Waste Management 22 (2002) 747–753

www.elsevier.com/locate/wasman

Glass recycling in cement production—an innovative approach

Guohua Chena,*, Harry Leea, King Lun Younga, Po Lock Yuea,
Adolf Wongb, Thomas Taob, Ka Keung Choib
a
Department of Chemical Engineering, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, China
b
Green Island Cement (Holdings) Limited, No. 7 Lung Yiu St., Tap Shek Kok, Tuen Mun, N.T. Hong Kong, China

Accepted 14 February 2002

Abstract
An innovative approach of using waste glass in cement production was proposed and tested in a laboratory and cement pro-
duction plant. The laboratory characterization of 32 types of glass show that the chemical composition of glass does not vary sig-
nificantly with its color or origin but depends on its application. The alkali content of glass, a major concern for cement production
varies from 0 to 22%. For the glass bottles mainly found in Hong Kong waste glasses, the alkali content (Na2O) ranges from 10 to
19% with an average around 15%. There is no significant change of the SO2 content in the gas exhaust of the rotary kiln when
about 1.8 t/h of glass bottles were loaded along with the 280–290 t/h raw materials. The content of NOx, mainly depends on the
temperature of the kiln, does not show significant change either. The SO3 content of the clinker is comparable with that obtained
without the loading of glass. The alkaline content shows a slight increase but still within three times the standard deviation obtained
from the statistical data of the past year. The detailed analysis of the quality of the cement product shows that there is not any
significant impact of glass for the feeding rate tested. # 2002 Elsevier Science Ltd. All rights reserved.
Keywords: Alkali; Nitrogen oxides; Rotary kiln; Solid waste; Sulfur oxides

1. Introduction aggregate substitute in concrete, road beds, pavement

and parking lots, as well as drainage medium, backfill or
In Hong Kong, there are 44,000 t of waste glass gen- landscaping purposes; or as raw materials to produce
erated from domestic sources per yr. In addition, there glass pellets or beads used in reflective paint for highways,
are 20,000 t of waste glass annually generated from to produce fiberglass; frictionators for lighting matches
commercial sources. Of the total 64,000 t of waste glass, and firing ammunition. Glass powders can be used as
only about 8000 t are recycled and reused although lubricants, core additives and fluxes in metal foundry
about 50,000 t of waste glass are technically recoverable. work fabrication, as flux/finders in the ceramics industry,
This waste glass is currently disposed of at the landfills in peening and cleaning metals, or simply as decorative
which are being filled much faster than expected. If the materials (Glass Packaging Institute). Due to the limited
current trend of waste generation and disposal continues, industrial bases in Hong Kong, all these beneficial usage
our landfills will be exhausted by 2015 (Environmental of waste glass seems impractical. One innovative idea of
Protection Department). reusing the waste glass in the production of cement was
Some of the used glass bottles can be reused after brought up by Green Island Cement (Holdings) Ltd (GIC).
thorough cleaning as currently practised by some local One main concern for using waste glass in cement
breweries (Environmental Protection Department). The production is the alkali content (Na2O 0–22% and K2O
refillable portion of the recycled glass takes about 12% 0–6%) in glass (Table 1; Wheeler, 1958), which may
in the USA. Other ways of recycling waste glass are to result in alkali-aggregate reaction (AAR) or more speci-
use them as cullet in glass production; as raw material fically the alkali-silica(te) reaction (ASR) and alkali-car-
for the production of abrasives as in sand-blasting; as bonate reaction (ACR) (Mebta, 1978). These reactions
may lead to cracking of concrete. Therefore, the top
* Corresponding author. Tel.: +852-2358-7138; fax: +852-2358-
quality cement requires the alkali content to be below
0054. 0.6 wt.% (Mebta, 1978). Currently, the alkali content
E-mail address: [email protected] (G. Chen). of cement produced by GIC is less than 0.3 wt.%.
0956-053X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0956-053X(02)00047-8
748 G. Chen et al. / Waste Management 22 (2002) 747–753

Table 1
Chemical composition of some typical commercial glasses

Glassa

1 2 3 4 5 6 7 8 9 10 11

SiO2 80.5 73.0 77.9 71 73.5 61.5 72.68 63.0 72.0 57.3 75.3 68.0
Na2O 3.8 – 4.1 14 17 8.8 13.25 7.6 11.0 1.9 5.7 22.0
K2O 0.4 4.5 2.3 0 1.5 5.3 – 6.0 – 0.5 0.8 –
PbO – 6.0 – – 23.6 – 21.0 – – – –
Al2O3 2.2 – 0.8 – – – – 5.0 3.2 6.2 3.0
B2O3 12.9 16.5 14.1 – – – – 12.0 23.6 7.6 –
CaO – – 0.8 5 6 – 12.76 0.3 – 4.7 1.1 7.0
MgO – – – 3.5 4.5 – 0.26 0.2 – 8.7 – –
Fe2O3 – – – 0.01 0.8 – – – – – 3.0
MnO – – – – – – 0.2 – – – –
Others – – – – – – As2O3 – – BaO –
0.05 3.5
a
1, Pyrex 7740; 2, Pyrex 7720 (Nonex); 3, Silex; 4, modern lime bulb glass; 5, lead glass; 6, flat drawn window glass (Libby-Owens, American); 7,
American lead bottle glass; 8, borosilicate thermometer glass; 9, Jena Supermax (German); 10, Jena Gerate (German); 11, Christmas tree ornament glass.

Therefore the company can afford to increase its alkali 5. The dried samples were ground into fine powders
content by 0.2 wt.% which means over 3200 t of alkali, with a mortar and pestle followed by a screening
calculated based on an annual cement production of 1.6 using a sieve tower with an aperture of 250 mm.
million t. Secondly, an equilibrium between alkali and 6. The fine powder was finally transferred into
sulphur trioxide inside kiln atmosphere has to be main- labeled test tubes and sealed with plastic caps. It
tained to attain a stable operation. Before implementing was ready for analysis.
this innovative idea, there are two issues that have to be
addressed: the alkali contents in the waste glass and the The chemical composition of different types of glasses
quality of the cement produced. Hence a research team were analyzed using the surface-scanning technique of
was formed between Hong Kong University of Science X-ray photoelectron spectroscopy (XPS). In XPS,
and Technology and GIC first to carry out the char- monochromatic X-ray beams are concentrated onto the
acterisation of the typical waste glasses generated in sample surface where the inner K or L shell electrons
Hong Kong and then to conduct the plant trial. This are excited. By locating the peaks of energies in the XPS
paper reports the results obtained. spectrum, the existence and atomic percentage of a
respective element are deduced (Skoog, 1985).
Before starting to analyze the chemical components of
2. Glass characterisation various glasses, the results obtained using the XPS
method are compared with the data shown in Table 1.
The waste glass characterization was carried out in The good comparison of the literature data and those
the following procedures: obtained by the XPS method are given in Table 2. The
slight variations are possibly associated with different
1. Different types of glasses were collected and sources of glass.
labeled with descriptions including the materi- After verifying the analytical technique, various glas-
als contained, names and industrial codes (if ses were analyzed with the description shown in Table 3.
any). As it shows that the majority of glasses used in Hong
2. All contents of the bottles were disposed of or Kong is from the packaging, more samples were taken
stored in other containers. Then the bottles were from this category. Since color usually was taken as a
cleaned with detergent and DI water. primary parameter for the reprocessing of glass, its
3. For bottles with narrow openings, the mouths of relation with the composition is analyzed along with the
the bottles were cut off by a diamond saw. For glass application and origin.
other types of glass and bottles, pieces of glass The chemical composition of all the glasses analyzed
were obtained by hitting the totally wrapped is listed in Table 4. For packaging glasses, their alkali
glass with a hammer. contents are high (around 10% or higher) and the var-
4. The glass samples were washed again together iation is quite small (8–14%). The sodium oxide con-
with the mortar and pestle, and dried in an oven tents range from 10 to 19 wt.%. The average alkali
at 100  C for about 20 min. oxide content is found to be 15%, with a standard
 G. Chen et al. / Waste Management 22 (2002) 747–753 749

Table 2
Comparison between literature data and analyzed results

Category Source O (%) Si (%) Na (%) K (%) Al (%) B (%) Ca (%) Mg (%)

Pyrex Literature 53.3 40.3 3.2 0.0 0.7 2.6 0.0 0.0
XPS 50.0 36.3 3.7 0.3 1.0 2.0 0.0 0.0
Light Literature 45.8 32.7 12.1 1.2 0.0 0.0 5.1 3.0
Bulb XPS 40.0 30.5 12.0 0.0 1.0 0.0 2.5 1.0
Window Literature 45.3 32.0 9.3 0.0 0.3 0.0 12.9 0.2
Glass XPS 41.0 30.0 12.0 0.0 1.0 0.0 3.0 2.0

Table 3
Description of the glass samples analyzed

Sample no. Category Company Description Location Color

1 Packaging Crystal Blackberry flavour mures, 355 ml USA Blue

2 Home made beer bottle Brown
3 San Miguel Beer, 640 ml Hong Kong Brown
4 Bargetto Winery, Soquel Chardonnay, 1994 USA Brown
5 Del Monte Pure juice, 946 ml USA Brown
6 Tsingtao Beer, 640 ml China Green
7 Appletiser Sparkling white grape juice, 200 ml South Africa Green
8 Spring Valley Apple juice, 250 ml Australia No color
9 Nescafe Instant coffee, classic, 50 g China No color
10 Sunquick Orange drink concentrate, 840 ml China No color
11 Blue girl Beer, 640 ml Germany No color
12 Coca Cola Coke, 192 ml Hong Kong No color
13 Ribena Blackcurrant drink, 250 ml Malaysia No color
14 Gatrorade Lemon, 500 ml Philippines No color
15 Lucozade Orange barley, 1 l Singapore No color
16 Snapple Kiwi strawberry cocktail, 473 ml USA No color
17 Welch’s White grape juice, 355 ml USA No color
18 Park’n shop Strawberry preserve, 340 g Hong Kong No color
19 Park’n shop Pickled cucumber, 230 g Taiwan No color
20 Lea & Perrins Worcestershire sauce, 142 ml England No color
21 Lee Kum Kee Premium soy sauce, 750 ml Hong Kong No color
22 Park’n shop Tomato ketchup, 340 g Malaysia No color
23 Domestic Ordinary drinking glass No color
24 Light bulb, 25 W No color
25 Thorn Light tube England No color
26 Philips Pearl matt light bulb, 40 W Holland No color
27 Iwaki Pyrex microwave glass container Japan No color
28 Philips Super VGA monitor (screen) Taiwan No color
29 Industrial Insulation fiber glass Light brown
30 Impactex Safety glass (BS 6202 AT) England No color
31 Laboratory Borosilica Light brown
32 Pyrex Test tube No color

deviation of 2%. Since packaging glass is expected to be usually Pyrex Borosilicate, the alkali metals had been
the majority of the recycled glass, the numbers obtained extracted out during their manufacturing. Although the
here can serve as good reference values for loading glass recycled from laboratory glass is more attractive to
design in the cement production. the cement production, the treatment of such glass is
The glass used in the domestic light bulb or tubes currently well established as chemical waste.
resemble those for packaging glass in terms of alkali Out of the 32 samples tested, nine different colors
contents, while the glass for microwave containers and were observed. Statistically, there is not a significant
computer monitors show distinctively low alkali con- difference found for the chemical compositions between
tents with sodium content being 5 and 7 wt.%, respec- transparent and colored glasses although color is an
tively. Insulation fibre shows the least sodium content, 1 important factor in the reprocessing of glass. Such a
wt.%. Laboratory glass also has low alkali content, 3 finding is not surprising because the coloring of glass
wt.%, which is expected because for these glasses, only requires a minute amount of metal oxides and
750 G. Chen et al. / Waste Management 22 (2002) 747–753

Table 4
Chemical composition of glass samples, by weight percent (%)

Sample no. O Si Na C Ca Al Mg K B F Cl Sr Zr Ba

1 Packaging 40 31 11 14 3 1 0 0 0 0 0
2 49 34 8 7 0 1 1 1 0 0 0
3 39 29 10 13 6 1 1 0 0 0 0
4 41 30 12 10 4 1 0 0 0 0 0
5 43 32 12 8 3 1 1 0 0 0 0
6 41 31 11 8 3 2 1 1 0 0 0
7 40 31 11 12 5 1 0 0 0 0 0
8 42 30 14 8 5 1 0 0 0 0 0
9 39 29 12 12 4 1 2 0 0 0 0
10 43 32 10 10 3 1 1 0 0 0 0
11 44 33 10 10 3 1 0 0 0 0 0
12 50 33 9 6 0 1 0 1 0 0 0
13 42 33 10 11 2 1 0 0 0 0 0
14 43 30 13 8 5 1 0 0 0 0 0
15 43 31 13 7 5 1 0 0 0 0 0
16 44 33 11 7 4 1 0 0 0 0 0
17 44 32 13 6 4 1 0 0 0 0 0
18 43 31 13 7 5 1 0 0 0 0 0
19 40 30 13 9 5 1 1 1 0 0 0
20 40 32 10 13 3 1 0 0 0 0 0
21 42 32 11 10 2 1 1 0 0 0 0
22 44 34 10 8 3 1 0 0 0 0 0
23 Domestic 40 30 13 12 3 1 0 0 0 0 0
24 40 30 13 12 2 1 2 0 0 0 0
25 42 29 15 7 3 2 2 0 0 0 0
26 40 31 11 14 3 1 0 0 0 0 0
27 52 36 5 6 0 1 0 0 0 0 0
28 39 28 7 9 1 1 0 5 0 0 0 4 1 4
29 Industrial 41 20 1 23 9 5 0 0 0 0 0
30 41 30 12 12 3 1 2 0 0 0 0
31 Laboratory 49 37 3 5 0 1 0 1 3 0 0
32 49 36 3 6 0 1 0 0 3 0 0

these metals are not alkali but nickel, cobalt, copper or Table 5
even iron. Besides the color, the origin of the glass does Sodium oxide content in different glasses (%)
not show any significant difference in the chemical Insulating fiber glass 1.5
composition, either. Laboratory test tube 4
Since in the cement industry the alkali content means Microwave glass plate 6.5
the alkali oxide content, it is more proper to convert the TV, computer monitor 9.5
Coca Cola bottles 12
alkali into alkali oxide. It should be noted that the
Wine bottles 12–17.5
number of moles of oxygen in glass is insufficient to Sauce bottles 12–17.5
convert all metals and silicon to their respective oxides. Soft-drink bottles 12–19
Some of these elements have to exist in their elemental Light bulbs 15
forms. We assume that all of the alkali content exists in Light tubes 20
its oxide form, i.e. Na2O and K2O. The average content
of sodium oxide is about 15 wt.% while that of potas-
sium oxide is still less than 1 wt.%. In cement terminol- 3. Comparison of class composition with cement
ogy, sodium oxide equivalent (Na2O)e or total alkali
content is often used, where (Na2O)e is equal to A detailed comparison between compositions of glass
0.658%K2O+%Na2O (Hewlett, 1998). Thus it is safe and Portland cement is given in Table 6. The glass
to say here that (Na2O)e is equivalent to average sodium composition is the averaged value of the data shown in
oxide content since the K2O is comparatively very low. Table 4 and the compositions of Portland Cement is
The sodium oxide contents of typical glasses are sum- from the American Standard of Test and Methods
marized in Table 5. It is remarkable to know that the (Annual Book of ASTM Standards). A simple compar-
sodium oxide content of the bottles found in Hong ison of the glass composition and that of Portland cement
Kong is significantly higher than that reported for the shows a significant difference. Beside alkali oxides, glass
American lead bottles as listed in Table 1 of class 7. contains approximately 67 wt.% of silica (SiO2) and
 G. Chen et al. / Waste Management 22 (2002) 747–753 751

Table 6
Comparison of components of glass and Portland cement

Element Glass Portland Cement plus

average cement recycled glass
(wt.%) average (wt.%) average (wt.%)

Oxygen (O) 42.59 37.53 37.60

Silicon (Si) 31.51 7.90 8.21
Sodium (Na) 11.22 0.15 0.30
Carbon (C) 9.28 0.20 0.32
Calcium (Ca) 3.43 48.82 48.22
Aluminum (Al) 1.10 3.70 3.67
Magnesium (Mg) 0.53 0.55 0.55
Potassium (K) 0.26 1.15 1.14
Boron (B) 0.00 0.00 0.00
Fluorine (F) 0.02 0.00 0.00

18% of other elements. For Portland cement, the major

Fig. 1. Flow chart of cement production line.
component is calcium silicates. The difference implies
that the chemical composition of the cement would be
more or less altered. However, this change is insignif- bottles passed through the orifice and entered the rotary
icant, as shown in the last column of Table 6. The kiln where they were mixed into the normal kiln feed
composition of cement produced with the addition of that consisted of roughly limestone (80%), flyash (9%),
glass was calculated based on the annual production silica sand (9%), and iron/slag (2%). The kiln feed and
capacity of 1.6 million t of cement and the load of crushed bottles were then heated up to a temperature of
21,000 t of waste glass with the average composition. approximately 1450  C on their 30 min journey through
The amount of glass that can be handled by GIC is the rotary kiln. A semi-product clinker was then formed
calculated based on a 0.2% margin of alkali content to and discharged from the kiln into the grate cooler for
increase and an alkali oxide content of 15% in the waste cooling. Lower alkaline kiln feed was prepared before
glass. the trial.
The data listed in Table 6 confirmed our expectation From the estimation made in the previous section,
that waste glass can be used as one of the raw materials GIC can handle over 2.6 t/h at an 8000 h/year schedule.
in cement production. It gives confidence to GIC to Two plant trials were conducted. For the sake of safety,
go ahead with the plant trial. The results are shown the first plant trial was set to have the glass feeding rate
subsequently. at 1 t/h for a period of 24 h. The glass was obtained
from a local glass collector. Due to the tight schedule of
the plant trial, the glass used in the experiment was
4. Plant test results mainly washed beer bottles packed in nylon bags. A
temporary hoisting and conveying system was set up at
There are at least two methods, which GIC can use to GIC to feed the bottles into the system. The total kiln
absorb waste glass into the cement manufacturing pro- feed rates (excluding glass bottles) were maintained at
cess. Waste glass can either be fed as raw material to 280–290 t/h for the trial period. It should be noted that
produce clinker (semi-product of cement) or be ground GIC specifically arranged a shipment of lower alkaline
with clinker to produce cement as one of the fillers. The silica sand to be used during the trial in order to main-
former option was tested with the belief that waste glass tain the alkaline/sulfur balance. The second trial was
would have less adverse impact on the cement quality carried out mainly to test the quality of the cement
this way. The high temperature in the kiln would also product. About 99 t of waste glass was fed to the pro-
sterilize any virus and destroy any carcinogenic pollu- duction line over 56 h operation with the average feed-
tants contained in the glass. ing rate of glass being 1.77 t/h. The raw materials
feeding for the second trial was as usual, i.e. without
4.1. Process description specially lowering alkali content.

Fig. 1 shows the flow chart of the cement production 4.2. Parameters analyzed
line employed by GIC. Bottles were fed into the mixing
chamber and fell down a distance of approximately 15 Two locations were selected to monitor the environ-
m to crush on the kiln orifice area. The temperature in mental impact, one at the gas exhaust before the elec-
the kiln orifice area was about 1100  C. The crushed trostatic precipitator (EP) and the other in the stack.
752 G. Chen et al. / Waste Management 22 (2002) 747–753

Table 7
The analytical results of the gas samples before EP

Baseline During trial Post

29 January 2000 29 January 2000 31 January 2000 31 January 2000 1 February 2000
a.m. p.m. a.m. p.m. p.m.

Pumping rate l/min 0.87 0.88 0.87 0.87 0.83

T ( C) 410.5 422.5 414.7 418.7 421
CO (ppm) 423 285 424 389 208
SO2 (ppm) 1.37 1.15 0.34 1.56 0.25
NO2 (ppm) 1.8 0.97 0.97 0.98 0.98
NO (ppm) 448 378 456 480 372
CO2 (%) 28.3 25.9 28.6 27.9 25.6
O2 (%) 4.1 4.7 4.2 3.7 5.4
NOx (ppm) 450 379 457 481 373

SO2 level was of key interest at the first location while mean before the trial started. Although the data pre-
NOx, SO2, O2 and OPAC levels were all important at sented in Tables 7–9 are for the first test, similar results
the second location. The SO3 and alkali content of kiln were obtained for the second test.
feed, third-stage material, fourth-stage material, EP The averaged sodium oxide content of all materials
dust and clinker were analyzed at an hourly interval were larger than the historical mean but the potassium
during the glass burning period. Comparison was made oxide content remains unchanged. Consequently, the
on chemical and physical parameters of cement with equivalent (Na2O)e or the total alkali content increased
and without waste glass feedings. noticeably (Table 10). The total alkali content of all
materials during glass burning period, however, was
4.3. Results and discussion also within  3 S.D. obtained from the data of 30 days
randomly chosen from the past year. It is useful to note
The monitoring stations show that there is no sig- that an increase in the sulphur supply is required in
nificant change in the NOx, SO2 and OPAC levels dur- order for the smooth operation of the kiln system.
ing the glass addition in comparison with the results No material built up due to increased alkaline content
obtained at normal working conditions as shown in could be developed or observed. Kiln operation and
Tables 7 and 8. The mean values in Table 8 were process control were normal during and after the trial
obtained from hourly emission monitoring for 21 h. The burn. No obvious change was found on clinker appear-
slight increase of SO2 content over that before the trial ance. The physical and chemical properties of cement
is probably due to the high sulphur content fuels used in using clinkers produced before glass feeding and during
order to balance the expected increase of alkali content. glass feeding were tested and shown in Table 11. The
This also explains why after the trial the SO2 content is difference between the values of alkali in Tables 10 and
even higher. However, the concentrations of the pollu- 11 resulted from the addition of fillers in the manu-
tants are well below the emission standards. facturing of cement. Table 11 clearly shows that there is
As for the processing material analysis, Table 9 shows no difference between the cement produced with and
that the SO3 of all materials during glass burning period without glass feeding both chemically and physically.
were within  3 times the standard deviation (S.D.)
obtained from the data of 29 days randomly chosen
Table 9
from the previous 7 months before the trial. The mean
The sulphur content in the burned material (% SO3)
values during the trial are averaged from the values
obtained every 6 h. They are comparable to those of the Before burning glass During burning

Mean S.D. Max Min Mean S.D. Max Min

Table 8
The test results of the gases in the stack KF 0.08 0.04 0.14 0.03 0.09 0.02 0.11 0.07
3rd-E 0.26 0.09 0.59 0.17 0.31 0.05 0.4 0.28
Before trial During trial Post trial
3rd-W 0.27 0.08 0.4 0.14 0.31 0.06 0.38 0.24
Mean S.D. Mean S.D. Mean S.D. 3rd-Avg 0.27 0.07 0.55 0.19 0.31 0.05 0.38 0.25
4th-E 0.72 0.14 0.95 0.49 0.82 0.16 1.09 0.68
NOx (mg/Nm3) 625.8 69.6 540.7 48.8 533.2 92.7 4th-W 0.72 0.16 0.98 0.46 0.78 0.1 0.91 0.68
SO2 (mg/Nm3) 0.4 0.3 0.7 0.2 1 0.2 4th-Avg 0.72 0.14 0.91 0.48 0.8 0.12 1 0.7
O2 (%) 10.6 1 10.4 0.1 10.9 0.6 EP 0.11 0.02 0.15 0.07 0.11 0.02 0.13 0.09
OPAC (mg/m3) 27.8 3.2 28.2 2.5 24.3 2.4 CLK 0.39 0.08 0.07 0.15 0.36 0.02 0.38 0.36
 G. Chen et al. / Waste Management 22 (2002) 747–753 753

Table 10
The total alkali content in the burned material (%)

Before glass addition During glass addition

First test Second test

Mean S.D. Max Min Mean S.D. Max Min Mean S.D. Max Mm

Na2O KF 0.10 0.03 0.19 0.05 0.13 0.03 0.19 0.11 0.08 0.01 0.10 0.06
3rd-AVG 0.11 0.03 0.17 0.08 0.15 0.05 0.25 0.12 0.11 0.01 0.13 0.09
4th-AVG 0.16 0.03 0.22 0.12 0.20 0.03 0.25 0.18 0.17 0.02 0.21 0.15
EP 0.08 0.03 0.17 0.04 0.13 0.05 0.22 0.07 0.08 0.01 0.09 0.07
CLK 0.14 0.04 0.25 0.07 0.23 0.02 0.26 0.20 0.24 0.06 0.35 0.18
K2O KF 0.28 0.06 0.45 0.18 0.27 0.02 0.29 0.23 0.24 0.02 0.27 0.21
3rd-AVG 0.75 0.16 0.24 0.50 0.86 0.21 1.20 0.64 0.76 0.08 0.89 0.65
4th-AVG 1.14 0.21 1.67 0.86 1.25 0.14 1.46 1.08 1.14 0.13 1.34 0.92
EP 0.28 0.06 0.42 0.19 0.28 0.07 0.38 0.21 0.31 0.03 0.37 0.27
CLK 0.38 0.06 0.54 0.28 0.31 0.03 0.35 0.28 0.34 0.04 0.42 0.28
(Na2O)e KF 0.28 0.07 0.42 0.20 0.31 0.04 0.37 0.26 0.23 0.02 0.28 0.20
3rd-AVG 0.61 0.12 0.96 0.44 0.72 0.15 0.91 0.55 0.60 0.05 0.68 0.52
4th-AVG 0.92 0.15 1.27 0.71 1.03 0.11 1.17 0.88 0.92 0.09 1.05 0.78
EP 0.27 0.06 0.45 0.19 0.32 0.06 0.38 0.24 0.28 0.02 0.33 0.27
CLK 0.39 0.06 0.53 0.29 0.44 0.03 0.46 0.39 0.46 0.07 0.58 0.37

Table 11 Two plant trials were conducted at a load of 1 t/h for 24 h

Impact on cement quality of glass feeding and 1.77 t/h for 56 h. The exhaust from the stack does not
Normal Cement with Statistical show significant changes in the concentration of pollutants.
cement glass addition, data in The SO3 content in the processing materials varies within
second test September  3 times S.D. obtained from the past year’s statistical
2000
data. The mean value of the SO3 is also comparable with
Na2O (%) 0.10 0.20 0.08–0.14 the historical mean. The cement produced using some glass
K2O (%) 0.32 0.31 0.30–0.42 as feed material does not show any chemical and physical
Total alkali (%) 0.31 0.40 0.28–0.42
SO3 (%) 2.49 2.47 2.25–2.79
differences compared with cement produced without glass
Free lime (%) 1.25 1.20 0.80–1.80 addition. The alkali content increased slightly but still
SiO2 (%) 20.7 21.3 20.4–21.6 within  3 S.D. obtained from the past year’s statistical
CaO (%) 65.2 65.1 64.8–65.6 data. No adverse effect on the rotary kiln plant operation
Fe2O3 (%) 3.33 3.37 3.30–3.42
Al2O3 (%) 4.96 5.37 4.79–5.63
was experienced during feeding and processing of glass.
MgO (%) 0.57 0.61 0.48–0.66
Tricalcium silicate (C3S) (%) 62.7 55.6 52.3–65.5
Blaine (m2/kg) 308 304 290–323 Acknowledgements
Fineness ( 325) (%) 93.5 94.4 90.8–94.6
The support from the Environmental Protection
EN compressive strength
2 day (MPa) 21.9 20.3 18.3–23.1
Department of Hong Kong SAR Government is greatly
7 day (MPa) 39.6 36.5 34.3–42.3 appreciated. The authors are grateful to the financial
28 day (MPa) 55.1 53.5 48.8–59.6 support from the Environmental Conservation Fund/
Woo Wheelock Green Fund.

References
5. Conclusion
Environmental Protection Department (EPD) 2002. Waste. Hong
Thirty-two types of glass typically found in Hong Kong SAR Government. Available: http://www.info.gov.hk/epd/.
Glass Packaging Institute (GPI) 2000. Recycling. Available: http://
Kong were analyzed in terms of their elemental compo- www.gpi.org/.
sition. The composition of the glass does not relate with Wheeler, E.L., 1958. Glass and Its Working Characteristics. Inter-
its color, nor its origin. It depends on the glass applica- science Publishers, New York.
tion. The results are consistent with reported values Mebta, P.K., 1978. Cement Standards—Evolution and Trends.
American Society for Testing & Materials, Philadelphia, USA.
except that the sodium oxide content of the glass bot-
Skoog, D.A., 1985. Principles of Instrumental Analysis, third ed.
tles, 12–19%, is significantly higher than that reported Saunders College Publishing, USA.
for the American lead bottle glass, 7.6%. The general Hewlett, P.C., 1998. Lea’s Chemistry of Cement and Concrete, fourth
composition of glass fits the requirement of the raw ed. Arnold, UK.
materials for cement production. Annual Book of ASTM Standards, Designation: C150-97.