Backfill material

From QueensMineDesignWiki
Jump to: navigation, search


Backfill binders


Portland cement and ground granulated blast furnace slag (GGBFS), also referred to as slag cement, are two principal binding agents that are used in the Canadian mining industry. In Europe, slag cement means a blend of Portland and GGBFS, while in North America slag cement refers to 100% GGBFS. Slag cement, however, is not available in all markets. At present, there are few slag grinding plants located in North America; therefore, volumes are low and distances from plant to mines can lead to costly transportation. Consequently, in many instances, Portland is the binder of practical choice at most mine sites.

Portland cement

Portland cement is also referred to as hydraulic cement because it sets and hardens by reacting chemically with water. The reaction, called hydration, produces heat as well as enabling the cement to interlock and adhere to nearby substances. It is produced by grinding cement clinker, essentially a hardened hydraulic calcium silicate manufactured in a kiln, to which gypsum has been added.

The principal chemical constituents of Portland cement are listed below. In addition, note that cement producers utilize a short form for writing the chemical formulas of the cement constituents.

Tricalcium silicate or

3CaO \bullet SiO_2 \rightarrow C_3S    called alite

Dicalcium silicate or

2CaO \bullet SiO2 \rightarrow C_2S     called belite

Tricalcium aluminate or

3CaO \bullet Al_2O_3 \rightarrow C_3A

Tetracalcium aluminoferrite or

4CaO \bullet Al_2O_3 \bullet Fe_2O_3 \rightarrow→ C_4AF

With slight chemical or physical variations, these constituents can be adjusted to produce Portland cements with special characteristics. Knowledge of cement chemistry is helpful when choosing a cement type for binding tailings backfill that may carry significant amounts of impurities.

Cement is classified into several product types, of which six have standards for use in concrete set out by the Canadian Standards Association (CSA), in CSA- A3000 Cementitious Materials Compendium. Along with other types of cementitious materials, these types are listed in Table 1.

Table 1: Types of cementitious material used in Canada and the U.S.A

Cementitous material type Canada/USA Name in CSA Application
GU/Type 1 General use hydraulic cement Used in most applications.
{MS or MH} /Type 2 Moderate sulphate – resistant or moderate heat of hydration hydraulic cement MS provides moderate resistance to sulphate attack.
MH provides moderate heat of hydration.
HE/Type 3 High – early – strength hydraulic cement Provides high strengths in concrete within a week.
LH/Type 4 Low heat of hydration hydraulic cement Used where the rate and amount of heat generated must be minimized.
HS/Type 5 High sulphate – resistant hydraulic cement Slow strength gain but high sulphate resistance
Type S/ Ground granulated blast furnace slag Used as a supplementary cementing material in most applications.
Type F Fly Ash Low CaO content (<8%), especially used as a supplementary cementing material in low heat applications.
Type CI Fly Ash Intermediate CaO content (8% - <20%), used as a supplementary cementing material in many applications.
Type CH Fly Ash High CaO content (20%>), used as a supplementary cementing material in many applications.

In Canada, the most common type of hydraulic cement used in the mining industry is general use cement (GU), which is known as Type 1, Normal Portland Cement in the USA.

Portland cement is manufactured under standards set out by the CSA in Canada, ASTM in the United States, AS in Australia, BSI in the United Kingdom, SABS in South Africa, DIN in Germany etc. In the European Union, member countries are transitioning to European Standards (EN) referred to as Eurocodes in place of national codes. Both chemical and physical requirements are specified in these standards. Quality control plays a significant role in the production of all cement in order to meet these standards.

General use cement

General use cement  is vulnerable to chemical attack whenever backfill high in sulphates occurs. Many base metal ores consist of sulphide bearing minerals. If tailings produced from these ores contain a high content of sulphides the potential for the occurrence of sulphate attack is high if water is abundant and GU is used as the binder. Sulphate attack can cause backfill strengths to deteriorate in approximately 120 days. The sulphates produce sulphoaluminates, which cause expansion and can destroy cohesion.

High sulphate-resistant cement

High sulphate-resistant cement (HS) is manufactured to minimize these effects. Cement manufacturers create HS resistance by limiting the presence of C3A in the cement. Consequently, HS effectively resists sulphate attack. However, consumers pay a premium for specialty cements like HS, largely because of the time required to switch production runs, but there are also higher energy and material cost considerations. An alternative to HS are blended cements with a high component of slag. Such blends can provide equivalent sulphate resistance. Blended cements and are discussed in more detail in the following section.

Portland cement has played an ever-increasing role in Canadian mine backfill since the late 1950’s when it was first introduced in Canadian mines (Weaver and Luka, 1970). Its diversity of application, not only in backfill, but in concrete construction such as shaft linings, surface plants, as well as in ground support and backfill barricades in the form of shotcrete, adds to its utility at a mining operation.

Ground granulated blast furnace slag

Ground granulated blast furnace slag (GGBFS) is a product of the iron and steel industry. Slag that is quenched rapidly forms amorphous silica, i.e., material with a high glass content. High glass content and the right chemistry are key characteristics of slag cement. The process of manufacturing slag into cement begins when molten slag is water-cooled and either pelletized or granulated. It is then ground in a ball mill and sized generally finer than GU. It should be noted that a key difference between iron blast furnace slag and slag obtained from a base metal operation is the presence of calcium and the glass content. Calcium and a high glass content are essential for any slag in order for it to be cementitious.

Quality control begins after cooling with analysis of the chemistry. Product fineness, glass content and compressive strength of slag cement mortar cubes are tests performed on a regular basis and used to compare with standards in order to maintain a consistent and reliable product.

Ground granulated blast furnace slag can be used as 100% of the binder in backfill but can take up to two weeks before it starts to set. More often, slag is blended with GU. A blend of 10% GU and 90% GGBF slag has produced very good results, relative to 100% GU, in the Canadian market (see Table 2). Characteristically, 100% GGBF slag cement has lower short term strengths for the first 14 days, but longer term strengths can far exceed GU beyond 28 days. Conversely, GU provides early strength, as shown in the table below.


Table 2: Strength results from a 70% pulp density mix using base metal tailings

Binder Type
(Tailings to Binder Ratio)

Day 3

Day 7
Day 14
Day 28
100% GU (10:1) 101 (0.70) 138 (0.95) 170 (1.17) 184 (1.27)
100% GU (30:1) 20 (0.14) 22 (0.15) 25 (0.17) 29 (0.20)
100% GGBF Slag (10:1) soft 10 (0.07) 182 (1.26) 292 (2.01)
100% GGBF Slag (30:1) soft 6 (0.04) 48 (0.33) 96 (0.66)
90% GGBF slag/ 10% GU (10:1) 38 (0.26) 106 (0.73) 141 (0.97) 165 (1.13)
90% GGBF slag/ 10% GU (30:1) 35 (0.24) 56 (0.39) 75 (0.52) 88 (0.61)
70% GU/ 30% “C” Fly Ash.(10:1) 88 (0.61) 114 (0.79) 131 (0.90) 139 (0.96)
70% GU/ 30% “C” Fly Ash.(30:1) 18 (0.12) 21 (0.14) 23 (0.16) 33 (0.23)

* average of three cylinder breaks, units of p.s.i. (MPa in brackets)

The slower hydration rate of slag cement and blends that are high in slag cement is a characteristic that can be favourable for paste fill. Should a distribution system become plugged while carrying a slag-based cement, operators have more time to clean the pipes before the mix sets up, compared to GU. Slag cement is also resistant to sulphate attack, equivalent to 100% MS cement (see Table 1) when blended in proportions with GU from 50% and greater.


One natural cementing material that has been utilized in Canada is pyrrhotite-rich fill. The Horne Mine, Sullivan Mine and the Mattagami Mine all utilized pyrrhotite to bind their backfills at one time. In the late 1950’s (Barsotti, 1978) and again in the 1980s, Inco Ltd. had also examined its stream of pyrrhotite tailings for its binding capacity. Based on these bodies of work it appears that the oxidation reaction required for cementing the tailings is very difficult to control, making for unreliable binding. When mining in a timely manner is essential, use of an unreliable binder is unacceptable, because it is potentially unsafe for people working around it and may disrupt production schedules.


In addition to the cements used in backfills, pozzolanic material such as non-ferrous slag and fly ash are being used in the industry. A Pozzolan is a material capable of reacting with lime in the presence of water at ordinary temperature to produce cementitious compounds (Lea, 1971). It should be noted that not all non-ferrous slag or fly ash is pozzolanic.

Non-ferrous slag

Thomas (1973 and 1979) reported on the testing of several non-ferrous slags. The air-cooled copper converter slag and air cooled dezinced lead smelter slag that he tested had no pozzolanic activity. However, upon remelting and quenching, Thomas found that a high pozzolanic activity number was achieved. Consequently, it was demonstrated that some quenched non-ferrous slag has the potential to replace some GU cement. McGuire (1978), of Falconbridge Ltd., showed that quenched, granulated nickel slag ground to the fineness of cement showed promise as a pozzolan, activated by Portland cement for binding hydraulic tailings. Barsotti (1978), at Inco Ltd., reported that finely ground air-cooled nickel reverberatory slag activated by Portland cement possessed pozzolanic properties. However, over 30 years have passed since this work was undertaken by these Sudbury, Ontario researchers and no commercial development has come from it. This lack of development may stem from the fact that Sudbury is located reasonably close to binder manufacturing plants, keeping material transportation costs low, thereby removing the cost savings incentive for these local solutions. In contrast, remote mines pay a premium in the form of higher material transportation costs for binder, adding significantly to the product cost and making local solutions more viable.

For mines where remote location causes transportation costs to be as much as the cost of the binder, ground non-ferrous slag may be an economical alternative. Thomas (1973) had reported on the use of ground copper reverberatory furnace slag (CRFS) at the Mt. Isa Mine, Australia. To this date Mt. Isa mine continues to successfully use 6% CRFS blended with 3% Portland Cement.

In 1986, the O’Kiep Copper Company Ltd., of South Africa, filed for a patent that describes how a non-ferrous quenched slag can be modified chemically by the addition of products enhanced in CaO to the molten slag, which is cooled, and then further enhanced with gypsum. The slag and gypsum are then ground to the fineness of Portland Cement. One of its mines, the Carolusberg mine had successfully used this binder for several years until the mine was closed in 1997 when its ore reserves were exhausted.

Fly ash

Fly ash, derived from the burning of coal, is a residual material produced by the burning of pulverized coal in boilers used to generate electricity. Upon burning, the non-combustible mineral matter in the coal melts. The finest portion is exhausted with the flue gasses and is recovered by electrostatic precipitators. This is what is referred to as fly ash. The coarser portion of this material is drawn to the bottom of the boiler and is referred to as bottom ash and boiler slag.

Two main types of fly ashes are distinguished and referred to as Class C Ash and Class F Ash. Generally, coals from eastern North America produce low lime fly ashes referred to as F ash, originating dominantly from bituminous coals. Coals from western North America, generally lignite and sub-bituminous coal, produce high lime ash, referred to as class C fly ash. According to CSA, an F Ash has less than 8% CaO and a maximum of 5% loss on ignition where a C ash has 8% or greater CaO and a maximum of 5% loss on ignition. The C Ash is further classified into CI which is 8 -<20% CaO and CH which is over 20% CaO. The strength performance of particular C and F fly ashes utilized to bind a rock fill in test cylinders has been tabulated in Table 3. In general, C ash pound for pound typically outperforms the strength of mixes using F ash in mine backfills. Table 2 shows the performance of a C Ash blend relative to GU, slag cement and slag blend. It should be noted that the performance of one C ash may be different from the performance of another due to chemical variances stemming from the wide definition of a C ash with respect to CaO content. Comparative testing for flow and strength is the best way to predict expected performance.

Table 3: Strength results from a 55% pulp density slurry with mine waste development rock. (45cm x 90 cm cylinders)
Binder type
(Binder to solids ratio 5%)
Day 28
Day 56
100% GU 197 (1.36) 335 (2.31)
65% GU/35% C Fly Ash 234 (1.61) 383 (2.64)
65% GU/35% F Fly Ash 127 (0.88) 265 (1.83)

(reference necessary)

  • average of three cylinder breaks, units of p.s.i. (MPa in brackets)

    Note the significant strength difference between the C and F ash blends.

In the Canadian mining industry, C Fly Ash is often blended with GU cement. Blends from 30%-50% C Ash with GU are common. It is not unusual for fly ash blends to outperform straight GU cement in 28 day strengths.

With respect to quality, fly ash has been used in concrete since the 1940’s and in Canadian hardrock mines, in backfill, since the early 1980s. Although there is no specific CSA standard for mining applications, there is a standard for use of fly ash in concrete: CSA A3000 Cementitious Materials Compendium. This standard covers both chemical and physical characteristics of both C and F fly ashes

The burning of wood products to produce steam for electricity, especially utilized in co-generation plants, produces a wood ash. This type of ash typically contains high amounts of residual carbon. This is reflected by high “losses on ignition,” which can be of the order of 40% of the weight of the ash. Consequently, wood ash tends not to be pozzolanic, and in general, makes poor binders for mine backfill applications as demonstrated by the strength data shown in Table 4.

Table 4: Test data from a 50% pulp density slurry with rock aggregate

Binder type
(Binder to solids ratio of 4.8%)
Day 7
Day 14
Day 28
Day 56
Day 90
100% GU 674 (4.65) 809 (5.58) 947 (6.53) 966 (6.66) 1159 (7.99)
50% Wood Ash, 50% GU 124 (0.86) 131 (0.90) 170 (1.17) 166 (1.14) 162 (1.12)

(reference necessary)

  • average of three cylinder breaks, units of p.s.i. (MPa in brackets)

    Note the effect on strength of the wood ash

Temperature effects

Cemented backfill, like concrete, needs the correct moisture and temperature to cure properly. Favourable temperatures for the curing of concrete ranges from 10◦C to 20◦C. Curing is all about maintaining adequate moisture and temperature in backfill at early stages so that it can develop the properties it was designed for. For backfill, these temperatures refer to the in-situ rock temperatures that surround the fill. In hot temperatures, a higher initial curing temperature will result in rapid strength gain and lower long term strength. With respect to mine backfill, the only practical controls are initial material temperatures, which should be as cool as possible and water content in the paste, which may have to be boosted if temperatures drive off water before the binder hydration reaction is complete. However, the additional water will reduce strengths by 15% or more depending on the volume. So there is a trade-off.

In cold temperatures, fresh cement paste will freeze when the temperature is below -4◦C. Below about -10◦C cement hydration ceases along with any backfill strength gain. If the paste freezes, its potential strength can be reduced by over 50% when non-freezing conditions are returned. If the paste temperature can be maintained at 10◦C or higher for a minimum of two days, the backfill will have a good chance of reaching its intended strength. In the case of rock fill trucked from surface and exposed to sub-zero temperatures, heated aggregate, water, heated dump boxes on trucks and the use of a concrete accelerator are strategies that have worked in cold temperatures.

Percentage binder addition

Although it is common in the mining industry for the binder to dry solids ratio to range anywhere from 1:8 (11%) to 1:30 (3%), the volume of binder necessary to cement a backfill is dependent on several variables.

Design strength

The forces to which the backfill will be subjected dictate the necessary design strength of the backfill. There is a direct relationship between backfill strength (cohesion and friction angle, (Thomas 1969, Thomas, E.G. and Vance, W.E., 1969) and binder addition. Consequently, higher design strengths will require higher binder addition rates.

Pulp density

It is well recognized in backfill technology that maximizing the pulp density of the fill can improve strengths (Weaver and Luka, 1970). A fill slurry that contains excess water will require greater volumes of binder in order to achieve the same strength as a mixture with a higher pulp density. One of the benefits often used to justify the introduction of paste fill, for example, is that it will require less binder than a conventional hydraulic fill. This benefit comes about because of the lower binder to water ratio found in the paste fill. However, if total plant tailings are utilized and the fill material is dominated by a fine particle size, for example, then additional water requirements due to the higher surface area of the smaller particles may reduce strengths.

Grain size distribution

Although the cohesion and friction angle of backfill are known to increase with increasing binder content, the lack of a certain grain size in tailings or rock fill can limit the effectiveness of the binder strength or affect friction losses, which could add a degree of variability to the fill rheology. For example, gold mine tailings are typically very fine and lacking in coarser particles, whereas alluvial sand, used in some backfilling operations, typically lacks fine particles. Both of these sources of backfill may potentially require much more volume of binder than an operation with a broader distribution of grain sizes in its backfill material to achieve the same strength. The source of the grain size problem is in the packing efficiency of the grains. It is well known in concrete technology that the amount of cement paste required to bind aggregate is greater as the volume of the voids increase. In essence, the better the packing, the more efficient the very fine grained binder will be at cementing particles. As noted in the section under Void ratio, McGeary (1961) determined an ideal grain size distribution for ternary and quaternary grained material. This theory provides a model that can be used to improve fill strengths, solely by adjusting the grain size distribution. Such a model can be effective when sources of backfill material of a specific grain size are available for addition to the fill. However, hydraulic tailings that are packed too tightly will suffer from inadequate percolation rates, allowing ponding and segregation of cement thereby nullifying the efforts of mixing. Paste fills, on the other hand, require no drainage, but may be prone to pipe plugging in the delivery system if friction factors increase appreciably due to a fluctuating grain size distribution of the fill material. Consequently, the grain size distribution of the aggregates plays a role in binder addition and additionally, the denser the packing, the greater the efficiency of the binder.

Type of cement

If early (for a plug) and or long term backfill strengths are required, the type of cement, i.e., Portland, slag or blended cement can affect the addition percent of the binder.

Tailings, sand and course aggregate

Most backfill practitioners need to know that the aggregate used in backfill does not generally fall under any criterion established in any national codes. Unlike the aggregate and sand used in concrete, which is highly regulated and used in an extraordinary range of construction applications from high-rise building to road surfaces to under water structures, mine backfill has a comparatively narrow and shorter-lived application.


Mine tailings are desired in backfill only because they are generally available. If a mill is or was near to the mine site, then its tailings are usually the lowest cost backfill option. With environmental costs and concerns mounting, the use of tailings in backfill is not only very practical but may save money compared to storing the tailings in surface impoundment areas. The same can also be said for the use of mine waste rock or development muck, which can potentially be used as rock fill underground but should be sorted on surface into piles according to chemistry, i.e. presence of sulphide minerals, and monitored for acid run off water. Tailings can be found in a range of grain size distributions as shown in Figure 1. Note that the mineralogy of an ore and gangue dictates how finely an ore must be milled in order to optimize recovery rates. In general, gold tailings tend to be finer than polymetalic tailings, which tend to be finer than nickel tailings. Tailings particles also tend to be angular in shape as a result of the milling process. This angularity characteristic can add up to 5 degrees of friction angle (roughly 35-39 degrees) compared to rounded particles (Thomas, 1979).


Natural sand may also be a suitable backfill material depending on availability, grain size and quantities in reserve. Knowing the origin of the sand—glacial, fluvial, alluvial or aeolian—is important because it essentially reveals how the sand was segregated. For example, aeolian sand tends to consist of uniform rounded particles of like size, while alluvial sand is also rounded, but often missing the finest fraction sizes. Regardless of the origin, a systematically collected unsegregated sample is recommended in order to determine a representative particle size distribution of the material. Screening may be required to remove undesirable material including organics or certain size factions. These deposits typically vary greatly over their breadth and depth in grain size distribution. The mapping of certain size factions may also be necessary to define and locate desired materials. The blending of material from different locations may be a strategy required to optimize the grain size distribution and stretch resources.


Course aggregate as used in rock fill is often sourced from development waste, crushed rock, stripped rock from open pit operations, course gravel or crushed slag. As with any backfill material, the grain size distribution plays a crucial role in strength development. Run of mine development waste can be lacking in particles under five cm., particularly when the rock is hard, contributing to poorly packed backfill that will underperform in terms of strength. In many underground operations, waste is hoisted or trucked to surface to be stockpiled and dumped down a raise to be made available as rock fill for underground use. Two comments are worth mentioning:

1.   Measure the grain size distribution at the bottom of the pass:
After being dumped down a waste pass, depending on the hardness and Rock Mass Rating, the rock, will have undergone a degree of comminution that will reduce the average grain size. Consequently, measuring the grain size distribution at the bottom of the pass is more fruitful than a measurement from surface.
2.  Measure the moisture content of the rock fill:
In addition to a different grain size distribution, the rock fill from the raise may have gained moisture if it was sitting on surface and exposed to the environment or it may be subject to water collecting at the bottom of the raise. It is recommended that the moisture content of the waste rock be measured and that the moisture content be deducted from the total water added to the paste slurry.


Potable water is always recommended. Many operations utilize some percentage of recycled process water for use in their backfill. Many gold operations use a sulphate-based chemical to capture cyanide and when process water is used in backfill these impurities remain in the water. Some mines have only saline water available for filling operations. There is no hard and fast rule on the effects of water quality on the backfill strength. The best advice is to run a series of lab strength tests comparing the strength of cylinders using mine water with those using potable water. Determine if there are any detrimental strength effects. Once an issue has been investigated and quantified, an appropriate course of action can be taken.

See also

Personal tools