Experiments with precipitation of silica from waste water at Hellisheiðarvirkjun

(1)

Experiments with precipitation of silica from waste water at

Hellisheiðarvirkjun

2014-‐12-‐15 Final report Instructors:

Vera Sólveig Ólafsdóttir Ingvi Gunnarsson

Examiner Per Alvfors Orkuveita Reykjavík

Kungliga Tekniska Högskolan Einar Jón Ásjörnsson

(2)

Abstract

This report is about the experiments with mixing of the separated water and the vacuum pump seal water at Hellisheiði power plant. This is done to prevent silica scaling and clogging in pipes and reinjection wells as well as eliminating vacuum pump seal water from the plant. The experiments were done in four stages: the first stage comprised of tests with different flows of separated water at 70°C, the second stage was carried out by mixing the separated water at 70°C and the seal water with different amounts of the seal water, the third and the fourth stages were like the first and second but with the separated water at 120°C. The results show that this method is good if the mixture is around 50/50 separated water and seal water, to control the silica scaling in the separated water and to be able to reinject the seal water with the separated water. This does not eliminate the silica scaling in all of the separated water because the amount of separated water is much more than the amount of seal water that comes from the plant.

(3)

Table of Equations

Equation 1 Dissolution of quartz ... 7

Equation 2 Formation of metal silicate no1 ... 7

Equation 3 Formation of metal silicate no2 ... 7

Equation 4 Polymerization reaction no1 (Gallup, 2011) ... 8

Equation 5 Polymerization reaction no2 (Gallup, 2011) ... 8

Equation 6 is for calculating the concentration of hydrogen sulphide ... 13

Equation 7 is for calculating the concentration of carbon dioxide by using theory. ... 13

Equation 8 is for calculation the concentration of carbon dioxide by using acid and base. ... 13

Equation 9 Calculation of concentration of silica by using the chlorine concentration. ... 15

Table of Figures

Figure 1 production process of Hellisheiðavirkjun (or_ffrl_web2, 2014) ... 6

Figure 2 Shows operation line for Hellisheiði power plant and the solubility lines for amorphous silica and quarts (Gunnarsson, 2014). ... 7

Figure 3 the effect of seal water on the saturation levels for silicon for different temperatures for the separated water. (Gunnarsson, 2014) ... 8

Figure 4 experimental setup. The separated water and the seal water are mixed and lead through the tank. ... 12

Figure 5 Amorphous silica scaling and polimerization with separated water at 70°C ... 15

Figure 6 Total silica in the separated water at 70°C and the mixture with seal water. ... 16

Figure 7 the iron in the separated water at 70°C and in the mixture with seal water. The red line is the corrosion threshold for the pipes ... 16

Figure 8 the pH as function of the mixing ratio for separated water at 70°C. The red line is the corrosion threshold for the pipes. ... 17

Figure 9 Amorphous silica scaling with separated water at 120°C ... 18

Figure 10 Total silica in the separated water at 120°C and the mixture with seal water. ... 18

Figure 11 the iron in the separated water at 120°C and in the mixture with seal water. The red line is the corrosion threshold for the pipes. ... 19

Figure 12 The pH as function of the mixing ratio for separated water at 120°C. The red line is the corrosion threshold for the pipes. ... 19

(5)

Introduction and motivation

The power plant

Orkuveita Reykjavíkur was first founded in 1909 with the sole purpose of distributing clean water to homes in Reykjavík. In 1921, power production was started in the river Elliðará in Elliðarárdal in Reykjavík and in 1930 the company started distributing heat in the form of warm water from the hot pools in Laugardal. Today the company runs the largest geothermal system in the world. The source of the water in the district heating system comes from, among other places, Nesjavellir power station that has 300 MW thermal input and Hellisheiði power station that has 133 MW thermal input (Starfsemi: Orkuveita Reykjavíkur, 2014).

Hellisheiði power plant is owned by Orkuveita Reykjavíkur, but in January 2014 its daughter company Orka náttúrunnar took over the running of the plant. This was done to accommodate changes in Icelandic law. Orka náttúrunnar runs one hydroelectric power plant and two geothermal power plants. Both of the geothermal power plants are situated in the Hengill area. The geothermal area for the Hellisheiði power plant is south of Hengill. The area is divided up into two production areas. The first one is the upper area that is situated above Hellisskarð and the second one is the lower area that is situated below Hellisskarð and Skarðsmýrafjall. Hellisheiði power plant can produce 303 MW of electricity and 133 MW of thermal energy. Hellisheiði power plant can expand its thermal capacity to 400 MW in the future if there is a need for hot water in the capital (Virkjanir: Orka náttúrunnar, 2014).

The process at Hellisheiði power plants (see Figure 1) starts at the boreholes. The holes are about 3 km deep and can be either vertically or directionally drilled. Directional drilling is environmentally friendly because it minimizes the impact on the surface. It is possible to drill about 1200 m from the drill side using directional drilling. The borehole liquid is a mixture of liquid and steam at a temperature of around 300 °C. Each drill hole has a muffler attached to it to dampen sound in case of an emergency stop in the production line. The liquid and steam mixture is then directed into a steam separator. There the steam and liquid are separated. The steam is used for electricity production and the liquid, now called the separated water, is used for production of hot water. The separated water is heavier then the steam, and will fall to the bottom and flow out of there while the steam will flow out at the top of the separator. The pressure of the steam will then be regulated by a control valve that can let out a part of the steam through a hood. The rest of the steam flows toward the steam dryer. The dryer eliminates any liquid that might still be present in the steam by filtration. After that, the steam flows into the electricity-‐generating unit. The unit is composed of a turbine and a generator. The steam will flow into the turbine. When the pressure is lowered the thermal energy gets released and that is what is producing the electricity in the generator. The electric energy is then put through a transformer to change the voltage from 220 V to 220000 V to prevent energy losses along the way to consumers in Reykjavík. Before the electric energy gets to users the voltage is changed back to 220 V in stages. The steam from the turbine flows into the steam condenser. The condenser works in conjunction with the cooling towers. The condenser uses cold water to cool the steam down. This cold water comes from a closed circuit with the cooling towers. The cooling towers get hot water from the condenser, cool it down and give the condenser the cold water back. The cooling towers provide the condenser with constant cool water so that the condenser can condense the steam. The condenser also heats fresh ground water up to 50°C to use in district heating. The flow of condensate that comes from the power plant is ca 500 kg/s. Part of the condensate is used in the operations of the vacuum pumps. The pumps are seven and the vacuum pump’s seal water that comes from them is ca 77 kg/s. The separated water from the separator goes through a control valve that determines how much the pressure fall is. With less pressure the separated water boils again and makes more steam that can be used for electricity production. The mixture then flows into a low pressure separator that separates the steam from the separated water. The low-‐pressure steam goes through the same process as the high pressure steam did. The description of the process is in the text

(6)

ground water for district heating. The separated water from the low-‐pressure separator is ca 600 kg/s. It flows towards the heat exchanger and on the way there is a control valve that controls how much of the separated water flow can get into the heat exchanger. There is also another control valve that can make the separated water bypass the heat exchangers altogether if necessary. In the heat exchanger, the ground water that has already been heated up to 50°C gets heated up further. The separated water is then reinjected into the ground in a reinjection well that is 2 km deep. The separated water then flows down into the earth to the ground water reservoir where it is heated up again and reutilized in Hellisheiði power plant. The ground water that is coming from the heat exchanger flows into a gas extractor to eliminate any corrosive gases in the water. The water is then flowing into a pump. Most of the time the water flows freely without help but if the load is heavy then the pump helps the water move. This is done so that the pipes can be smaller in diameter and therefore more compact. The water only cools 1°C as it flows into hot water utility tanks (they are similar to a reservoir for hot water) at Reynisvatnsheiði. From there the water flows to the consumer (or_ffrl_web2, 2014; Gunnarsson, 2014).

Figure 1. Production process of Hellisheiðavirkjun (or_ffrl_web2, 2014)

There are three types of wastewater coming from Hellisheiði power plant. One is the separated water, the second is seal water and the last one is the condensate water. The amount of silica in the separated water at the plant is around 15000 tonns per year. The separated water therefore has a lot of potential for silica scaling. By mixing the separated water and the seal water together it is possible to influence the amorphous silica scaling. The silica scaling is influenced in two ways by the mixing. First because of the low pH of the seal water and the high pH of the separated water the pH of the

(7)

Background

One of the main problems in geothermal power production is the scaling of silica. It is a limiting factor of how much heat can be extracted from geothermal water. When heat is extracted from high-‐ temperature geothermal water (>200 °C), silica moves from being in equilibrium with quartz to being over saturated amorphous silica. This is what happens at Hellisheiði power plant as can be seen in Figure 2. The green line is the solubility line for quartz and the blue line is the solubility line for the amorphous silica and finally the red line is the operational line for the plant. When the plant separates the steam from the water then the chemicals in the separated water get concentrated. The solubility of the chemicals in the separated water decreases with decreasing temperature. At the beginning, the silica is in equilibrium with quartz, but as soon as the heat is extracted the water gets oversaturated with quartz and under saturated with amorphous silica. Then the plant continues to use the heat and the silica begins to be oversaturated amorphous silica. When quartz dissolves in water a silicic acid is formed (see Equation 1). For high pH like in the separated water in Hellisheiði the acid has a tendency to disassociate into 𝐻!_{and
𝐻}

!𝑆𝑖𝑂!!. If there are metal ions in the water it is

possible to get metal silicate complexes (see Equation 2 and Equation 3(Gallup, 2011)).

𝑆𝑖𝑂!+ 2𝐻!𝑂 ↔ 𝑆𝑖 𝑂𝐻 !! Equation 1. Dissolution of quartz

𝐻!𝑆𝑖𝑂!!+ 𝑀!!↔ 𝑀𝐻!𝑆𝑖𝑂!!! where M=Aluminium (Al) and iron (Fe) Equation 2. Formation of metal silicate no1

𝐻_!𝑆𝑖𝑂_!!+ 𝑀(𝑂𝐻)_! ↔ 𝑀(𝑂𝐻)_!𝐻_!𝑆𝑖𝑂_!! where M=Aluminium (Al) and iron (Fe)

Equation 3. Formation of metal silicate no2

Figure 2. Operation line for Hellisheiði power plant and solubility lines for amorphous silica and quarts (Gunnarsson, 2014).

There are two processes that can occur in the amorphous silica silicate water. The first process is that the water will precipitate amorphous silica on to the surface. This means that if the over saturated amorphous silica water is injected straight into the injection well the probability of clogs in the well rises considerably and therefore shortens the lifespan of the well. The second process is that the silica will polymerize and form colloids. These colloids are less likely to precipitate and clog the injection wells (Gunnarsson & Arnórsson, 2005). If the flow in the pipes is turbulent it is more likely that silica will precipitate onto the surface but if the flow is put into a tank where it is quiet then it is more likely to polymerize (Gunnarsson, Ívarsson, Sigfússon, Thrastarson, & Gíslason, 2010). It is therefore economical to make the wastewater polymerize to get as much as possible of the silica as

(8)

The polymerisation for the amorphous silica and silicates is described in Equation 4 and Equation 5. The mechanism for the polymerisation on is not completely known. In the experiments that have been done, reaction orders between 1 and 8 have been found. The problem is that there is no way to measure the silica, instead the molybdate active silica and the total silica are measured and the difference is considered to be polymeric silica (Gunnarsson & Arnórsson, 2005).

2𝑆𝑖 𝑂𝐻 !!↔ 𝑂𝐻 !𝑆𝑖𝑂𝑆𝑖 𝑂𝐻 !+ 𝐻!𝑂

Equation 4. Polymerization reaction no1(Gallup, 2011)

2𝑀 𝑂𝐻 !!+ 𝑥𝑆𝑖 𝑂𝐻 ! ↔ 𝑀!𝑂!∙ 𝑥𝑆𝑖𝑂!+ 2𝑥 + 3 𝐻!𝑂

Equation 5. Polymerization reaction no2(Gallup, 2011)

The kinetics of these two processes depend on a few factors such as the pH of the solution, ionic strength, temperature and amorphous silica oversaturation. Therefore, the success of making the geothermal waste water polymerize and reducing the risk of clogging depends on the rate of colloid formation, the rate of colloid precipitation and the rate of amorphous silica precipitation. It is possible to affect the polymerization process by changing the pH of the solution. This can be done by adding acid or base to the solution. It is also possible to affect the ionic strength of the solution by adding salt. Neither of these solutions is good economically because of the amount of waste water from a geothermal power plant (Gunnarsson & Arnórsson, 2005).

It is possible to theoretically calculate the silica saturation levels for different temperatures for the separated water by looking at Figure 3 where the silica saturation has been plotted against the amount of seal water

Figure 3. Effect of seal water on the saturation levels for silicon for different temperatures for the separated water.(Gunnarsson, 2014)

Other work done in the field

As mentioned before, there are many ways to avoid silica scaling and therefore there are different ways that geothermal power plants deal with their silica scaling problem. Some of them use: 1) hot brine injection at or near amorphous silica saturation, 2) adjusting the brine pH, 3) aging of geothermal waste water, 4) crystallization, 5) Removing the silica with controlled precipitation with metals, 6) making silica precipitated by cationic surfactants, 7) diluting the separated water with either the condensate or fresh water, 8) evaporation, 9) using a reducing agent, 10) using organic

(9)

Ageing of geothermal waste water

The idea is to allow the geothermal wastewater time to polymerize the over saturated monomeric silica in a retention tank or in a pond. This reduces the amorphous silica oversaturation that is known to scale in the injection wells. The polymeric silica with low ionic strength has less tendency to scale in the injection well than the monomeric silica. For this process to work an understanding of the rate of polymerization is necessary. An experiment was done to find out the time needed to lower the monomeric silica concentration to a concentration close to the equilibrium with amorphous silica. Sufficient polymerization time was found to be between one and four hours depending on the pH of the water. In this experiment there was an induction period, a period that the monomeric silica seemed stable. This period is thought to be the result of a measurement error when the monomeric silica is starting to polymerize. The speed of polymerization also depends on the ionic strength. If the ionic strength was higher the polymerization was faster, but higher ionic strength also means faster deposition of silica. Therefore, this is not considered a good method for geothermal wastewater with high ionic strength. This method is usable for high temperature geothermal wastewater if the pH is above 6. If the pH is below 6 the polymerization time will be too long for this method to be considered economical (Gunnarsson & Arnórsson, 2005).

Environmental aspects

Surface disturbance

Geothermal fields are often to be found in beautiful unspoiled nature, even in or near national parks where man has not made any marks yet. National parks can be very special, with unusual and special vegetation. It is possible that they can be of importance for the tourist industry as well as having an historical importance. In other cases the geothermal area can be in a rich agricultural area or in forests. This is the case in for example Japan, Indonesia, USA, New Zealand and even more countries. The construction of a power plant that utilizes geothermal energy will disturb the surface so the decision to start such a power plant has to be made with consideration to nature and it has to take into consideration what the land is used for today. The drill site is usually small, up to 2500 m2_{.

The}

site is often kept as small as possible by directional drilling of the wells. That way more than one well can be drilled from one drill site. Sometimes the surface will change because ponds are drained and other times because of the drill, use and depletion of the geothermal reservoir as well as due to the necessary infrastructure and the power plant buildings changes on the site. Some consider the buildings to disturb the scenery that was there before the construction. It is possible to paint the buildings and the pipelines to blend into the landscape. This is done to minimise the disturbance they have on the scenery. It depends on how far the power plant is from populated areas how long the power transmission lines have to be. In some cases where the power would be transported over great distances, power consuming industries can be built close to the geothermal power plant. Iceland has done this with its aluminium production. The extraction of fluids can cause hot springs or geysers to disappear entirely, change them to fumaroles or shift them to another location. The land used in geothermal power plants can therefore be a lot bigger than just the land used for the buildings and the wells (Kristmansdóttir & Ármannsson, 2003; Bayer, Rybach, Blum, & Brauchler, 2013).

Not all of the changes are bad for the tourist industry. Everyone can for example use the roads that are made. In the Hengill area, where Hellisheiði and Nesjavellir power plants are situated, there are now hiking trails that Orkuveita Reykjavíkur has been instrumental in making in collaboration with local authorities (or.is, 14). There is one tourist attraction in Iceland that came to be because of the Svartsengi power plant. As soon as they activated the steam holes the separated sea created a lagoon. Today this lagoon is called Bláa lónið, or in English the Blue lagoon, and it is one of Iceland’s most popular tourist attraction (hsorka.is, 2014).

(10)

Noise

There is almost always some noise when the construction crew is working and then there is the drilling noise, though these are temporary until the buildings are up and the drilling is done. The noise coming from these rarely exceeds the 90 dB permissible exposure limit given by the national institute for occupational safety and health in the USA. Once the plant is operational, there will be noise from the discharging boreholes. This noise can exceed 120 dB but with a noise muffler installed on the boreholes this is kept below the 65 dB limit that the US Geological Survey has set for such an operation (Kristmansdóttir & Ármannsson, 2003).

Chemical pollution

There are two ways that a geothermal power plant can pollute the environment. One is air pollution by steam, the other is ground pollution by liquid. There can be many harmful chemicals in both the steam and the liquid. In the steam, the polluting chemicals can be carbon dioxide, hydrogen sulphide, methane, mercury, radon, ammonia and boron. In the liquid, they can be hydrogen sulphide, mercury, ammonia, boron, arsenic, lead, cadmium, iron, zinc, manganese, lithium and aluminium (Bayer, Rybach, Blum, & Brauchler, 2013; Kristmansdóttir & Ármannsson, 2003).

Carbon dioxide

In geothermal power production the carbon dioxide emissions come mostly from degassing of the magma. Reported carbon dioxide emissions can wary greatly, from 4-‐740 g/kWh with a weighted average of 122 g/kWh. At Kizildere geothermal power plant in Turkey, the carbon dioxide content is very high, or more than 1300 g/kWh if the steam consumption of the plant is 10.96 kg/kWh. Instead of releasing all of the carbon dioxide into the atmosphere they make industrial grade carbon dioxide. In geothermal regions, carbon dioxide is emitted naturally and it depends on the geothermal field how much is emitted. It is therefore interesting to look at the changes in emissions when the geothermal energy is used. The Larderello field in Italy has observed a decrease of carbon dioxide emissions. The Svartsengi field in Iceland has however around six times higher emissions after the start of power production than before. At the Ohaaki field in New Zealand no change in carbon dioxide emissions was found. At Wairakei in New Zealand, a doubling of carbon dioxide emissions was observed (Bayer, Rybach, Blum, & Brauchler, 2013).

Sulphur

There is a small amount of sulphur dioxide in the gaseous emissions, but the main source of sulphur compounds in the gaseous emissions is hydrogen sulphide. Hydrogen sulphide is the compound that can cause local environmental concern because of the bad odour and the potential toxicity (Bayer, Rybach, Blum, & Brauchler, 2013). The faculty of civil and environmental engineering at the University of Iceland did a study of the hydrogen sulphide concentrations inside a 30km radius of the two geothermal power plants closest to Reykjavík, the capital city of Iceland. They found that the odour exceeds the national health limit of hydrogen sulphide, which is 50 µg/m3_{for
a
24
hours}

running average. This is significantly higher than the 11 µg/m3_{limit
for
mean
odour
that
the
world}

health organisation has set. Research has shown that terrain and wind contribute to how the hydrogen sulphide disperses from the power plant (Olafsdottir, Gardarsson, & Andradottir, 2014). In 2011, there was a 140% increase of sulphur pollution in Reykjavík and the suspect is Hellisheiði power plant. The power plant emitted 6,96 g/kWh of hydrogen sulphide that year. Total hydrogen sulphide emissions have been decreasing since the 1970s even though there is more geothermal use (Bayer, Rybach, Blum, & Brauchler, 2013).

(11)

If the wastewater is pumped into them the ecosystem of these streams, rivers or lakes can be seriously affected. There are a few ways to reduce the effects of wastewater on the ecosystem of the streams, rivers or lakes. One is to pump the wastewater into a cooling pond. The reason why this is not used much is that the pond tends to get bigger with time, and also because it is possible that the pond itself will pollute the environment. Another way is to reinject the wastewater into the holes. This will conserve a large amount of the energy that otherwise would be lost (Kristmansdóttir & Ármannsson, 2003). Orkuveita Reykjavíkur wants to reuse the wastewater. In cold climates like Iceland, the wastewater can be used for house heating, and after the house heating it can still be warm enough for snow melting and ground heating. In warm climates it is possible to use heat pumps to cool the air. This is an environmental friendly way to deal with the heat; by using the heat further, the heat emissions to the environment from the power plant will be less (Kristmansdóttir & Ármannsson, 2003; Bayer, Rybach, Blum, & Brauchler, 2013).

Physical effects of fluid withdrawal

If the fluid withdrawal in a geothermal power plant is more than the natural inflow of water then there will be subsidence. This phenomenon is recorded in most geothermal power plants. How much the subsidence is varies between geothermal power plants. At the Svartsengi geothermal power plant in Iceland it is 10 mm/year, but at the Wairakei power plant in New Zealand it is 400-‐450 mm/year. Over the next years there is going to be a very noticeable difference. The subsidence can cause damages on the infrastructure of the geothermal power plant itself. There can be some positive effects from subsidence, for example it can lead to local wetlands and therefore new habitats. By withdrawing water it is also possible to draw so much that the ground water table lowers. That can cause contamination of the aquifer water with corrosive water. Another possibility is that springs and/or fumaroles can vanish. But the most dangerous effect of subsidence is when it forms a steam pillow that can lead to a major explosion. In the past such an explosion has killed people. It is possible to compensate at least to some extent for the loss of water by reinjecting the spent water into the reservoir. The reinjection can cause other problems though (Kristmansdóttir & Ármannsson, 2003; Bayer, Rybach, Blum, & Brauchler, 2013).

Materials and methods

The separated water and the seal water from the plant were mixed into a tank that mimics pipelines. There were 4 different sites on the tank that were connected to the experimental container with pipes so that it was easier to get to the water from the tank for testing. These sites where placed before the tank, close to the beginning of the tank, around the middle and close to the end of the tank. The water is then tested on-‐site for silicon monomer, hydrogen sulphide, total carbonate, ferrous iron and total iron. The last two where only done once. This is done on-‐site because most of the concentrations change with time.

(12)

Figure 4. Experimental setup. The separated water and the seal water are mixed and lead through the tank.

Silicon monomer

The silicon monomer has a tendency to form polymeric silicon when the monomeric silicon is above the concentration limit for amorphous silica. The polymeric silicon is not detected by this method. First 10ml of deionised water was put into a plastic bottle and 0,25 ml of 20% sulphuric acid (H2SO4)

was added to the water to make the sample acidic. 0,25 ml of the sample or standard was then added. One drop of 0,1 N iodine (I) was added to react with the hydrogen sulphide (H2S) that was in

the sample. This was done so that the hydrogen sulphides in the sample did not interfere with the measurements of the silicon monomer. 2-‐3 drops of 0,05 N sodium thiosulfate pentahydrate (Na2S2O3 ∙ 5 H20) were added to react with the rest of the iodine that did not react with the hydrogen

sulphide. 1,25 ml of 10% ammonium molybdate ((NH4)6Mo7O24 ∙ 4H2O) was added to react with the

silicon monomer. After preparation, the sample needed to wait for 10 minutes before it was measured in a spectrophotometer at 410nm. The results for the standards were than plotted into a graf that showed their concentrations against the absorbance and from the plot it was possible to read the concentrations for the samples (Gunnarsson, 2013).

Hydrogen sulphide

First 5 ml of 5 M sodium hydroxide solution (NaOH) was put into an Erlenmeyer flask. This is done to make the solution strongly alkaline. There are at least three reasons why the solution is made such. Reason number one is the colouring for the titration endpoint. In a strong alkaline solution, the colour with dithizone is yellow and it makes a sharp pink endpoint. Reason number two is that mercury sulphite (HgS) precipitates in a strongly alkaline solution. The third reason is that the hydrogen sulphide will not evaporate during the experiment. 5ml of acetone ((CH3)2CO) was added;

this was so that the dithizone would dissolve. Dithizone is not a water-‐soluble material but when it has been dissolved in acetone the acetone is water-‐soluble. The sample was then added. The sample size was not constant; the more hydrogen sulphide there was in the sample the smaller sample size was taken. When there was a large amount of hydrogen sulphate in the sample and the sample size was too big, the change in colour was more gradual without a sharp endpoint as was needed. Deionised water was then added to around 50 ml. This is to make it easier to see the end point. The dithizone was added next. This gave the colour change. This sample was than titrated with 0,001 M mercury acetate (Hg(CH3COO)2) to a pink endpoint. The pink endpoint comes from the fact that when

(13)

𝐻!𝑆 𝑝𝑝𝑚 =

𝑚𝑙 0,001𝑀 𝐻𝑔 𝐶𝐻!𝐶𝑂𝑂 !∙ 34

𝑚𝑙 𝑠𝑎𝑚𝑝𝑙𝑒

Equation 6. For calculating the concentration of hydrogen sulphide

Total carbonate

There are two ways that were done to some extent to find out the total carbonate. The first one was always used, but the second one was also used to some extent. In the first method, the sample was weighted and the pH was accurately adjusted to 8,30 with hydrochloride acid (HCl) or sodium hydroxide (NaOH). If the pH was higher than 8,30 hydrochloride acid was used to lower the pH level and if the pH was lower than 8,30 sodium hydroxide was used to increase the pH level. The next step was to use hydrochloric acid to lower the pH level to 3,80. The amount of hydrochloride acid was documented and based on this information the carbonate was calculated with equation 2.

𝐶𝑂!=

𝑚𝑙 ℎ𝑦𝑑𝑟𝑜𝑐ℎ𝑙𝑜𝑟𝑖𝑑𝑒 𝑎𝑐𝑖𝑑 ∙ 4400

𝑔 𝑠𝑎𝑚𝑝𝑙𝑒 − 6,97 + 1,182 ∙ 𝑝𝑝𝑚𝐻!𝑆 + 0,0088 ∙ 𝑝𝑝𝑚𝑆𝑖𝑂!+ 0,100 ∙ 𝑝𝑝𝑚𝐵

Equation 7. For calculating the concentration of carbon dioxide by using theoretical calculations.

To get an accurate value for carbonate the equation takes into account that the hydrochloride acid reacts with the hydrogen sulphide, the total silicon dioxide, the boron and to some extent the water in the sample. The second method starts the same as the first, the sample is weighted and pH was adjusted to 8,30. The hydrochloride acid was then used to lower the pH to 3,80, the air was then bubbled through the sample for 10 minutes. By doing this some of the interfering compounds get bubbled away with the air. The last step was that the sample is titrated from 3,80 to 8,30 with sodium hydroxide. In this case the amount for both hydrochloride acid and sodium hydroxide was documented. It is then possible to calculate the carbonate with equation 3 (Arnórsson, 2000).

𝐶𝑂!=

𝑚𝑙 ℎ𝑦𝑑𝑟𝑜𝑐ℎ𝑙𝑜𝑟𝑖𝑑𝑒 𝑎𝑐𝑖𝑑 − 𝑚𝑙 𝑠𝑜𝑑𝑖𝑢𝑚 ℎ𝑦𝑑𝑟𝑜𝑥𝑖𝑑𝑒 ∙ 4400

𝑔 𝑠𝑎𝑚𝑝𝑙𝑒 − 1,182 ∙ 𝑝𝑝𝑚𝐻!𝑆

Equation 8. For calculation the concentration of carbon dioxide by using acid and base.

Total silica

To be able to find the amount of silica the sample size of 0,5 ml was chosen for both the standard or the sample. The sample should contain ca 0,1 – 0,5 mg of silicon dioxide (SiO2). Iodized water was

added so that the total volume was 5ml. To make the sample acidic, 0,5 ml of 20% sulphuric acid is then added. To that, 1 ml of 1,5 M Hydrofluoric acid (HF) was also added. This mixture was heated in a 60°C hot water bath for ca 40 minutes so that the hydrofluoric acid could react with the silicon dioxide to form silicon tetrafluoride (SiF4). When this was done the samples were cooled down to

room temperature. Next 2,5 ml of 0,25 M aluminium sulphate (Al2(SO4)3) was added and the samples

were then heated again for ca 20 minutes and cooled again to room temperature. Next 15 ml of deionised water was added and the sample was shaken to mix it well. 2,5 ml of 10% ammonium molybdate solution was then added to the mixture. After preparation of a sample, it needed to wait for 15 minutes before it was measured in a spectrophotometer at 410 nm. The results for the standards were than plotted into a graf that showed their concentrations against the absorbance and from the plot it was possible to read the concentrations for the samples (Gunnarsson, 2014).

Total iron

The standards were made by mixing iron (III) chloride (FeCl3) in hydrochloric acid (HCl) and diluting it

to get 10-‐500 ppb of iron (Fe). Standards were not done for a lower concentration because the spectrophotometer was not accurate for such a low concentration.

The procedure was to start with about 45 ml of a filtered sample or standard in a 50ml brown glass bottle. 0,06 ml of 6M hydrochloric acid, 0,10 ml 2 M Sodium hydroxide (NaOH) and 1 ml of 0,40 M ascorbic acid (C H O) solution were added to the filtered sample. The ascorbic acid solution was

(14)

least 30 s. 1 ml of TPTZ (2,4,6-‐tripyridyl-‐1,3,4-‐triazine) solution was then added; this was done to allow ferrous ion (Fe+2_{)
to
form
a
violet
complex
with
the
TPTZ.
The
ammonium
acetate
(NH}

4C2H3O2)

solution was added to the mixture; this was done to prevent reduction of ferric to ferrous iron after the preparation. The mixture was then mixed well and the bottle filled up with the filtered sample or the standard solution, the same standard of sample that was put in it to begin with. This mixture was then shaken well. The last thing was to measure the samples and the standards in the spectrophotometer at 595 nm. The results for the standards were then plotted up into a straight line. From the equation of the line it was possible to calculate the concentrations in the unknown samples (Arnórsson, 2000).

Ferrous iron Fe

+2

The standards that are used are the same as for the total iron. The procedure starts by taking about 45 ml of ether standard or unknown sample and putting it in a 50 ml brown glass bottle. 0,06 ml of 6 M hydrochloric acid solution, 0,10 ml of 2 M Sodium hydroxide solution and 1ml of TPTZ are also added to the solution. The ferrous iron forms a violet complex with the TPTZ if the pH of the solution is somewhere between 3,5 and 5,8. Next, 1 ml of ammonium acetate solution is added and the mixture is shaken well. The bottle is then filled up with an unknown sample or a standard solution. This is then mixed well and the samples are measured in a spectrophotometer at 595 nm. The results for the standard were then plotted into a straight line. By finding the equation of the line it is possible to calculate the concentration for the unknown samples (Arnórsson, 2000).

Results and discussion

Results with the separated water at 70°C

Table 1 shows that the concentrations of the minerals that are in the seal water and the separated water. Silica (Si) and chlorine (Cl) have a high concentration in the separated water. It also shows that there is almost no silica (Si) or chlorine (Cl) in the seal water. The fact that there is such low concentration of silica (Si) in the seal water is why it works as a diluter and that the seal water also has a low concentration of chlorine (Cl). Chlorine is used because it has shown almost no tendency to precipitate in the separated water. It is therefore possible to say that the chlorine is stable and use the concentrations of chlorine when checking if there is precipitation of amorphous silica onto a surface.

Particles Seal water

Separated water Si [ppm] 0,21 682 Na [ppm] 0,22 177 K [ppm] 0,4 31,4 Ca [ppm] 0,018 0,62 Mg [ppm] 0,002 0,009 Fe [ppm] 0,023 0,014 Al [ppm] 0,013 1,8 Cl [ppm] 0,55 152,4 B [ppm] 0,025 1,06 Li [ppm] 0,01 0,19

(15)

respective valves at the tank. When the total Silica gets lower through the tank it is because of precipitation of amorphous silica onto the surface. When the amorphous silica gets lower, is because of both of the processes, the precipitation and the polymerization. The red dots show the precipitation and the polymerisation of the amorphous silica at their respective valves. The amorphous silica falls faster than the total silica, therefore it is known that the amorphous silica is polymerizing. The total silica drop in the experiments that were done with only separated water at 70°C with different flows showed a drop somewhere between 8-‐16 ppm. This means that the silica scaling will be somewhere between 150-‐300 tons per year. This silica could then be clogging the injection wells as well as clogging the pipes in Hellisheiði power plant.

Figure 5. Amorphous silica scaling and polymerization with separated water at 70°C

When the seal water is mixed with the separated water it is diluting the separated water. Because there is no clorine in the seal water, it is possible to compare the silica consentrations with the clorine concentrations. Through this comparison, it is possible to see if the silica concentrations for valves S1, S2 and S3 (see Equation 9 ) are being diluted. Based on this information, it is possible to construct lines like those in Figure 6. The dots are the concentration of total silica in the respective valves. The concentrations of chlorine and silica should lower after the mixing but only as much as the dilution allows it to and no further. If the concentrations of silica are lower than the concentration calculated for the silica, it indicates that the amorphous silica has precipitated on to the surface. As can be seen by looking at Figure 6, there is no precipitation when mixing the seal water and the separated water together at 70°C for the separated water.

𝑆𝑖 != _𝐶𝑙𝐶𝑙! !! 𝑆𝑖 !!

Equation 9. Calculation of concentration of silica by using the chlorine concentration.

Where [Si] stands for silica concentration, [Cl] stands for chlorine, S0 stands for the experimental valve before the tank and i stands for S1, S2 or S3 which indicate the experimental valves.

(16)

Figure 6 Total silica in the separated water at 70°C and the mixture with seal water.

There is a limit on how much it is possible to lower the pH of the mixture. If pH is lowered too much the pipes will start to corrode. The corrosion threshold for the pipes is considered to be at pH six. When the pH is lower than six the iron concentration gets exponentially higher according to Figure 7.

Figure 7 the iron in the separated water at 70°C and in the mixture with seal water. The red line is the corrosion threshold for the pipes

By plotting the pH as the function of the mixing ratio (Figure 8) it is possible to see that mixing the seal water and the separated water at 70°C to around 50/50 would be without too much corrosion in the pipes.

(17)

Figure 8 the pH as function of the mixing ratio for separated water at 70°C. The red line is the corrosion threshold for the pipes.

Results with separated water at 120°C

The mineral composition for the seal water and the separated water at 120°C can be seen in Table 2. The Silica and the chlorine have low concentrations in the seal water and high concentrations in the separated water like it is for the separated water at 70°C.

Particles Seal water Separated water

Si [ppm] 1,21 710 Na [ppm] 0,54 182 K [ppm] 0,45 33,6 Ca [ppm] 0,038 0,573 Mg [ppm] 0,00719 0,00615 Fe [ppm] 0,0562 0,0538 Al [ppm] 0,01241 1,42 Cl [ppm] 0,2 164 B [ppm] 0,0042 1,04 Li [ppm] 0,01 0,199

Table 2. The mineral composition in the seal water and the separated water, with the separated water at 120°C

When looking at Figure 9, the amorphous silica concentration gets higher than the concentration for the total silica. The total silica should be higher or at the same level. There are two possible explanations for this. One is that while doing the experiments at different velocities for the separated water, there was some boiling in the tank. That is why there is only one experiment done with high heat separated water only; the plan was to experiment with few velocities. It is possible that the heat in that experiment was too high without the water actually being partly in gas form. Another explanation is that there was some mistake done when doing the experiments for the amorphous silica. If that is the case then it should be safe to assume that the total silica is equal to the amorphous silica. In both of these cases it should be safe to look at the total silica concentration and see that the total silica concentration is going down, which indicates that there is amorphous silica scaling in the tank.

(18)

Figure 9 Amorphous silica scaling with separated water at 120°C

Figure 10 is done the same way as Figure 6, the only difference is that the separated water is at 120°C. The total silica concentration is found to be in line with the calculated total silica. This means that there is no precipitation onto a surface when the separated water is at 120°C.

Figure 10 Total silica in the separated water at 120°C and the mixture with seal water.

By looking at Figure 11 it is possible to see that the concentration of iron gets higher the closer to pH six it gets. It is not as obvious that it is an exponential curve as it was for the separated water at 70°C, but it is still possible to see the increase in the concentration for the iron the closer it gets to pH six.

Experiments with precipitation of silica from waste water at Hellisheiðarvirkjun