The Use of Perlite and Rhyolite in Concrete Mix Design: Influence on Physical-Mechanical and Environmental Performance

Abstract

During the last decades, the ever-growing evolution of the construction industry has led to a significant increase in demand for increasingly high-performing construction materials both in terms of mechanical characteristics and sustainability. Focusing on concrete, several researchers have designed different mixes to improve mechanical properties such as compressive strength, workability and durability, and in many of the proposed mixes, the use of industrial waste stands out both for their ability to improve the mechanical properties of concrete and for the importance of their reuse from a sustainability point of view. In this paper, the use of two waste materials, perlite and rhyolite, in concrete mix design was studied in detail, considering their influence on the compressive strength at 7 and 28 days of curing. The waste materials were introduced in the mix design as substitutes for cement in percentages of 15% and 30% in weight. In addition, perlite was micronized to two different particle sizes, 20 μm and 63 μm, respectively, according to what is already used in concrete within perlite in the mix design. The behavior of the structural concrete containing perlite and rhyolite was compared in terms of compressive strength, Young modulus and produced equivalent CO₂ with that of a standard C25/30 reference concrete, and with that of a mix design created using other waste materials, namely fly ash, metakaolin and silica fume, considering cement replacements that are always at 15% and 30% by weight. Moreover, ultrasonic testing and rebound hammer tests were run to evaluate a possible relationship between the physical-mechanical properties of the design mixes and their volumetric and surface characteristics.

1. Introduction

During the last few years, the constant evolution of the construction industry and the increasing attention to sustainability has led the building industry sector to study new construction materials, which would respect not only the performance in terms of structural and durability requirements but also the minimum environmental criteria established by the different countries and their regulations. For this reason, the use of industrial waste in the production of construction materials, such as concrete, has been characterized by an outstanding development. Concrete is the most widely used building material, although its production process is one of the most energy-consuming, causing high CO₂ emissions in the atmosphere. For this reason, the study of waste materials for use in concrete mix design is ongoing, in order to reduce the exploitation of raw materials and the impact on the atmosphere.

Several researchers have investigated the possibility of using industrial waste in concrete mix design, evaluating the advantages and disadvantages in terms of both mechanical properties and sustainability, the latter especially in terms of CO₂ reduction [1,2,3,4,5,6,7]. One of the most important industrial wastes used in concrete mix design is fly ash [8,9]. Nuaklong et al. [10] investigated the strength and post-fire performance of fiber-reinforced alkali-activated fly ash concrete containing granite industry waste. Ho and Huynh [11] analyzed the mechanical properties and the durability of high-strength concrete (HSC) containing fly ash sourced in Vietnam. Magalhães et al. [12] evaluated the influence of Brazilian fly ash on the concrete compressive strength and modulus of elasticity.

Another industrial waste used in concrete mix design is silica fume, used both as cement replacement and as an additive [13,14,15]. Vijayan et al. [16] presented a detailed review of the application of silica fume in concrete mix design, highlighting the advantages of its use in several application fields. In Tak et al. [17], the effects of using silica fume as partial replacement of cement in concrete are investigated, determining that the ideal percentage of silica fume for concrete is approximately 11%. Chishi and Gautam [18] studied the possibility of using silica fume in green cement concrete production, focusing the attention on the sustainability aspect.

In the last years, the use of metakaolin as partial replacement of cement in concrete mix design has undergone an important development as shown by the recent literature [19,20,21]. Kumar et al. [22] analyzed the effects of metakaolin on the mechanical properties and durability of mortar and concrete, showing that metakaolin is a highly reactive pozzolanic material that reacts with calcium hydroxide and produces additional cementitious compounds enhancing the compression strength. Asghari et al. [23] studied the influence of metakaolin on the mechanical properties of self-consolidating concrete, highlighting that metakaolin could be a valuable partial substitute of cement in concrete mix design. In Sankar and Ramadoss [24], the influence of metakaolin on the concrete compressive strength at 28 days is evaluated using statistical techniques such as linear regression, multi-logistic regression, nonlinear regression and response surface methodology.

Fly ash, silica fume and metakaolin are the industrial wastes usually considered in the literature as substitutes for cement in mix designs of structural concrete, and they also find wide application in practice, being admitted by several national and international regulations. Two other possible waste materials to consider are perlite and rhyolite, which however, have been less studied so far compared with the above-mentioned industrial wastes.

Perlite is a natural, amorphous siliceous volcanic rock widely accessible in the world, extensively used in expanded form as aggregate material in concrete or mortar since the 20th century because of its light weight, thermal insulation skill and fire resistance [25,26,27,28]. More recently, some authors have investigated the pozzolanic effect of perlite powder added to concrete and evaluated the pros and cons of its use [29,30]. Stratoura et al. [31] examined the effect of perlite on concrete durability, demonstrating that the mix design provides adequate strength and durability for structural applications. Ramezanianpour et al. [32] investigated the effect of calcined perlite powder, used as a partial replacement of ordinary Portland cement, on the compressive strength and concrete resistance to chloride ions. El Mir and Nehme [33] investigated the use of waste perlite powder, provided by production of the expanded perlite process, at a high content as filler material in self-compacting concrete. Bakhshi et al. [34] analyzed the potential optimization of lightweight-reinforced concrete containing expanded perlite under impact loading. Stefanidou et al. [35] investigated the use of waste perlite products in lime-based mortars and grouts.

Rhyolite is an extrusive igneous rock with a very high silica content. Its use as a concrete component is not as common as perlite. Elsheikh et al. [36] tested the feasibility of using coarse aggregate composed of andesite and rhyolite to produce a green high-strength self-compacting concrete. Karahüseyin and Erdoğdu [37] performed an experimental study where mortar mixtures were produced by substituting cement with rhyolite and metakaolin at different ratios by weight of cement in view of preventing delayed ettringite formation in the long term. Abdelfattah et al. [38] investigated the applicability of perlite and rhyolite tuff as additives materials in lightweight aggregates (LWAs) production, explaining how the high amorphous content can improve the bloating properties of expandable clays and the properties of the LWAs produced. In Baki et al. [39], the pozzolanic activities of ground trachyte and rhyolite in mortars were investigated comparatively by several methods, concluding that the ground trachyte may be utilized as a pozzolanas in the cement industry, while the ground rhyolite does not meet some limits prescribed by the related standard.

It is important to highlight that perlite and rhyolite are not generally used in the structural concrete mix design as partial replacement of cement because they do not have the same physical-mechanical characteristics as other industrial waste materials such as metakaolin, fly ash and silica fume, already widely studied and adopted by the product standards.

Furthermore, as reported in [40], the use of rhyolite with reactive silica in the concrete mix design may cause an alkali-silica reaction. For this reason, in some cases, calcium nitrate is added [41].

The results achieved up to now demonstrate that the use of perlite secondary products and rhyolite in building materials such as mortar and concrete is feasible. However, the possibility of their use as a partial replacement of cement in structural concrete mix design still remain an open issue. Despite the growing interest in using perlite and rhyolite in concrete, there is a limited amount of research specifically addressing their effectiveness as partial cement replacements in structural concrete, particularly with respect to their impact on both mechanical and environmental performance. This study aims to fill this gap by providing a detailed evaluation of their properties in concrete mix design.

In this frame, this paper presents the results of an extensive campaign of laboratory tests aimed at evaluating the influence of the use of perlite and rhyolite as partial replacement of cement in structural concrete mix design. Effects on the compressive strength and modulus of elasticity are discussed and compared with the results obtained using (i) fly ash, (ii) metakaolin and (iii) silica fume as other cement replacement materials [42,43,44]. The results are also compared with a reference mix (M0) of standard C25/30 concrete. In particular, two different cement substitution percentages, 15% and 30%, are considered when evaluating the compressive strength after 7 and 28 days, and two different particle sizes, 20 μm and 63 μm, are considered when using perlite. Moreover, two non-destructive tests, ultrasonic testing [45,46] and the rebound hammer test [47], have been carried out on the specimens to evaluate whether possible inner or superficial properties could affect the mechanical and physical properties of the concrete. The results show that the compressive strength, and to a lesser extent the modulus of elasticity, are influenced by the type of industrial waste used, by the percentage of substitution and by the degree of micronization. In particular, while the mix designs with perlite, especially in low percentages of substitution, allow us to achieve physical-mechanical performances comparable with those of the other mixes analyzed and adequate for structural concrete, the mix designs containing rhyolite offer low performances, making the use of rhyolite as a substitute for cement still questionable. In addition, an analysis of the equivalent CO₂ produced by the mix designs considered was carried out, highlighting that the use of industrial waste generally leads to a reduction, often significant, of harmful emissions, and that perlite offers a good compromise between mechanical performance and reduction of the impact on the atmosphere.

After this brief introduction, Section 2 presents the experimental program, followed by the results in Section 3. The analysis and discussion are covered in Section 4, with the environmental impact evaluation in Section 5. Finally, some conclusive remarks are drawn in Section 6.

The main goal of this research work is to evaluate the possible use of perlite secondary products and rhyolite in the mix designs of structural concrete as partial replacement of cement to improve the sustainability level of the construction.

2. Experimental Program

2.1. Materials

The experimental program provides the casting of class C25/30 concrete, using several waste materials as a binder to partially replace cement [48]. A Portland-limestone cement type CEM II/B-LL 42.5 R has been used [49], along with a coarse aggregate having a maximum size equal to 25 mm. The water absorption and specific gravity of aggregate are 0.81% and 2.7, respectively. The characteristics of fine aggregate (sand) are as follows: maximum size 1.5 mm, water absorption 1.42%, specific gravity 2.67 and fineness modulus 2.8.

The selected waste/by-product materials are:

(i)

Perlite (PE): It is a natural, lightweight, inert and fireproof volcanic rock used for a wide variety of end uses, from cosmetics to construction. In expanded form, perlite offers thermal insulation, fire resistance and other desirable properties when used in Portland cement-based plaster [25]. Perlite is a non-crystalline glass with a typical composition including 70–75% silicon dioxide (SiO₂), 12–15% aluminum oxide (Al₂O₃), 3–5% potassium oxide (K₂O), 3–4% sodium oxide (Na₂O) and traces of iron oxide, magnesium oxide and calcium oxide. In this study, the perlite has been used in its natural form, micronized into two different particle sizes: 20 μm (μ20) and 63 μm (μ63), according to previous research works.

(ii)

Rhyolite (RIO): It is a highly silicic, fine-grained, light-colored volcanic or extrusive igneous rock. It is a felsic rock with mainly quartz, alkali feldspar, plagioclase and minor ferromagnesian minerals content. During the mining of raw perlite, there is a possibility of encountering rhyolite, which is similar to perlite but has higher silica content and lower water content. Its use in construction is not as common as perlite, so this aspect will be further explored in this work.

(iii)

Fly ash (FA): It is a waste product derived from the coal combustion process, generally used for electricity production. Its composition can vary depending on the source and composition of the coal being burned, and it includes silicon dioxide (SiO₂), aluminum oxide (Al₂O₃) and calcium oxide (CaO). Nowadays, it is widely used in concrete production as a component of cement clinker.

(iv)

Silica fume (SF): It is a by-product of the production of elemental silicon or ferrosilicon alloys in electric arc furnaces, and because of its extreme fineness and high silica content, silica fume is a very effective pozzolanic material commonly used as binder in the concrete mix design.

(v)

Metakaolin (CA): it is a cementitious material representing one of the other kinds of supplementary cementitious and most effective pozzolanic materials used for reducing the cement content from concrete and mortar. Its production requires less energy for the de-carbonation of limestone, and it releases a smaller amount of CO₂ compared with cement production [50].

Two different substitution percentages, 15% and 30%, with respect to the cement weight, have been used to assess the impact of the waste materials in the mix design on the concrete compressive strength, measured after 7 and 28 days of curing.

Table 1. Mix designs.

ID	w/c	Cement	Water	Effective Water	Fine Aggregates	Coarse Aggregates	Additive
[-]	[-]	[kg]	[kg]	[kg]	[kg]	[kg]	[kg]
M0	0.505	366.267	185.000	186.667	893.333	893.333	3.307
(a) Reference Mix.
ID	w/c	Cement	Water	Effective water	Fine aggregates	Coarse aggregates	Additive	PE
[-]	[-]	[kg]	[kg]	[kg]	[kg]	[kg]	[kg]	[kg]
PE15μ63	0.505	311.200	185.000	186.667	886.667	886.667	3.086	54.798
PE30μ63	0.505	256.387	185.000	186.667	886.667	886.667	3.306	109.880
PE15μ20	0.505	311.200	185.000	186.667	886.667	886.667	3.306	54.798
PE30μ20	0.505	256.387	185.000	186.667	886.667	886.667	5.993	109.880
(b) 15% and 30% PE substitution—particle size 20 μm (μ20) and 63 μm (μ63).
ID	w/c	Cement	Water	Effective water	Fine aggregates	Coarse aggregates	Additive	RIO
[-]	[-]	[kg]	[kg]	[kg]	[kg]	[kg]	[kg]	[kg]
RIO15	0.505	311.111	184.876	186.728	877.469	877.469	0	54.938
RIO30	0.505	256.424	184.776	186.931	859.774	859.774	3.663	109.896
(c) 15% and 30% RIO substitution.
ID	w/c	Cement	Water	Effective water	Fine aggregates	Coarse aggregates	Additive	FA
[-]	[-]	[kg]	[kg]	[kg]	[kg]	[kg]	[kg]	[kg]
FA15	0.505	311.200	185.000	186.667	877.333	877.333	5.154	54.798
FA30	0.505	256.387	185.000	186.667	860.000	860.000	7.318	109.880
(d) 15% and 30% FA substitution.
ID	w/c	Cement	Water	Effective water	Fine aggregates	Coarse aggregates	Additive	SF
[-]	[-]	[kg]	[kg]	[kg]	[kg]	[kg]	[kg]	[kg]
SF15	0.505	318.667	187.333	186.667	877.333	877.333	6.852	56.000
SF30	0.505	256.387	185.000	186.667	860.000	860.000	31.446	109.880
(e) 15% and 30% SF substitution.
ID	w/c	Cement	Water	Effective water	Fine aggregates	Coarse aggregates	Additive	CA
[-]	[-]	[kg]	[kg]	[kg]	[kg]	[kg]	[kg]	[kg]
CA15	0.505	311.200	185.000	186.667	877.333	877.333	3.704	54.798
CA30	0.505	256.566	185.185	186.531	856.933	856.933	3.367	109.764
(f) 15% and 30% CA substitution.

summarizes the different mix designs (referred to 1 m³ of fresh concrete) investigated in this research work, where w/c is the water/cement ratio.

A total of 78 concrete specimens (six for each mix design, three for each ageing), having dimensions 150 × 150 × 150 mm³ according to [48], have been manufactured in the Material Testing Laboratory of the University of Cagliari—Italy. After the cast, the specimens have been stored in a humidity- and temperature-controlled room until the compression tests, which have been performed according to [44] by means of an MC8 Series multifunctional control console (Figure 1) having a maximum load capacity of 3000 kN.

Before the compression tests, aiming to understand if the mix designs could somehow affect physical/mechanical characteristics of concrete, two non-destructive tests (NDTs) were performed after 7 and 28 days of curing. Specifically, ultrasonic testing (UT) and the rebound hammer test (RHT) were carried out on each specimen according to the reference standards [45,47], respectively. Data processing and calculations were performed by means of Microsoft 365 Excel 2025 software.

2.2. Ultrasonic Testing

UT has been performed by means of the direct method (Figure 2) according to [46]. Pulses of longitudinal stress waves, generated by an electro-acoustical transducer (emitter), that is held in contact with one side of the concrete specimen, are received and converted into electrical energy by a second transducer (receiver), located on the opposite side of the specimen. The transit time T that the waves need to run from the emitter to the receiver is measured.

The ultrasonic pulse velocity V (m/s) is calculated from the ratio:

$$V=L/T,$$

(1)

where L (m) is the wave path length, assumed as the distance between the emitter and the receiver (in this case, the side length of the cube, L = 0.015 m), and T (s) is the transit time measured by the instrumental set.

The pulse velocity V of longitudinal stress waves in a concrete mass is related to its elastic properties and density according to the following relationship [46]:

$$V=\sqrt{{\displaystyle \frac{E_{dyn}(1-\mathit{\mu })}{\rho \left(1+\mathit{\mu }\right)\left(1-2\mathit{\mu }\right)}}} ,$$

(2)

where: E_dyn is the dynamic modulus of elasticity, μ is the dynamic Poisson’s ratio and ρ is the density.

UT is applicable to assess the uniformity and relative quality of concrete, to indicate the presence of voids and cracks and to evaluate the effectiveness of crack repairs. Moreover, by applying Equation (2), it is possible to evaluate E_dyn when ρ and μ are known parameters.

The pulse velocity is independent of the dimensions of the test object provided reflected waves from boundaries do not complicate the determination of the arrival time of the direct one. To avoid this problem, the measurements of the transit time were taken on two opposite surfaces of the cube and the transducers were positioned along a straight line passing through the center of the surfaces. Additionally, to avoid measurements errors due to a bad coupling of the transducers to the concrete surfaces, a specific ultrasonic gel was applied. The device used was a PunditLab by Proceq equipped with two 54 kHz transducers for emitting and receiving the ultrasonic waves.

2.3. Rebound Hammer Test

The RHT provides an easy and speedy assessment of the uniformity of concrete on-site and, with suitable correlation, it can give an estimate of on-site compressive strength. The test can be used for comparative testing, referenced against concrete with known strength or against a concrete, where it has been shown that it has come from a defined volume of concrete with a population verified as conforming to a particular strength class [48]. The principle of the test is based on a mass propelled by a spring that strikes a plunger in contact with the surface of the structural element or specimen to be tested. When the plunger is pressed against the concrete surface, a spring-controlled mass strikes the surface, and the rebound distance is measured, providing an estimate of surface hardness. The measured value, the rebound number (RN) or rebound index (RI), is obtained from a graduated scale on the device. Concrete with lower strength and stiffness tends to absorb more energy, resulting in a lower rebound value. Tests were carried out according to [47] on two opposite sides of each specimen that were smooth and without rough texture or high porosity (Figure 3). The RN has been calculated as the median of a minimum of nine valid measurements of each side of the specimen. A type N concrete hammer having an impact energy of 2207 Nm has been used.

3. Results

3.1. Compression Tests

Before testing, the 78 specimens were measured and weighed.

Table 2. Sample weights.

Ageing Day	Sample	M0	PE15μ63	PE15μ20	PE30μ63	PE30μ20	RIO15	RIO30	FA15	FA30	SF15	SF30	CA15	CA30
		[kg]
7	1a	7.70	7.94	7.69	7.88	8.05	8.11	8.00	7.80	7.86	8.10	7.91	8.00	7.78
	2a	7.75	7.70	7.68	7.71	8.05	7.88	8.10	7.85	7.77	8.00	7.67	8.10	7.94
	3a	7.82	7.79	7.70	7.78	8.03	7.05	8.10	7.72	7.87	7.60	7.84	8.30	7.90
28	1b	7.77	7.95	7.67	7.81	7.85	7.99	8.04	7.76	7.95	8.05	7.45	8.24	8.05
	2b	7.55	7.99	7.66	7.77	8.00	8.06	8.24	7.80	7.90	7.95	7.65	8.30	7.80
	3b	7.77	7.73	7.66	7.89	8.17	7.90	8.13	7.73	7.93	8.00	8.20	8.15	7.93
Mean		7.73	7.85	7.68	7.81	8.03	8.00	8.10	8.10	7.87	8.02	7.79	8.18	7.90
SD		0.095	0.125	0.016	0.069	0.103	0.092	0.105	0.105	0.066	0.057	0.259	0.120	0.111
COV		1.23%	1.59%	0.21%	0.88%	1.29%	1.16%	1.30%	1.30%	0.84%	0.71%	3.32%	1.47%	1.41%

reports the samples weight at 7 and 28 days, while

Table 3. Density at 7 days.

	M0	PE15μ63	PE15μ20	PE30μ63	PE30μ20	RIO15	RIO30	FA15	FA30	SF15	SF30	CA15	CA30
	ρ [kg/m3]
Mean	2298	2314	2279	2308	2383	2374	2390	2308	2321	2341	2313	2410	2333
SD	17.86	35.92	2.96	25.32	3.42	35.35	17.11	19.43	16.32	78.31	36.57	45.26	24.71
COV	0.78%	1.55%	0.13%	1.10%	0.14%	1.49%	0.72%	0.84%	0.70%	3.35%	1.58%	1.88%	1.06%

and

Table 4. Density at 28 days.

	M0	PE15μ63	PE15μ20	PE30μ63	PE30μ20	RIO15	RIO30	FA15	FA30	SF15	SF30	CA15	CA30
	ρ [kg/m3]
Mean	2280	2338	2271	2318	2372	2365	2422	2300	2346	2370	2301	2439	2349
SD	37.64	37.06	1.71%	18.10	47.44	23.77	34.08	10.41	8.48	14.82	115.07	22.37	37.05
COV	1.65%	1.59%	0.08%	0.78%	2.00%	1.00%	1.41%	0.45%	0.36%	0.62%	5.00%	0.92%	1.58%

summarize the density at 7 and 28 days of curing, respectively. The labeling used in these tables follows that of Table 1, and SD is the standard deviation and COV is the coefficient of variation.

The density of the specimens assumes little dispersed values (Figure 4), included in a range of 234 kg/m³, with the mean equal to 2327 kg/m³. The low dispersion of the values indicates that the influence on the density of the type of waste aggregate, the percentage of substitution and the micronization is overall very limited.

Compression tests have been carried out for each mix design after 7 and 28 days of curing.

Table 5. Compressive strength at 7 days.

	M0	PE15μ63	PE15μ20	PE30μ63	PE30μ20	RIO15	RIO30	FA15	FA30	SF15	SF30	CA15	CA30
n°	Compressive Strength [MPa]
1a	36.9	32.3	30.8	23.8	24.7	26.5	22.3	31.0	20.9	31.9	22.6	45.7	46.5
2a	36.4	28.6	30.0	20.0	25.3	26.8	21.0	32.7	20.4	35.7	17.7	46.3	43.1
3a	36.9	32.3	29.5	20.3	23.6	25.7	21.7	31.5	20.1	33.6	27.5	39.0	44.8
Mean	36.7	31.1	30.1	21.4	24.5	26.3	21.7	31.7	20.5	33.7	22.6	43.7	44.8
SD	0.289	2.136	0.656	2.113	0.862	0.569	0.919	0.874	0.404	1.904	4.900	4.053	2.404
COV	0.79%	6.88%	2.18%	9.89%	3.51%	2.16%	4.25%	2.75%	1.97%	5.64%	21.68%	9.28%	5.37%

and

Table 6. Compressive strength at 28 days.

	M0	PE15μ63	PE15μ20	PE30μ63	PE30μ20	RIO15	RIO30	FA15	FA30	SF15	SF30	CA15	CA30
n°	Compressive Strength [MPa]
1a	42.6	41.1	38.0	31.9	35.3	34.4	30.6	39.3	31.6	45.3	38.9	45.6	46.9
2a	41.4	41.2	38.4	32.8	34.8	35.2	27.6	38.7	29.8	47.2	39.6	44.2	49.2
3a	41.5	39.8	37.3	33.8	35.3	35.2	29.1	42.3	30.7	46.6	39.5	49.3	52.1
Mean	41.8	40.7	37.9	32.8	35.1	34.9	29.1	40.1	30.7	46.4	39.3	46.4	49.4
SD	0.666	0.781	0.557	0.950	0.289	0.462	2.121	1.929	1.273	0.971	0.379	2.635	2.606
COV	1.59%	1.92%	1.47%	2.89%	0.82%	1.32%	7.29%	4.81%	4.15%	2.09%	0.96%	5.68%	5.27%

summarize the obtained results.

The data show that as the percentage of cement replacement increases, the compressive strength generally decreases, except for the mixes CA15, CA30 and SF15, whose compressive strength increases. Focusing on PE and RIO, comparing the mean compressive strength of the control mix M0 at 7 days and 28 days of curing with that of mixes containing perlite and rhyolite, a noticeable reduction in strength both at 7 and 28 days is observed. Specifically, the reduction in compressive strength for PE30μ63 and PE30μ20 is 41.8% and 33.2% at 7 days and 21.5% and 16.0% at 28 days, respectively, while the one for RIO30 is 41.1% and 30.4%.

Figure 5 illustrates the trends of the mean compressive strength for the different mixes compared with the values at 7 and 28 days obtained for the reference mix M0, where R is the ratio between the mean compressive strength of the considered mix design and that of M0.

The mix design with rhyolite shows a mean compressive strength at 28 days close to 35 MPa for RIO15 and close to 30 MPa for RIO30. Anyhow, the mean compressive strength characterizing RIO30 at 28 days is the lowest value in comparison with the results of all the other mixes. The FA30 mix shows the lowest mean compressive strength at 7 days. On the contrary, the FA15 mix shows a mean compressive strength at 7 days, which is very close to the reference mix M0 strength (R = 0.96). The best performance in terms of compressive strength is reached, both at early and standard ageing, by the mixes PE15μ63, SF15, FA15, CA15 and CA30, showing R > 0.8.

This fact highlights how cement replacement percentages higher than 15% lead to a significant reduction in the compressive strength of the concrete.

3.2. Non-Destructive Tests

As described in the previous Section 3.1, before compression tests, each sample was tested by UT and RHT at each age of casting.

Measures of T have been acquired, and then, according to the Equations (1) and (2), mean V and E_dyn have been calculated, the latter by assuming μ = 0.2 [46]. The results are reported in

Table 7. Ultrasonic pulse velocity (V).

Ageing Time	Sample	M0	PE15μ63	PE15μ20	PE30μ63	PE30μ20	RIO15	RIO30	FA15	FA30	SF15	SF30	CA15	CA30
		[m/s]
7 days	1a	4267	4286	4380	4310	4255	3984	3958	4392	4208	4484	4202	4121	4208
	2a	4342	4380	4274	4310	4261	4093	3995	4373	4178	4386	4110	4280	4121
	3a	4317	4167	4418	4317	4213	3958	3976	4425	4132	4310	4310	4155	4211
	Mean	4309	4277	4357	4312	4243	4012	3976	4397	4173	4394	4207	4185	4180
	SD	38	107	75	4	26	72	18	26	38	87	100	84	51
	COV	0.88%	2.49%	1.72%	0.08%	0.61%	1.78%	0.46%	0.59%	0.91%	1.99%	2.39%	2.00%	1.22%
28 days	1b	4317	4360	4484	4444	4298	4149	4093	4178	4093	4405	4190	4950	4267
	2b	4360	4329	4511	4329	4360	4208	4038	4049	4161	4438	4132	4870	4267
	3b	4360	4329	4317	4360	4360	4178	4065	3932	4127	4425	4208	5119	4121
	Mean	4346	4339	4437	4378	4340	4178	4065	4053	4127	4423	4177	4980	4219
	SD	25	18	106	60	36	29	28	285	34	16	39	127	85
	COV	0.58%	0.42%	2.38%	1.36%	0.83%	0.70%	0.68%	7.22%	0.83%	0.37%	0.94%	2.56%	2.01%

and

Table 8. Dynamic modulus of elasticity.

Ageing Time	Sample	M0	PE15μ63	PE15μ20	PE30μ63	PE30μ20	RIO15	RIO30	FA15	FA30	SF15	SF30	CA15	CA30
		[m/s]
7 days	1a	37,393	38,890	39,333	39,041	38,871	34,327	33,417	40,130	37,107	43,435	37,238	36,228	36,729
	2a	38,954	39,384	37,402	38,199	38,982	35,199	34,468	40,034	36,173	41,038	34,543	39,560	35,956
	3a	38,855	36,065	40,083	38,656	38,016	33,625	34,468	40,306	35,835	37,656	38,843	38,213	37,113
	Mean	38,401	38,113	38,939	38,632	38,623	34,384	34,118	40,157	36,372	40,710	36,875	38,000	36,599
	SD	874.19	1790.83	1383.09	421.63	528.58	788.24	607.01	137.81	658.64	2903.64	2172.74	1676.60	589.43
	COV	2.28%	4.70%	3.55%	1.09%	1.37%	2.29%	1.78%	0.34%	1.81%	7.13%	5.89%	4.41%	1.61%
28 days	1b	38,607	40,309	41,130	41,139	38,670	36,684	35,914	36,126	35,512	41,659	34,877	53,851	39,093
	2b	38,281	39,929	41,572	38,830	40,562	38,051	35,823	34,093	36,473	41,753	34,834	52,496	37,879
	3b	39,396	38,630	38,060	40,005	41,424	36,778	35,823	31,846	36,473	41,768	38,712	56,960	35,911
	Mean	38,761	38,623	40,254	39,991	40,219	37,171	35,853	34,021	36,152	41,727	36,141	54,436	37,627
	SD	573.57	880.47	1912.56	1154.68	1409.17	763.45	52.35	4683.39	554.97	58.73	2226.72	2288.90	1605.95
	COV	1.48%	2.22%	4.75%	2.89%	3.50%	2.05%	0.15%	14.42%	1.54%	0.14%	6.16%	4.20%	4.27%

, respectively.

Figure 6 shows the trend of V considering the ratio R_v between the considered mix design and the reference mix M0 at the same age.

It can be noticed that the variation of R_v calculated at 7 and 28 days is characterized by quite low values, which means low differences in V between the tested mixes both at early and standard ageing. Variation of mean V with respect to the reference mix is rather low too, with a maximum reduction of 7.7% (RIO30) and 9.0% (FA15) at 7 and 28 days, respectively, and with a maximum amplification of 2.0% (FA15) and 12.7% (CA15) at 7 and 28 days, respectively. Therefore, only the CA15 and FA15 mixes highlight quite significant differences, as shown by the slopes of the lines, which describe the trend of the normalized indices over velocity. The CA15 mix design reaches a value of R_v = 1.15 at 28 days, starting from R_v = 0.97 at 7 days.

As can be seen from Figure 7, there is a strong correlation between V and E_dyn, with a correlation coefficient close to 0.9. This is because E_dyn is derived from V through Equation (2), in which the influence of density is limited since, as seen in Figure 6, density is highly homogeneous within the various mix designs. Therefore, the behavior of E_dyn with respect to the reference mix M0 is not very different from that of V. Variation of mean E_dyn with respect to M0 is limited, with a maximum reduction of 11.1% (RIO30) and 12.3% (FA15) at 7 and 28 days, respectively, and with a maximum amplification of 5.7% (SF15) and 7.1% (SF15) at 7 and 28 days, respectively. The RIO30 and FA15 mixes offer the lowest performance in terms of both E_dyn and V.

The RN was evaluated as the median of the readings on two sets of 15 measuring points chosen on two opposite sides of each specimen according to [47], and the results are resumed in

Table 9. Rebound number (RN).

Ageing Time	Sample	M0	PE15μ63	PE15μ20	PE30μ63	PE30μ20	RIO15	RIO30	FA15	FA30	SF15	SF30	CA15	CA30
7 Days		[-]
	1a	24	24	24	22	23	23	23	27	22	25	25	32	29
	2a	26	24	24	20	22	28	28	30	22	26	22	29	27
	3a	22	24	24	20	24	25	26	30	20	26.5	24	28	27
	Mean	24	24	24	21	23	25	26	29	21	26	24	30	28
	SD	2.00	0.289	0.289	1.155	0.866	2.566	2.517	1.732	1.155	0.764	1.528	1.803	1.155
	COV	8.33%	1.21%	1.21%	5.59%	3.77%	10.20%	9.80%	5.97%	5.41%	2.96%	6.45%	6.11%	4.17%
28 Days	1b	28	32	28	28	31	30	32	29	28	33	29	29	32
	2b	30	32	28	29	30	30	28	34	28	33	27	30	34
	3b	28	30	28	32	31	31	30	38	27	32	28	30	32
	Mean	29	31	28	30	31	30	30	34	28	33	28	30	33
	SD	0.866	1.041	0.00	2.291	0.500	0.577	2.00	4.25	0.577	0.577	1.041	0.577	1.155
	COV	3.04%	3.34%	0.00%	7.77%	1.64%	1.90%	6.67%	12.76%	2.09%	1.77%	3.74%	1.95%	3.53%

. Furthermore, Figure 8 shows the trends of the RN at 7 and 28 days, considering the ratio of R_RN between the actual mix design and the reference mix M0 at the same age.

It can be observed that the variation of R_RN at 7 and 28 days is particularly significant for the mixes CA15 and PE30μ63. For CA15, the R_RN index starts from the highest value, 1.23, at 7 days and decreases to 1.04 at 28 days, whereas for PE30μ63, R_RN starts from the lowest value, 0.86, at 7 days and increases up to 1.04 at 28 days. The FA15 mix design shows the highest R_RN value, 1.17, at 28 days compared with the other mixes design. The different behavior showed by the mixes could be ascribed to the different surface hardness of the related specimens. Some of them (i.e., PE30µ63, FA30, PE30µ20, SF30) show, at early curing, a lower superficial hardness with respect to the specimen of the reference mix; other ones instead (i.e., RIO15, RIO30, SF15, CA30, FA15, CA15) at the same age show higher hardness than the reference mix specimen, so at 28 days, when they more or less reach the same strength or higher than the M0 mix, the R_RN values increase or decrease depending on both the final value of the strength and the surface hardness with respect to the reference value. This aspect could also be understood from Table 9, where the RN always increases from 7 to 28 days of ageing for each mix, as expected, while R_RN exhibits lower or higher values than one depending on the mix design; nevertheless, the lower reduction of R_RN at 28 days of curing never exceeds 2.3%, while the higher increase is nearly 17%. Furthermore, a weak, direct relationship is observed between the RN and the compression strength, as shown in Figure 9.

4. Discussion

A synthesis of the results presented in the previous section is shown in Figure 10, Figure 11, Figure 12 and Figure 13, where the average value is denoted by a cross while the standard deviation is reported as a bar for each mix and for each property. It is possible to highlight the following aspects.

Regarding the compressive strength, see Figure 10; when the percentage of cement replacement using industrial waste materials increases, a decrease in concrete compressive strength occurs for all mix designs, except for the mixes CA15, CA30 and SF15. Focusing on PE, the mix designs PE15μ63, PE15μ20, PE30μ63 and PE30μ20 are characterized by a compressive strength lower than that obtained for the reference mix M0 both at 7 and 28 days. The loss of mean strength at 28 days is limited for mixes with a lower replacement percentage (PE15μ63, PE15μ20), for which it does not exceed 10%, while it becomes significant for mixes with a 30% replacement (PE30μ63, PE30μ20), for which it goes beyond 20%. Overall, the behavior of PE15 mixes is intermediate between that of mixes with CA and SF, which offer high strength, even higher than M0, and that of mixes with FA. However, in all cases, the mean strength of PE mixes remains above 30 MPa, an acceptable value for use in structural concrete. Focusing on RIO, the RIO15 and RIO30 mixes show the lowest strength at 28 days compared with all the other mix designs. The loss of mean strength compared with M0 is over 16% for the RIO15 mix and over 30% for the RIO30 mix, whose final mean strength does not reach 30 MPa, making its use in structural concrete questionable.

Regarding the dynamic modulus of elasticity, see Figure 11; due to the rather limited variability of the values, it is not possible to identify with certainty an increase or decrease trend either with the percentage of cement replacement variation nor with respect to reference mix design M0. However, it is interesting to note that, using the empirical relationship (3) proposed by the British testing standard BS8110 Part 2 [49] for the evaluation of the static modulus of elasticity E_t at 28 days as a function of E_dyn, the mean E_t of the mixes M0, PE15μ63, PE15μ20, PE30μ63, PE30μ20, RIO15 and RIO30 takes the values 29 GPa, 31 GPa, 31 GPa, 31 GPa, 31 GPa, 27 GPa and 26 GPa, respectively, highlighting the similarity between M0 and PE and the worse performance of the RIO mix designs, which once again makes their use in structural concrete questionable.

$$Es \left(GPa\right)=1.25 Edyn \left(GPa\right)-19,$$

(3)

Regarding the ultrasonic velocity, see Figure 12; the values determined for the various mixes are quite similar to each other and to the reference mix and show minimal variability also in relation to the percentage of cement replacement. This may indicate that the specimens in general have a similar degree of volumetric homogeneity, not influenced by the presence of waste materials in the mix design. This consideration is supported by the poor dispersion of the density values of the tested specimens. The high degree of correlation between velocity and dynamic modulus of elasticity makes the qualitative estimation of the stiffness through non-invasive measurement of velocity acceptable from an engineering perspective.

Regarding the rebound number see Figure 13; in general, the values at 28 days are included in the range of 27–33 and therefore present a fairly limited variability. It is possible to highlight that the observed variability can be attributed to the variation of surface conditions of the specimens in terms of roughness and hardness.

5. Considerations About CO₂ Production

Together with studies of mechanical properties of concrete created using industrial waste, another important aspect that led to the ever-increasing use of industrial waste materials in concrete production is represented by the reduction in the quantities of carbon dioxide released into the atmosphere, according to the values of equivalent CO₂ production reported in the main international database [50,51,52].

Table 10. Carbon dioxide emitted by the materials considered in this research.

Material	Equivalent CO2 Produced
[-]	[kg/ton]
Cement	693 [50]
Fine aggregates	33 [50]
Coarse aggregates	33 [50]
Metakaolin	248 [51]
Fly ash	93 [52]
Silica fume	14 [52]
Perlite	30 [52]
Rhyolite	30 [52]

reports the equivalent CO₂ produced unit value (kg/ton) that characterizes the materials considered, and

Table 11. Carbon dioxide emitted by the mix designs considered in this research.

Mix Design	Equivalent CO2 Produced	Δ
[-]	[kg/ton]	[%]
M0	312.783	-
PE15μ63	275.826	−11.8%
PE15μ20	275.826	−11.8%
PE30μ63	239.493	−23.4%
PE30μ20	239.493	−23.4%
RIO15	275.161	−12.0%
RIO30	237.744	−24,0%
CA15	287.155	−8.2%
CA30	261.579	−16.4%
FA15	278.662	−10.9%
FA30	244.655	−21.8%
SF15	279.524	−10.6%
SF30	235.975	−24.6%

summarizes the equivalent CO₂ produced considering the mix designs reported in Table 1. Δ indicates the reduction of equivalent CO₂ produced to realize the considered mix design with respect to M0.

It could be observed that the use of industrial waste materials as a partial replacement of cement in concrete mix design leads to a significant reduction in equivalent CO₂ production with respect to the reference mix design, representing an important aspect from the sustainability point of view. The maximum reduction of equivalent CO₂ produced was obtained for the SF30 mix design, which showed a compressive strength decrease at 28 days of about 6% (Table 6). The minimum abatement of equivalent CO₂ produced was obtained for the CA15 mix design, which instead increased the compressive strength at 28 days by about 10% (Table 6). The FA mixes offered a good balance both in terms of compressive strength (with an increase of about 4% and 26,6% for FA15 and FA30, respectively) and in terms of equivalent CO2 production reduction (10.9% and 21.8%). Focusing on PE15 mixes, it can be noted that they showed a reduction in equivalent CO₂ produced of 11.8%, limiting at the same time the compressive strength reduction to less than 10% after 28 days (Table 6). PE30 mix designs showed a higher reduction of equivalent CO₂ produced, which reached 23.4%, and a 28-day compressive strength reduction, which went beyond 20% (Table 6). Regarding RIO mixes, they entailed a great equivalent CO₂ produced reduction at 12% for RIO15 and 24% for RIO30, respectively, but showed the highest compressive strength reduction at 28 days, which went beyond 16% and 30%, respectively, as discussed in Section 4. For these reasons, the feasibility of using rhyolite as a partial substitution of cement in concrete mix design is currently questionable and needs further research.

6. Conclusions

This paper illustrated an experimental campaign aimed at analyzing the physical-mechanical performances of structural concrete specimens that incorporate various waste materials as partial replacements for cement. This study also evaluated the environmental impact in terms of equivalent CO2 produced. The focus was on the effect of perlite and rhyolite, two waste materials still little studied for this specific use, on the mix design of structural concrete. The performances of the specimens containing perlite and rhyolite were compared with those of specimens produced with a standard C25/30 reference (M0) concrete, as well as the performances of specimens containing other commonly used industrial wastes, such as fly ash, silica fume and metakaolin. The waste materials were introduced in the mix design in two percentages of replacement for cement, 15% and 30%; moreover, perlite was introduced in two different particle sizes, 20 μm and 63 μm. The impact of these materials on both the internal and surface conditions of the specimens was analyzed by using ultrasonic and rebound hammer tests.

From the obtained results, the following main observations can be drawn.

• The 28-day compressive strength of specimens with waste materials is generally lower than that of reference specimens, except for those containing CA and SF15. Specimens with PE achieve an average strength well above 30 MPa, an acceptable value for use in structural concrete, as do specimens with RIO15. Specimens with RIO30 instead offer an average strength significantly lower than 30 MPa, making its use in structural concrete questionable. In general, it is observed that the use of PE and RIO in cement replacement percentages higher than 15% is not convenient from the point of view of compressive strength. The use of RIO mixes also determines lower performances in terms of modulus of elasticity.
• The rebound hammer test generally produces results weakly correlated to the compressive strength and largely influenced, as expected, by the conditions of hardness and surface roughness of the specimens. The ultrasonic test provides globally poorly dispersed velocity values, indicating the low influence of the mix design on the internal characteristics of the specimens and highly correlated with the modulus of elasticity; this makes the qualitative estimation of the stiffness through non-invasive measurement of velocity acceptable from an engineering perspective.
• The use of industrial waste generally leads to a reduction, with respect to the reference mix design, of the equivalent CO₂ produced; the greatest reduction is guaranteed by SF30, RIO30 and PE30 mixes, while the least reduction is guaranteed by CA15. In terms of physical-mechanical performance, the best balance between mechanical performance and reduction of the polluting impact is offered by the SF30 mix design, which marks a final mean compressive strength of about 39 MPa and a reduction in the equivalent CO₂ produced of about 24%; moreover, if a compressive strength of around 33–35 MPa is acceptable, an excellent compromise is achieved by the PE15 mix design, which marks a reduction in the equivalent CO₂ produced of around 23%.

In summary, the use of perlite as a partial substitute for cement in structural concrete offers a good compromise between maintaining mechanical properties and reducing the equivalent CO₂ compared with the reference mix design, in line with other waste materials already used for the same purpose. Rhyolite, on the other hand, despite an appreciable reduction in equivalent CO₂, causes an excessive drop in mechanical properties compared with the reference mix; therefore, its use as a partial substitute for cement in structural concrete appears, at the moment, questionable.

Therefore, further research is essential to expand the sample size and gain deeper insights into the long-term performance and suitability of both perlite and rhyolite in concrete applications, particularly with respect to their mechanical behavior and environmental impact.

An important further development of the research work is the evaluation of the durability of structural concrete created using perlite and rhyolite within the mix design to consider their possible use in the realization of structural elements exposed to aggressive environmental conditions.