 µm. This fiber length is commonly used in natural fiber-based composite materials. In this study, CF has a specific gravity of 1.44 g.cm^{−3}.

Types of fiber | Percentage of fiber usage (%) | Length of fiber (µm) | Cementitious matrix | Reference |
---|---|---|---|---|

Coir fiber | 0, 1, 2, and 3 | 1521 | Mortar | This study |

Date palm | 0, 2, 3 | 15000–60000 | Concrete | [ |

Date palm | 50 | 63 to 5000, and the estimated main size is 2000 | Lime | [ |

Sugarcane bagasse | 1 and 2 | 10000 | Cement | [ |

Hemp | 0.75, 1.5 and 3 | 40000–45000 | Concrete | [ |

^{−1}, as these wavelengths are associated with Si–O. The peak at 500–1100 cm^{−1} also indicates the Si–O bond [^{−1} peak, indicating the silicate units (SiO_{4}) [_{2} contributed by the raw material (cement and WGs) [^{−1}. In this phase, there was a higher increase in sulfate due to the ettringite formation through gypsum reaction with tricalcium aluminates in cement [

The prominent crystal peak was identified as quartz (SiO_{2}) with high intensity at 2θ = 26.68°. This peak appeared due to the contribution of SiO_{2} derived from cement and WGs. The crystalline silica increased with the increasing percentage of WGs. The C–S–H amount increased from 10% to 15% of the cement replacement. It was due to the pozzolanic activity of WGs, which utilized the available C–H in the mortar and produced a C–S–H gel. The increase in compressive strength of the mortar confirmed it. Meanwhile, adding 5% WGs did not increase the amount of C–S–H due to insufficient WGs to react with C–H, decreasing the compressive strength. It was confirmed in the statistical model optimization. Adding WGs by 12.776% and CF by 2.344% provided the best compressive strength optimization. Other patterns were not clearly detected in XRD. The number of quartz appearing was more due to the bonding between cement and WGs. From the XRD test, it can be concluded that raw materials affected the formation of C–S–H gel.

Run | Variable codes | Extracted variables | Compressive strength response (MPa) | ||
---|---|---|---|---|---|

X_{1} |
X_{2} |
WGs (%) | CF (%) | ||

1 | 1 | 1 | 15 | 3 | 16.21 |

2 | −1 | 1 | 5 | 3 | 16.11 |

3 | 0 | 0 | 10 | 2 | 18.41 |

4 | 0 | 1.41 | 10 | 3.41421 | 15.21 |

5 | 0 | 0 | 10 | 2 | 18.12 |

6 | −1 | −1 | 5 | 1 | 16.41 |

7 | 1 | −1 | 15 | 1 | 20.52 |

8 | 0 | 0 | 10 | 2 | 17.86 |

9 | 0 | −1.414 | 10 | 0.585786 | 18.17 |

10 | −1.414 | 0 | 2.92893 | 2 | 16.72 |

11 | 1.414 | 0 | 17.0711 | 2 | 19.51 |

12 | 0 | 0 | 10 | 2 | 17.59 |

13 | 0 | 0 | 10 | 2 | 18.32 |

The statistical summary model suggested by the Design Expert software and the statistical data for model selection is shown in ^{2}.

Source | Sequential |
Lack of Fit |
R-Square | Adjusted R² | Predicted R² | |
---|---|---|---|---|---|---|

Linear | 0.1216 | 0.0002 | 0.3439 | 0.2127 | 0.1562 | |

2FI | 0.1339 | 0.0002 | 0.4959 | 0.3278 | 0.165 | |

Source | Value |
---|---|

Std. Dev. | 0.2396 |

Mean | 18.98 |

C.V. % | 1.26 |

R² | 0.9851 |

Adjusted R² | 0.9745 |

Predicted R² | 0.9143 |

Adeq. Precision | 25.8918 |

^{2} close to 1 (R^{2} = 0.9851) with Adj-R^{2} = 0.9745 and predicted-R^{2} = 0.9143. This model is a good fit because the R^{2} is close to 1. Hence, the correlation is accurate, with the actual and the prediction values being almost similar. In contrast, the other models have larger deviations.

where Y is compressive strength (MPa), X_{1} is WGs concentration (%), dan X_{2} is CF concentration (%). The mathematical model above explains the effect of each variable, WGs and CF, either linearly or quadratic, and the interaction of the two variables.

The normal distribution of the residuals plot was assessed with the help of a graphical method referred to as the normal probability plot, as shown in

A numerical ramp was performed to assess the optimum compressive strength of WGs and CF-based mortar, as shown in

The statistical model showed that the optimum compressive strength was for the addition of 12.776% WGs and 2.344% CF. The addition of excess WGs and CF will reduce the compressive strength. Raw characteristics influence the increase and decrease in the compressive strength of mortar. Fiber with characteristics, cellulose content of 49.15% (see

Adding CF increases the compressive strength of the mortar, resulting the mortar resisting crack propagation. However, adding a percentage above it causes the fibers to agglomerate and decreases compressive strength. The addition of WGs in the matrix is needed as its silica content makes the matrix denser, but adding a higher percentage of WGs disrupts the hydration process of the cement and reduces the compressive strength. The combination of WGs and CF of 12.776% and 2.344% is essential as a mortar constituent since the silica content can play a role in the hydration process, and CF can act as reinforcement in mortar.

Based on the CCD, the ANOVA (analysis of variance) calculation shows the linear and quadratic individual effects of WGs and CF and the interactions between those variables.

Source | Sum of squares | df | Mean square | F-value | Characteristics | |
---|---|---|---|---|---|---|

Model | 26.58 | 5 | 5.32 | 92.62 | <0.0001 | Significant |

A-WGs | 0.1295 | 1 | 0.1295 | 2.26 | 0.1767 | |

B-coir fiber | 9.15 | 1 | 9.15 | 159.40 | <0,0001 | |

AB | 4.10 | 1 | 4.10 | 71.46 | <0.0001 | |

A² | 10.75 | 1 | 10.75 | 187.29 | <0.0001 | |

B² | 3.92 | 1 | 3.92 | 68.28 | <0.0001 | |

Residual | 0.4017 | 7 | 0.0574 | |||

Lack of fit | 0.3036 | 3 | 0.1012 | 4.13 | 0.1023 | Not significant |

Pure error | 0.0981 | 4 | 0.0245 | |||

Cor total | 26.98 | 12 |

The validation results show that the model for the compressive strength response of mortar is appropriate and applicable. The F and

This study used WGs as a partial replacement for sand and CF as an additional reinforcing agent in mortar mix. To obtain the optimal mix proportions of the constituent materials and build a mathematical model, RSM has been used to predict the mortar compressive strength. The proposed mathematical model was found significant using the

WGs and CF-based mortar microstructure show that silica from WGs enhanced the C–S–H phase, and CF worked as bridging in resisting crack growth. The combination of WGs and CF is important as mortar constituents because it contains the required amount of silica in the hydration of cement and as reinforcement in mortar. Based on this study, compositing WGs and CF in mortar can be beneficial to develop construction materials and saving the environment through effective non-biodegradable waste management.

Waste Glass

Coir Fiber

Response Surface Methodology

Central Composite Design

Scanning Electron Microscope

Fourier-Transform Infrared Spectroscopy

X-Ray Diffraction

This research was funded by the Ministry of Education, Culture, Research, and the Technology, Indonesia for Matching Fund (Kedaireka) Scheme in 2022 with Contract No. 155/E1/KS.06.02/2022.

The authors declare that they have no conflicts of interest to report regarding the present study.