Electrochemical chloride extraction effect on blended ...

Advances in Cement Research Volume 23 Issue 5 Electrochemical chloride extraction effect on blended cements Ismail and Muhammad

Advances in Cement Research, 2011, 23(5), 241–248 http://dx.doi.org/10.1680/adcr.2011.23.5.241 Paper 1000036 Received 07/05/2010; revised 07/09/2010; accepted 15/09/2010 Thomas Telford Ltd & 2011

Electrochemical chloride extraction effect on blended cements Mohammad Ismail

Bala Muhammad

Professor, Faculty of Civil Engineering, Universiti Teknologi Malaysia, Johor Bahru, Johor, Malaysia

Senior Lecturer, Faculty of Civil Engineering, Universiti Teknologi Malaysia, Johor Bahru, Johor, Malaysia

Use of high magnitude current for electrochemical chloride extraction (ECE) has led many researchers to look at the side effects on ordinary Portland cement (OPC) but not blended cements. To simulate the effect of applying electrochemical chloride extraction on blended cements experimentally, 100% OPC, 70% OPC + 30% pulverised fuel ash and 60% OPC + 40% ground granulated blast-furnace slag were prepared and subjected to electrolysis between embedded steel cathodes and external anodes of activated titanium mesh. The cathodic current density applied was 5 A/m2 over a treatment period of 12 weeks. The effect of ECE on the physical and microstructural properties of the specimens was studied using microhardness and mercury intrusion porosimetry techniques. Eventually, removal of chloride ions was noted to be more effective in OPC in comparison with blended cements. In addition, although a slight increase in the total porosity was observed near the steel cathode in the OPC matrix, the reverse was, however, the case with the blended cements.

Introduction Corrosion of reinforcement is one of the major problems causing undue deterioration in reinforced concrete structures. Steel corrosion is normally induced by chlorides from deicing salts, seawater or from admixtures included during mixing (Minagawa and Hisada, 2008). Chloride extraction from concrete is accomplished by applying an anode and electrolyte to the concrete surface and passing direct current between the anode and the reinforcing steel (as cathode). Principally, electrochemical chloride extraction (ECE) is similar to cathodic protection, except that the current in the former is about 50 to 500 times greater than that used in the latter. It is also worth noting that while chloride extraction is a short-term treatment, cathodic protection is normally intended to remain in operation throughout the life span of the structure (Abdelaziz et al., 2009). In the process of ECE, negatively and positively charged ions (Cl , OH , Naþ and K þ ) are observed to migrate toward the anode and cathode respectively (Orellan Herrera et al., 2006). Accumulation of these ions combined with hydroxyl ions, which are also generated at the cathode during the electrochemical process, to form alkali hydroxides have been suggested to cause softening in the cement matrix around the steel (Ali and Rasheeduzzafar, 1993). A popular method to assess softening of cement (binder) around the reinforcement after the ECE process is by looking at the bond strength (Ding et al., 2009; Guo et al., 2008; Ihekwaba et al., 1996; Weichung et al., 2006; Yao and Lei, 2009). The microhardness measurement technique (MMT) was, however,

used in some works (Bertolini et al., 1996; Sergi and Page, 1994). For instance, a small but statistically significant decrease in hardness of the hardened cement paste near the steel cathode (, 1 mm in distance) was observed by Sergi and Page (1994) when polarisation was applied simulating cathodic protection conditions (20 mA/m2 ) for a period of 110 days. Bertolini et al. (1996), however, found no significant differences between the cement hardness near the cathode and the hardness in the bulk, especially when compared statistically. While there are many studies on ECE in OPC concrete, little work has been done on blended cement. The aim of this study was to evaluate the efficiency of ECE in removing chloride ions from blended cement as compared to OPC matrix. The physical and microstructural changes around the cathode due to the ECE process were observed through MMT and mercury intrusion porosimetry (MIP). Possible softening in the hardened pastes of both OPC and blended cements was evaluated.

Experimental programme Materials and mix proportions Ordinary Portland cement, pulverised fuel ash (PFA) and ground granulated furnace slag (GGBS) were used as cementitious materials throughout. The chemical compositions of these materials are shown in Table 1. A mild steel plate was used as the cathode and a mixed metal oxide-activated titanium mesh was employed as the external anode. The chemical composition of the mild steel is shown in Table 2. Three distinct mixes were 241

Advances in Cement Research Volume 23 Issue 5

Electrochemical chloride extraction effect on blended cements Ismail and Muhammad

Material

OPC PFA GGBS

Chemical composition: % CaO

SiO2

Al2 O3

MgO

SO3

Na2 O

K2 O

LOI

65.3 1.45 40.09

20.6 48.2 35.51

5.32 32.2 12.59

1.2 0.66 9.11

3.03 0.52 0.150

0.09 2.58 0.54

0.75 0.98 0.24

0.77 3.84 1.17

Table 1. Chemical composition of cementitious materials and mild steel plate

Material Composition: %

C

Si

Mn

P

S

Cr

Mo

Ni

Al

Cu

Sn

Fe

0.06

0.01

0.29

0.004

0.03

0.04

0.01

0.03

.012

0.02

0.05

bal.

Table 2. Chemical composition of mild steel plate

employed and these include 100% OPC and two blended mixes; 70% OPC + 30% PFA and 60% OPC + 40% GGBS. However, for ease of discussion, OPC + PFA and OPC + GGBS are simply referred to as PFA and GGBS, respectively. The cement or binder to water ratio was 0.5 throughout. Specimen preparation Prior to mixing, 1% NaCl by weight of cement was dissolved in the mixing water. Fresh cement pastes were cast in specially made plastic moulds of internal size measuring 50 mm 3 50 mm 3 100 mm. Steel plates fixed to the moulds by special plastic holders were buried in the mix perpendicular to the longer side of the specimen. Compaction was achieved through mechanical vibration for 1 min in all cases. In order to avoid evaporation of water the specimens were covered with plastic and stored at room temperature and 80 5% relative humidity. Specimens were demoulded after 18 h and then wrapped in wet cloths followed by curing in air-tight containers for 8 weeks. Fabrication of test reservoir Polarisation was carried out in specially fabricated plastic reservoirs with capacity of six specimens each. Specimens were wrapped individually in cloth leaving one face uncovered for the anode. Eventually, the titanium anodes and the steel plates (cathodes) were connected to the positive and negative poles of a galvanostat, respectively. About 5 l of saturated calcium hydroxide solution was then introduced into the reservoirs with specimens placed parallel to their longitudinal dimensions at equal distances apart. A single current density (5 A/m2 ) was applied for 12 weeks.

involves slicing specimens into sections at equidistant positions as shown in Figure 1, followed by grouping as well as pressing in the pore-expression device. The pore fluids were analysed for OH , Naþ , K þ , SO4 2 and Cl ionic concentrations as a function of distance from the steel cathode. Spectrophotometric methods were employed to determine chloride and sulfate concentrations (Vogel, 1978). After being dried at 1058C, the powders collected from the cutting operations were dissolved in hot nitric acid and then analysed for total chloride contents. Concentrations of sodium and potassium ions in the extracted pore solutions were, however, determined by flame photometry.

Cathode 50

Surface anode

9

18 27

36 45 54 63 72 81 90

Chemical analysis At the end of specimen treatment, pore fluid was extracted from three samples using the pore-pressing technique. This process 242

Figure 1. Cutting arrangements for pore-pressing

Microhardness measurement The MMT was conducted by using sub-specimen sections which were obtained through cutting the middle of the main specimens at right angles to the cathode. The sub-specimens were ground to assume flat shapes using ordinary water as lubricant. These were then polished to 9 ìm with alumina and ethanediol (C2 H5 OH), followed by another polishing to 1 ìm with diamond paste. Microhardness was measured using a Buehler Micromet 4. Points at six different distances from the steel cathode were considered; 0.1, 0.5, 1, 5, 10 and 25 mm. However, for a given distance ten readings were taken across the surface resulting in a total of 60 measurements on each of the specimens. Readings were taken around the middle part of the polished sections, so that the edges were avoided. Indeed, readings taken for each depth from the surfaces were averaged and represented by a single point on the curve. Eventually, statistical analysis was carried out to compare differences in hardness with varying distance from the surface. Two points were selected for this purpose; 0.1 mm and at least 10 mm from the steel cathode. Mercury intrusion porosimetry measurement Pore size distribution (PSD) was examined on fragmental pieces (about 30 mg) from each cement paste. Prior to examination, fragments were placed in bottles containing propanol for 1 week. The idea was to remove any available water from the specimens. The specimens were then dried in a stream of cool air using a hair drier and placed in a desiccator. However, the desiccators were evacuated after a week in order to remove the alcohol.

1600

OH concentration: mM/l

Hydroxyl ion concentrations were determined by titration with standard nitric acid.


Cathode

OPC

1400

PFA

1200

GGBS

1000 800 600 400 200 0 0

10

20

30

40

50

60

Distance from surface: mm

Figure 2. Hydroxyl ionic concentration in OPC and blended cements

14·5

OPC

Cathode

PFA 14·0

GGBS

13·5

pH


13·0 12·5 12·0 0

10

20 30 40 Distance from surface: mm

50

60

Figure 3. pH in OPC and blended cements

Test results Ionic concentration profiles Figure 2 shows the concentration of hydroxyl ions in OPC and the two blended mixes. At a distance of 50 mm from the surface, hydroxyl ionic concentration in the vicinity of the cathode was observed to reach 1400, 1200 and 540 mM/l for OPC, PFA and GGBS, respectively. This rise in OH concentration around the cathode results in a pH of more than 14, particularly, with respect to OPC and PFA, as shown in Figure 3. The pH of GGBS cement was also elevated to about 13.7 near the cathode. Therefore, hydroxyl ionic concentration of this magnitude could be detrimental as it motivates higher pH values capable of triggering alkali–silica reaction (ASR). In fact, a previous study on cement close to the cathode indicated a strong possibility of ASR in concrete known to contain even mildly reactive aggregates which, under normal conditions, may be considered harmless (Page et al., 1994; Tritthart, 1996). Figures 4(a) and 4(b) show the free and total chloride concentration profiles. More chloride was driven out from the OPC when compared with the blended cements. This is not surprising since the presence of slag or pozzolana in blended cements often limits

chloride diffusion more effectively than OPC alone, which is attributed to a number of reasons such as fineness of pore structure, greater tortuosity and pore surface interactions (Page et al., 1981). Even though, in this case, the influence of an electric current exists, similar mechanisms resulting from the introduction of PFA and GGBS might restrict ionic migration. This is evident in the free and total sodium concentration as well as free and total potassium concentration results shown in Figures 5 and 6, respectively. In fact, very little increase in the potassium ionic concentration was observed near the cathode for the GGBS cement specimen, as indicated in Figure 6. Furthermore, equilibrium of free and total chloride was also evident in this case, as shown in Figure 7. However, as all the points appear to lie on a single straight line, there is barely any evidence in the present case to suggest that the blended cements have a superior chloride binding capacity, as suggested by Holden et al. (1983). 243



Free and total K content: mM/g

Cathode OPC

0·15

PFA GGBS

0·10 0·05 0 0

10

20 30 40 50 Distance from surface: mm (a)

0·6

Cathode

0·5

OPC-F

0·4

OPC-T PFA-F

0·3

PFA-T

0·2

GGBS-F

0·1

GGBS-T

0 0

10

20 30 40 50 Distance from surface: mm

60

Figure 6. Free and total K þ ionic concentration in OPC and blended cements

Cathode

0·35

Total Cl content: mM/g

60

0·30 0·25 0·20

0·150

OPC

0·15

0·05

GGBS

0 0

10

PFA

0·125

OPC PFA GGBS

0·10

20 30 40 50 Distance from surface: mm (b)

60

Figure 4. Chloride concentrations in OPC and blended cements: (a) free; (b) total

Free Cl: mM/g

Free Cl content: mM/g

0·20

0·100 0·075 0·050 0·025 0

Free and total Na content: mM/g

0

Cathode

0·7

0·10

0·15

0·20

0·25

0·30

Total Cl: mM/g

0·6

OPC-F

0·5

OPC-T

0·4

Figure 7. Free chloride content as a function of total chloride content

PFA-F

0·3

PFA-T

0·2

GGBS-F

0·1

GGBS-T

0 0

10

20 30 40 50 Distance from surface: mm

60

Figure 5. Free and total Naþ ionic concentration in OPC and blended cements

Meanwhile, the effects of pozzolanic materials on chemical binding of chloride ions have also been investigated. Presence of PFA or GGBS in blended cement, for instance, was observed to be responsible for low concentration of C3 A (Kropp and Hilsdorf, 1995). This could affect the overall binding capacity of the mix since C3 A is considered to be the most important compound for chemical binding of chloride ions to form Friedel’s salt (3CaO.Al2 O3 :CaCl2 :10H2 O). But, contrary to the formation of low C3 A in pozzolanic blended cements, calcium–silicate– 244

0·05

hydrates (CSH) and CSH-like phases were observed to be on the increase. However, increase in CSH phases could expedite chloride adsorption due to surface forces. This is a more loosely form of bound chloride in comparison to chlorides bound by C3 A (Maslehuddin et al., 1996; Orellan Herrera et al., 2006). The relationship between sulfate concentration and pH is shown in Figure 8. Indeed, the increase in pH of OPC and GBBS is seen as the main cause of the elevated sulfate ionic concentration near the steel cathode. Microhardness The MMT results for specimens before and after ECE process are presented in Figures 9–11. Figure 9 shows the microhardness of OPC with considerable scatter at regions close to the cathode and this is common to both treated and untreated specimens. The overall microhardness in the treated specimen was, however, observed to be lower when compared with the untreated specimens, and this is also shown in Table 3.


1000


50

OPC

Knoop hardness: Hk

PFA GGBS

SO4: mM/l

100

40 30 20 0 A/m2 5 A/m2

10

10

0 5

0

1 13·0

13·5

14·5

14·0

pH

15 10 20 Distance from cathode: mm

25

Figure 10. Microhardness profiles of PFA

Figure 8. Sulfate ionic concentration in pore solution of OPC and blended cements

50

Knoop hardness: Hk

40 50

Knoop hardness: Hk

40

30

30 20 0 A/m2 10

20

5 A/m2

0 0 A/m

10

5

0

2

15 10 20 Distance from cathode: mm

25

5 A/m2

Figure 11. Microhardness profiles of GGBS 0 0

5

10

15

20

25

The reduction in microhardness of the treated specimen might be attributed to the influence of ECE. This is further strengthened by the fact that hardness of the cement matrix at regions close to the steel cathode, where ECE is considered most effective, was

Distance from cathode: mm

Figure 9. Microhardness profiles of OPC

Cement type OPCtreated OPCuntreated PFAtreated PFAuntreated GGBStreated GGBSuntreated

Distance: mm

Current: A/m2

Mean

Variance

SD

P (T < t ) two-tail

0.1 > 10

5

21.90 27.18 29.87 26.57 30.90 27.20 35.77 26.90 29.53

48.51 37.68 77.20 82.60 59.55 43.13 40.94 97.45 82.19

6.96 6.14 8.79 9.09 7.72 6.57 6.40 9.87 9.07

0.0009 , 2.0 0.0003 , 2.0

0.1 > 10 0.1 > 10

0 5 0 5 0

0.03 , 2.0 0.76 , 2.0 1.9 3 106 , 2.0 0.003 , 2.0

Table 3. Microhardness: effect of distance and current density

245

observed to be slightly less than the hardness in the bulk. In fact, a current density of about 2 A/m2 was previously reported to be the cause of the formation of microcracks in cement paste containing 6% salt within 8 weeks (Miller, 1994). Similarly, there appeared to be a marked decrease in the microhardness of PFA near the cathode, as shown in Table 3. However, contrary to OPC, the microhardness profile of the treated PFA overlapped the untreated curve almost exactly, as indicated in Figure 10. Presence of fly ash in the PFA paste could be the obvious reason. Indeed, the capacity of treated PFA to maintain almost similar microhardness to that of the untreated material could be attributed to greater inter-particle compaction with a consequent slow percolation rate, which is normally the case with a matrix containing finer substances such as fly ash. In fact, according to Hussain and Rasheeduzzafar (1994), a distinct advantage of fly ash-blended cement matrix over plain cement matrix is greater density and impermeable microstructure. This outcome further suggests a higher capacity of PFA towards resisting the influence of ECE as against the situation witnessed in treated OPC. However, GGBS behaved in a different manner by depicting somewhat higher microhardness value, especially at regions close to the cathode, as shown in Figure 11. In addition, the treated GGBS specimen, unlike in the previous cases, yielded a higher microhardness value than the untreated sample, although this lasted for only about 14 mm from the cathode. Indeed, it may not be a surprise that the treated GGBS was barely affected by ECE, since the action of the applied current, which normally causes an increase in the concentration of cations also known to be responsible for cement weakening, may be prevented by the GGBS substances. This is in line with the fact that GGBS cements are known to exhibit inherent protection qualities against chloride attacks (Ali and Rasheeduzzafar, 1993). Mercury intrusion porosimetry The total intruded volume of mercury, which represents the porosity of all specimens treated under MIP, is shown in Table 4. Obviously, a reduction in porosity with increasing current density is evident. Figures 12–14 show the MIP results of the OPC and the two blended cements.

Cement type OPC PFA GGBS

Current density: A/m2 0 5 0 5 0 5

Time: weeks 12 12 12 12 12 12

Table 4. Mercury intrusion porosimetry: effect of ECE treatment on porosity

246

Cumulative intrusion volume: cm3/g


0·20 0·16 5 A/m2 0·12

Untreated

0·08 0·04 0 0·001

0·1 10 Pore diameter: µm

1000

Figure 12. Cumulative pore size distribution of OPC at 50 mm from the surface



0·20 0·16 0·12

5 A/m2 Untreated

0·08 0·04 0 0·001


1000

Figure 13. Cumulative pore size distribution of PFA at 50 mm from the surface

For the OPC shown in Figure 12, treated samples were more porous than untreated specimens. In fact, the total porosity of the untreated specimens near the steel was observed to be 0.1096 cm3 /g, whereas that of the treated samples yielded up to 0.1339 cm3 /g. Thus, as a result of the ECE treatment, there was an increment of about 22.2% porosity. However, the blended cements behaved differently. The treated samples were denser than the untreated specimens. There was no obvious change in

Distance from surface: mm 50 50 50 50 50 50

Total porosity: cm3 /g 0.1096 0.1339 0.1769 0.1678 0.1777 0.1442




0·20

specimens, particularly at regions close to the cathode. However, the total porosity reduced with distance from the cathode in both OPC and the blended cements.

0·16 0·12

Acknowledgements

5 A/m2 0·08

The authors would like to thank the Ministry of Science and Technology Malaysia and Research Management Center Universiti Teknologi Malaysia for providing financial support that led to the successful completion of this research work.

Untreated

0·04 0 0·001


1000

Figure 14. Cumulative pore size distribution of GGBS at 50 mm from the surface

the pore size distribution when current was induced, as can be seen from the graphs shown in Figures 13 and 14, but the total intruded volume as presented in Table 4 indicates that the treated samples were relatively denser. With respect to increase in pore size distribution, Locke et al. (1983) suggested that as the pH in the solution increases, the stability of the CSH gel decreases, therefore becoming more soluble. This, according to Locke, may possibly lead to more empty spaces and an increased porosity in the OPC paste. The fact that the trend was reversed for the blended cements suggests that the gel formed in blended cements does not easily dissolve with increase in pH. Indeed, in the present case, the pore structure appeared to become generally tighter. However, the total porosity reduced with increase in distance from the cathode.

Conclusions The following conclusions can be drawn from the present study. (a) Removal of chloride ions was more effective in purely OPC mixes when compared with PFA- and GGBS-blended cements. As a result, more chloride was driven out from the OPC. This is possibly related to the denser matrix or finer pore structure associated with the blended cements which lowers permeability thereby limiting chloride diffusion. (b) Changes were observed between microhardness in the treated and untreated samples. Although microhardness in the treated OPC was comparatively lower than that of the untreated specimens, similar observations on the microhardness in PFA and GGBS showed equivalent and higher values, respectively. These outcomes suggest higher capacities in PFA and GGBS towards resisting the influence of ECE in comparison with the situation witnessed in OPC. (c) Although, the pore structure of the cements around the cathode did not appear to change substantially after ECE as determined by MIP, increase in porosity was observed near the steel cathode in the OPC. Indeed, there was a porosity increment of about 22.2% between treated and untreated OPC

REFERENCES

Abdelaziz GE, Abdelalim AMK and Fawzy YA (2009) Evaluation

of short and long-term efficiencies of electrochemical chloride extraction. Cement and Concrete Research 39(8): 727–732. Ali MG and Rasheeduzzafar (1993) Cathodic protection current accelerates alkali silica reaction. ACI Materials Journal 90(3): 247–252. Bertolini L, Yu SW and Page CL (1996) Effect of electrochemical chloride extraction on chemical and mechanical properties of hydrated cement paste. Advances in Cement Research 8(31): 93–100. Ding Z, Xing F and Li WH (2009) Research of influencing factors in reinforced concrete subjected to electrochemical chloride extraction. Journal of Wuhan University of Technology 31(7): 19–21, 34. Guo YX, Gong JX and You ZG (2008) Experimental study of characteristics of concrete experienced electrochemical extraction of chlorides. Dalian Ligong Daxue Xuebao/Journal of Dalian University of Technology 48(6): 863–868. Holden WR, Page CL and Short NR (1983) The influence of chloride and sulphates on concrete durability. In Corrosion of Reinforcement in Concrete Contruction (Crane AP (ed.)). Society of Chemical Industry, London, UK, pp. 143–149. Hussain SE and Rasheeduzzafar (1994) Corrosion resistance performance of fly ash blended cement concrete. ACI Materials Journal 91(3): 264–272. Ihekwaba NM, Hope BB and Hansson CM (1996) Pullout and bond degradation of steel rebars in ece concrete. Cement and Concrete Research 26(2): 267–282. Kropp J and Hilsdorf HK (1995) Performance Criteria for Concrete Durability. E & FN Spon, London, UK, Rilem Report 12. Locke CE, Dehghanian C and Gibbs L (1983) Effect of impressed current on bond strength between steel bar and concrete. Proceedings of Corrosion 83, NACE, Anaheim, USA, paper no. 178, pp. 178.1–178.16. Maslehuddin M, Page CL and Rasheeduzzafar (1996) Temperature effect on the pore solution chemistry in contaminated cements. Magazine of Concrete Research 48(5): 5–14. Miller JB (1994) Structural aspect of high powered electrochemical treatment of reinforced concrete. In Corrosion and Corrosion Protection of Steel in Concrete (Swamy RN (ed.)). Academic Press, Sheffield, UK, pp. 1499–1514. 247



Minagawa H and Hisada M (2008) Influence of content and origin

alkali-silica reaction in reinforced concrete: stage 2. In Project Report Transport Research Laboratory, TRL, vol. 62. Bridges and Ground Engineering Resource Center, Transport Research Laboratory, Wokingham, Berkshire, UK. Tritthart J (1996) Electrochemical chloride removal – a case study and laboratory tests. Proceedings of the 4th International Symposium on Corrosion of Reinforcement in Concrete Construction (Page CL, Bamforth PB and Figg JW (eds)). Society of Chemical Industry, London, UK, pp. 433–447. Vogel AI (1978) Vogel’s Textbook of Quantative Inorganic Analysis, 3rd edn. Longmans, London, UK. Weichung Y, Jiang JC and Chi CH (2006) Selecting an adequate procedure for the electrochemical chloride removal. Cement and Concrete Research 36(3): 562–570. Yao W and Lei F (2009) Experimental study on electrochemical chloride extraction from mortar. Tongji Daxue Xuebao/ Journal of Tongji University 37(3): 363–368.

of chloride ion on repairing effect of desalination method. Proceedings of the 3rd ACF International Conference-ACF/ VCA, Ho Chin Minh City, Vietnam. Vietnam Institute for Building Materials, Hanoi, Vietnam, pp. 1041–1047. Orellan Herrera JC, Escadeillas G and Arliguie G (2006) Electrochemical chloride extraction: Influence of C3 A of the cement on treatment. Cement and Concrete Research 36(10): 1939–1946. Page CL, Short NR and Eltarras A (1981) Diffusion of chloride ions in hardened cement paste. Cement and Concrete Research 11(3): 395–406. Page CL, Yu SW and Bertolini L (1994) Some potential side effects of electrochemical chloride removal from reinforced concrete. Proceedings of the UK Corrosion and Eurocorr 94. The Institute of Materials London, UK, vol. 3, pp. 228–238. Sergi G and Page CL (1994) The effects of cathodic protection on

WHAT DO YOU THINK?

To discuss this paper, please submit up to 500 words to the editor at www.editorialmanager.com/acr by 1 January 2012. Your contribution will be forwarded to the author(s) for a reply and, if considered appropriate by the editorial panel, will be published as a discussion in a future issue of the journal. 248